Language, Texts, and Society: Explorations in Ancient Indian Culture and Religion (Cultural, Historical and Textual Studies of Religions) 9780857284310, 0857284312

Table of contents :
Half Title Page......Page 1
Series Page......Page 2
Main Title Page......Page 3
Copyright Page......Page 4
Contents......Page 5
Preface......Page 7
Abbreviations......Page 9
I. Young Svetaketu: A Literary Study of an Upanisadic Story......Page 13
II. dharmaskandhah and brahmasamsthah: A Study of Chandogya Upanisad 2.23.1......Page 53
III. Orgasmic Rapture and Divine Ecstasy: The Semantic History of ananda......Page 75
IV. Amrta: Women and Indian Technologies of Immortality......Page 101
V. Power of Words: The Ascetic Appropriation and the Semantic Evolution of dharma......Page 121
VI. Semantic History of Dharma: The Middle and Late Vedic Periods......Page 137
VII. Explorations in the Early History of Dharmasastra......Page 155
VIII. Structure and Composition of the Manava Dharmasastra......Page 179
IX. Caste and Purity: A Study in the Language of the Dharma Literature......Page 217
X. Rhetoric and Reality: Women’s Agency in the Dharmasastras......Page 247
XI. Manu and Gautama: A Study in Sastric Intertextuality......Page 261
XII. Manu and the Arthasastra: A Study in Sastric Intertextuality......Page 275
XIII. Unfaithful Transmitters: Philological Criticism and Critical Editions of the Upanisads......Page 287
XIV. Sanskrit Commentators and the Transmission of Texts: Haradatta on Apastamba-Dharmasutra......Page 301
XV. Hair and Society: Social Significance of Hair in South Asian Traditions......Page 321
XVI. Abhaksya and Abhojya: An Exploration in Dietary Language......Page 351
XVII. Food for Thought: Dietary Rules and Social Organization in Ancient India......Page 367
References......Page 395
Index......Page 413

Citation preview

LANGUAGE, TEXTS, AND SOCIETY

Cultural, Historical and Textual Studies of Religions The volumes featured in the Anthem Cultural, Historical and Textual Studies of Religions series are the expression of an international community of scholars committed to the reshaping of the field of textual and historical studies of religions. Titles in this series examine practice, ritual, and other textual religious products, crossing different area studies and time frames. Featuring a vast range of interpretive perspectives, this innovative series aims to enhance the way we look at religious traditions.

Series Editor Federico Squarcini, University of Firenze, Italy

Editorial Board Piero Capelli, University of Venezia, Italy Vincent Eltschinger, ICIHA, Austrian Academy of Sciences, Austria Christoph Emmrich, University of Toronto, Canada James Fitzgerald, Brown University, USA Jonardon Ganeri, University of Sussex, UK Barbara A. Holdrege, University of California, Santa Barbara, USA Sheldon Pollock, Columbia University, USA Karin Preisendanz, University of Vienna, Austria Alessandro Saggioro, Sapienza University of Rome, Italy Cristina Scherrer-Schaub, University of Lausanne and EPHE, France Romila Thapar, Jawaharlal Nehru University, India Ananya Vajpeyi, University of Massachusetts Boston, USA Marco Ventura, University of Siena, Italy Vincenzo Vergiani, University of Cambridge, UK

LANGUAGE, TEXTS, AND SOCIETY EXPLORATIONS IN ANCIENT INDIAN CULTURE AND RELIGION

Patrick Olivelle

Anthem Press An imprint of Wimbledon Publishing Company www.anthempress.com This edition first published in UK and USA 2011 by ANTHEM PRESS 75-76 Blackfriars Road, London SE1 8HA, UK or PO Box 9779, London SW19 7ZG, UK and 244 Madison Ave. #116, New York, NY 10016, USA Copyright © Patrick Olivelle 2011 The author asserts the moral right to be identified as the author of this work. Graphics and layout © Federico Squarcini and Stefano Miniati Cover photography © Clelia Pellicano All rights reserved. Without limiting the rights under copyright reserved above, no part of this publication may be reproduced, stored or introduced into a retrieval system, or transmitted, in any form or by any means (electronic, mechanical, photocopying, recording or otherwise), without the prior written permission of both the copyright owner and the above publisher of this book. British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library. Library of Congress Cataloging-in-Publication Data A catalog record for this book has been requested. ISBN-13: 978 0 85728 431 0 (Hbk) ISBN-10: 0 85728 431 2 (Hbk) This title is also available as an eBook.

Contents

Preface

7

Abbreviations

9

I. Young ‡vetaketu: A Literary Study of an Upanißadic Story

13

II. DharmaskandhåΔ and brahmasa∫sthaΔ: A Study of Chåndogya Upanißad 2.23.1

53

III. Orgasmic Rapture and Divine Ecstasy: The Semantic History of ånanda

75

IV. Amr¢tå: Women and Indian Technologies of Immortality

101

V. Power of Words: The Ascetic Appropriation and the Semantic Evolution of dharma

121

VI. Semantic History of Dharma: The Middle and Late Vedic Periods

137

VII. Explorations in the Early History of Dharma†åstra

155

VIII. Structure and Composition of the Månava Dharma†åstra

179

IX. Caste and Purity: A Study in the Language of the Dharma Literature

217

X. Rhetoric and Reality: Women’s Agency in the Dharma†åstras

247

6

LANGUAGE, TEXTS AND SOCIETY

XI. Manu and Gautama: A Study in ‡åstric Intertextuality

261

XII. Manu and the Artha†åstra: A Study in ‡åstric Intertextuality

275

XIII. Unfaithful Transmitters: Philological Criticism and Critical Editions of the Upanißads

287

XIV. Sanskrit Commentators and the Transmission of Texts: Haradatta on Åpastamba Dharmasütra

301

XV. Hair and Society: Social Significance of Hair in South Asian Traditions

321

XVI. Abhakßya and Abhojya: An Exploration in Dietary Language

351

XVII. Food for Thought: Dietary Rules and Social Organization inAncient India

367

References

395

Index

413

Preface

The credit –or the blame– for this collection of essays goes to Federico Squarcini. It is he who suggested the publication of some collected papers of mine during my visit to Bologna for the defense of his doctoral dissertation. The papers collected in this volume span about a decade from 1995 to 2004. During the previous two decades the focus of my scholarly work was the ascetic traditions of India, principally those associated with the Brahmanical tradition. My papers from that period are being published in a separate volume. The last decade, coinciding broadly with my move from Indiana University, Bloomington, to the University of Texas at Austin, saw a shift in my focus. The invitation to translate the Upanißads from the Oxford University Press spurred me to work more closely with the late Vedic literature, resulting in several articles of this volume (I-IV, XIII). My long-standing interest, however, has been the Indian legal tradition represented by the Dharma†åstras, an interest that goes back to my teacher at the University of Pennsylvania, Professor Ludo Rocher, and sustained by my close association with my friend and colleague, Professor Richard Lariviere. In the late 1990s I undertook the edition and translation of the earliest extant legal texts, the four Dharmasütras (Oxford, 1999; and Motilal Banarsidass, 2000), and then the critical edition of the Månava Dharma†åstra (Oxford, 2004 and 2005). Work on the Dharma†åstric material resulted in several articles included in this volume (V-XII, XIV). Another long-standing interest of mine has been the social construction of the human body, the ways in which the human body is conceived, constructed, and manipulated by culture. Struggling over the years to write a book on this topic, I have only succeeded in producing a series of articles included here (XVXVII). These essays, therefore, span not only a relatively long period

8

LANGUAGE, TEXTS AND SOCIETY

of time; they also represent several scholarly interests and pursuits over that period. Collections of papers –Kleineschriften– most often lack a theme or a focus. By necessity, papers included in such volumes are written on different occasions, for different publications and audiences. This collection is no different. If there is a unifying theme here, it is the search for historical context and developments hidden within words and texts. An early word study on sa∫nyåsa (included in the companion volume) convinced me that words, and therefore the cultural history represented by those words, that we take for granted as having a continuous and long history are often new and even neologisms and thus provide important clues to cultural and religious innovations. My book-length study on the å†ramas, as well as the short pieces included in this volume, such as those on ånanda and dharma, again seek to find cultural innovation and historical changes within the changing semantic fields of key terms. Closer examination of other terms taken for granted as central to “Hinduism”, such as dvija, †åstra, †ruti, smr¢ti, and purußårtha, will, I am sure, provide similar results. Indian texts have often been studied in the past as disincarnate realities providing information on an ahistorical and unchanging culture. This volume is a small contribution towards correcting that method of textual study. Many influences have shaped my work and interests over the years; many friends and colleagues have given me generously of their time, knowledge, and intellectual companionship. I can here acknowledge only a few. I have already mentioned Ludo Rocher and Richard Lariviere. In 1984 a cowboy named Gregory Schopen joined me at Indiana University and followed me to the University of Texas in 1991. His brilliant scholarship, penetrating questions, and iconoclastic attitude have influenced the questions I ask and the way I approach textual data. More recently, Joel Brereton, Stephanie Jamison, and Oliver Freiberger have been my conversation partners. My student Mark McClish prepared the index. To all of them, and to untold others, a heart-felt Thank You. My wife Suman has been a collaborator in all my research endeavors, especially those involving the painstaking reading of manuscripts. My daughter Meera, now a wonderful young woman, bore with patience and good humor the strange activities of her parents. Patrick Olivelle Austin, March 2005

A

9

Abbreviations

Åpastamba Dharmasütra Aitareya Årañyaka Aitareya Bråhmaña Åpastamba Dharmasütra Åpastamba ‡rautasütra ņvalåyana Gr¢hyasütra Aitareya Upanißad Atharva Veda Atharvaveda Sa∫hitå, Paippalåda recension Atharvaveda Sa∫hitå, ‡aunaka recension Baudhåyana Dharmasütra BU version of the ‡vetaketu story Baudhåyana Gr¢hyasütra Baudhåyana Dharmasütra Bharadvåja ‡rautasütra Bhagavad Gœtå. Böhtlingk, O. and Roth, R. Sanskrit-Wörterbuch. 7 vols. 185575. Reprint: Delhi: Motilal Banarsidass, 1990. BU Br¢hadårañyaka Upanißad. BU(K) Br¢hadårañyaka Upanißad, Kåñva recension C* CU version of the ‡vetaketu story CU Chåndogya Upanißad G Gautama Dharmasütra GDh Gautama Dharmasütra

A AÅ AB ÅpDh Åp‡r ņGr¢ AU AV AV(P) AV(S) B B* BauGr¢ BDh Bhar‡r BhG BR

10

GoB Hir‡r IU JB JU K* Kåt KaU KeU KS KßB KßU M MaU MBh MDh MtU MS MuU N NSm PårGr¢ PMS PU Råm R¢V ‡A ‡åõGr¢ ‡B ‡B(K) ‡B(M) SBE SU TÅ TB TS TU Va

LANGUAGE, TEXTS AND SOCIETY

Gopatha Bråhmaña Hirañyake†i ‡rautasütra Œ†å Upanißad Jaiminœya Bråhmaña. Jaiminœya Upanißad. KsU version of the ‡vetaketu story Kåtyåyana-Smr¢ti Ka™ha Upanißad Kena Upanißad Kå™haka Sa∫hitå Kaußœtaki Bråhmaña Kaußœtaki Upanißad Månava Dharma†åstra (Manusmr¢ti) Måñ∂ükya Upanißad Mahåbhårata Månava Dharma†åstra (Manusmr¢ti) Maitråyañœya (Maitrœ) Upanißad Maitråyañœ Sa∫hitå Muñ∂aka Upanißad Nårada Smr¢ti Nårada Smr¢ti Påraskara Gr¢hyasütra Pürva Mœmå∫så Sütra Pra†na Upanißad Råmåyaña R¢g Veda. ‡åõkhåyana Årañyaka ‡åõkhåyana Gr¢hyasütra ‡atapatha Bråhmaña ‡atapatha Bråhmaña, Kåñva recension. ‡atapatha Bråhmaña, Mådhyandina recension. Sacred Books of the East, Oxford. ‡vetå†vatara Upanißad Taittirœya Årañyaka Taittirœya Bråhmaña Taittirœya Sa∫hitå Taittirœya Upanißad Vasiß™ha Dharmasütra

Abbreviations

VaDh VeS Vi ViDh Vkh VS VS(M) Y YDh YV ZDMG

Vasiß™ha Dharmasütra Vedånta Sütra Vißñu Dharmasütra (Vißñu Sm®ti) Vißñu Dharmasütra (Vißñu Sm®ti) Vaikhånasa Dharmasütra Våjaseneyi Sa∫hitå of the White Yajurveda. Våjasaneyi Sa∫hitå, Mådhyandina recension Yåj∞avalkya Dharma†åstra (Yåj∞avalkya Sm®ti) Yåj∞avalkya Dharma†åstra (Yåj∞avalkya Sm®ti) Yajurveda Zeitschrift der Deutschen Morgenländischen Gesselschaft

11

12

LANGUAGE, TEXTS AND SOCIETY

A

13

I Young ‡vetaketu : A Literary Study of an Upanißadic Story*

Of the many interesting individuals we encounter in the vedic literature, ‡vetaketu, the son of Uddålaka Åruñi,1 comes across as one of the most colorful and true-to-life characters, not least because he is frequently depicted as the vedic equivalent of a spoiled little brat. Although he appears with some frequency in vedic and later literature both as a young man and as a mature adult, his character is most fully developed and exploited for literary-cum-theological purposes in the story of young ‡vetaketu’s2 encounter with a king, a story that has become famous because it contains the important doctrines of “five fires” and the two paths along which the dead travel.

1. Versions of the ‡vetaketu Story We have three versions of the ‡vetaketu story in the Upanißads: Br¢hadårañyaka 6.2.1-8 (B*), Chåndogya 5.3 (C*), and Kaußœtaki 1 (K*).3 The aim of this paper is to examine the divergent ways in Originally published in the Journal of the American Oriental Society 119 (1999): 46-70. MBh 1.3.20 identifies Åruñi as a student of the seer Dhaumya Åyoda. Åruñi got the name Uddålaka because Åyoda sent him to stop the break in a dike. Unable to close the breach he put his body into it and managed to stop the leakage. He got up when his teacher called him and the breach was renewed. Åyoda gave him the name Uddålaka, “Puller-of-the-Stop”. In the Uddålaka Jåtaka (Jåtaka 487) he is said to have got his name because his mother gave birth to him near an Uddåla tree. 2 I have called him “young ‡vetaketu” to distinguish his character as a youth from that of the mature ‡vetaketu (see below §4). 3 For ease of reference I shall name these versions, as well as their putative authors, B*, C*, and K*, respectively. I skirt the issue of whether these authors are the same as the authors of the respective Upanißads, especially because these documents are probably composite works with a series of authors and editors. I deal later (§§2.1.1-3) with the liter* 1

14

LANGUAGE, TEXTS AND SOCIETY

which the authors of these versions develop the character of young ‡vetaketu and to explore the possible theological and/or literary reasons for those divergences. Of the three versions, B* and C* follow each other rather closely, while K* represents a distinctly different redaction. The king’s name in the first two is Pravåhaña Jaivali, and in the latter, Citra Gåõgyåyani (or Gårgyåyañi). These versions have been studied repeatedly by scholars, whose principal, if not sole, aim has been either to establish which of the versions is the oldest and may have served as the archetype for the others, or to reconstruct a hypothetical archetype underlying all the version.4 Renou (1955) has rightly cast doubt on whether the priority of any of the existing versions can be established; indeed, it is highly doubtful that an analysis of these versions will ever provide us with a single clear archetype. Such archetypes are most easily constructed when, as in the case of manuscript transmissions, the changes introduced into the versions are unconscious and accidental, disclosing the genealogy of the manuscripts. The versions of the ‡vetaketu story, I will argue in this paper, are not accidental creations but deliberate literary inventions. Although the archeology of texts has become somewhat unfashionable lately, my objection has less to do with its merits than with the fact that, as a result, a much more significant, interesting, and (most importantly) feasible project –namely the literary study of these texts– has been ignored. Biblical scholars have taken a leadership role in exploiting the literary study of sacred texts; they have asked different types of questions and thereby obtained new insights into the literary and theological motives underlying the composition of biblical text.5 Close attention to language, style, narrative strategy, and choice of words helps us understand what the author is aiming to do, what message, subtle and otherwise, he is attempting to impart to his readers or listenary contexts within which the authors developed these versions, the contexts within which they should be examined. B* includes both the Kåñva and the Mådhyandina recensions, whose differences are minimal and will not affect our study. The doctrine of the five fires and the path to the gods are recorded also in the JB I.45-46, 49-50 (see Bodewitz 1973: 11023; Schmithausen 1994). 4 Renou (1955), in a balanced study of B* and C*, finds that with regard to the listing of the five questions and the final answer (CU 5.9) “la tradition de Ch. [=CU] est indiscutablement plus sûre” (p. 97), while “l’itihåsa est mieux articulé dans BÅ”, but is forced to conclude “La conclusion qui semble s’imposer est que ni l’une ni l’autre version n’ont conservé le texte primitif” (p. 100). Bodewitz (1973: 110-14) and Schmithausen (1994) have focused on the versions of the doctrine of five fires, including the one in JB I.45-46, again with the intension of discovering mutual influences and ultimately the archetype behind all. Söhnen (1981) has focused on the story involving ‡vetaketu, his father, Uddålaka Åruñi, and the king, a story that forms the preamble to the doctrine of the fires. Her conclusion is that K* is the source of the other two versions, and that B* frequently uses C* as the model. More recently, Bronkhorst (1996), focusing again on the question of historical priority, has also come down in favor of the priority of C* vis-à-vis B*. 5 See, for example, Robert Alter, The Art of Biblical Narrative (New York: Basic Books, 1981).

Young ‡vetaketu

15

er.6 Scholars whose main goal is to uncover the most ancient versions of texts often tend to ignore later versions, even though it is these very versions that provide insights into the religious, intellectual, and social history behind the texts. To pay attention only to the oldest version of a text is as shortsightedas an archeologist looking only at the lowest stratumof a dig or a paleontologist only interested in the oldest bones. The story is told not just in the oldest but in the changes we can see from the older to the newer. Likewise, the literary study of texts can also become historically significant when we know the material the authors were working with. Historical and literary study of texts, therefore, need not be antagonistic to each other; they are interdependent and complementary.7

2. Context and Sources We have to address two issues at the outset. First, what were the sources at the disposal of the authors of B*, C*, and K* in composing their respective narratives? Second, what is the literary context within which these narratives are to be located and studied and which may shed light on the authors’ theological and literary objectives? The second is related to the first in that a considerable part of the immediate context of the narratives is shared by B*, C*, and K* and is found also in other vedic texts (Jaiminœya Bråhmaña and ‡åõkhåyana Årañyaka), raising the possibility of tracing at least some of the source material (as opposed to a single archetype) used by the authors. The following is a schematic view of the literary context: I

Contest between faculties

II

Mantha rite

III

‡vetaketu story

IV

Five fires

V

Paths after death: two versions V.1 V.2

BU 6.1, CU 5.1.1-2.3, ‡Å 9.1-7 BU 6.3, CU 5.2.4-9, ‡Å 9.8 BU 6.2.1-8, CU 5.3, KßU 1 BU 6.2.9-14, CU 5.4-9, JB I.45-46 (first part), ‡B 11.6.2.6-10 JB I.46 (second part), 49-50, KßU 2-7 BU 6.2.15-16, CU 5.10

Since BU and CU follow each other closely, we are fortunate to have for each of these sections at least one other independent parallel which can serve as a check in uncovering possible sources. So, for 6 Given the oral nature of the vedic texts, I have regularly refered to the audience of the Upanißads as “listeners”. 7 If we can determine with some certainty the chronology of B*, C*, and K* and whether the later authors were aware of the prior versions, we would be able to give a intertextual dimension to our literary study, casting considerable light on the theological and

16

LANGUAGE, TEXTS AND SOCIETY

example, in I, CU and ‡Å list only five faculties and place II immediately after I, whereas BU lists semen as the sixth faculty and places II after V. We can, therefore, assume that these two features are innovations introduced by the author of BU, and we can ask what may have motivated him to do this (see below 2.1.1). Likewise, the omission of IV by the author of KßU can be seen as an innovation, since IV is found in JB, as well as in BU and CU. It is, moreover, likely, as both Bodewitz (1973: 113) and Schmithausen (1994) have noted, that the JB provides clues to the sources that may have been used by BU and CU, permitting us to see what innovations may have been introduced by the respective authors. It is also likely that V.2, the doctrine of the two paths –to gods and to ancestors– as an innovation shared by BU and CU, goes back to a source they shared, while V.1, the passage to heaven of JB, later recast in KßU, was probably the older sequel to the doctrine of five fires (Bodewitz 1973: 113-14). This leaves us with III, the story of ‡vetaketu, which forms the preamble to IV and V.2 in BU and CU, and to V.1 in KßU, but which is missing in the parallel passage of JB. In her pioneering and detailed study of this episode, Söhnen (1981) has analyzed all three versions, paying close attention to the language, style, and selection of words. Hers is in some ways a literary study of this story, but her analysis is aimed at discovering the historical priority of the respective versions. That aim sometimes biases her judgments, as when she takes brevity or “logical consistency” as an indicator of historical priority (1981: 199). Söhnen takes K* to be the oldest version and the probable source of B* and C*, and in many areas she thinks C* has preserved an older version than B*. When a passage of B* or C* is in agreement with K* we can readily accept that it probably goes back to an original source and that the author of the other version has introduced something new and ask why he may have done so. I am, however, not convinced that there is compelling evidence to claim that K* is either the oldest version or the model for B* and C*. Söhnen has shown that K* is brief and its narrative structure is logical and simple. But does that necessarily make it older? Simplicity and logic can be imposed on a rambling story by a narrator just as, or even more, easily than a simple and logical narrative can be turned into a disjointed one. If, as seems likely, the author of KßU omitted IV, though found in his sources, then he might well have made other drastic changes to the narrative sequence that he deemed necessary for his own literary or theological purposes. What I propose to show is that each version has its own narrative logic from the viewpoint of the respective author, and the additions, subtractions, and modifications can be viewed as part of the narrative strategy of each author.8 literary history of these texts. Unlike their biblical counterparts, however, it is impossible to establish a definite chronology of Upanißadic passages. 8 In explaining these upanißadic passages, Bodewitz (1973, 275) notes: “One should

17

Young ‡vetaketu

It appears likely that of the five text fragments I have isolated above, the fragments I and II existed as a separate unit (which I will call I-II*) as evidenced by ‡Å, and likewise the fragments IV and V form a unit (which I will call IV-V*) as evidenced by JB, a unit which may have contained other material.9 It also seems likely that in this unit the path after death was at first represented by V.1, since it is found in both JB and KßU. At some point IV-V* was recast with an introductory story containing three protagonists: a royal person, ‡vetaketu, and his father.10 This recast unit (which I will call III-V*) was the source of the KßU version. The recast unit appears to have been further modified by replacing V.1 with V.2 and by combining it with I-II*. Now, it is possible that this last version (which I will call IV*) was the work of the author of either BU or CU,11 in which case we must assume that the one borrowed this version from the other. Given the discrepancies between the two versions, and the partial agreement of each with other versions of these fragments, especially with K* in fragment III, it appears more probable that the BU and CU versions are modeled on a version of I-V* that is now lost. Let me present this hypothetical relationship and derivations of the five text fragments: ‡Å 9.1-7

I-II*

IV-V*

JB I.45-46, 49-50

III-V*

KßU 1

I-V* BU 6.1-3 CU 5.1-10

2.1 Theological and Literary Intent In analyzing their theology and the narrative strategy, I find that the author of BU intends to teach a theology of sexual intercourse as a fire sacrifice, while the author of CU pursues a theology of the fire sacrifice offered to one’s breath (pråñågnihotra). The clue to the literary intents of these authors, I believe, is found in the concluding bear in mind that several disconnected passages have been brought together in these upanißads.” That may well be true as far as the origin and the original meanings of the text fragments are concerned, but what I propose to show is that they were not put together haphazardly as an anthology but woven into a literary composition with clear literary and theological motives. 9 In the JB, for example, between the path of those who return (JB 46 first part) and the path of those who do not return (JB 49, second part, and 50), there is the funeral rite (JB 46, second part, 47-48, 49, first part). Another peculiar feature of the JB version is that the doctrine is not ascribed to the Kßatriyas: see Bodewitz 1973: 110-49; 1996: 52. 10 Section IV, the five fires, is also given within the story of the encounter between King Janaka and Yåj∞avalkya in ‡B 11.6.2.6-10. 11 To be precise, I am speaking here only of the authors of Chapter 6 of the BU and Chapter 5 of the CU, even though I think that the same author/editor was responsible also

18

LANGUAGE, TEXTS AND SOCIETY

sections that they have appended to I-V*, sections that deal with sexuality and offering food to the breaths, respectively. The intent of the author of the KßU is more difficult to determine; it appears that his purpose was somewhat narrow and limited to recasting the path after death of V.1 into a narrative of an epic or puråñic type describing a man’s journey to the world of Brahman. Bodewitz (1973: 250-51, 269-75) has objected vigorously, and I think rightly, to Varenne’s (1960) indiscriminate attempt to trace the pråñågnihotra in all these upanißadic texts. Bodewitz, however, is principally interested in examining the “original” intent of these passages, an intent that he discovers by comparing their different versions. Within that context, clearly not all the passages of the fifth chapter of the CU deal with the pråñågnihotra. Bodewitz, and before him Frauwallner (1953, 49f), likewise, find a “water doctrine” (Wasserlehre) as the underlying teaching of the five fires and the path to heaven. This may well be true with regard to the possible original intent and context of these doctrines. Clearly not all the text fragments comprising the sixth chapter of the BU were intended in their original contexts to teach the theology of sex as a sacrifice. The literary study of these texts, however, aims at discovering not an “original” meaning but the literary intent of the author who brought these diverse passages into a narrative unity. Further, it is not necessary that each passage directly espouse the theology; but, together, they are building blocks in the overall literary strategy. Thus, for example, Bodewitz (1973: 269-70) correctly observes that the contest of breaths has nothing directly to do with pråñågnihotra; nevertheless, the supremacy of breath that it establishes sets the scene in the CU for the detailed exposition of the theology of pråñågnihotra in the final section of the fifth chapter. It is within this specifically literary context that I claim that the authors of the BU and the CU intend to teach the theology of sex as sacrifice and the theology of the pråñågnihotra, respectively.

2.1.1 Br¢hadårañyaka Upanißad We know that the author of BU has drastically modified I-II*. For the most part, the structure and content of I-II* are identical in ‡Å and CU, and we can assume that they present more or less the original I-II*.12 I will ignore the numerous minor differences between the BU and ‡Å/CU versions and concentrate here on a few that provide an insight into the author’s aims in constructing his narrative. The for Chapter 5 of the BU and for at least Chapter 6 of the CU. 12 For an examination of I-II as it occurs in ‡Å, BU, and CU, see Bodewitz 1973: 26975. He observes (pp. 274-75): “Note how the myth on the pråñåΔ and the deities forms the introduction to the magic rite and how (mythical) speculations on the pråñåΔ are applied to practical purposes in this årañyaka [‡Å] text. . . CU. 5.1-2 forms a unity and deals with

Young ‡vetaketu

19

author of BU places the mantha rite, which is longer and more complex here than in the parallel versions, after the teaching on the five fires and the two paths (III-V), breaking thereby the natural continuity between the two in I-II*; adds a sixth faculty, semen (retas), together with its power, fecundity or procreation (prajåti), both in the contest and in the mantha rite; adds a sentence containing “When a man knows this . . .” (ya eva∫ veda) to each statement (BU 6.1.1-6) about the powers of the faculties; combines into a single question the query by breath about his food and clothing (and recasts this segment of the narrative); and, lastly, transfers the saying ascribed to Satyakåma Jåbåla from the end of the contest to the end of the mantha rite, and ascribes that saying to a series of teachers and pupils. These changes, I believe, reveal the author’s deliberate strategy to recast the series of text fragments I-V in order to present a theology and (in the final section of Chapter 6) rituals relating to sex and sexual intercourse.13 His theology presents sexual intercourse as a sacrifice. The centerpiece of this theology is given at the beginning of BU 6.4.1, which presents semen as the quintessence of all reality: Of these beings here, the essence is clearly the earth; of the earth, the waters; of the waters, the plants; of the plants, the flowers; of the flowers, the fruits; of the fruits; man; of man, semen. eßå∫ vai bhütånå∫ pr¢thivœ rasaΔ pr¢thivyå åpa apåm oßadhaya oßadhœnå∫ pußpåñi pußpåñå∫ phalåni phalånå∫ purußaΔ purußasya retaΔ.

Then Prajåpati, the creator, sought to prepare a base (pratiß™hå) for the semen and produced the woman. Prajåpati himself provides the primordial divine model for sex; after creating the woman, he stretched out from himself the elongated stone for pressing Soma and impregnated her with it (BU 6.4.2). The Soma stone functions as a penis, establishing a clear link between intercourse and the Soma sacrifice. The author (BU 6.4.3) elaborates his sexual theology by drawing a parallel between the sexual organ of a woman and a sacrificial altar: Her vulva is the sacrificial ground; her pubic hair is the sacred grass; her labia majora are the Soma-press; and her labia minora are the fire the mantha rite [...]. The parallel ‡åõkhÅ. 11 [probably a typo for 9] agrees with CU. The version of BU. has inserted the pa∞cågnividyå, which in CU. comes after the mantha and is omitted in ‡åõkhÅ.” Bodewitz (1973: 273-74) is right in rejecting Deussen’s (1897, 132) view that the contest between the breaths is a later interpolation. Bodewitz (n. 33 on p. 286) concludes that the BU version is less original and that “the whole mantha passage in BU. makes the impression of a later elaboration.” 13 The BU is a document belonging to the White Yajurveda. In the context of the term ånanda, I have noted elsewhere (Olivelle 1997: 172; see below p. 98) that the sexual meaning of ånanda is most prominent in the literature of the Yajurveda, including the BU. Theological speculation about sex and the use of sexual terminology in theological discourse appear to have been a special feature of the Yajurvedic tradition.

20

LANGUAGE, TEXTS AND SOCIETY

blazing at the center. A man who engages in sexual intercourse with this knowledge obtains as a great a world as a man who performs a Soma sacrifice. tasyå vedir upastho lomåni barhi† carmådhißavañe samiddho madhyastas tau mußkau Ù sa yåvån ha vai våjapeyena yajamånasya loko bhavati tåvån asya loko bhavati ya eva∫ vidvån adhopahåsa∫ carati.

In the light of this sexual theology, we can see the reason why the author of BU introduces semen as the sixth and last human faculty in the contest among faculties and in the mantha rite, setting the scene at the very outset for the elaboration of that theology.14 At the end of his narrative of the contest, he uses a phonetic-cumetymological argument to establish the identity of breath, ana (the greatest of the faculties), with food, anna. This identity is also given in the CU narrative, but because the CU separates the two questions regarding food and clothing, the section ends with the drinking of water and the saying ascribed to Satyakåma. The BU, on the other hand, ends on a high note: etam eva tad anam anagna∫ kurvanto manyante (“they think that they are thus making the breath not naked”). With the repetition of ana (= anna) and the alliterated anagna, the author uses a subtle strategy to recall to the listener’s mind that breath is the same as food. By placing the narrative of the five fires immediately after this, he is able to produce a further identification: in the fourth fire food is converted into semen, meaning that semen is the essential form of food (BU 6.2.12). When it rains (third fire), food (anna) is produced; the listener is bound to think here of plants (oßadhi) because plants grow when it rains and then, through the medium of flowers and fruit, become human food, as described in the above passage on the essences. Food is eaten by a man, i.e., within the sacrificial metaphor used, offered in the mouth of the man (fourth fire). In his body the essence of the food is extracted as semen, which he deposits in a woman (fifth fire). Note that in a man the mouth acts as the sacrificial fire, whereas in a woman it is the vagina. The author has subtly taken us from breath, through food, to semen and sexual intercourse. I want to argue further that the author of the BU may have visualized not just the offering in the fifth fire but the offerings in all five fires as a kind of sexual intercourse. A significant passage in the Aitareya Årañyaka (2.1.3) presents a sequence similar to that of the five fires where each subsequent element of the sequence is considered the semen of each preceding: “Next, the creation of semen. The semen of Prajåpati is the gods; the semen of the gods is rain; the semen of rain is 14 There are precedents for this. The account of the internal agnihotra in ‡Å 10 mentions six faculties, including semen. In the BU itself semen is frequently enumerated among the faculties and bodily parts: 2.5.2-7; 3.2.13; 3.7.16-23.

Young ‡vetaketu

21

plants; the semen of plants is food; the semen of food is semen; the semen of semen is the creatures” (athåto retasaΔ sr¢ß™iΔ Ù prajåpate reto devå devånå∫ reto varßa∫ varßasya reta oßadhaya oßadhœnå∫ reto ‘nnam annasya reto reto retaso retaΔ prajåΔ). Thus, for example, in the first fire we can visualize rain as the product of the offering by gods, on the one hand, and as the ejaculated semen of the gods, on the other. This is not far-fetched, because in the description of the path to the fathers from which people return back to this earth (BU 6.2.16), the crucial element is the moon. People reach the moon making it swell, thereby becoming food. Gods feed on that food and emit them once again. Although in the BU description the ejection (ejaculation) of the food/people by the gods in the form of rain is mediated by their passage through the sky and the wind, the Aitareya Årañyaka version makes a direct connection between the gods’ seed and rain. In placing the mantha rite (II) after the fire doctrine (IV-V), the author has made another transition, this time from “knowing” to “doing”, from knowledge to ritual. I noted above that BU adds a statement containing the phrase ya eva∫ veda (“who knows thus”) to each description of the faculties, and the section on the contest ends with the statement that “when a man knows in this way that breath is food, nothing he eats becomes an improper food, nothing he accepts becomes an improper food” (na ha vå asyånanna∫ jagdha∫ bhavati nånanna∫ pratigr¢hœta∫ ya evam etad anasyånna∫ veda). Likewise, the section about the five fires and the two paths deals with knowledge. It begins with ‡vetaketu’s ignorance and the request by his father Åruñi for the knowledge (vidyå) that Brahmins have never had. The narrative of the two paths begins with what happens to people who know the fire doctrine (te ya evam etad viduΔ), and ends with what happens to those who do not know these two paths (atha ya etau panthånau na viduΔ). The mantha rite (BU 6.3), on the contrary, introduces the listener to the ritual side of this knowledge: rites performed with knowledge become productive. The BU description of this rite is the longest. It begins with gathering the necessary ritual items, including “every type of herb and fruit” (sarvaußadha∫ phalånœti) and a bowl made of Udumbara (fig.) wood. Fruit and the Udumbara bowl are not mentioned in ‡Å or CU. The introduction of herbs or plants and fruits both connect this rite to the food that is breath and to the food that is offered in the mouth of the man in the previous sections, and anticipates the next section (BU 6.4.1) in which the author presents the sequence of plants, flowers, fruits, man, and semen. Udumbara is connected with vitality, food, and strength, especially in the ‡B.15 Udumbara is said to be the 15 “It is of Udumbara wood, for him to obtain food and strength –the Udumbara means food and strength: therefore it is of Udumbara wood.” ‡B 3.2.1.33 (Eggelin’s tr.). This type of statement is frequent in the ‡B: 3.3.4.27; 4.6.9.22; 5.4.3.25-26; 7.4.1.38; 7.5.1.15; 9.2.2.3, etc.

22

LANGUAGE, TEXTS AND SOCIETY

sap (rasa) and to represent all trees.16 The connection between rasa and semen is common in the vedic literature,17 and Udumbara points to fecundity and fertility. The crushing and squeezing of the herbs in the mantha rite recalls the crushing of Soma and its sexual symbolism. The herbal juice is mixed with ghee (another symbol of semen) by offering a portion of a spoonful of ghee in the fire and pouring the remainder into the juice. Finally, the mixture is sipped while reciting the Såvitrœ verse. We have here a nice parallel between the offering of ghee in the sacrificial fire and the offering of the juice mixture in the mouth, as in the fourth fire of the preceding section. The fertility aspect of the mixture is highlighted by the saying attributed to a series of teachers and pupils: “Even if one were to pour this mixture on a withered stump, it would sprout new branches and grow new leaves.” As we saw, the author of the BU has moved this statement from the end of the contest to the end of the mantha rite. The wording is also changed from “saying this to a withered stump” of the other versions to “pouring the mixture on a withered stump.” Knowledge and saying are replaced by a rite, and the fertility aspect of the mixture is highlighted. Finally, the BU inserts this concluding statement: “There are four things made of Udumbara wood: Udumbara spoon, Udumbara cup, Udumbara kindling stick, and the two Udumbara stirring sticks. There are ten types of cultivated grains: rice, barley, sesame, bean, millet, mustard, wheat, lentil, pea, and legume. After grinding these, he pours curd, honey, and ghee on them, and offers an oblation of ghee.” The rite intended here is unclear; is it an allusion to a new rite or a summation of the rite just concluded? Are the ten types of grain a gloss on “every type of herb”? In any case, the mention of grain is a good opening to the next section (6.4.1 cited earlier) that presents semen as the essence of plants/flowers/fruits. After the statement about sex as a Soma sacrifice (BU 6.4.1-3), the chapter concludes with a series of six rites, all connected with sex: rite when one spills semen (BU 6.4.4-5);18 rite at seeing one’s reflection in 19 water (BU 6.4.6); rites for intercourse with women (6.4.6-11); rite 16 See ‡B 6.7.1.13. “Then they together lay hold of an Udumbara (branch) saying, ‘Sap and strength I lay hold of ‘. The Udumbara is strength and food. In that the gods distributed sap and strength, then the Udumbara came into being. Therefore thrice a year it ripens.” AB 5.24 (tr. Keith, modified). 17 See Olivelle 1997: 166; see below p.91. 18 This rite is naturally connected with the statement at 6.4.2 that Prajåpati created the woman to be the proper receptacle for semen. Depositing semen anywhere else, either through masturbation or emission in sleep, is viewed as a depletion of one’s virility which has to be ritually recaptured. 19 This rite is out of place in this series. The text reads: atha yady udake åtmåna∫ pa†yet tad abhimantrayeta mayi teja indriya∫ ya†o draviña∫ sukr¢tam iti (“If, moreover, he sees himself in water, let him address it thus: ‘May vigor, virility, fame, wealth, and merit remain in me’”). I wonder whether åtmånam here stands for retas, in which case this rite concerns ejaculating semen in water. This equation is not unprecedented. At AU 2.1 we read: puruße

Young ‡vetaketu

23

against a wife’s lover (BU 6.4.12); rite during intercourse with one’s wife (6.4.13-22); and rites at pregnancy and birth (BU 6.4.23-28).20 An interesting sub-text running through these rites is the fear of losing virility and merit by engaging in sexual activity. Women are said to appropriate to themselves the merits of a man who engages in sex without knowing its nature as a Soma sacrifice (BU 6.4.3). And Uddålaka Åruñi is said to have exclaimed: “Many are the mortals of Brahmin descent who, engaging in sexual intercourse without this knowledge, depart this world drained of virility and deprived of merit” (BU 6.4.4). The theology of sex as sacrifice is intended to safeguard against the dangers of sex, an ancient Indian way of practicing “safe sex”. Finally, there is the repeated mention of Uddålaka Åruñi. In the mantha rite the author of BU places Uddålaka at the head of a series of teachers and pupils who repeated the saying about the potency of the mixture: Yåj∞avalkya,21 Madhuka Paiõgya, Cüla Bhågavitti, Jånaki Åyasthüña, and Satyakåma Jåbåla. The ‡Å and CU mention only the last. Again the statement about many Brahmins departing drained of their virility (BU 6.4.3) is ascribed to Uddålaka. The same Uddålaka is the Brahmin whom the author presents earlier as having received the knowledge of the five fires from Pravåhaña Jaivali. Although the evidence is not compelling, I wonder whether the author is, on the one hand, interpreting the five fires as a theology of sexual intercourse as a sacrifice, a theology that was known at first only to Kßatriyas, and, on the other, presenting Uddålaka as the first Brahmin to learn this secret and to teach it to other Brahmins. If this is right, then we can see the “logic” of the author of BU in his rearrangement of the early sections of this chapter. ha vå ayam ådito garbho bhavati yad etad retaΔ Ù tad etat sarvebhyo ‘õgebhyas tejaΔ sa∫bhütam åtmany evåtmåna∫ bibharti (“At the outset, this embryo comes into being within a man as semen. This radiance gathered from all the bodily parts he bears in himself as himself”). Here one’s semen is viewed as one’s self that a man deposits in a woman. The placing of the semen in the woman in sexual intercourse is taken as a man’s first birth, while the birth of the son is his second birth. For an examination of this passage and the related statement at Taittirœya Upanißad 2.7, see Olivelle 1997: 165; see below p.91. 20 I think the author here intends to distinguish sex with one’s wife from sex with other women. The latter is dealt with in the third set of rites (BU 6.4.6-11), where the question is how to deal with a woman who refuses to have sex (including bribing and beating), how to ensure that the woman loves you, and how to make sure that she does not become pregnant or does become pregnant. The former deals with sex with one’s own wife and includes rites to ensure different types of sons and daughters (on this, see Wezler 1993), rite of intercourse, rites at delivery, and birth rites. 21 It is significant that in the sixth chapter of the BU Uddålaka is presented both here and in the genealogy (BU 6.5.3) as the teacher of Yåj∞avalkya, who looms large as the teacher par excellence in the earlier chapters. In the third chapter Uddålaka (BU 3.7) is among a series of prominent theologians that Yåj∞avalkya defeats in a series of debates. Uddålaka Åruñi appears in the genealogy of ‡Å 15 and in all likelihood belonged to a R¢gvedic †åkhå, whereas Yåj∞avalkya is credited with the composition of the White Yajurveda (BU 6.5.3).

24

LANGUAGE, TEXTS AND SOCIETY

2.1.2 Chåndogya Upanißad In Chapter 5 the author of CU pursues, I believe, a theology of the fire sacrifice as an offering to one’s breaths (pråñågnihotra). The arrangement of material of sections I and II, we saw, is identical in ‡Å and CU, an arrangement that is probably original to I-II*. The author of CU had a much easier time than his BU counterpart and did not have to recast this section because it fitted nicely into his literary structure. It starts with the assertion of the supremacy of breath over all other vital functions. He does not introduce the sixth function, semen; indeed, he is quite happy with the number five, both here and in the five fires. It permits him to lead naturally to the grand finale in CU 5.19-23 where mouthfuls of food are offered to the five breaths: out-breath, inter-breath, in-breath, linkbreath, and up-breath (pråña, vyåna, apåna, samåna, udåna). In the section on the contest (I), the major new element is the addition that follows immediately after the faculties offer their own powers to breath (CU 5.1.13-14): Surely, people do not call these “speeches”, or “sights”, or “hearings”, or “minds”. They call them only “breaths”, for only breath becomes all these. na vai våco na cakßü∫ßi na †rotråñi na manå∫sœty åcakßate Ù pråñå ity evåcakßate Ù pråño hy evaitåni sarvåñi bhavanti Ù (CU 5.1.15; cf. BU 1.5.21).

Alluding to the common vedic practice of calling all vital faculties pråña, the author asserts the absolute supremacy of breath. The section on the contest ends with a saying ascribed to Satyakåma Jåbåla: “Even if one were to say this to a withered stump, it would sprout new branches and grow new leaves” (CU 5.2.3). The antecedent of enat “this” is unclear, but the power of the saying is undoubtedly related to the acknowledgement of the supremacy of the breath over the other faculties. The mantha rite is brief both in ‡Å and in CU. This rite is also the weakest point in my argument for taking the pråñågnihotra as providing an overarching structure to the CU narrative. The author of the CU has not adapted the mantha narrative to further his literary-cum-theological purpose; perhaps he did not see the need for such adaptation because the rite itself shows the power of making offerings to the five faculties that had earlier been identified with breath. The connection between the contest and the mantha rite, on the one hand, and the internal fire offering consisting of eating, on the other, is established also in the ‡Å where the latter (‡Å 10) immediately follows the former (‡Å 9). Unlike in the CU, however, the offerings are made here not to the five breaths but to the six faculties (with the addition of semen), relating the offerings directly to the contest between and the mantha offerings to these faculties.

Young ‡vetaketu

25

The next two sections of the CU, containing two sets of instructions by two Kßatriyas, Jaivali (CU 5.3-10) and A†vapati (5.11-24), are interesting in that they present these teachings as Kßatriya secrets unknown to Brahmins. There is no dispute that the latter contains a clear enunciation of the theology of pråñågnihotra.22 The former does not teach this doctrine directly, but I think that the author is using the doctrine of the five fires to set the scene for the doctrine of offering food to the five breaths. Besides the obvious refrain of the number five –five faculties, five fires, five breaths–, the central element of the five offerings is the offering of food in the mouth of the man (fourth fire). This is clearly not the same as pråñågnihotra, but the author, I think, is drawing a parallel between this and the offering to breaths in the concluding section.23 Both involve putting food in the mouth, and in both the mouth is the sacrificial fire; the CU (5.18.2) explicitly equates the mouth with the åhavanœya fire, in which oblations to gods are offered. One must realize that the pråñågnihotra is not an offering in breaths conceived of as fires, although the breaths are often homologized with fires, but the oblations to the breaths (conceived of as the deities to whom the oblations are intended) offered in the fire of the mouth. This is apparent in the mantras used at these offerings: “To outbreath, svåhå!” etc. The author is here drawing a parallel between the fire doctrine and the pråñågnihotra, without equating the one with the other. Such parallelisms, sometimes based on much slimmer connections such as phonetic similarity of words (for example, ana-breath and anna-food that we encountered earlier), abound in the Upanißads (Olivelle 1996a: liii-liv). The author appears to be drawing the attention of the reader to this parallelism in the concluding statements of the two sections. He is the only one to propose a rider to the two path model, making moral conduct a factor in the after-death condition of a man: Now, people here whose behavior is pleasant can expect to enter a pleasant womb, like that of a woman of the Brahmin, the Kßatriya, or the Vai†ya class. But people of foul behavior can expect to enter a foul womb, like that of a dog, a pig, or an outcaste woman. (CU 5.10.7)

He cites a verse on the five great sins that causes a man to fall, the last of which is association with a person who commits such a sin (CU 5.10.9). The concluding statement of the section states: A man who knows these five fires in this way, however, is not tainted with evil even if he associates with such people. Anyone who knows this becomes pure and clean and attains a good world. (CU 5.10.10) 22 For an examination of this passage and parallels in other vedic texts, see Bodewitz 1973: 263-69. 23 In the ‡B (11.6.2.6-10) the doctrine of the five fires are actually introduced as a secret teaching (the secret essence) of the fire sacrifice (agnihotra).

26

LANGUAGE, TEXTS AND SOCIETY

At the conclusion of A†vapati’s discourse on the offerings to breaths, the author likewise picks up the theme of immunity from sin and stain in the case of a person who performs those offerings: When someone offers the daily sacrifice with this knowledge, all the bad things in him are burnt up like the tip of a reed stuck into a fire. Therefore, even if a man who has this knowledge were to give his leftovers to an outcaste, thereby he would have made an offering in that self of his which is common to all men. (CU 5.24.3-4)

3. The Story of ‡vetaketu I now turn to the ‡vetaketu story and present below a concordance of parallel passages in the three versions, divided for convenience into narrative units. I have separated each unit of the narrative sequence and numbered them sequentially. Söhnen (1981: 179) has conveniently divided the story into three narrative components contained in all three versions: A) dialogue between ‡vetaketu and Jaivali or Citra; B) dialogue between ‡vetaketu and his father, Uddålaka; and C) dialogue between Uddålaka and Jaivali or Citra. There is a clear structure to these three units, each opening with the arrival of a person into the presence of another: ‡vetaketu to Jaivali or Citra; ‡vetaketu to Uddålaka; and Uddålaka to Jaivali or Citra.

Br¢hadårañyaka (B*)

Chåndogya (C*)

Kaußœtaki (K*)

A) Dialogue between ‡vetaketu and Jaivali (Citra) 1 †vetaketur ha vå åruñeyaΔ pa∞cålånå∫parißadam åjagåma ‡vetaketu, the son of Åruñi, came to the audience hall of the Pa∞cålas.

2 sa åjagåma jaivali∫ pravåhaña∫ paricårayamåñam He came to Jaivali Pravåhaña while he was being waited upon.

citro ha vai gåõgyåyanir yakßyamåña åruñi∫ vavre Ù sa ha putra∫ †vetaketu∫ prajighåya yåjayeti Citra Gåõgyåyani, when he was preparing to perform a sacrifice, chose Åruñi. He [Åruñi] sent his son, ‡vetaketu, telling him: “Perform the sacrifice.”

27

Young ‡vetaketu

Br¢hadårañyaka (B*)

Chåndogya (C*)

Kaußœtaki (K*)

3 tam udœkßyåbhyuvåda kumårå3 iti Seeing him, he [Jaivali] greeted him: “Young man!”

ta∫ ha pravåhaño jaivalir uvåca kumåra Pravåhaña Jaivali said to him: “Young man,

ta∫ håsœna∫ papraccha gautamasya putra When he [‡vetaketu] was seated, he [Citra] questioned him: “Son of Gautama,

4 sa bho3 iti prati†u†råva He replied: “Yea!” 5 anu†iß™o nv asi pitreti anu två†ißat piteti “Have you been educated did your father educate by your father?” you?” 6 om iti hovåca He said: “Yes.”

anu hi bhagava iti “He did, indeed, your honor.”

7 vettha yathemåΔ prajåΔ prayatyo vipratipadyantå3 iti “Do you know how these creatures, when they depart, go in different ways?”

vettha yad ito ‘dhi prajåΔ asti sa∫vr¢ta∫ loke yasmin må prayantœti dhåsyasy anyatamo vådhvå “Do you know to where tasya måloke dhåsyasœti creatures depart from is there a closed door in the world in which you will here?” place me, or does it have another road—so you won’t place me in a false world?”

8 neti hovåca He said: “No.”

na bhagava iti “No, your honor.”

9 vettho yathema∫ loka∫ punar åpadyantå3 iti “Do you know, then, how they return again to this world?”

vettha yathå punar åvartanta3 iti “Do you know how they return again?”

10 neti haivovåca He just said: “No.”

na bhagava iti “No, your honor.”

11 vettho yathåsau loka eva∫ bahubhiΔ punaΔ punaΔ prayadbhir na sa∫püryatå3 iti “Do you know, then, how the world over there is not filled up with the great many people who continuously depart in this manner?”

vettha pathor devayånasya pitr¢yåñasya ca vyåvartanå3 iti “Do you know how the path to the gods and to the fathers take different turns?”

12 neti haivovåca He just said: “No.”

8 na bhagava iti “No, your honor.”

sa hovåca nåham etad veda He [‡vetaketu] said: “I do not know it”.

28

LANGUAGE, TEXTS AND SOCIETY

Br¢hadårañyaka (B*)

Chåndogya (C*)

13 vettho yatithyåm åhutyå∫ hutåyåm åpaΔ purußa våco bhütvå samutthåya vadantœ3 iti “Do you know, then, which oblation it is at whose offering the water, taking on a human voice, rises up and speaks?”

vettha yathåsau loko na sa∫püryata3 iti “Do you know how the world over there is not filled up?”

14 neti haivovåca He just said: “No.”

na bhagava iti “No, your honor.”

15 vettho devayånasya vå vettha yathå pa∞camyåm pathaΔ pratipada∫ pitr¢yåña åhutåv åpaΔ purußavacaso sya vå yat kr¢två devayåna∫ bhavantœti vå panthåna∫ pratipadyante “Do you know how at the pitr¢yåña∫ vå Ù api hi na r¢ßer fifth offering the water vacaΔ †rutam Ù dve sr¢tœ takes on a human voice?” a†r¢ñava∫ pit≤ñåm aha∫ devånåm uta martyånåm Ù tåbhyåm ida∫ vi†vam ejat sameti yadantarå pitara∫ måtara∫ ceti “Do you know, then, the access to the path to the gods or the path to the fathers; that, when done, they get on the path to the gods or on the path to the fathers? For have you not heard the seer’s words? ‘Two paths mortals have, I have heard: to fathers and to gods. By these travel all that live between the father [heaven] and the mother [earth].’” 16 nåham ata eka∫cana vede- naiva bhagava iti “Not at all, your honor.” ti hovåca He said: “I don’t know even one of these.” 17 athaina∫ vasatyopamantrayå∫ cakre Then he [Jaivali] invited him to stay.

athånu kim anu †iß™o ‘vocathå yo hœmåni na vidyåt Ù katha∫ so ‘nu†iß™o bruvœteti “Then, why did you say that you had been educated? How can a man who does not know these call himself educated?”

Kaußœtaki (K*)

29

Young ‡vetaketu

Br¢hadårañyaka (B*)

Chåndogya (C*)

18 anådr¢tya vasati∫ kumåraΔ pradudråva Spurning (the invitation) to stay, the young man ran off.

Kaußœtaki (K*) hantåcårya∫ pr¢cchånœti But let me ask my teacher.”

B) Dialogue between ‡vetaketu and Uddålaka 19 sa åjagåma pitaram sa håyastaΔ pitur ardham sa ha pitaram åsådya He came back to his father. eyåya He approached his father Crestfallen, he came back to his father’s place. 20 ta∫ hocåva iti våva kila no bhavån purånu†iß™ån avoca iti He said to him: “Thus indeed, I dare say, did you once announce that we were educated!”

ta∫ hovåcånanu†ißya våva kila må bhagavån abravœd anu två†ißam iti He said to him: “Without actually teaching me, I dare say, your honor told me ‘I have taught you.’”

21 katha∫ sumedha iti “What’s the matter, my clever boy?” 22 pa∞ca må pra†nån råjanyabandhur apråkßœt tato naika∫cana vedeti “That excuse for a prince asked me five questions. I didn’t know even one of them.”

pa∞ca må råjanyabandhuΔ pra†nån apråkßœt teßå∫ naika∫canå†aka∫ vivakum iti “That excuse for a prince asked me five questions. I could not explain even one of them.”

papracchetœti måpråkßœt katha∫ pratibravåñœti and asked him: “He asked me this. How should I answer him?”

sa hovåca yathå må tva∫ tadaitån avado yathåham eßå∫ naika∫cana veda Ù yady aham imån avedißya∫ katha∫ te nåvakßyam iti He said: “As you report them to me, I do not know even one of them. If I had known them, how could I have not told them to you?”

sa hovåcåham apy etan na veda He [father] said: “Even I do not know this.

23 katame ta iti “What were they?” 24 ima iti ha pratœkåny udåjahåra “These,” he said and quoted the opening lines. 25 sa hovåca tathå nas tva∫ tåta jånœthå yathå yad aha∫ ki∫ca veda sarvam aha∫ tat tubhyam avocam He [father] said: “Whatever I know, all that I have taught you; that’s how, son, you should know me.

30

LANGUAGE, TEXTS AND SOCIETY

Br¢hadårañyaka (B*)

Chåndogya (C*)

26 prehi tu tatra pratœtya brahmacarya∫ vatsyåva iti “But, come; we shall both go there and live as students.”

sadasy eva vaya∫ svådhyåyam adhœtya haråmahe yan naΔ pare dadati Ù ehy ubhau gamißyåva iti Within the very sacrificial arena let us, after we have performed our vedic recitation, gather what others may give us. Come, we shall both go.”

Kaußœtaki (K*)

27 bhavån eva gacchatv iti “You, sir, can go on your own.”

C) Dialogue between Uddålaka and Jaivali (Citra) 28 sa åjagåma gautamo yatra pravåhañasya jaivaler åsa That man, Gautama, came to where Pravåhaña Jaivali was.

sa ha gautamo råj∞o ‘rdham eyåya That man, Gautama, came to the king’s place.

29 tasmå åsanam åhr¢tyodakam åhårayå∫cakåra Ù atha håsmå arghya∫ cakåra After bringing him a seat, he [Jaivali] had some water brought for him. Then he offered him the arghya water.

tasmai ha pråptåyårhå∫ cakåra When he arrived, he [the king] paid him homage.

30 sa ha pråtaΔ sabhåga udeyåya In the morning, when he was in the assembly hall, he got up. 31 ta∫ hovåca vara∫ bhagavate gautamåya dadma iti He said to him: “We will grant a wish to the Honorable Gautama.”

ta∫ hovåca månußasya bhagavan gautama vittasya vara∫ vr¢ñœthå iti To him he said: “Honorable Gautama, choose a wish among human riches.”

sa ha samitpåñi† citra∫ gåõgyåyani∫ praticakrama Carrying firewood in his hand, he [Gautama] went up to Citra Gåõgyåyani,

31

Young ‡vetaketu

Br¢hadårañyaka (B*)

Chåndogya (C*)

32 sa hovåca pratij∞åto ma eßa varaΔ Ù yå∫ tu kumårasyånte våcam abhåßathås tå∫ me brühœti He said: “You have promised me this wish; explain to me the words that you spoke before the young man.”

sa hovåca tavaiva råjan månußa∫ vittam Ù yåm eva kumårasyånte våcam abhåßathås tåm eva me brühœti He said: “Keep your human riches to yourself, king. Explain to me the very words that you spoke before the young man.”

Kaußœtaki (K*)

33 sa hovåca daiveßu vai gau- sa ha kr¢cchrœ babhüva tama tad vareßu månußåñå∫ He [Uddålaka?] became brühœti distressed. He said: “That, Gautama, is in the category of divine wishes. Mention (one) from among human (wishes).” 34 sa hovåca vij∞åyate håsti hirañyasyåpåtta∫ goa†vånå∫ dåsœnå∫ pravåråñå∫ paridhånasya Ù må no bhavån bahor anantasyåparyantasyåbhyavadånyo bhüd iti He said: “It is well known that I have my share of gold, cows, horses, slave girls, blankets, and clothes. Do not be stingy, sir, (in giving me) more, (in giving me) the infinite and the boundless.” 35 sa vai gautama tœrthenecchåså iti “Then, Gautama, you will have to request it in the correct manner.” 36 upaimy aha∫ bhavantam iti “I come to you, sir, as a pupil.” 37 våcå ha smaiva pürva upayanti With words alone did the people of old come as pupils.

ta∫ ha cira∫ vasety åj∞åpayå∫ cakåra He commanded him: “Stay longer.”

upåyånœti (and said): “Let me come to you as a pupil.”

32

LANGUAGE, TEXTS AND SOCIETY

Br¢hadårañyaka (B*)

Chåndogya (C*)

Kaußœtaki (K*)

ta∫ hovåca yathå må tva∫ gautamåvadaΔ yatheya∫ na pråk tvattaΔ purå vidyå bråhmañån gacchati Ù tasmåd u sarveßu lokeßu kßatrasyaiva pra†åsanam abhüd iti Ù tasmai hovåca He said to him: “As you have told me, Gautama, before you this knowledge has not reached Brahmins in the past. In all the worlds, therefore, government has belonged exclusively to royalty.” He [king] said to him.

ta∫ hovåca brahmårgho ‘si gautama yo na månam upågåΔ Ù ehi vy eva två j∞apayißyåmœti He [Citra] said to him: “You have proved yourself worthy of the formulation of truth, Gautama, you who have not succumbed to pride. Come, I’ll make you perceive it clearly.”

38 sa hopåyanakœrtyovåsa He lived (there) recognized as one who has come as a pupil. 39 sa hovåca tathå nas tva∫ gautama måparådhås tava ca pitåmahå yatheya∫ vidyetaΔ pürva∫ na kasmi∫†cana bråhmaña uvåsa Ù tå∫ tv aha∫ tubhya∫ vakßyåmi Ù ko hi tvaiva∫ bruvantam arhati pratyåkhyåtum iti He [Jaivali] said: “As before now this knowledge has not resided in any Brahmin, so may you, Gautama, or your grandfathers not cause us harm. But I will tell it to you, for who can refuse you when you speak like that.”

3.1 ‡vetataketu in B* In this first section of my analysis I focus on B* because the author through a finely nuanced narrative has put into sharp focus the character traits of the three individuals: the impolite, ignorant, and arrogant ‡vetaketu, the fatherly and magnanimous Pravåhaña Jaivali, and the loving, patient, and humble Uddålaka Åruñi. Not all aspects of my analysis of the narrative dynamic and the author’s use of the language are equally compelling; some are speculative. But together they reveal the author’s clear and deliberate literary strategy to contrast the arrogance of ‡vetaketu with the fatherly solicitude of Jaivali. A) ‡vetaketu and Jaivali. ‡vetaketu comes24 into the audience hall25 of the Pa∞cålas (1).26 Unlike K*, neither B* nor the parallel in 24 B* always uses the verb å-√gam (see 1, 2, 19, 28; also gacchatu at 27), while C* always uses å-√i (see 1, 19, 28). I am not sure what to make of these different choices. The CU (5.1.711) appears to prefer the compounds of √i also in the passage on the contest between faculties (where BU [6.1.7-12] always uses compounds of √gam) and at CU 6.1.2. But in the A†vapati story CU uses √gam (5.11.2, 4). 25 B* uses the term parißad, while C* has samiti and K* sadas. I have not been able to ascertain a reason for their choice of different words or whether they reflect regional differences. The term parißad, however, acquired a technical meaning in the legal literature, where it refers to a conclave of normally 10 Brahmins that would decide points of law and prescribe penances (Gautama Dharmasütra, 9.49; Baudhåyana Dharmasütra, 1.1.7-8). ‡a∫kara (on BU 6.3.1) interprets this term in its technical sense (see note 27). 26 My references to B*, C*, and K* are to numbers given in the above concordance of the narrative units.

Young ‡vetaketu

33

C* gives an explicit reason for the visit.27 The narrative sequence of B* leading up to Jaivali’s question as to whether ‡vetaketu has been educated by his father is absent in C* and provides an insight into the literary strategy of B*. In C* Jaivali is not directly introduced (we must assume that he was present in the assembly and that it was to visit him that ‡vetaketu came there) and questions the young man abruptly, even haughtily. In B*, on the other hand, the questioning is preceded by three narrative units: ‡vetaketu comes up to Pravåhaña Jaivali28 while the latter was being waited upon (2); Jaivali notices him and greets him (3); ‡vetaketu returns the greeting (4). In B* ‡vetaketu not only enters the audience hall but goes directly up to Jaivali, and he does so while Jaivali is “being waited upon” (paricårayamåñam; 2 ). This grammatical form, the middle present participle of the causative of pari- √car, is not found elsewhere in the vedic literature. The non-causative forms of the verb are used regularly especially with reference to the service of, that is, putting firewood into the ritual fire,29 which is equivalent at the human level to serving food. A clearly sexual meaning is attached to the term in the only other occurrence of a causative form. In the Ka™ha Upanißad (1.1.25), Death promises Naciketas lovely girls of a sort unobtainable by men: “I’ll give them to you; you’ll have them wait on you (paricårayasva).” At CU 4.4.4, moreover, the mother of Satyakåma Jåbåla tells the boy that she had him when she was a maid and had many relationships: bahv aha∫ carantœ paricåriñœ, where the latter term assumes a sexual connotation at least by association. What the author of B* is trying to signal here, I think, is that ‡vetaketu did not know his manners and barged into the presence of Jaivali during an inappropriate moment, either because he was being entertained by women or because he was being served his meal, or both. Jaivali notices him (udœkßya), but instead of having him thrown out, the king in a fatherly and respectful manner greets (abhyuvåda) him, saying “Young man” (kumåra; 3). The term abhyuvåda can connote respect and/or affection,30 as does kumåra, the term also for the son of ‡a∫kara, commenting on the BU, says that ‡vetaketu out of arrogance came with the intention of defeating the parißad of the Pa∞cålas, as well as the parißad of the king. He is, however, silent on this point in his commentary on the parallel passage of the CU. 28 B* calls him Jaivali Pravåhaña, and C* Pravåhaña Jaivali. No special significance, I think, can be ascribed to this difference; B* calls him Pravåhaña Jaivali at 28. 29 Såmavidhåna Bråhmaña 3.6.2; Gopatha Bråhmaña, 1.2.3, 7; CU 4.10.1,2,4; TÅ 1.32.1. The term is used with reference to bodily and cosmic powers (conceived of as children) serving some other power (regarded their parents) in AÅ 2.1.7. The terms paricaritå and paricaran at CU 7.8.1 also probably refer to a student’s duty to serve the fire or the teacher. At CU 8.8.4 Virocana, satisfied with the partial definition of the Self (åtman), tells the other demons that it is the body (åtman) that should be extolled (mahayya) and cared for (paricarya). See also TS 6.1.11.6. 30 Respect is indicated, for example, at CU 4.1.8; 4.2.1; 4.2.14. Frequently, however, the term is used merely with reference to one person talking to or greeting another: BU 2.4.14; 3.2.3; 3.8.8; 4.5.15; CU 4.5.1; 4.6.2; 4.7.2; 4.8.2; KeU 3.4; PU 6.1; 4.2; 2.2. Sometime 27

34

LANGUAGE, TEXTS AND SOCIETY

a king. ‡vetaketu responds to this greeting with a simple bhoΔ (4). This term may not necessarily indicate disrespect,31 but two factors make me think that the author is once again signaling the incivility of the young man. First, it comes after the discourteous intrusion into Jaivali’s private space. Second, I think that in ancient India, as in modern America or Europe, good upbringing required a younger person to use a “sir” or “madam” equivalent in addressing an older and respectable person. You don’t simply say “yes” or “no” to a superior. It is remarkable that throughout B* (4, 6, 8, 10, 12, 14, 16) ‡vetaketu replies to questions with a curt “yes” or “no”, whereas in the parallel passages of C* he uses the honorific bhagavaΔ (“lord” or “sir”). The term bhoΔ here may, therefore, mean something like the colloquial “Hey!”32 This brief exchange sets the scene for Jaivali’s opening question (5). Comparing the wording of this question in B* and C*, Söhnen (1981, 187) notes the older anu två†ißat (aorist, in tmesis) of C* in contrast to the younger past participial construction anu†iß™o nv asi of B*. I doubt that these constructions warrant her conclusion that C* is older than B*; aorist forms are indeed found elsewhere in the B* narrative: avocaΔ (20),33 apråkßœt (22). In my view, the author’s use of anu†iß™a is deliberate; it evokes in the listener’s mind the related word †iß™a, which means not just a learned man, but a man who is a paragon of deep learning, correct speech, good behavior, and proper etiquette.34 On a listener who probably knew some version of the episode already these subtle points would not have been lost: Jaivali is posing for young ‡vetaketu a double-entendre and putting a doublethis term is used when a teacher calls a pupil (CU 4.1.2; 4.9.1; 4.14.1), or when a father greets a son (KaU 1.10), where affection is probably implied. 31 ‡a∫kara himself notes that bhoΔ was an inappropriate form of address uttered in anger: bho3 ity apratirüpam api kßatriya∫ pratyuktavån kruddhaΔ san. The meaning of apratirüpa is not altogether clear. Ånandagiri is off the mark, I think, when he explains that bhoΔ is said to a teacher and not to a Kßatriya, because the latter is lower in status: bho3 iti prativacanam åcårya∫ praty ucita∫ na kßatriya∫ prati tasya hœnatvåt. Why whould a man in anger respond with a greeting of reverence? The Mahåbhårata (3.186.33), however, appears to indicate that bhoΔ was used in a disrespectful manner, as opposed to the obsequious årya. In describing the social upheavals in the Kaliyuga, it says: bhovådinas tathå †üdrå bråhmañå† cåryavådinaΔ, which van Buitenen translates: “The serfs (=Südras) will say ‘Hey you!’, the brahmins will say ‘Pray, sir!’” I think van Buitenen has captured well the subtle nuances of bhoΔ (Hey you!) and årya (Pray, sir!). 32 I am not sure whether ‡vetaketu’s use of bhoΔ/bhavån elsewhere in B* (20, 27) in place of bhagavaΔ/bhagavån (6, 8, 10, 12, 14, 16, 20) of C* is intended to make a similar point. Although the latter is more respectful, the former normally does not carry disrespectful overtones. Söhnen (1981: 198), in presenting the parallel versions, has ignored the difference in B* and C* with respect to the honorific title. 33 Söhnen (1981: 188) reads avocad, but this reading, though grammatically “correct”, has no basis in either the Kåñva or the Mådhyandina recension. The anomaly of a third person subject (bhavån) with the second person verb was already noted by Whitney (1890: 417). 34 In the Taittirœya Upanißad (1.11) anu†åsati is used with reference to the teacher’s final admonition to a student about how he should behave after he returns home and clearly refers to points of good behavior and etiquette. On the practice of the †iß™as as the standard for both the correct use of language and the correct modes of behavior, see M. Deshpande,

Young ‡vetaketu

35

edged question: did your father impart to you learning and did he train you in basic norms of etiquette and good behavior? The subtle irony of the question is, as expected, lost on the brash ‡vetaketu, who replies with a curt “yes” (o∫; 6), again without any honorific title. Parpola (1981) has shown the widespread use of o∫ as a particle of assent even outside the ritual context. I am not sure whether in normal usage o∫ was used to say “yes” by an inferior to a superior.35 The use of o∫, however, is quite rare in conversations, in contrast to the ubiquitous tathå (“OK”). Given its rarity, its usage was possibly “marked” and carried a particular connotation.36 In any case, this curt answer stands in sharp contrast to the obsequious anu hi bhagavaΔ of C*. Jaivali then asks ‡vetaketu five questions, all beginning with vettha (“do you know”). Starting with the second question, Jaivali adds the emphatic u (vettho),37 which may convey something like “do you at least know”. To each question ‡vetaketu answers with a curt “no”. Paralleling the emphatic u of the questions, answers two to four have the additional eva (haivovåca), possibly conveying something like “just” or “simply”. The final answer is longer, but still without an honorific title. With each impolite answer, the author instills in his listeners the image of ‡vetaketu as “not educated”; he is neither an anu†iß™a nor a †iß™a. The sequel to this exchange is significant. In contrast to the cutting words of Jaivali in C*, the author of B* presents the king as solicitous of ‡vetaketu’s welfare and inviting the young man to stay (17). “Staying” (vasati) here probably refers to a student’s residence with a teacher.38 Jaivali in effect tells him to stay so he can teach him –teach him the answers to the questions and proper manners. But the haughty young man spurns the kind invitation, the term anådr¢tya again suggesting lack of politeness and humility. And he runs away. B) ‡vetaketu and Uddålaka. The section opens with ‡vetaketu coming back to his father (19) and blurting out this rather testy and sarcastic accusation: iti våva kila no bhavån purånu†iß™ån avocaΔ (20). The expression våva kila39 is found in both B* and C*. The iti at the begin“Historical Change and the Theology of Eternal (Nitya) Sanskrit,” in his Sanskrit & Prakrit: Sociolinguistic Issues (Delhi: Motilal Banarsidass, 1993), 53-74. 35 I have not been able to research this point. One of its rare uses with this meaning in the Upanißads occurs at BU 3.9.1 (and in the parallel at ‡B 11.6.3.4) where Såkalya says “o∫” to Yåj∞avalkya’s answers, but in a dismissive way because the answers are only superficially true, and he goes on to repeat the same question. In my own native Sinhala, however, the parallel term oÓw is used mostly among intimates and friends, while other more respectful terms, such as eÓheÓy, are used when addressing superiors. 36 I am grateful to Stephanie Jamison for these observations. 37 This particle, however, is lacking in the Mådhyandina version. 38 ‡a∫kara, however, interprets the word to mean that the king was going to perform the customary hospitality rites such as giving water to wash the feet. 39 For the meaning of kila as referring to something the listener should know or is

36

LANGUAGE, TEXTS AND SOCIETY

ning of B* stands in contrast to the explicit, but prosaic, ananu†ißya (“without teaching me”) of C* and, I think, means something like “this is the way”, referring perhaps to the father’s former boast (to his friends and family?) about how learned his son was. ‡vetaketu uses the same word that Jaivali had used earlier (5), anu†iß™a, and in using it reveals that he had not grasped the full implication of the term. The listener by now knows what ‡vetaketu should have known; he clearly is no anu†iß™a. As in the question of Jaivali, so here C* uses the verbal forms that do not have quite the resonance of the past participle. Söhnen (1981: 189), mistakenly I believe, thinks that the use of the plural (naΔ and anu†iß™ån) in B* indicates that ‡vetaketu is here speaking also on behalf of his classmates. This is in all likelihood a majestic plural, and the author uses it possibly to signal the arrogance of ‡vetaketu in using such a pompous form especially in talking to his father.40 The father is baffled by this outburst and cannot quite follow the point. He tries to soothe the angry young man, calling him sumedha (literally, “[a man] with a fine understanding”!). The irony in this choice of words will not be lost on the listener. This repartee is absent in C*, which goes directly from ‡vetaketu’s initial accusation to his report about the five questions he could not answer. The report is almost identical in both versions, except for the final veda, “I did (not) know”, in B*,41 compared with a†aka∫ vivaktum, “I could (not) explain” in C* (20). In both texts ‡vetaketu uses what ‡a∫kara42 and Söhnen (1981: 188) have correctly recognized as a disparaging epithet, råjanyabandhu, to refer to Jaivali. In C*, however, it is a reflection of his justifiable anger, whereas in B* it is a reflection of his arrogance. The version of B* continues with the father asking what the questions were and ‡vetaketu enumerating them briefly (23-24). Uddålaka’s answer differs substantially in the two versions (Söhnen 1981, 189-90). In B* the father does not address directly the issue of the questions his son had failed to answer; his reply is directed at ‡vetaketu’s implied accusation that his father had withheld information from him. The father, in effect, says that he is not the type of man to withhold information from his own son. The author of B* paints the picture of a father deeply hurt by his son’s cutting words and unfair accusations, further strengthening the listeners’ perception of ‡vetaketu as not only haughty and impolite, but also without feeling or filial piety. After that B* has a narrative unit (26-27) absent in C* and only partially found in K*. In both B* and K* the father asks the son to come along to visit the king and receive instruction from him. B* uses generally known, see Emenau 1968-69; Daalen 1988. 40 It is possible that the use of the plural in B* is merely stylistic, because this author uses the same expression with reference to the father at 25, where C* uses the singular. 41 This word is used in C* when the father tells the son that he too does not know the answer to these questions (25). 42 paribhavavacanam etad råjanyabandhur iti.

Young ‡vetaketu

37

the term brahmacarya signaling that Uddålaka formally intended to become a student of Jaivali. ‡vetaketu’s reply (found only in B*) is in character; he tells the father to go on his own. The proud young Brahmin is perhaps unwilling to be a pupil of a Kßatriya. With this final refusal ‡vetaketu exits the narrative. C) Uddålaka and Jaivali. The last section of the narrative opens with Uddålaka going to Jaivali (28), who receives him with great respect (29). While C* states briefly that Jaivali paid his respects (arhå∫ cakåra),43 B* carefully notes each act of homage: Jaivali first offers him a seat, then gets his servants to bring water for the guest, and finally offers him the arghya water. These three items are part of the traditional rite of receiving honored guests.44 The mention of arghya would evoke in the listener the elaborate ritual associated with it. Jaivali, in the typical manner of a generous king, then declares that he will grant a wish to his guest (31). Uddålaka uses this promise to ask what the king had said to his son (32). The wording of the request, which is nearly identical in both B* and C* is somewhat unclear (Söhnen 1981: 191-97): is he asking Jaivali to repeat the questions he posed to ‡vetaketu, or to repeat the entire conversation he had with ‡vetaketu, or to teach him the answers to the questions, or to teach him the full doctrine pertaining to those questions? Repeating the questions would have been pointless, because he knew them already. In the context of C*, the second option may imply that Uddålaka is defending his son and putting the king on the spot for having been so arrogant with the young man (Söhnen 1981, 191-92); but as B* narrates the incident, Jaivali had nothing to be ashamed of, and it would be out of character for Uddålaka to be combative. Furthermore, in the subsequent conversation both Jaivali and Uddålaka understand the question as a request for instruction (3439). Since the subsequent conversation does not answer the questions point by point (especially in B*), the likelihood is that this is an oblique request to tell him the doctrine underlying the questions, what Jaivali would have told ‡vetaketu had he been modest and prudent enough to accept Jaivali’s invitation to stay. In the pattern of most upanißadic teachers,45 Jaivali hesitates and tries to wiggle out: “That, Gautama, is in the category of divine wishes. 43 With minor variations, the same expression is used when A†vapati receives the Brahmins: CU 5.11.5. 44 See, for example, ņvalåyana Gr¢hyasütra 1.24.7 and parallels in other Gr¢hyasütras. The Påraskara Gr¢hyasütra (1.3.1) specifies six persons to whom the arghya reception is due: teacher, officiating priest, father-in-law, king, friend, and Snåtaka, and goes on to describe (1.3.4f) the hospitality rite in detail. The arghya is perfumed water containing rice grains, flower petals, and the like. 45 See, for example, Naciketas and Death in KaU 1.12-29, Raikva and Jåna†ruti in CU 4.1-3, stories of Satyakåma Jåbåla and Upakosala in CU 4.4-15, Prajåpati and Indra/Virocana in CU 8.7-12, Yåj∞avalkya and Janaka in BU 4.3.1.

38

LANGUAGE, TEXTS AND SOCIETY

Mention (one) from among human (wishes)” (33). Uddålaka humbly begs Jaivali not to begrudge him the higher boon, the knowledge Jaivali possesses (34). He tells Jaivali that he already has quite enough wealth, and characterizes the knowledge that he is seeking as “more”, “unending”, and “uncircumscribed” (bahu, ananta, aparyanta). The exchange between Jaivali and Uddålaka is quite different in C* and projects quite different images of both individuals (see below §3.2). The narrative in B* continues with Jaivali telling Uddålaka that if he wants knowledge he should request it in the proper manner (tœrthena, 35 ).46 I do not think the author is hinting here at Jaivali’s arrogance; he is simply asking that the imparting of instruction be done in the proper way, that is, by Uddålaka formally becoming a student of Jaivali; this certainly is the way Uddålaka understands the statement and I think it echoes the general belief that only the knowledge imparted by one’s teacher is productive.47 Uddålaka immediately says, “I come to you, sir, as a pupil,” using the technical term upaimi (36).48 The narrator continues with an explanation of this ritual for becoming a student: “with words alone did the people of old come as pupils” (37). Söhnen (1981: 195) thinks that this is a later gloss that found its way into the text. That is clearly possible, but it could equally well have been introduced by the author of B* who felt the need to explain a procedure that his listeners may have found somewhat odd.49 The narrator then emphasizes that Uddålaka lived with Jaivali openly as his student (38). This entire section (35-38) is missing in C*, which has in its place Jaivali’s command that Uddålaka stay longer (35). Clearly the intentions of the two authors diverge widely on this point, the former highlighting Uddålaka’s studentship and the latter ignoring it completely. The term tœrthena is used at TS 2.6.8.4 with a similar meaning. Note, for example, Satyakåma’s plea to his teacher (CU 4.9.3): “I have heard from people of your eminence that knowledge leads one most securely to the goal only when it is learnt from a teacher.” 48 The verb upa- √i is used in the formula uttered by a student at BU 2.1.14-15; CU 4.4.3; KßU 1.1. 49 Bronkhorst (1996: 592-95), arguing against Boris Oguibénin (Three Studies in Vedic and Indo-European Religion and Linguistics [Poona: Bhandarkar Oriental Research Institute, 1990]), who uses this passage quite inappropriately to draw historical conclusions regarding the rite of initiation, thinks that both the words Uddålaka is supposed to have spoken and the explanation have been inserted by the author of B* because they are absent in C* (note, however, that a similar initiation with similar words but without firewood is found at BU 2.1.14, whereas in the parallel passage at KßU 4.19 firewood is introduced). Indeed, the fact that he felt compelled to offer an explanation of this formula argues in favor of its antiquity; the author of B* found it in his source and felt compelled to explain it. Surely, it is implausible that the author inserted the formula and then went on to explain a difficulty that he himself had created! It is easier to assume that the author of C* omitted it for his own theological or literary reasons. The fact that K* has the standard samitpåñiΔ (“firewood in hand” [28], appearing also at KßU 4.19) may indicate that that was the way its author dealt with the problem rather than that this expression was found in the original and was omitted by the author of B*. The differences in the narratives of these three versions, however, have to be seen not as peepholes into ancient history but as windows into the literary and theological motives of the narrators. 46 47

Young ‡vetaketu

39

The final narrative unit consists of Jaivali’s response (39). Here also there are significant differences between B* and C*, and the use of the correlatives yathå/tathå in B* and yathå/yathå in C* makes the syntax unclear. Söhnen (1981: 196) provides the best interpretation, which I have followed. Both versions state that the knowledge Jaivali is about to impart has not been known before to Brahmins. In B* Jaivali foresees an angry curse by the Brahmin Uddålaka or one of his deceased ancestors and uses the above statement to forestall it. In C* Jaivali says that he is repeating what Uddålaka himself had said. The author of B* concludes with a rhetorical question by Jaivali: “for who can refuse you when you speak like that.” Söhnen (1981: 196) thinks that this is a reference to the magical power of Uddålaka’s request. It may also refer, however, to the humility that Uddålaka exhibited in sharp contrast to his son, a humility explicitly recognized in K*.

3.2 ‡vetataketu in C* I turn now to the version in C*. I will be brief here because I have already noted many of its differences from B* in the preceding section. The author of C* presents the three main characters of the narrative in quite a different light. The sharp edge given to ‡vetaketu’s character in B* is blunted here; ‡vetaketu’s anger, for example, is not the result of his own arrogance but an understandable reaction to the haughty demeanor and the cutting words of Jaivali.50 The latter appears as a haughty king picking on a young boy and ordering around an older Brahmin. Uddålaka’s humility is not stressed, and, although he wants to get the knowledge from the king, he does not stoop to becoming his student. A) ‡vetaketu and Jaivali. When ‡vetaketu comes to the assembly of the Pa∞cålas, Jaivali abruptly asks him, “Young man, did your father educate you” (5)51 ‡vetaketu replies politely: “He did, your honor” (6). I have remarked already on the use of bhagavaΔ by ‡vetaketu in C*. It is impossible to determine whether the author of C* introduced this honorific title into ‡vetaketu’s answers or whether it was the author of however, is used commonly throughout B* who deleted it. The title, 52 the CU in similar contexts. In the long instruction by his father (CU 6), moreover, ‡vetaketu always prefaces his answers and questions with 50

In this I fully concur with Söhnen’s (1981: 191) assessment. The opening word of Jaivali’s remarks is the same in B* and C*: kumåra. In B*, however, this is a greeting, while in C* it is the beginning of the question. This makes it likely that one of the authors, probably B*, has changed the wording of his source, which probably began with the initial kumåra. In K* also the king’s question begins with the vocative gautamasya putra. 52 The student Satyakåma Jåbåla’s response to various creatures who instruct him and to his own teacher (CU 4.4-9: a total of 12 times); Upakosala to his teacher Satyakåma Jåbåla (CU 4.14.2-3); a king to Ußasti (CU 1.11.1-2); Jåna†ruti to Raikva (CU 4.2.2, 4); Nårada to Sanatkumåra (CU 7.1.1-25); Indra and Virocana to Prajåpati (CU 8.7-12). 51

40

LANGUAGE, TEXTS AND SOCIETY

bhagavaΔ. Irrespective of whether the author of C* introduced this title into the narrative or found it in his sources, the effect is to make ‡vetaketu a polite young man instead of the spoiled brat of B*. The most significant difference between B* and C* is in the narration of the events that took place after Jaivali had asked the five questions. In place of the kind invitation of B*, Jaivali is here presented as humiliating the boy: athånu kim anu †iß™o ‘vocathå yo hœmåni na vidyåt (17). The repetition of the tmesis (anu) with atha and the interrogative particle kim at the beginning adds intensity to the question; Jaivali is twisting the knife. The author uses the past participle anu †ißta here, but its meaning is clearly restricted to learning; the relative clause makes it plain that the reason why ‡vetaketu cannot call himself anu †iß™a is because he does not know the answers to Jaivali’s questions. ‡vetaketu is hurt, crestfallen, and enraged –all of which emotions can be contained in the term åyasta (19). ‡vetaketu’s emotional state is caused by the humiliating words of Jaivali. In the eyes of the author of C*, these words, as Söhnen (1981: 191) has noted, justify the young man’s anger. ‡vetaketu then comes to his father, with nothing said about his rejection of Jaivali’s invitation.53 Söhnen (1981: 188) remarks: “Die Erregung ‡vetaketus scheint mir allerdings in der ChU-Fassung nach dem Tadel des Königs wesentlich besser motiviert als in der Br¢U-Fassung nach der Einladung.” One may ask how one knows which narrative contains the better motivation for ‡vetaketu’s anger, except by using the questionable strategy of “If I were ‡vetaketu...” This remark reveals the difference in method and goals between Söhnen’s study and mine. The question I would ask is not whether the one or the other narrative gives a “better motive” for ‡vetaketu’s anger, but why the two authors provide two different motives for his anger. The very fact that, as Söhnen says, in C* ‡vetaketu had better reason to be angry makes his anger understandable and excusable. The author of B*, on the other hand, shows that ‡vetaketu not only had no reason to be angry but in fact spurned the kind invitation of a gentle king. In B* ‡vetaketu was angry because he was a spoiled brat! B) ‡vetaketu and Uddålaka. ‡vetaketu accuses his father of having told him a lie: without teaching him fully, the father had told him that he had so taught him (20). Without stopping for the father’s response, ‡vetaketu tells why he thinks so: he was unable to answer five questions (22). Uddålaka then tells his son that, “as you report them to me,”54 he himself does not know any of them; and he concludes with the rhetor53 Söhnen (1981: 188) is not entirely accurate in saying, “Die Reaktion ‡vetaketus is freilich die gleiche wie die in der ChU-Fassung: er nimmt die Einladung nicht an, sondern läuft (ärgerlich) zu seinem Vater.” In fact, C* does not contain any invitation that ‡vetaketu could have refused. 54 Both Deussen (1897: 141) and Renou (1955: 97) detect a lacuna in the C* version.

Young ‡vetaketu

41

ical questions: “If I had known them, how could I have not told them to you?” (25)55 In both B* and C* the father gently rebukes the son; he should know better than to accuse the father of lying or cheating. So ends the encounter between father and son in C* –unlike B*, there is no hint of what the father intends to do, no invitation to the son to come with him and visit Jaivali. From a purely narrative perspective the dialogue in C* is disappointing, because the participants are insufficiently characterized. I think the changes in C*, especially the omission of the invitation that they both go to Jaivali, are deliberate modifications introduced by the author. The invitation was in all likelihood present in his sources, because it is found also in K* (26). His motive, I think, was to make ‡vetaketu, who will appear again as a model student in the very next chapter of CU, not appear in too bad a light. C) Uddålaka and Jaivali. When Uddålaka arrives, Jaivali receives him with respect (arhå∫ cakåra; 29); unlike in B*, there are no details of his reception. The narrative then moves to the next morning; we have to assume that Uddålaka spent the night there (30).56 “In the morning, when he was in the assembly hall, he got up” (sa ha pråtaΔ sabhåga udeyåya; 30): now it is unclear what the antecedent of the pronoun “he” (saΔ) is. Is it Uddålaka or Jaivali? A similar lack of clarity is found in the subsequent statement, “He became distressed” (sa ha kr¢cchrœ babhüva; 33). I think that in both cases the pronoun refers to Uddålaka. In the entire C* an initial pronoun, both the nominative (saΔ) and the accusative (tam), is always followed by the enclitic particle ha, placing some stress on the pronoun.57 In the first section, the pronoun (irrespective of whether it is the subject or the object of the sentence) always refers to ‡vetaketu (3, 19).58 From the time the father is introduced (30), however, the pronoun invariably refers to the father. It is most likely, therefore, that in the two doubtful cases also this pattern is applicable. Jaivali then asks him to choose a wish consisting of “human riches” (31), eliminating the verbal contest about human and divine wishes of B* (31-36). Uddålaka’s reply (32) is derisive of the king’s offer: “Keep your human riches to yourself”. Instead of the humble and Söhnen (1981: 189, n. 23) thinks that the reference is to the five questions that ‡vetaketu had mentioned and suggests (following her view that C* is prior to B*) that the elaboration in B* may have been intended precisely to fill such a perceived lacuna. 55 A nearly identical expression is found in a similar context in PU 4.1. ‡vetaketu, too, uses a nearly identical expression with reference to his teachers, who, he assumes, did not know what his father had just told him, “for had they known, how could they have not told it to me?” yad dhy etad avedißyan katha∫ me nåvakßyan (CU 6.1.7). 56 In the story of King A†vapati’s instruction of a group of Brahmin also, a story that follows the Jaivali episode, the king tells the Brahmins to wait till tomorrow (CU 5.11.7). 57 “This position [of ha] near the opening of a new passage is likely to draw attention to the first word of a paragraph or sentence” (Hartman 1966: 82; cited in Bronkhorst 1996: 592). 58 The particle ha in the very first sentence also draws attention to ‡vetaketu (1).

42

LANGUAGE, TEXTS AND SOCIETY

obsequious individual of B*, Uddålaka is presented here as a spirited Brahmin willing to confront a king. He wants Jaivali instead to tell him exactly what he told ‡vetaketu. The same ambiguity that I pointed out in the parallel passage of B* exists here also. The next statement, however, puts a further wrinkle in C*. It states that “He became distressed” (sa ha kr¢cchrœ babhüva; 33). Söhnen (1981: 192), as almost all translators,59 takes the pronoun here, mistakenly I think, as referring to Jaivali. Not only do all other pronouns in this section refer to Uddålaka, we also have in 33-34 an exact parallel to 28-29, and 30-31. In all these the first sentence begins with sa ha (subject) and the second with ta∫ ha or tasmai ha (direct or indirect object), and both pronouns refer to the same individual –Uddålaka. The only difference here is that two sentences (32, 33) begin with sa ha, with ta∫ ha following both. I think that all three pronouns refer to Uddålaka: he blurts out his angry retort to the king (32) and became distressed (33).60 The reason for both was his perception that Jaivali was not going to reveal his secret knowledge. Here, unlike in B*, Uddålaka’s humility is not given prominence. Jaivali, according to C*, then commands Uddålaka to stay longer (35). The wording here parallels that of Jaivali’s invitation to ‡vetaketu in B* (17), but here there is an imperative (cira∫ vasa –“stay longer”) and åj∞åpayå∫ cakåra (“he commanded”) replaces upamantrayå∫ cakre (“he invited”) of B*. As before, the author of C* presents Jaivali as a haughty king ordering about a Brahmin. The motif of a teacher delaying the instruction of a pupil is a common one and is found frequently in the CU itself.61 Another unique feature of C* is the omission of Uddålaka’s becoming a student of Jaivali found in both B* and K*. It appears that the author of C* did not like Brahmins formally becoming students of non-Brahmins, a feature found also in other narratives of the CU. At CU 1.8.8, for example, the same Pravåhaña Jaivali instructs two Brahmins without initiating them. At CU 5.11.7 a group of Brahmins comes to A†vapati with firewood in hand, a sure sign of seeking to be students. The author says explicitly, however, that A†vapati instructed them without initiating them.62 59 So Max Mller, Deussen, Böhtlingk, Geldner (1928: 133), Hume, I myself in my 1996 translation, and also ‡a∫kara. Geldner (1928: 133, n. 139) explains that the king was embarrassed because he did not really want to teach his secret doctrine to Uddålaka. 60 Both Geldner (1928: 133) and Söhnen (1981: 192) take kr¢cchœ babhüva to mean “became embarrassed”, possibly under the influence of Böhtlingk’s (1889) translation: “Der Frst gerieth in Verlegenheit”. So also Monier-Williams’ dictionary, citing this CU passage, while Böhtlingk and Roth’s Wörterbuch gives the meaning “ungehalten”. Söhnen goes on to propose that the reason for Jaivali’s embarrassment is Uddålaka’s request that he repeat what he had said to ‡vetaketu. He was embarrassed to repeat the haughty and cutting words he had said earlier. I am not convinced by this interpretation. 61 See above note 45. 62 This attitude stands in sharp contrast to that of BU where Brahmins are initiated by Kßatriyas on two occasions (BU 2.1.14-15; 6.2.7).

Young ‡vetaketu

43

This section of the narrative concludes with Jaivali telling Uddålaka that the knowledge he is about to impart was never before known to Brahmins (39). The final phrase of this speech, I think, hints once again at the pride of Jaivali: he tells that government has always belonged to Kßatriyas because they alone possessed this secret knowledge.

3.3 ‡vetataketu in K* In contrast to B* and C*, ‡vetaketu plays a minor role in K*. Its narrative is brief and the characters are not well developed. The author uses the introductory story as a peg on which to hang his teaching about the way to heaven. A) ‡vetaketu and Citra. K* is the only version that sets the scene and gives the reason for ‡vetaketu’s visit. Citra is about to perform a sacrifice and has chosen Åruñi (i.e., Uddålaka) to officiate. Uddålaka, however, sends his son instead (1). It is unclear whether this was perceived as a snub by Citra; but when ‡vetaketu arrives Citra questions the young man. The reading here is undoubtedly corrupt, and there are numerous variants.63 Broadly, however, the question pertains to ‡vetaketu’s knowledge of the diverse ways people go after death. ‡vetaketu replies: “I do not know it” (8). Note that here, as in B*, the honorific title bhagavaΔ is missing. He adds, “But let me ask my teacher,” who in this case also happens to be his father. No mention is made of ‡vetaketu’s emotional state, but there is no indication that he was offended by the question. The emphasis in K* is on the ignorance of Uddålaka, the representative of Brahmins, rather than on the arrogance of ‡vetaketu. The author plays on the rivalry for knowledge between Brahmins and Kßatriyas. The “inversion of the norm” that requires Kßatriyas to be instructed by Brahmins is a theme we find also in another episode of the KßU (4.1-19) where the proud Brahmin Bålåki is defeated and then instructed by Ajåta†atru. B) ‡vetaketu and Uddålaka. ‡vetaketu then goes to his father and tells him in a matter-of-fact way: “He asked me this. How should I answer him?” (22). He expresses no anger at his father for not having taught him this point or chagrin at his own ignorance. ‡vetaketu wants to know how to answer Citra so he can get on with the business of the sacrifice. 63 For the variants, see Frenz 1968-69; Olivelle 1998a. I follow Frenz in reading måloke dhåsyasi (“So you won’t place me in a false world”). The term aloka means more than a false world; it is a “non-world”, i.e., where a person cannot exist. For a detailed study of the concept loka, see J. Gonda, Loka: World and Heaven in the Veda (Amsterdam: N.V. NoordHollandsche Uitgevers Maatschappij, 1966).

44

LANGUAGE, TEXTS AND SOCIETY

The father tells him, however, that he himself does not know it (25) and invites the son to come along with him to Citra (26). The wording here is somewhat unclear. Why the two should perform their private vedic recitation (svådhyåya) within the sacrificial enclosure (sadas)64 beforehand is not explained. Is there sarcasm in the statement “let us gather what others may give us”? And does pare mean not just others but outsiders, that is non-Brahmins? The final sentence, however, is clear: “Come, we shall both go.” ‡vetaketu’s reaction to this invitation is left unstated, but the next section begins with only Uddålaka going to Citra. ‡vetaketu drops out of the narrative silently; his refusal, which is given prominence in B*, is also passed over in silence. C) Uddålaka and Citra. The opening of this section finds Uddålaka coming to Citra carrying firewood in his hands, a clear signal in this literature that one is placing himself as a pupil under a teacher (28). He says only one word to Citra: upåyåni, a technical expression we have already encounterd in B*: “Let me come to you as a pupil” (36). Citra’s response is also brief (39). He says that Uddålaka has proved himself worthy of brahman (that is, the formulation of truth contained in the subsequent teaching),65 because he has “not succumbed to pride” (yo na månam upågåΔ). Söhnen (1981: 182, n. 18) prefers the reading yo måm upågåΔ (“who has come to me”) because it parallels Uddålaka’s upåyåni (“I come to you”).66 I agree with Böhtlingk’s (1898, 84) assessment that this is probably a gloss aimed at improving the text. The reference to Uddålaka not succumbing to pride recalls the parallel statement of Jaivali in B*: “for who can refuse you when you speak like that” (39). The allusion to the humility of Uddålaka in K*, coupled with the fact that ‡vetaketu did not accompany his father to receive instruction from Citra, is further evidence that the sources of the three versions must have contained some reference to ‡vetaketu’s pride.

4. Conclusions What conclusions can we draw from this literary study of the ‡vetaketu story? At the most obvious level, we gain an insight into and an appreciation of these Upanißadic authors as creative writers; and that is important in itself. The Upanißads have been generally studied 64

Given the sacrificial context of ‡vetaketu’s journey, this term has generally been translated as the sacrificial enclosure. Söhnen (1981: 182, n. 16) suggests a more general meaning of assembly (parallel to parißad and samiti of B* and C*). The term svådhyåya here may also be an oblique reference to studentship. 65 For the meaning of brahman as “formulation of truth”, see P. Thieme, “Brahman,” ZDMG 102(1952): 91-129. 66 Although Söhnen does not identify the source of her variant reading, Böhtlingk (1898: 84) ascribed it to the “Bombay Edition” without further details. The Venkateshwar Press edition of thirty-eight Upanißads (1910) also has this reading.

Young ‡vetaketu

45

for their philosophical insights. A study such as this, hopefully, will encourage us to read these wonderful documents also as works of literature. This study also throws some light into the literary structure and theological intent of the BU and CU. Within the compass of this already lengthy paper, it is not possible to examine these issue in detail. Here I will only sketch briefly some that merit further study. I want to guard, however, against sweeping judgments or conclusions; the ones I propose are tentative at best and need to be confirmed and supported by further studies.

4.1 Br¢hadårañyaka Upanißad The BU consists of three distinct sections comprising chapters 1-2, 3-4, and 5-6, each section concluding with a genealogy of teachers. Given the likelihood that these sections were composed by different authors and the uncertainly about the types of textual changes that may have been introduced by the editor(s) who brought these sections together, it is difficult to speak about the literary structure or theological intent of the BU as a whole. Nevertheless, some common features emerge, and it is useful to examine them. The BU is dominated by the larger-than-life figure of Yåj∞avalkya; this is especially true of the second section, appropriately called Yåj∞avalkyakåñ∂a. Yåj∞avalkya is associated closely with Janaka, the great king of Videha, and Witzel (1987: 198-99) has shown that the BU originated probably in the frontier region of Videha. A motif evident throughout the text is the humiliation of proud Brahmins, especially the learned Brahmins from Kuru-Pa∞cåla, the ancient center of Brahmanical culture. Clearly there is a literary effort to establish Videha as a rival center of theological learning, with Yåj∞avalkya as the leading theologian. In the first section eight Brahmins from KuruPa∞cåla are defeated and humiliated in debate by Yåj∞avalkya. Then in a conversation with Janaka, he dismisses derisively the opinions of six other prominent theologians. In the first section, moreover, a Brahmin with the appropriate name Dr¢pta Bålåki (Bålåki, the Proud) is humiliated as a ignorant babbler by king Ajåta†atru (BU 2.1); and in the story we have just examined, the ignorance of Uddålaka and his son, ‡vetaketu, is revealed by another king, Pravåhaña Jaivali. The contribution of Kßatriyas to Upanißadic lore has been a much debated topic (Olivelle 1996a, xxxiv), and the BU, more than any other Upanißad, uses numerous literary strategies to proclaim the victory of the Yåj∞avalkya-Janaka alliance over the Kuru-Pa∞cåla establishment. To some degree, I think, the narrative strategy of the ‡vetaketu story is linked to that strategy. The authors of the BU, moreover, are better story tellers than those of the CU; they are better at creating their characters and at using humor, irony, and sarcasm to good effect. Bålåki, the Proud, who comes to Ajåta†atru boasting “Let me tell you brahman,” is

46

LANGUAGE, TEXTS AND SOCIETY

forced to eat humble pie and to come to him with firewood in hand to become his pupil (BU 2.1). Note Ajåta†atru’s sarcastic comment each time Bålåki attempts to define brahman again: må maitasmin sa∫vadiß™håΔ– which may roughly translate into today’s vernacular “Give me break! Don’t bring me that nonsense!”. Likewise, in the Yåj∞avalkya episode (BU 3.1.1-2) we see humor and sarcasm when he tells his pupil to drive the cattle away; to the complaint of the other Brahmins: “How dare he claim to be the most learned?”, Yåj∞avalkya replies derisively that all they are after are the cows! The penchant for the dramatic is also evident in the way this encounter between Yåj∞avalkya and the Kuru-Pa∞cåla Brahmins ends: their leader is cursed by Yåj∞avalkya and his “head burst asunder”! This, according to Insler (1989-90), appears to have been a literary innovation that left its mark on later literature. The ‡vetaketu story we have examined further corroborates this assessment of the literary ability of the author(s) of BU. Within the context of the literary structure and theological intent of the entire BU, there may have been a further reason for the author to depict ‡vetaketu and his father, Uddålaka Åruñi, in the worst possible light. Uddålaka appears first among the group of eight Kuru-Pa∞cåla Brahmins who were defeated by Yåj∞avalkya. Within that group he does not appear as the leader, but he was probably perceived by the author of BU as a leading theologian.67 More importantly, the author considered him to be the teacher of his hero Yåj∞avalkya.68 Defeating his teacher was one way to establish the supremacy of Yåj∞avalkya.69 ‡vetaketu, as the son of his teacher, was not only a contemporary and possibly a classmate of Yåj∞avalkya, but also, according to general Brahmanical practice, a person to whom the latter owed respect and obedience.70 This may be one reason why the author is keen to portray these two individuals in a very unflattering way in the story of young ‡vetaketu.

4.2 Chåndogya Upanißad Unlike the BU, the CU contains no genealogy of teachers71 that would indicate an independent section. Yet, I think that chapters 1At CU 5.11.2-3 five eminent theologians decide to consult Uddålaka Åruñi regarding the nature of åtman and brahman, indicating that he had an established reputation as an eminent theologian. 68 See above, note 21. 69 This motif is found elsewhere in Indian literature. The Buddha himself wanted to convert his former teachers first. In medieval times, Råmånuja is viewed as defeating and then converting his former Advaita teacher, Yådava Prakå†a. 70 A student was required to obey and respect his teacher’s son just as the teacher himself: see Gautama Dharmasütra 2.31; 3.7. 71 Except the brief one at 8.15: “All this Brahmå told to Prajåpati; Prajåpati to Manu; and Manu to his children.” 67

Young ‡vetaketu

47

3 form a separate section,72 as do chapters 4-5, which establish the preeminence of wind/breath, and 6-7, which present three episodes of instruction, at the human (Uddålaka to his son) and the divine (Nårada to Sanatkumåra) levels, and finally the creator god Prajåpati’s own instruction to the leaders of gods and demons, Indra and Virocana. Here too, then, we may have multiple authors and it is difficult to speak about the literary structure or theological intent of the CU as a whole. As in the BU, however, certain common features emerge. A central theme in the entire Upanißad is the importance of vedic studentship (brahmacarya), which is highlighted in the very last passage about the return of the student from the teacher’s house, his marriage, and his obligation to father children and to live a virtuous life (CU 8.15). The CU has more dialogues between teachers and pupils than any other Upanißad, including the BU. Two other motifs run through the document. The first is that teachers are usually brusque and reluctant to reveal what they know; most often they will give half-answers containing half-truths, as illustrated in the long instruction of Indra and Virocana by Prajåpati (CU 8.7-12). The burden falls on the student to get around these obstacles, to be persistent, and, most importantly, to ask the right questions. A student needs humility, persistence, and basic intelligence to detect a half-truth and to press the teacher to reveal the truth more fully. The second is that knowledge can come from unexpected and unlikely places. So, the great humanitarian Jåna†ruti has to beg the comic character Raikva of uncertain ancestry to instruct him (CU 4.1-2); Jåbåla is taught by bulls and birds (CU 4.4-9) and Upakosala by the sacred fires (CU 4.10-15); Baka is taught by a dog (CU 1.12), and, of course, Brahmins are taught by kings (CU 1.8-9; 5.11-24). In spite of all this, however, the authors of the CU prove to be a rather conservative lot. They take care to inform the listeners that even though Brahmins may from time to time receive knowledge from Kßatriyas, they are never formally initiated as their students. The CU is willing to go only so far in “inverting the norm” that Brahmins are the teachers within society. Pravåhaña Jaivali teaches his two Brahmin friends without initiating them as pupils (CU 1.8.8). In the story of King A†vapati, the author states explicitly that the king did not initiate them: “So the next morning they returned to him carrying firewood in their hand. Without even initiating them as students, he said this to them” (CU 5.11.7). This pattern is repeated in the encounter between Uddålaka and Jaivali: CU omits the initiatory words of Uddålaka, whereas both BU and KßU include them (see B* 35-36). Bodewitz (1996b, 52) has drawn attention to 72 For arguments, including the lack of the key terms udgœtha and såman in the rest of the Upanißad, see Olivelle 1996b: 212-13; see below, pp, 66-67.

48

LANGUAGE, TEXTS AND SOCIETY

the fact that in the JB, which like the CU belongs to the Såmaveda, the account of the two paths lacks the Kßatriya motif. Further research would be needed before we can say whether this is a feature common to Såmavedic schools. It would have been out of character for the authors of CU to ridicule ‡vetaketu. Unlike the authors of BU, moreover, they had no reason to do so. For them, Uddålaka was a great teacher, as evidenced by the long and elaborate teaching ascribed to him in chapter six. ‡vetaketu clearly made a mistake in not perceiving that Jaivali possessed a secret knowledge; he was not smart enough to understand that knowledge is found in unlikely places and not humble enough to seek it from an arrogant king. ‡vetaketu is here an example of a “stupid student”, in contrast to the good students portrayed by Satyakåma Jåbåla and Upakosala; but he is not a bad kid and knows how to be polite to the king. It is his father, however, who is presented as the exemplar of the good student, willing to go even to a Kßatriya to obtain knowledge. In the very next chapter (CU 6.1.1-7) we have a very different scene. Here too, ‡vetaketu is said to have studied for twelve years and came back with a swollen head. His father wants to teach him humility and exposes his ignorance. Instead of acting like the spoiled brat of the earlier story, here ‡vetaketu becomes a “good student”, able to confess his ignorance and to learn from his teacher. We may have been able to draw wider and more significant conclusions from the differences in the telling of the ‡vetaketu story in BU and CU if we had more and better information about their authors and the circumstances of their composition. For ancient India, however, we have to be thankful for small mercies.

5. Postscript. The Continuing Saga of ‡vetaketu The image of young ‡vetaketu as proud and impetuous is found in stories outside the episode we have examined. Indeed, in a story appearing in the very next chapter of the CU, ‡vetaketu and his father are the central figures. The story opens with the father sending ‡vetaketu to a teacher for his studies (CU 6.1.1). The young man was 12 years old when he goes away and “after learning all the Vedas, returned when he was 24, swell-headed, thinking himself to be learned, and arrogant” (caturvi∫†ativarßaΔ sarvån vedån adhœtya mahåmanå anücånamånœ stabdha eyåya; CU 6.1.2). In the rest of the story, however, nothing further is said about his arrogance.73 73

The motif of the arrogant son proud of his meager learning is found also in the famous story of Bhr¢gu, son of Varuña. Bhr¢gu is sent to the realm of death by his father to receive wisdom and humility: JB I.42-44 (see Bodewitz 1973: 102-09); ‡B 11.6.1.1-13. 74 The Kaußœtaki Bråhmaña (26.4), in presenting views about how to rectify a ritual flaw, gives those of Åruñi and ‡vetaketu, and in the very next paragraph the view of Jåtükarña.

Young ‡vetaketu

49

In the Jaiminœya Bråhmaña (2.329) and the ‡åõkhåyana ‡rautasütra (16.29.6-11) there are brief and somewhat unclear references to a story that must have been current at the time. A man named Jala, the 74 son of Jåtükarña, performed a sacrifice and obtained the office of royal chaplain (purohita) among the kings of K农, Videha, and Kosala. In the JB version it was Åruñi (probably the father of ‡vetaketu) who gets Jala to perform the sacrifice. ‡vetaketu becomes jealous of Jala and furious with his father, whom he reproaches for not having any ambition himself and for working to make others prosperous. Young ‡vetaketu also makes an appearance in the story about Aß™åvakra recorded in the Mahåbhårata (3.132-134). Uddålaka has a son, ‡vetaketu, and a daughter, Sujåtå, as well as a brilliant student named Kaho∂a. Uddålaka bestows on Kaho∂a all his learning and his daughter. Sujåtå becomes pregnant and the child, ‡vetaketu’s nephew, is arrogant while still in the womb and tells his father: “you have spent all night studying but still haven’t got it right!” Kaho∂a is furious and curses the child to be born “crooked in eight ways” (hence the name aß™åvakra). Sujåtå prods her husband to go to King Janaka to obtain wealth, but his minister, Bandin, belonging to the süta (bard) class, defeats Kaho∂a in debate and has him drowned in the sea. Uddålaka asks his daughter not to tell the child about his father’s death, and Aß™åvakra grows up believing that Uddålaka is his father and that ‡vetaketu is his brother. When the child is twelve and was sitting one day on Uddålaka’s lap, ‡vetaketu, true to character, becomes jealous and in a rage “grabbed his hand and pulled the crying Aß™åvakra away, yelling, ‘This is not your father’s lap’ “ (MBh 3.132.16). Outside the Brahmanical texts, young ‡vetaketu (Setaketu) appears in a Påli Jåtaka tale (Jåtaka 377; see Lders 1914). Once upon a time the future Buddha was a famous teacher with five hundred pupils. The seniormost of his pupils was Setaketu born of a Brahmin family from the north and very proud of his caste. One day an outcaste Cañ∂åla happened to cross his path. Setaketu cursed the man, ordering him to stand leeward so the wind would not blow from the outcaste to him. The Cañ∂åla quickly moved to the windward making Setaketu even more irate. The Cañ∂åla then challenges Setaketu: if he cannot answer the outcaste’s question, he will put Setaketu between his feet. The young Brahmin is unable to answer the question and is put between the outcaste’s feet. The other pupils tell their teacher, the future Buddha, about the incident. Questioned by the teacher, Setaketu admits it and once again vents his anger at the outcaste. The teacher admonishes him not to be angry with that wise man, and teaches him humility by saying that what Setaketu has not seen or heard or understood is far greater than what he has. In spite of this, Setaketu is chagrined by the fact that he had been put between the feet of a Cañ∂åla and in a huff leaves the teacher. The character of young Setaketu of this Jåtaka story matches that of ‡vetaketu drawn by the author of B*.

50

LANGUAGE, TEXTS AND SOCIETY

The brashness of young ‡vetaketu comes across also in a few brief anecdotes scattered in the vedic literature. As opposed to the general opinion that students should not eat honey, ‡vetaketu ate honey while he was a student, saying that honey is the residue of the triple Veda (‡B 11.5.4.18). In another episode ‡vetaketu says that he is going to get himself initiated for a whole year. His father asks, “Do you know the fording footholds of the year?” And the son replies confidently, “I know” (‡B 12.2.1.9; GoB 1.3.3). Another adventurous young man appearing in the Upanißads is Naciketas. The story of his running foul of his irascible father sets the scene for the instruction of Naciketas by the god of death in the Ka™ha Upanißad. This story was originally part of the Kå™haka Bråhmaña and a fragment of it is preserved in the Taittirœya Bråhmaña (3.11.8.1-6). During a sacrifice at which all the father’s possessions are given as gifts to the priests, the son irritates the father by pestering him with the question: “To whom will you give me?” The father in exasperation says “To Death”. So Naciketas goes to the house of Death. The Taittrœya Bråhmaña, as well as the opening of the KaU (1.1) identifies the father of Naciketas as U†an Våja†ravasa. It appears, however, that the characters of this father-son pair became associated with ‡vetaketu and Uddålaka. In making his wish that when he returns home his father be well disposed and without anger, for example, Naciketas refers to his father as Gautama (KaU 1.10), which is the lineage name of Uddålaka. Death then addresses Naciketas as the son of Uddålaka Åruñi (KaU 1.11). In this study I have used the expression “young ‡vetaketu”; the sources present quite a different image of the mature ‡vetaketu. He is frequently cited as an authority in ritual matters.75 He also appears among a group of learned Brahmins, including the famous Yåj∞avalkya, a group that King Janaka questions about the daily fire sacrifice (agnihotra; ‡B 11.6.2.1-6). Yåj∞avalkya’s connection to ‡vetaketu is interesting. In the genealogy of teachers at BU 6.5.3 and at BU 6.4.3 ‡vetaketu’s father Uddålaka is said to have been the teacher of Yåj∞avalkya. If true, ‡vetaketu and Yåj∞avalkya would not only be contemporaries but possibly classmates. The Åpastamba Dharmasütra (1.13.19-20) gives ‡vetaketu’s view (rejected by Åpastamba) that a householder may spend two months each year as a student with his former teacher, with the added boast “For by this method I learned more than during the earlier period of studentship”. In another passage (1.5.4-6), after saying that among 75 See KßB 26.4; ‡B 3.4.3.13; 11.2.7.12; JB I.249. ‡vetaketu also makes an appearance at ‡B 10.3.4.1, where his father poses a riddle to the officiating priest ‡vetaketu has selected for a sacrifice. 76 For a discussion of the problems created by this statement, see G. Bhler, The Sacred Laws of the Åryas, Sacred Books of the East, II (Oxford, 1879) xl-xlii. I am, however, skeptical about drawing historical conclusions, such as the date of the Åpastamba Dharmasütra, from this sort of literary reference.

Young ‡vetaketu

51

people of later times (avareßu) seers are not born, Åpastamba states that some become seers on account of their learning (†rutarßi) and gives ‡vetaketu as an example.76 In the Mahåbhårata also the mature ‡vetaketu appears as a learned Brahmin: as a priest, together with his father Uddålaka, at Janamejaya’s snake sacrifice (MBh 1.48.7), and visiting the palace of the god ‡akra (MBh 2.7.10). The image of ‡vetaketu as an arrogant and irrascible young man puffed up by a little learning has endured in the Sanskrit literary tradition. ‡a∫kara’s commentary on the BU shows that he understood the literary intent of the author accurately. According to ‡a∫kara, the unstated reason for ‡vetaketu’s visit to the king was his desire to defeat the learned assembly (parißad) of the king, and ‡a∫kara finds the root of this impetuous desire in ‡vetaketu’s arrogance (garva). And it is anger, ‡a∫kara observes, that makes ‡vetaketu give impolite answers to the gentle king. This image of ‡vetaketu is also reflected in the writings of the fourteenth-century polymath Mådhava/Vidyårañya. He gives ‡vetaketu as an example of a man who becomes proud (darpa) of his learning, citing not his encounter with the king but his treatment of his father in CU 6.1.2-3: “‡vetaketu, learnt all the Vedas in a very short time and in his arrogance did not behave properly even in the presence of his father.”77 In contrast to the mature ‡vetaketu, considered almost a seer by Åpastamba, the young ‡vetaketu has remained the quintessential "spoiled little brat" of ancient Indian literature.

77 †vetaketur alpenaiva kålena sarvån vedån adhœtya darpeña pitur api purato vinaya∫ na cakåra (Vidyårañya, Jœvanmuktiviveka, ed. S.S. Sastri and T.R.S. Ayyangar [Adyar: Adyar Library and Research Center, 1978], 60).

52

LANGUAGE, TEXTS AND SOCIETY

53

A

II DharmaskandhåΔ and brahmasa∫sthaΔ: A Study of Chåndogya Upanißad 2.23.1*

I I.1 Chåndogya Upanißad 2.23.1 has been singled out for special attention by Indian theologians and exegetes, as well as by modern scholars, because they consider it to be the vedic basis for and the earliest evidence of the central Brahmanical institution of the four å†ramas. I give here the Sanskrit text together with Hume’s English translation,1 marking each phrase with a Roman numeral to facilitate reference to each: A.

trayo dharmaskandhåΔ Ù i yaj∞o ‘dhyayana∫ dånam iti prathamaΔ Ù ii tapa eva dvitœyaΔ Ù iii brahmacåry åcåryakulavåsœ tr¢tœyo ‘tyantam åtmånam åcåryakule ‘vasådayan ÙÙ iv sarva ete puñyalokå bhavanti Ù v

B.

brahmasa∫stho ‘mr¢tatvam eti Ù vi

A.

i. There are three branches of duty. ii. Sacrifice, study of the Vedas, and alms-giving –that is the first. iii. Austerity, indeed, is the second. iv. A student of sacred knowledge dwelling in the house of a teacher, settling himself permanently in the house of a teacher, is the third.

Originally published in the Journal of the American Oriental Society 116 (1996): 205-19. I give Hume’s translation because it is representative of how both the native tradition and modern scholarship have, by and large, understood and interpreted this passage. The division of the text represented by letters and Roman numerals, of course, is my own. * 1

54

LANGUAGE, TEXTS AND SOCIETY

B.

v. All these become possessors of meritorious worlds. vi. He who stands firm in Brahma attains immortality.

I.2 Arguing against the position of the “Jaiminœyas,” that is, the followers of the Mœmå∫så tradition, ‡a∫kara (in his commentary on the VeS 3.4.17-20) cites this passage as the vedic basis for the å†rama system and for the legitimacy of the celibate modes of life, especially that of the wandering ascetic. Commenting on VeS 3.4.20, he presents two possible interpretations of our text: A) The three branches of dharma include all four å†ramas– the first branch comprises the å†rama of a householder, and the third, that of a vedic student, while the second refers in common to the å†ramas of both a forest hermit (vaikhånasa) and a wandering ascetic (parivrå™). According to this hypothesis, the brahmasa∫sthaΔ refers to a person in any of the four å†ramas who is firmly established in Brahman.2 B) The three branches refer only to the first three å†ramas, the second branch comprising only the forest hermit. According to this hypothesis, the brahmasa∫sthaΔ refers to the fourth å†rama, that is, the wandering ascetic. Dismissing the first alternative, ‡a∫kara argues in favor of the second and concludes: “Therefore, those who belong to the first three å†ramas obtain worlds earned by merit. Only the remaining one, that is, the wandering ascetic, obtains immortality” (tasmåt pürve traya å†ramiñaΔ puñyalokabhåjaΔ pari†ißyamåñaΔ parivrå∂ evåmr¢tatvabhåk). In the commentary on the CU (2.23.1) ascribed to ‡a∫kara, however, the author expresses a very different view.3 At the beginning of the commentary (p. 100), he says that the second branch of dharma is “a hermit or a wandering ascetic, although not an ascetic who is established in Brahman but only one who is established in just the dharma of his å†rama, because the vedic passage assigns immortality to one who is established in Brahman” (tåpasaΔ parivrå∂ vå na brahmasa∫sthaΔ å†ramadharmamåtrasa∫sthaΔ brahmasa∫sthasya tu amr¢tatva†ravañåt dvitœyaΔ dharmaskandhaΔ). Later in the commentary, however, he rejects the notion that the second branch could include a wandering ascetic, taking the latter to be simply the brahmasa∫sthaΔ (p. 103). Although the text is somewhat unclear and ‡a∫kara can be seen as contradicting himself, he appears to make a distinction here between a wandering ascetic who belongs to the fourth å†rama (see p. 106 of the commentary) 2

The same interpretation is put forward by an opponent (pürvapakßa) in ‡a∫kara’s commentary on the CU 2.23.1 (pp. 101-102). 3 A similar view is expressed also in his commentary on the BU 3.5.1. Both the CU and the BU commentaries are generally accepted as authentic works of ‡a∫kara. For a discussion of ‡a∫kara’s divergent viewes on the question whether renunciation constitutes an å†rama or is beyond all å†ramas, see Olivelle 1993: 223-27.

DharmaskandhåΔ and brahmasa∫sthaΔ

55

and one who is brahmasa∫sthaΔ and therefore beyond all å†ramas. Indeed, he calls such a person by the technical term atyå†ramin (p. 105; see Sprockhoff 1981: 82; Olivelle 1993: 222-32). Thus, in his commentary on the CU 2.23.1, ‡a∫kara appears to include all four å†ramas in the three branches of dharma; in any case, he considers the brahmasa∫sthaΔ as beyond the four å†ramas. Bhåskara, in his commentary on VeS 3.4.20, cites “the view of some,” a view that is identical to ‡a∫kara’s, and goes on to present a contrary interpretation similar to that of ‡a∫kara’s opponent. Although he presents the latter as the view of “others”, the arguments he puts forward in its favor make it clear that this second view is Bhåskara’s own. Here he argues that the “three skandhas” comprise all the å†ramas and that the final statement regarding a brahmasa∫sthaΔ applies to a person in any å†rama. Råmånuja, in his Bhåßya, as well as the Vedåntasåra (on VeS 3.4.18-19), basically follows Bhåskara’s view. The native tradition, however, is divided on the interpretation of our text. The Mœmå∫sists object strongly to the opinion of people such as ‡a∫kara and claim that at best this passage contains a “reference” to the å†ramas and not an injunction establishing them.4 Indeed, an opinion presented anonymously by Råmånuja (on VeS 3.4.19), but probably representing the Mœmå∫sist position, claims that all five virtues listed under the three dharmaskandhåΔ– sacrifice, vedic recitation, gift giving, austerity, and brahmacarya– belong to the householder. Thus, the three would represent three different emphases in the life of a householder rather than three separate å†ramas. Even though there are differences with regard to specific points, most modern scholars from Max Mller, in his early and influential translation of the CU (SBE I, 1879), and Paul Deussen, in his monumental translation of sixty Upanißads in 1897, to R. E. Hume (1931) and P. V. Kane (1962-75) have, by and large, disregarded the Mœmå∫sist views and followed the lead of ‡a∫kara in interpreting this passage as a reference to the å†rama system.5 These scholars, moreover, by and large ignore the problematic status of the brahmasa∫sthaΔ highlighted by the native interpreters and take the expression as referring simply to the å†rama of a wandering ascetic. Sprockhoff in many of his writings (principally 1981) has vigorously rejected the traditional view that CU 2.23.1 is a reference to the å†rama system. On the whole, he is right, and I will not repeat all his arguments here.6 Even though many, including Sprockhoff, have 4 I have discussed this point at length in my study of the å†rama system and translated there the lengthy argument of the Mœmå∫sists on this point: see Olivelle 1993: 239-43. 5 For a discussion and bibliography, see Sprockhoff 1981: 80-82, especially n. 176. In my early essay on the subject of the å†ramas (Olivelle 1974: 33-34) I, too, followed this interpretation; my current view is presented in Olivelle 1993: 108-09. 6 I expressed some reservations elsewhere (1993: 109), noting that there is “one aspect of the threefold division that is significant: it presents different duties that a person can fol-

56

LANGUAGE, TEXTS AND SOCIETY

referred to and frequently discussed this difficult passage, it has not been subjected to the close reading and analysis that it deserves. That is the aim of this paper. I.3 The native exegetical tradition and modern scholars alike have focused on phrases ii-iv and vi, considering them to be the most problematic. They have assumed that the meanings of the compounds dharmaskandhåΔ and brahmasa∫sthaΔ are clear and evident, even though the native tradition itself had raised some questions regarding the meaning and reference of brahmasa∫sthaΔ (see IV.1). This confidence, I believe, is misplaced; the meanings of these compounds, especially of the former, are far from clear. An adequate understanding of their meanings, I believe, is the key to a proper assessment of the entire passage. Modern scholars, moreover, have discussed this passage almost exclusively as an independent and self-contained unit. Yet, it is embedded within the second chapter of the CU. In this regard, modern scholarship represents a step backward, because, at least in its interpretation of the compound brahmasa∫sthaΔ, the native tradition has taken into account its context within the CU. Important light, I believe, can be thrown on the significance of this passage if we study it within the context of the Chåndogya discussion on the udgœtha and the såman (see IV.2).

II II.1 ‡a∫kara (on CU 2.23.1, p. 100) presents a lucid explanation of the phrase trayo dharmaskandhåΔ: “Its meaning is this: (there are) three–i.e., numbering three–skandhas of dharma, i.e., divisions of dharma” (trayaΔ trisa∫khyåkå dharmasya skandhåΔ dharmapravibhågå ity arthaΔ). He thus takes skandha to mean “division” or “classification” (pravibhåga) and interprets the compound dharmaskandhåΔ as a dependent determinative (tatpurußa). Råmånuja is not as explicit, but he too clearly takes skandha to mean a division (vibhåga, on VeS 3.4.19) and the compound as a tatpurußa. The received wisdom regarding these two points has never been questioned by modern scholarship. Max Mller (1879) translated the compound as “branches of the law.” This rendering has been followed with remarkable faithfulness by subsequent translators.7 Even Sprockhoff (1981: 80), who is usually cautious about accepting low, at least one of which is a permanent mode of life, as divisions or branches of dharma. This is precisely the kind of thing that the å†rama system does, a system that also presents itself [...] as a division of dharma.” This conclusion also will have to be modified somewhat in the light of the present study (see V.3). 7 Böhtlingk (1889b) “Das moralische Verdienst hat drei Stufen; “ Deussen (1897) “Zwige der [religiösen] Pflicht”; Senart (1930) “branches de la rägle religieuse”; Hume (1931) and Radhakrishnan (1953) “branches of duty”.

DharmaskandhåΔ and brahmasa∫sthaΔ

57

traditional interpretations, translates the compound “Zweiges des dharma.” Following this interpretation, both the native commentators and modern scholars have seen the ordinal numbers in phrases ii-iv as referring to these “branches.” ‡a∫kara is explicit, attaching the compound dharmaskandhaΔ to each ordinal, e.g., iti eßaΔ prathamaΔ dharmaskandhaΔ (“This here is the first branch of dharma”). But this interpretation runs into a serious problem in phrase v, which begins sarva ete (“all these”). Now, if we were to follow the traditional interpretation to its logical conclusion, “these” should refer back to the “three branches of dharma.” Such an interpretation, however, does not sit well with the context; “branches” cannot come to possess or to attain worlds. All interpreters, ancient and modern alike, therefore, are forced to shift gears here and understand “these” as referring to people following the three branches of dharma. ‡a∫kara, as usual, is quite explicit and clear: sarva ete trayo ‘py å†ramiñaΔ yathoktair dharmaiΔ puñyalokå bhavanti –“All these, i.e., all three å†ramins [those belonging to an å†rama], by following the prescribed duties become possessors of worlds earned by merit.” None of these interpreters has noticed or acknowledged this shift from branches to people.8 Now, it is possible to argue that such a shift of reference in phrase v may have been influenced by phrase iv. Whereas in phrases ii and iii the “second” and “third” refer to impersonal acts such as sacrifice and austerity, in phrase iv the “third” is the vedic student (brahmacårin), that is, a person. This shift from acts to a person, one may argue, is carried over into phrase v, where “all these” would refer to people such as a vedic student and, by extension, to people who undertake the “first” and the “second” branches of dharma. Determining a shift here from impersonal branches of dharma to persons who follow those branches is further complicated by the grammatical gender of the Sanskrit term skandha. Although, as we shall see, the term is declined as both a masculine and a neuter noun in the vedic literature, in the CU it is clearly masculine. There is, thus, no way to distinguish grammatically a reference to skandha from a reference to a person, such as a vedic student. Nevertheless, I think that the shift from skandha as the referent of the ordinal numbers in phrases ii-iv to persons as the referent of the pronoun “these” (ete) is unjustified. If we read these sentences without an exegetical or theological eye, the natural way to understand them is to see the referents of “first,” “second,” “third” and “these” as the same; and that referent is the original statement: trayo dharmaskandhåΔ. Yet I believe that the natural tendency of both ancient 8 Kane (1962-75; II: 420-21) circumvents the grammatical problems by appending parenthetical qualifiers to the ordinal numbers: “the first (is constituted by) sacrifice [...], the second is (constituted by the performance of) tapas [...].”

58

LANGUAGE, TEXTS AND SOCIETY

commentators and modern scholars to take “these” as referring to people rather than to impersonal “branches” is correct. If the reference of phrases ii-v is the same, namely, dharmaskandhåΔ, while the reference of phrase v is persons, then I believe that the reference of all these phrases should be persons and we need to rethink the accepted meaning of the compound dharmaskandhåΔ. II.2 Before dealing with this compound, however, I want to examine briefly the meaning of the term skandha. The term occurs in the vedic literature both as a masculine noun (skandhá) and as a neuter (skándhas). As a masculine noun the word occurs frequently in the enumeration of the bodily parts of a sacrificial animal.9 Within these contexts it has been generally translated as “shoulder.” The term probably does not mean shoulder in the strict sense (i.e., a∫sa) but rather the upper part of the back below the neck and between the shoulders, that is, the upper (in the case of a human) or front (in the case of an animal) back torso of the body.10 That it did not mean “the two shoulders” is confirmed by the curious fact that skandha is used regularly either in the plural or in the singular (e.g., TS 5.7.18.1), whereas all pairs of bodily parts, such as eyes and arms, are given in the dual. The plural is also used when speaking of splitting the skandhåΔ of enemies (AV 6.135.1; 12.11.6), while skandhyåΔ (also in the plural) is used with reference to the flesh of this area (AV 6.25.3; AB 7.1). The neuter skándhas in the vedic literature invariably refers to the trunk of a tree. Thus the AV (10.7.38), in describing how the gods gathered around the yakßa, uses the simile of “the branches of a tree gathered around its trunk” (vr¢kßasya skandhaΔ parita iva †åkhåΔ), where the branches (†åkhåΔ) are clearly distinguished from the trunk (skandhas) to which they are attached. Thus, in a homage to trees the TS (7.3.20) states: vanaspatibhyaΔ svåhå mülebhyaΔ svåhå tülebhyaΔ svåhå skandhobhyaΔ svåhå †åkhåbhyaΔ svåhå parñebhyaΔ svåhå pußpe bhyaΔ svåhå phalebhyaΔ svåhå. In this enumeration of the parts of a tree the skandhas (“trunk”) is clearly distinguished from †åkhå (“branch”). I think that Mayrhofer (1953-76, 506) is simply wrong in taking skándhas to mean “Baumast, Zweig / bough, branch of a tree.”11 The R¢V 1.32.5 that he cites in support clearly refers to trunks rather than to branches of trees: ahan vr¢tra∫ vr¢tratara∫ vya∫sam indro vajreña mahatå vadhena Ù skandhå∫sœva kuli†enå vivr¢kñåhiΔ †ayata upapr¢k pr¢thivyåΔ ÙÙ 12 Maurer (1986) translates: “With the 9

AV 9.12.3; 10.9.20; MS 3.15.6; TS 5.7.18.1; TS 7.3.16.1. The term is also used in connection with the bodily parts of the primeval Purußa: AV 10.2.4. It is used with an identical meaning in ‡B(K) 4.4.4.1; ‡åõÅ 2.3. 10 Såyaña, in his commentary on AV 6.25.3, supports this meaning by explaining grœvåbhyo ‘dhaΔprade†aΔ skandhaΔ (“skandha is the area below the neck”). 11 The Petersburg Lexicon, on the contrary, gives the meaning of the neuter skándhas as “der Teil des Baumstammes, an den sich die Aeste ansetze, Baumstamm.” 12 Cited in MS 4.12.3; TB 2.5.4.3.

DharmaskandhåΔ and brahmasa∫sthaΔ

59

thunderbolt, his great weapon, Indra slew Vr¢tra, the arch-Vr¢tra, the shoulderless: like a tree-trunk split asunder with an axe, the serpent lay flat on the ground.”13 Clearly the comparison is between the fallen Vr¢tra lying on the ground and trees (rather than branches) felled by an axe.14 This neuter meaning of the term was absorbed into the masculine in the later language, which, for the most part, lost the neuter form. One can see the similarity between the upper torso of a human and, to a lesser extent, of an animal, and the trunk of a tree, especially of a single-stemmed tall tree, like the coconut.15 The earliest reference that Böhtlingk and Roth (BR, s.v.) can give for the meaning of skandha as “part” or “division” is our passage, CU 2.23.1. All other examples refer rather to a section that is actually a mass or an aggregate, as when the term is used both with reference to a division of an army and to a section or chapter of a treatise, the latter meaning being recorded also in Buddhist Sanskrit and Påli (khandha or khandhaka). Edgerton (1953) is right in listing “mass, bulk” as its primary meaning; other meanings within the strictly Buddhist vocabulary, such as the five skandhas that constitute a human being and the various divisions of dharma into skandhas (for the compound dharmaskandha, see II.4), I think, are related to that primary meaning. Even if one were to grant that the term takes on a technical meaning of “division” within the Buddhist scholastic vocabulary (see II.4), I think it would be inappropriate to extend that meaning to a vedic text such as the CU. The vocabulary of the CU is vedic, and we should interpret the CU use of skandha within the context of its vedic usage. On these grounds, then, skandha in this passage should be taken to mean the upper back or torso (if we follow the metaphor of a human/animal body), or, what I consider is less likely, a tree trunk. II.3 The analysis of the compound dharmaskandhåΔ is complicated by the fact that it is found nowhere else, other than the Buddhist scholastic literature. We have already noted (II.1) that both ancient 13 Geldner’s (1951-57) German translation also takes skandha to mean a tree-trunk: “Indra erschlug den Vr¢tra, den größten Feind, den Schulterlosen (?) mit der Keule, seiner großen Waffe. Wie Baumstämme, die mit der Axt gefällt sind, liegt der Drache platt auf der Erde.” 14 The connection between the neuter and masculine meanings also supports the meaning of tree-trunk; if the neuter meant a branch, one would more clearly see a similarity with the arms rather than with the torso. The various derivatives of the term in other Indo-Aryan languages given by Turner 1966 also support this meaning. 15 Indeed, the coconut tree is called skandhataru. Later Sanskrit regularly uses skandha in connection with carrying a load, a child, or an object such as an ax (see the many examples given in Petersburg Lexicon s.v.), but here too the reference is not strictly to the “shoulder” but to the upper body in general, including the shoulder area. Even today people in South Asia carry heavy loads not on the shoulder but on the upper back, which makes the person hunch forward.

60

LANGUAGE, TEXTS AND SOCIETY

and modern interpreters of the CU take this compound to be a tatpurußa. This analysis was probably influenced by their assumption that skandha meant branch or division. If we take that term to mean torso or upper body, however, such an analysis of the compound would sound strange; dharma is often depicted as having four legs,16 but can it have three torsos? Within the context of the entire passage, it makes much better sense, I believe, to take dharmaskandhåΔ as a possessive or exocentric compound (bahuvrœhi) rather than a tatpurußa. The compound would then mean “those who have dharma for their torso”17 or “people whose torso is the dharma.” The entire phrase can then be translated: “There are three (kinds of) people who have dharma for their torso.” Since skandha means the strongest and the most prominent part of the body, the expression refers to people for whom dharma is the most important thing in their life. Let us see how this interpretation of the compound fits with the immediate context of the passage. Now, the masculine ordinal numbers in phrases ii-iv can refer either to a skandha of dharma (tatpurußa analysis) or to a person whose skandha is or consists of dharma (bahuvrœhi analysis), because the compound is masculine under either analysis. Phrases ii-iv are thus neutral with respect to the two interpretations. Even though phrase iv appears to favor the bahuvrœhi analysis because the “third” is a person, namely a vedic student, naming a person as constituting a class is, nevertheless, a common phenomenon in Brahmanical literature. Thus, for instance, side by side with the statement: catvåra å†ramå gårhasthyam åcåryakula∫ mauna∫ vånaprasthyam –“There are four å†ramas: the householder’s state, (residence) at the teacher’s house, the muni’s state, and the state of a forest hermit” (ÅpDh 2.21.1), we also have: catvåra å†ramåΔ brahmacårigr¢hasthavånaprasthaparivråjakåΔ –“There are four å†ramas: student, householder, forest hermit, and wandering ascetic” (VaDh 7.1-2). The context of phrase v, on the other hand, fits much better with the bahuvrœhi analysis than with the tatpurußa. We have already seen the problem that previous commentators and scholars have run into in interpreting sarva ete –“all these”; this expression refers to persons, while, according to the tatpurußa analysis the previous sentences deal with divisions of dharma. According to my interpretation, phrase v states that all the three persons mentioned in phrases i-iv, persons whose skandha consists of dharma, will obtain worlds earned by merit. The most important contextual evidence in support of the bahuvrœhi analysis, however, comes from the final phrase. As I have indi16

See, for example, MBh 3.148.21. “Torso” not only sounds strange but also is not a totally accurate rendering of skandha. I have not been able to find an exact English equivalent referring specifically to the region of the upper back immediately below the neck and between the two shoulders! 17

DharmaskandhåΔ and brahmasa∫sthaΔ

61

cated in the transcription of CU 2.23.1 (see I.1), it contains basically two sections: A, consisting of phrases i-v; and B, consisting of phrase vi. The final phrase stands in stark contrast to both the opening statement (phrase i) and the concluding statement (phrase v) of A: the brahmasa∫sthaΔ is contrasted with the dharmaskandhåΔ,18 and amr¢tatva is contrasted with puñyaloka. Now, brahmasa∫sthaΔ is a tatpurußa compound with an adjective (sa∫stha) as the final member, making the compound adjectival and referring to a person: a brahmasa∫sthaΔ is a person who is grounded or established in, or devoted to brahman. It would thus be natural that its counterpart in phrase i also refer to persons; contrasting a person established in brahman to divisions or branches of dharma would be highly unusual. The bahuvrœhi analysis of the compound dharmaskandhåΔ provides a much better context for this purpose. The intent of the author is clear and in keeping with similar statements in this and other Upanißads: persons devoted to activities prescribed by dharma end up in a world subject to redeath and rebirth, whereas those devoted to what these texts deem as the highest knowledge or truth reach final immortality.19 II.4 An examination of similar compounds may give us further insights into the import of dharmaskandhåΔ. First, we find several parallel expressions in which the first member is dharma and the second is either the body or a prominent part of the body, e.g., dharmanåbha (“one whose navel is the dharma”), dharmanetra (“one whose eye is the dharma”), dharmåõga (“one whose body is the dharma”), dharmåtman (one whose self is the dharma”). We find that such compounds are usually bahuvrœhis. The first three of these examples are used as names of people and epithets of Vißñu. The compound dharmåtman (“one whose åtman, i.e., the body or the self, is the dharma) occurs in the Bhagavad Gœtå 9.31. More significant, however, are the compounds that have -skandha as the final member. Schwarz (1974-78) in his Reverse Index lists 55 such compounds, if we exclude the five that begin with a particle and six that are strictly Buddhist.20 Of these I have been able to trace 52 (including dharmaskandha), a number that, I believe, is statistically significant.21 Seventeen of these are names or epithets of people, divine beings, Buddhas, and Bodhisattvas,22 and one, Ta™iskandha, is Some see the brahmasa∫sthaΔ as contrasted with sarva ete. This is certainly true, but sarva ete itself merely recapitulates the initial dharmaskandhåΔ. 19 See, for example, CU 4.15.5-6; 5.10; BU 6.2.15-16; KsU 1.2-7; MuU 1.2.5-12; PU 1.9-10. 20 These Buddhist terms will be examined later. The only two compounds ending in -skandhaka listed by Schwarz are not nominal compounds: praskandhaka and vyåskandhaka. 21 I was unable to trace three others: asitaskandha-, srastaskandha-, and ku™håraskandha-. 22 Nine are names or epithets of Buddhas or Bodhisattavas: Upacitaskandha, Amitaskandha, Sa∫vr¢taskandha, Sudr¢dhaj∞ånara†mijålabimbaskandha, Dharmadhan18

62

LANGUAGE, TEXTS AND SOCIETY

a place name. With the possible exception of the last, all are bahuvrœhi compounds, and skandha refers to the upper part or torso of their bodies: “one whose torso is/consists of [...].” Nine examples are names of various trees, and these compounds are also bahuvrœhis with the meaning “(the tree) whose trunk is [...].23 Four are bahuvrœhis used as names or epithets of various animals, with the meaning “one whose upper body is [...].24 Four are used as epithets of humans; these again are bahuvrœhis with the meaning “whose upper body is [...].25 Of the rest, seven are tatpurußa compounds in which skandha has the meaning of a herd, a multitude, or a heap.26 It is apparent that most of these compounds are bahuvrœhis and that the term skandha regularly means the upper body, the trunk of a tree, or a multitude. Nowhere do we encounter the meaning “division” or “branch” outside the context of a treatise, where the meaning, as we have seen, is probably related to “collection” or “aggregate.” Finally, we have the compound dharmaskandha. It is noteworthy that outside of the CU this compound, to my knowledge, occurs only in Buddhist literature, both Påli (dhammakhanda) and Sanskrit, where it has several specific and technical meanings. The compound is found both as the title of various texts (Lamotte 1958: 163, 203-6) and as an expression meaning “article or item of doctrine” (Edgerton 1953, s.v.). It is beyond the scope of this paper to study these technical meanings in detail. I think that Edgerton is right in seeing the basic meaning of “mass” or “agglomeration” behind the term skandha in its use within several compounds. So, we have duΔkhaskandha (“entire mass of misery”), referring to life as a whole. This is clearly its meaning in upådånaskandha, which refers to the five aggregates that form the human person: rüpa, vedanå, sa∫j∞å, sa∫skåra, and vij∞åna. A similar meaning is found in the several religious agglomerations such as †œlaskandha, samådhiskandha, praj∞åskandha, and a†ikharåbhaskandha, ‡ålendraskandha, Acalaskandha, Mahårciskandha, and Mr¢gapatiskandha. Names of other beings are: Gajaskandha, Bhütadhanaskandha, Senaskandha, Jayaskandha, Tatraskandha, Kapiskandha, Veñœskandha, Madhuskandha. 23 They are: dœrghaskandha, dr¢∂Ûhaskandha, sthülaskandha, kr¢ßñaskandha, ghanaskandha, kharaskandha, kålaskandha, godhåskandha, and guruskandha. 24 They are: dhœraskandha (buffalo), pœtaskandha (hog), pr¢thuskandha (boar), and råjaskandha (horse). 25 They are: si∫haskandha (“person with the upper body of a lion”), mahåvr¢kßagalaskandha (“one whose neck and upper body are like a mighty tree”), vipulaskandha (“with broad upper body”), and vr¢ßaskandha (“one with the upper body of a bull”). 26 Thus, we have turaõgaskandha and hayaskandha (“herd or troop of horses”), suraskandha (“class of demons”), kariskandha (“herd of elephants”), påpaskandha and puñyaskandha (“heap of sins or merits”), and naraskandha (“multitude of people”). Of the remaining, three refer to titles or sections of treatises (triskandha, †åkhåskandha, and sa∫hitåskandha), one refers to the section of a foot between the ankle and the heel (a∫hriskandha, i.e., the stem of the foot), one to the trunk of a banana tree (kadalœskandha, which is a kind of illusion), and two refer to the regions of the wind (våyuskandha and våtaskandha); these probably refer to the area where wind is accumulated and from which it blows.

DharmaskandhåΔ and brahmasa∫sthaΔ

63

vimuktiskandha. The reference of dharmaskandha varies. Sometimes it refers to the above four religious agglomerations, but more often the reference is to 84,000 topics of dharma (Lamotte 1958: 162-63). These are the distinct points or articles of doctrine that were supposed to have been preached by the Buddha. Here we come closest to the meaning of “division” or “section.”27 Even if within this technical Buddhist vocabulary skandha may have acquired a meaning of a unit of classification, I think it is highly unlikely that such a meaning can be transferred to the CU, which, even if it were not pre-Buddhist, can hardly be expected to share a technical vocabulary with Buddhist scholasticism.

III III.1 If we accept that the compound dharmaskandhåΔ means “persons who have the dharma for their torso,” then the translation of the phrases ii-iv, which refer back to that compound, would also have to be modified. The ordinal numbers prathamaΔ, dvitœyaΔ, and tr¢tœyaΔ should refer not to impersonal divisions of dharma but to the three kinds of people mentioned in phrase i. Such an interpretation fits nicely with phrase iv, which deals with the vedic student: “The third (kind of person) is a vedic student [...].” The preceding two phrases, however, pose a problem because they refer to virtues or activities; it is awkward to state, for example, that “the second (kind of person) is austerity.” A possible clue to the interpretation of these two phrases is provided by phrase ii: yaj∞o ‘dhyayana∫ dånam iti prathamaΔ. The iti that marks off the three activities, I think, does not merely mark the end of a list but may have a more pregnant meaning. It may indicate the mental attitude of this kind of person; that is, he is devoted to, or thinks highly of, or believes in the efficacy of “sacrifice, vedic recitation, and gift giving.” The iti also puts a closure to the list of activities that such a person thinks is important; to put it colloquially, “Sacrifice, vedic recitation, and gift giving –that’s it.” This closure is emphasized in phrase iii with the emphatic eva. The second kind of person is devoted only to a single activity. It is possible that the iti of the preceding phrase marking the mental attitude is understood here also in the manner of an anuvr¢tti within sütra literature. Indeed, the sütra-like brevity of this passage makes such an interpretation plausible. III.2 Given the sütra-like brevity of phrases i-iii and even of phrases v and vi, however, the prolixity of phrase iv dealing with the vedic student is somewhat jarring. Böhtlingk (1889b: 99), followed by 27

Even here, I would venture to argue that the meaning is rather a collective mass, similar to a division of an army, than a meaning that is synonymous with the normal Sanskrit word for division, namely bheda.

64

LANGUAGE, TEXTS AND SOCIETY

Senart (1930: 28), considers the section after tr¢tœyaΔ, that is, atyantam åtmånam åcåryakule ‘vasådayan, as a gloss and relegates it to the notes. I think Böhtlingk is right. I would go a step further; åcåryakulavåsœ does not add any new information to brahmacårœ but only clarifies its meaning, as a gloss would normally do. We can, thus, see two levels of gloss: an earlier one that glossed brahmacårœ immediately (that is, before tr¢tœyaΔ) as åcaryakulavåsœ, and a later gloss that attempted to explain the meaning of the entire passage and, hence, was appended at the end.28 There are five activities listed in phrases ii-iv as belonging to the dharmaskandhåΔ: sacrifice, vedic recitation, giving gifts, austerity, and the life of a vedic student. Now, these are activities that are also mentioned in other Upanißadic passages as conducive to attaining various worlds after death. In BU(K), 6.2.16, three of these activities –sacrifice, gift giving, and austerity (ye yaj∞ena dånena tapaså lokån jayanti)– are engaged in by people destined to rebirth. In the parallel passage in the CU 5.10.3 those destined to be reborn engage in giving gifts to priests and offering sacrifices to gods (iß™åpürte). Even clearer is BU(K) 4.4.22: tam eta∫ vedånuvacanena bråhmañå vividißanti yaj∞ena dånena tapasånå†akena (“ It is he that Brahmins seek to know by means of vedic recitation, sacrifice, gift giving, austerity, and fasting”29). The Mådhyandina version of this passage (SB[M] 14.7.2.24) reads: tam eta∫ vedånuvacanena vividißanti brahmacaryeña tapaså †raddhayå yaj∞enånå†akena ca (“It is he that they seek to know by means of vedic recitation, vedic studentship, austerity, faith, sacrifice, and fasting”). Here, gift giving is replaced by faith30 and a new activity, brahmacarya, is introduced.31 In listing these five activities, therefore, the author of CU 2.23.1 is not breaking any new ground; these are commonly known and valued activities of virtuous Brahmins. The use of the expression dharmaskandhåΔ, however, indicates that the three kinds of people listed here gave themselves totally to their respective activities. In other words, we are not dealing here with activities that people may undertake now and then, or which may be pursued for brief periods of time. The people the CU has in mind are those who have made these activities their “torso,” people who are totally devoted to these 28

Unlike Böhtlingk, however, in my recent translation of the Upanißads (Olivelle 1996) I retain these glosses. The received text contains them, and we do not have the right, I believe, to reject them unless we are engaged in establishing an ur-text. 29 I think anå†akena (“fasting”) is merely a gloss on tapaså. ‡a∫kara, Bhåskara, and Råmånuja cite this passage as the basis for the statement in VeS 3.4.26. 30 I think †raddhå in these texts is closely associated with giving gifts and may here be a synonym of dåna. See the expression †raddhådeya in CU 4.1.1, and, of course, the common term †råddha that involves feeding ancestors by means of the gift of food to Brahmins. See also BU 3.9.21; CU 8.8.5; TU 1.11.3. 31 Similar enumerations of central virtues/activities are found elsewhere: dama dåna dayå (BU 5.2.3); tapas, dåna, årjava, ahi∫så, satyavacana (CU 3.17.4).

DharmaskandhåΔ and brahmasa∫sthaΔ

65

activities. This is possibly the reason why someone felt the need to add the second gloss to phrase iv; he wanted the reader to know that the brahmacårin this passage refers to is not a student who may spend a few years with a teacher but one who devotes his entire life to this enterprise. That this was in the mind of native interpreters is evident from ‡a∫kara’s comment that one should here understand a naiß™hika (i.e., a perpetual student) because “a temporary student does not attain a world earned by merit through vedic studentship, because it is aimed at acquiring facility in private vedic recitation” (upakurvåñasya svådhyåyagrahañårthatvåt na puñyalokatva∫ brahmacaryeña).

IV IV.1 We turn finally to the last phrase: brahmasa∫stho ’mr¢tatvam eti –“A person who is steadfast in brahman reaches immortality.” I have already drawn attention to the contrast that the author draws between brahmasa∫sthaΔ and dharmaskandhåΔ. In attempting to unravel the meaning of the latter expression, I focused on the second member, -skandha; the meaning of dharma, although not always transparent, is made clear by the activities listed in phrases ii-iv. In exploring the sense of brahmasa∫sthaΔ, on the other hand, we must focus on the meaning of brahman. Modern scholars, by and large, have ignored the context of the CU and either explicitly or implicitly taken this term to refer to the ultimate and absolute entity “Brahman,” most using the capital ‘B’.32 Senart (1930: 28) translates “qui est établi en brahman,” but his note here is representative of the scholarly consensus: “La connaissance de brahman, l’dentification avec brahman donne seule le salut, l’immortalité, le nirvåña.” Sprockhoff (1981: 80) also renders the brahmasa∫sthaΔ as “Der im Brahman Feststehende” without comment, leading the reader to surmise that he is referring to the absolute Brahman. In interpreting this expression, the native commentators have actually done a better job than modern scholars by taking into account the context of the first few chapters of the CU within which this passage and this expression are embedded. I will examine these chapters in greater detail later on, but it appears that all native interpreters regarded the central theme of these chapters as the worship of or meditation on the sacred syllable O· (prañavopåsana). In ‡a∫kara’s commentary on CU 2.23.1, both he and his hypothetical opponent (pürvapakßin) are in agreement in taking brahmasa∫sthaΔ as a man who is totally devoted 32 So, Max Mller (1879), Deussen (1897), Hume (1931), and Radhakrishnan (1953). Kane (1962-75, II: 420-21) takes it to mean one “who has knowledge of brahma and holds fast by it.” Böhtlingk (1889b) uses the term “Heilige” without specifying the meaning; the same term is also used under brahmasa∫stha in BR. Monier-Williams in his dictionary renders the compound as “wholly devoted to Brahma or sacred knowledge.”

66

LANGUAGE, TEXTS AND SOCIETY

to O· (prañavasevaka).33 The immortality that such a person attains is called “the result of serving O·” (prañavasevåphala). Bhåskara and Råmånuja (on VeS 3.4.18), likewise, ascribe to the “Jaiminœyas” the view that brahmasa∫sthaΔ refers to the meditation on brahman by means of O· (prañavena brahmopåsanam). Bhåskara (on VeS 3.4.20) presents this also as his own view.34 IV.2 In attempting to define the literary context of a passage such as CU 2.23.1 we face several difficult issues. First, individual sections in the CU, including our passage, are, in all likelihood, not original compositions of the editor who put together the version of the CU we have today. This becomes apparent when we consider the several passages common to the CU and the BU. Should we review this pre-CU literary context in assessing the meaning of our passage? On the one hand, this is impossible, because that context is irretrievably lost; on the other hand, it is unnecessary, because, although interesting and possibly important in itself, our quest, nevertheless, is not for an ur-meaning but for the meaning that the passage assumes within the literary context of the extant CU. What is important here is the intent of the editor, as manifested in the CU itself, in his literary decision to incorporate this passage into the second chapter of the CU. A second issue relates to the very composition of the CU. Now, it is generally accepted that the BU was put together in at least three stages; the three sections comprising chapters 1-2, 3-4, and 5-6, respectively, probably existed as independent texts before they were incorporated into the extant BU.35 Whether a similar development took place in the composition of the CU is less clear. Several factors, however, indicate that the first three chapters form a unit distinct from the remaining five. The central focus of this unit is the fivefold såman chant and its central part, the udgœtha,36 and in a special way the identity of the udgœtha with the sacred syllable O·. The very 33 ‡a∫kara in his commentary on the VeS (3.4.18-20), however, does not mention prañava in discussing brahmasa∫sthaΔ, and it appears likely that there he takes brahman to mean the absolute. Here again we see a real difference in interpretation between the two texts, leading to doubts about the identity of the authors (see above n. 3). 34 Båskara’s (p. 205) explanation of the expression is significant: brahmasa∫sthatå ca brahmaniß™hocyate så ca sarvå†ramåñåm avi†iß™å Ù na co∫kårålambana∫ brahmånucintanam itareßåm a†akyam iti †akya∫ kartum ! “Being established in brahman means grounding in brahman. And we say that it is common to all å†ramas, for it is not possible to make the act of reflecting upon brahman carried out with the help of O· something that people in other å†ramas are unable to do.” 35 There are several indicators of this development, including the repetition of the episode of Yåj∞avalkya and his wife, Maitreyœ (BU 2.4 and 4.5) and the genealogies of teachers appended to each of the three sections. So, in the first stage individual passages and episodes were composed; in the second, editors wove them into the three sections; and in the third, the three sections were brought together and possibly changed somewhat to form the BU more or less as we now have it. 36 For the five (or sometimes seven) parts of a såman and the importance of the udgœtha, see Kane 1962-75, II: 1166-74; Olivelle 1996a, note to CU 2.2.1.

DharmaskandhåΔ and brahmasa∫sthaΔ

67

terms såman and udgœtha, as well as the syllable O·, almost disappear in the vocabulary of the remaining chapters.37 I think it is justifiable to take the first three chapters of the CU as the immediate literary context for interpreting CU 2.23.1. Indeed, this was the context within which the native tradition represented by ‡a∫kara, Bhåskara, and Råmånuja sought to understand this passage. The entire CU, of course, will provide a more extensive context. IV.3 What, then, is the meaning of brahman in the first three chapters of the CU? The term is used once in the first chapter (CU 1.7.5) twice in the second (CU 2.23.1; 2.24.1; both in compounds), and twenty-two times in the third.38 I will divide these twenty-five occurrences of brahman into three groups: A. brahman as a text or something verbal; B. brahman identified with a particular entity; and C. where the meaning is not altogether clear. A. In the two sections CU 1.6 and 7, r¢c and såman are identified first with pairs of cosmic entities (adhidaivatam) and then with pairs of vital human powers (adhyåtmam). In both categories, there is a hierarchical ascent to the highest of each, namely, the sun and the eye, respectively. CU 1.7.5 goes on to identify the “person one sees within the eye” with five39 types of ritual utterances: r¢c, såman, uktha, yajus, and brahman. The first four, it is clear, are intended to form two pairs: r¢c and såman (this pair is the focus of the entire passage), uktha (R¢gvedic recitation) and yajus. So, the basic vedic formula, that is, the r¢c, is contrasted to the såman and the yajus, the two other types of vedic formulae. The fifth, brahman, is clearly a similar ritual formula, but as the fifth surpasses the other two pairs. What this brahman is, the text does not make immediately clear. The context, however, is the identity between udgœtha and O·; this identity is stated boldly at the beginning (CU 1.1.1) and repeated at 1.4.1 and 1.5.1. Indeed, immediately after saying “the udgœtha is O·, and O· is the udgœtha” (1.5.1), the text identifies O· with the sun. Then, at 1.6.8 the CU connects the person in the sun with the r¢c and the såman, and identifies him with the udgœtha, that is, with O·. At 1.7.5, the CU identifies the person in the sun with the person in the eye. The person in the sun is first connected with the pair r¢c-såman, and then presented as transcending them in its identity with the udgœtha. The person in the eye is first connected with two pairs r¢c-såman and ukthayajus, and then presented as transcending them in its identity with the brahman. It appears justifiable, then, to take brahman in CU 1.7.5 as identical with the udgœtha, that is, with O·. 37 The syllable O· appears only once at CU 8.6.5. ‡åman is found at CU 4.17.2-3,6 and 6.7.2 always within the context of the three vedic texts. Udgœta is completely absent. 38 CU 3.5.1, 2; 3.10.1, 3; 3.11.2, 3, 5, 6; 3.12.7; 3.13.6 (twice); 3.14.1, 4; 3.18.1(twice), 2, 3, 4, 5, 6; 3.19.1,4. In addition, the masculine brahmå is mentioned once at CU 3.11.4 as the original teacher of the doctrine. 39 I will examine later (see V.1) the significance of five and other numbers within the first three chapters of the CU.

68

LANGUAGE, TEXTS AND SOCIETY

Turning now to the third chapter of the CU, in its first eleven sections the sun is compared to a honeycomb. It contains eastern, southern, western, northern, and upper honey cells. Connected with each are bees which gather the honey and a flower from which it is gathered. These bees and flowers are identified with different ritual texts and/or utterances.40 The bees are viewed as incubating their respective flowers;41 the result is honey, that is, the nectar of the sun. Different classes of deities feed on this nectar, using different mouths. All this can be presented graphically: Chåndogya Upanißad 3 BEES FLOWER DEITY MOUTH

East r¢c r¢gveda Vasus Agni

South yajus yajurveda Rudras Indra

West såman såmaveda Ådityas Varuña

North atharvåõgirasa itihåsapuråña Maruts Soma

Up åde†a brahman Sådhyas brahman

The image of a bee within a flower works well as a metaphor for r¢c, yajus, and såman, which are discovered within their respective collections, R¢gveda, Yajurveda, and Såmaveda. The metaphor is unclear with reference to atharvåõgirasa and itihåsapuråña, unless this is an indication of an early meaning of itihåsapuråña which contained the atharvåõgirasa formulae. We are, however, most interested in the fifth column, where åde†a42 corresponds to the bee, and brahman, to the flower (CU 3.5.1-2). Although here too the metaphor is unclear,43 we can safely conclude that both åde†a and brahman are here either texts or verbal expressions. Indeed, after describing the five kinds of “immortal waters” that proceed in the five directions from the five types of incubations, the author states: “These, clearly are the very essence of the essences, for the essences are the Vedas, and these are their essence. These are, moreover, the immortal nectar of nectars, for the nectars are the Vedas, and these are their nectar.” The author thus takes at least the five flowers, and possibly also the bees that do the incubating, to be “Vedas.” The fifth flower, that is, brahman, therefore, must also fall within the category “Veda.” The immediate context, however, does not specify that brahman here is O·. The mouth of the Sådhyas, moreover, is also said to be brahman. The meaning here is not altogether clear, but it appears that the Sådhyas do not drink the nectar, which is the essence of brahman, 40

For a similar comparison, see TU 2.3.1. Bees resting within a flower is like a hen sitting on the eggs. Incubation (tapas) is a common theme in vedic creation stories. 42 Åde†a in the Upanißads means “a rule of substitution,” as when it is said that “brahman is the sun.” Here “sun” is a substitute for brahman. For a detailed examination of this term, see Thieme 1968a. For åde†a as a class of texts, see also TU 2.3.1. 43 If we are to carry the metaphor of the vedic formulae and vedic texts over to åde†a and brahman, then åde†a would be individual statements or formulations, while brahman would be a collection in which those åde†as are contained. 41

DharmaskandhåΔ and brahmasa∫sthaΔ

69

with an external mouth; the mouth and the drink are the same, that is, brahman. At the conclusion of this section on honey, the CU (3.11.2) cites a verse: “There, surely, it has never set, nor even risen. By this truth, O gods, let me not be stripped of brahman,” and concludes: “When someone knows this very brahmopanißad, for him the sun neither rises nor sets, for him it is always day.” The term upanißad here probably means “ a hidden teaching” or more precisely “a hidden connection.”44 But what is the meaning of brahman? That it refers to a verbal expression is clear both from its earlier use in CU 3.5.1-2 and from the statement that immediately follows at CU 3.11.4, a statement that gives a list of teachers from Brahmå to Uddålaka Åruñi who taught this brahman and concludes: “So, a father should impart (prabrüyåt) this brahman only to his eldest son or to a worthy pupil.” As Thieme (1952: 119, n. 3) has noted, when brahman is the direct object of verbs such as √vac, √vad, and √brü, it is not correct even from a grammatical point of view to translate such statements as “speak about brahman.” Here, the father speaks brahman, that is, a particular formulation, just as he would speak or recite a particular r¢c or såman. All this is nothing new; I have merely attempted to show that the evidence of the CU supports the conclusion drawn by Thieme (1952) with reference to the earlier vedic literature that brahman has always a verbal reference and that it means a “formulation of truth” (Wahrheitsformulierung). B. I now turn to the sentences in which an identity is expressed in the formula: “Brahman is X.” There are nine such statements in the first three chapters of the CU.45 Now, such statements, where one thing is said to be another, is given the technical term åde†a, that is, a rule of substitution. That is, when a homology is established by saying, for example, “Brahman is the sun” (CU 3.19.1), the sun can be substituted for brahman. The close relationship between åde†a and brahman is seen in the earlier example where the bee is called åde†a and the flower, brahman (CU 3.5.1). In these åde†as also, therefore, brahman refers to a verbal formula. This is made explicit in the parallel drawn between gåyatrœ and brahman at CU 3.12.2 and 7 within their respective åde†as. In this respect, brahman functions in a manner identical to r¢c, and såman, which are also used in a variety of åde†as: “r¢c is this earth, and såman is fire” (CU 1.6.1).46 The question that is answered in these åde†as, therefore, is the cosmic or bodily counterparts (bandhu)47 of central elements of the ritual, here ritual utterances that constitute the vedic texts. On the meaning of the term upanißad, see Renou 1946 and Falk 1986. CU 3.12.7; 3.14.1, 4; 4.18.1(twice), 2-6; 3.19.1, 4. 46 The åde†as of r¢c and såman are found frequently in the Upanißads: see CU 1.6.15; 1.7.1-4; BU 1.5.5; 5.13.1-3. Similar åde†as are found with references to other verbal formulae: CU 2.1-21 regarding the parts of a såman; CU 3.12 regarding gåyatrœ. 47 On this central concern of the vedic thinkers, see Gonda 1965a and Smith 1989. 44 45

70

LANGUAGE, TEXTS AND SOCIETY

C. Among the occurrences of brahman whose meaning is unclear, I will ignore CU 3.13.6, which speaks of the “courtiers of brahman” (brahmapurußåΔ). At CU 2.24.1 we have the compound brahmavådinaΔ. Thieme’s (1952: 119, n. 3) observation noted above and the parallel expression satyavådin make it clear that this compound does not refer to people who discourse on brahman; they “speak brahman,” which clearly indicates that brahman must refer to something said, that is, a verbal expression. The final “uncertain” use is in the compound under discussion, namely, brahmasa∫sthaΔ. IV.4 What, then, is the meaning of brahman in this compound within the context of the first three chapters of the CU? The negative is easier to establish: it is very clear that brahman here does not mean the absolute Brahman of later theologies such as Advaita Vedånta. Positively, it is also clear that the term in general refers to some verbal formulation, such as the “formulation of truth” of Thieme. We have seen that the first three chapters of the CU aim at establishing the identity of udgœtha and O·. We have also seen that the term brahman always refers to some verbal formulation, and that at least in some instances brahman means O·.48 Are the native commentators right, then, in taking brahman in the compound brahmasa∫sthaΔ as simply O·? The answer, I believe, is yes. This conclusion is supported by the passage (CU 2.23.2-3) that immediately follows it and provides its immediate context: prajåpatir lokån abhyatapat Ù tebhyo ‘bhitaptebhyas trayœ vidyå sa∫pråsravat Ù tåm abhyatapat Ù tasyå abhitaptåyå etåny akßaråñi sa∫pråsravanta bhür bhuvaΔ svar iti ÙÙ2 ÙÙ tåny abhyatapat Ù tebhyo ‘bhitaptebhya o∫kåraΔ sa∫pråsravat Ù tad yathå †aõkunå sarvåñi parñåni sa∫tr¢ññåny evam o∫kåreña sarvå våk sa∫tr¢ññå Ù o∫kåra eveda∫ sarvam o∫kåra eveda∫ sarvam ÙÙ3 ÙÙ Prajåpati incubated the worlds, and, when they had been incubated, the triple Veda sprang from them. He incubated the triple Veda, and, when it had been incubated, these syllables “bhür, bhuvaΔ, svar” sprang from it. He incubated these syllables, and, when they had been incubated, the syllable O· sprang from them. As all the leaves are bored through by a pin, so all words are bored through by O·. This whole world is nothing but O·.49

The syllable O· comes into existence through a process of incubation that gives rise to more and more refined products: from the material worlds emerge the three Vedas, their verbal counterparts; 48 I am, of course, not taking into account here the many places in the later Upanißads which explicitly identify brahman with O·: see, for example, TU 1.8.1; KaU 2.15-16; PU 5.2; MaU 1-2; 49 A more extensive description of this incubation, which, however, stops with the three syllables, is given at CU 4.17.1-7.

DharmaskandhåΔ and brahmasa∫sthaΔ

71

from the three Vedas emerge the three sacred sounds, that also stand for the three worlds; and from these sounds O· comes into being, which is the entire world (sarvam). This passage appears to be a commentary on how a brahmasa∫sthaΔ attains immortality; the reason is that O·=brahman is “All,” which in vedic thought is equivalent to immortality.50 The connection between these two sections of CU 2.23 is evident also in the use of the term loka, “world.” In 2.23.1 the dharmaskandhåΔ become puñyalokåΔ, that is, they obtain/possess “worlds earned by merit.”51 It is implied that the amr¢tatva that a brahmasa∫sthaΔ attains is beyond these worlds; compare the “region beyond the sun” that is attained by the twenty-second syllable of the såman (CU 2.10.5; below V.1). That loka here refers to the three standard worlds of earth, intermediate region, and sky is clear from the use of the term in 2.23.2. Prajåpati incubates the “worlds” (lokån) and the triple Veda sprang from them; clearly, the plural lokån here refers to the three words.52 Just as amr¢tatva, so O· is the fourth that transcends the three worlds. Indeed, the syllable O· is characterized as “immortal and free from fear/danger” (amr¢tam abhayam).53 The statement at CU 1.4.4-5 in many respects parallel the statement on brahmasa∫sthaΔ: eßa u svaro yad etad akßaram etad amr¢tam abhayam Ù tat pravi†ya devå amr¢tå abhayå abhavan ÙÙ 4ÙÙ sa ya etad eva∫ vidvån akßara∫ prañauty etad evåkßara∫ svaram amr¢tam abhaya∫ pravi†ati Ù tat pravi†ya yadamr¢tå devås tadamr¢to bhavati ÙÙ 5 ÙÙ This syllable –the immortal and the fearless– is that very sound. Upon entering that syllable, the gods became immortal and free from fear. A man who utters this syllable with that knowledge enters this very syllable, the sound that is immortal and free from fear. As the gods became immortal by entering it, so will he.

V V.1 To understand the entire import of CU 2.23.1, finally, we need to consider the significance of the numerology that permeates the first 50 See Gonda 1955. A more extensive process in the emergence of O· as the quintessence of the essences which are the Vedas is given at CU 1.1.2. At CU 1.1.5-6 O· emerges from the sexual union of r¢c and såman. 51 Loka is also connected with those who return to this world in BU 6.2.16 cited earlier. 52 The connection between the three Vedas and the three worlds and the cosmic occupents of these worlds (R¢gveda = earth, fire; Yajurveda = intermediate region, wind; Såmaveda = sky, sun) is established elsewhere as well: CU 4.17.1-3. For a systematic analysis of these correspondences, see Smith 1994: 129-32, 289-90, 292-93. The three sacred utterances bhüΔ, bhuvaΔ, and svaΔ, moreover, refer to the three worlds (see ‡B(M) 11.1.6.35). The connection between these and the three Vedas is established in AA 1.3.2; TU 1.5. 53 This is a common characterization also of brahman in the CU: 4.15.1; 8.3.4; 8.7.4; 8.8.3; 8.10.1; 8.11.1.

72

LANGUAGE, TEXTS AND SOCIETY

three chapters of the CU. Our passage establishes, first, a triple category (dharmaskandhåΔ) and posits another, brahmasa∫sthaΔ, that transcends the three (3 + 1). Upon further analysis, we see that the three actually consist of five, because the first is further subdivided into three. Thus, we get a further grouping of 5 +1. Adding one to a pre-existing set, especially to a set of three, to produce a more complete whole and to signal the uniqueness of what is added is a common feature in India (Gonda 1976; Malamoud 1982; Olivelle 1993: 107). In the first three chapters of the CU, however, this feature appears as a central theme. I list below fourteen clear examples of this feature in these chapters: 1.1.2-3: O· emerges as the eighth, the highest = 7 + 1 1.1.9: three Vedas continue because of O· = 3 + 1 1.2.1-7: breath in mouth beyond five vital functions = 5 + 1 1.4.1-4: three Vedas plus O· = 3 + 1 1.6-7: five homologies of r¢c and såman, plus “person” = 5 + 1 1.8-9: space beyond seven cosmic entities = 7 + 1 1.12: twelve stobhas, plus hu∫ as thirteenth and unexplained = 12 + 1 2.2-7: five parts of såman related to 5, and to an additional 1 = 5 + 1 2.10: seven-fold såman consists of 3 x 7, plus 1 (= 22) = 3x7 + 1 2.11-21: ten types of såman plus the highest = 2x5 + 1 2.23.1: three dharmaskandhåΔ (= five) plus brahmasa∫sthaΔ = 3/5 + 1 2.23.2-3: O· emerges from thee worlds, three Vedas, three syllables = 3x3+ 1 3.1-11: five quarters, Vedas and nectars, plus zenith/brahman = 5 + 1 3.13: five brahmapurußåΔ and brahman = 5 + 1

The numbers to which 1 is added are, for the most part, 3, 5, and 7, or multiples of these. The number 3 corresponds to the three Vedas, to the three worlds, and to the three syllables that stand for the three worlds (bhüΔ, bhuvaΔ, svaΔ). The numbers 5 and 7 correspond to the såman with five or seven parts; five, of course, is a number indicating fullness even in other contexts.54 In four of these fourteen instances, the additional one is O·; in three others it is brahman, which, as we have seen, is often a synonym of O· in these chapters. Just as O· is called “the quintessence of all essences, the highest, the ultimate, the eighth” (CU 1.1.3), so space, which is also the eighth (CU 1.9.1-2), is called “the most extensive (parovarœyas) udgœtha, without limit.” As the equivalent of udgœtha, space is equal to O·; space is also said to be brahman (CU 3.12.7; 3.18.1). In explaining the cosmic equivalences of the five or the seven parts of a såman, the author inserts an additional explanation that transcends the others. Thus, after relating the fivefold såman to five cosmic sets of five, he presents a sixth set consisting of 54 “There are five utterances[...]; the sacrifice is fivefold, the animal victim is fivefold, the seasons of the year are five: this is the one measure of the sacrificial rite, this is its completion” ‡B(M) 1.5.2.16. See, Olivelle 1993: 54.

DharmaskandhåΔ and brahmasa∫sthaΔ

73

the five vital functions (CU 2.7.1-2), calling it “the most extensive” (parovarœyas), the same term used to describe the udgœtha that is space. Likewise, after relating the ten types of the fivefold såman to ten (5x2) cosmic, ritual, and bodily sets, the author presents an additional såman “woven upon the Whole (sarvam; CU 2.11-21);” and here the set of five consists not of five single items as in the preceding but of three items each: triple Veda, triple world, fire-windsun; stars-birds-glittering specks; snakes-Gandharvas-ancestors. This last såman transcends the rest, because through this a person can “become the whole,” which is another way of stating that he becomes immortal. In a similar way, the “person in the sun/eye” is the sixth, beyond the five cosmic and bodily equivalences of r¢c and såman (CU 1.6-7), and the knowledge of this permits the priest to obtain all desires through his singing. The “breath in the mouth,” likewise, is presented as the sixth beyond the five vital functions (CU 1.2.7-14), the sixth, which is homologized with the udgœtha and the sun (CU 1.3.2), the knowledge of which permits the priest to secure all his desires through his singing. In all these the added item is the highest and brings the highest reward, often immortality, to the person who knows it. We see this paradigm also in the counting of the syllables of a sevenfold såman at CU 2.10.1-5. After demonstrating that each of its seven parts contains three syllables and is, therefore, equal to the other parts, the author states that the total syllable count of the såman is twenty-one. But actually the names for the seven parts contain twenty-two syllables, and he sees in this the secret of transcendence. The twenty-second is beyond the sevenfold såman: “With twenty-one of those one reaches the sun [...]. With the twenty-second one conquers what is beyond the sun. That is the vault of the sky, a place free from sorrow.” V.2 Our passage on the dharmaskandhåΔ and brahmasa∫sthaΔ fits nicely into this paradigm, and this may have been one of the reasons why the editor of the CU included it in this section. The brahmasa∫sthaΔ is the one added to the prior group of three (if we take the three dharmaskandhåΔ) or five (if we take individual dharmas), and the added one is connected to immortality. The parallel with the passage that follows at CU 2.23.2-3 is clear. As the brahmasa∫sthaΔ and amr¢tatva are the additional categories that transcend the three dharmaskandhåΔ and the three puñyalokåΔ, so O· and sarvam are the additional categories that transcend the three sets of threes: three worlds, three Vedas, and the three sacred utterances.

VI VI.1 Let us finally return to the question we started out with: the connection, if any, between CU 2.23.1 and the å†rama system. The perception of such a connection was grounded in the interpretation

74

LANGUAGE, TEXTS AND SOCIETY

of dharmaskandhåΔ as a tatpurußa compound meaning “branches (divisions) of dharma.” We have seen that the preponderance of evidence suggests that this interpretation is wrong. Further, the focus of this passage is not on the three dharmaskandhåΔ but on the final statement about brahmasa∫sthaΔ, in keeping with the common feature of the first three chapters, whereby one is single out, generally O·, that transcends a prior set, generally a set of three or five. In other words, the intent of the author is not to point out that there are three dharmaskandhåΔ but to show that the brahmasa∫sthaΔ transcends them and attains immortality. The “three” here, moreover, should not be given undue significance; it is a number we come across repeatedly in the first three chapters. It is clear, then, that this passage has nothing to do with the later theological construction of the four å†ramas. We are thus returning to the position of the Mœmå∫sists and rejecting the interpretations offered by the Vedåntists (‡a∫kara, Bhåskara, Råmånuja) and most modern scholars. Nevertheless, even though dharmaskandhåΔ does not refer to a division of dharma, paralleling the dharmasya caturdhå bhedam (“the fourfold division of dharma”) of the BDh 2.11.9, which introduces the four å†ramas (Olivelle 1993: 109), I still believe that this enumeration of three kinds of people devoted to dharma, people who select a particular element of dharma to pursue all their lives, is the same type of theological thinking and classification that finally gave birth to the å†rama system. VI.2 In the light of the above study, I offer this new translation of CU 2.23.1: A.

i. There are three types of persons whose torso is the Law (dharma). ii. The first is one who pursues sacrifice, vedic recitation, and gift-giving. iii. The second is one who is devoted solely to austerity. iv. The third is a celibate student of the Veda (living at his teacher’s house; that is, a student who settles himself permanently at his teacher’s house).55 v. All these gain worlds earned by merit.

B.

vi. A person who is steadfast in brahman reaches immortality.

55

The section within parentheses probably consists of two later glosses.

A

75

III Orgasmic Rapture and Divine Ecstasy: The Semantic History of ånanda*

I.1 Ånanda1 is one of the most common terms in the religious vocabulary of the Bråhmañical/Hindu traditions both in Sanskrit and in the vernaculars, both in the monistically inclined traditions, such as Advaita Vedånta, and in the bhakti traditions. The term points to the intense feeling of joy that devotees experience in their loving devotion and service of god, and mystics, in their meditative trance or samådhi. Within Advaita and related traditions, it represents a central and essential “attribute” of Brahman. The term has found its way into the names or titles of religious figures, often in compounds, such as Brahmånanda and Ånandatœrtha.2 In many of the Indian religious traditions, mokßa, the final goal of human existence, has been defined as ånanda. * I want to thank Joel Brereton, Stephanie Jamison, and Mark Southern for sharing their deep knowledge of the vedic language with me and for helping me decipher several difficult passages. Originally published in the Journal of Indian Philosophy 25 (1997): 153-80. 1 It is clear that etymologically ånanda is derived from the root √nand ; but much less clear is exactly how the noun is derived. As van Buitenen has pointed out, the compound verb ånandate is not found in the early literature. He is right, I think, in regarding it “as a verbal noun nanda with prefixed å, and thus belong to a fairly large group that often goes unrecognized: å indicate the place where the verbal action occurs, for example, å†rama, where one toils; åråma, where one enjoys oneself; åkara, where things are scattered; ålaya, where things lie, etc. The word ånanda thus implies a locus: that in which one finds bliss, be it a son, the fulfillment of a wish, [...] Ånanda then is not just a free-floating unfocused bliss, a state of beatitude; it has an implied object” (van Buitenen 1979, 327; for an extensive discussion of the etymology and meaning of the similar word å†rama, see Olivelle 1993, 8-24). Although the implication of a locus is correct and is evident in some of the early usages that locates ånanda in the sexual organ or penis, in the later language the term assumes the general meaning of happiness and joy. Nevertheless, van Buitenen’s insight is important for teasing out the early semantics of ånanda. 2 W. Schwarz lists 96 compounds with ånanda as the final member, most of which

76

LANGUAGE, TEXTS AND SOCIETY

One of the most ancient, and perhaps the most significant, examples of the centrality of ånanda in theological discourse is the Brahmsütras ascribed to Bådaråyaña. After an introductory statement that brahman is the source of the universe (1.1.5-11), Bådaråyaña devotes eight sütras to demonstrating that brahman/åtman is defined in the Upanißads as ånanda.3 Although the expression does not occur in the writings of ‡aõkara, the compound saccidånanda (being-consciousness-bliss) became in time both within and outside the Advaita tradition a short-hand definition of brahman.4 Given the prominence of ånanda in the theological, as well as the non-technical, religious vocabulary of India at least from about the fifth century C.E., it is somewhat surprising that, with the exception of van Buitenen’s (1979) study, not much work has been done into the semantic history of this term. For heuristic purposes, I will divide the pre-Brahmasütra period into 1. Early Vedic (consisting of the RV, AV, and the mantra portions of the YV); 2. Middle Vedic (consisting of the Bråhmañas); 3. Late Vedic (consisting of the Årañyakas and Upanißads); and 4. Post-Vedic (principally Buddhist and epic literature). In this paper my focus will be principally on the first three periods, introducing evidence from the fourth only to point to the possible later semantic history of the term.

I.2 To begin with, however, I want to discuss briefly the conclusions drawn by van Buitenen (1979) in his pioneering study. He acknowledges the explicit sexual connotations of the term in a wide spectrum of vedic texts, including the Upanißads, but rejects what he calls the “reductionist” fallacy of equating ånanda with orgasmic thrill (1979, 326).5 He concludes his survey by pointing out the semantic multivocity of the term: appear to be names or titles: Reverse Index of Old Indian (Wiesbaden: Otto Harrassowitz, 1975), II: 219-20. 3 It is impossible to date the Brahmasütra with any precision or certainly, but we would not be too off the mark in assigning it to a period between the 2nd and 5th centuries C.E. Although a detailed analysis of texts from this period is required to arrive at a definitive conclusion, I think that this treatment of ånanda in the Brahmasütras was a watershed in the history of the term. It clearly influenced all the “Vedåntic” traditions, both the Advaita and the devotional. The Advaita use of the term may also have influenced the Tantric definition of the liberating experience as ånandå: Abhinavagupta, Tantråloka, 5.27-53;. Muller-Ortega 1989, 197-98. 4 See Deussen 1912, 212; Nakamura 1983, 486. According to Deussen, the expression first occurs in the Nr¢si∫hatåpanœya Upanißad, which is not of great help, because we know nothing about the date of that Upanißad. 5 Van Buitenen’s fear of “reductionist” tendencies in the study of religion was perhaps influenced by his connections to the Chicago school of “history of religion”. On the reductionist debate in the study of religion (where reductionism is often used as a derogatory term) and on the reductionist imperative if we are to engage in any explanatory endeavor, see Segal 1983; Preus 1987.

Orgasmic Rapture and Divine Ecstasy

77

In following the course of the uses of the word ånanda we have seen it pause at the landmarks in the development of religion and thought. It was the high joy of drinking the soma and of offering it, the climax of the ritual building of the universe, the unhindered happiness of gods, the orgasm that begets a son in one’s image as a metaphor of one’s selfrenewal as one of the gods, the joyous knowledge of oneself and the eldest brahman, and the bliss that is the brahman and the åtman (van Buitenen 1979, 330).

Van Buitenen is right to point out that ånanda has a range of meanings, consonant with the range of conditions under which a person experiences joy or pleasure. However, he makes no effort to examine which of these meanings may have provided the foundation for its technical use within the theological vocabulary of the Upanißads, a use that formed the basis of its later use in Vedåntic discourse. His reluctance to assign a central role to the sexual dimension, which (and which alone) he qualifies as “a metaphor”, coupled with his fear of “reductionism”, prevented him from seeing the explicit and unambiguous connection between ånanda as orgasmic rapture and ånanda as the experience of brahman/åtman.

II.1 Ånanda is a somewhat rare word in the early vedic literature. In the ¥V it occurs in only two verses, both in a single hymn to Soma (9.113.6, 11). The first reads: yatra brahmå pavamåna chandasyå∫ våca∫ vadan Ù gråvñå some mahœyate somenånanda∫ janayann–indråyendo pari srava ÙÙ6

Van Buitenen translates: “O thou now purified, Soma in whom the brahman priest, while speaking the words of the hymns, rejoices with the pressing stone, generating bliss through Soma–swirl around for Indra, O drop!” From this translation it appears that the verse is a complete sentence, which it is not, and that the relative pronoun yatra refers to Soma (perhaps seeing it as in apposition to the locative some). In fact, the co-relative yatra is left dangling until we reach the word tasmin in the next verse (¥V 9.113.7): yatra jyotir ajasra∫ yasmi∫l loke svar hitam Ù tasmin må∫ dhehi pavamånåmr¢te loke akßita–indråyendo pari srava ÙÙ7 6 Geldner translates: “Wo, o Pavamåna, der Hohepriester in gebundener Rede sprechend mit dem Preßstein (in der Hand) bei Soma sich erhaben fhlt, durch den Soma Wonne wirkend, da fließe usw [fr Indra ringsum ab, o Saft!].” By syntactically connecting the final refrain with the rest of the sentence, Geldner also appears to make this a complete sentence. In other verses, however, he translates the refrain as a separate sentence. The word mahœyate, we will see, is often associated with ånanda, and in these contexts probably refers to an internal feeling of exultation. 7 Geldner translates: “Wo das ewige Licht ist, in welche Welt die Sonne gesetzt ist, in

78

LANGUAGE, TEXTS AND SOCIETY

Verse 6 begins a group of six verses, all containing the opening word yatra. In verses 7-11 the pronoun clearly refers to the heavenly world of light and immortality. It appears that a parallel is drawn between the sacrificial spot, the place where the soma is crushed with the stone, of verse 6, and the immortal world in heaven of verse 7, a connection that is quite common in the vedic literature. As in verse 6 the priest is said to “generate ånanda” at the place where soma is crushed, so in the final verse of this group (RV 9. 133.11) the heavenly world is said to contain ånandas (in the plural): yatrånandå† ca modå† ca mudaΔ pramuda åsate Ù kåmasya yatråptåΔ kåmås tatra måm amr¢ta∫ kr¢dhi–indråyendo pari srava ÙÙ8

The supplicant asks that he be made immortal in the place (yatra) where every wish is fulfilled and where there are ånandåΔ, modåΔ, mudaΔ, and pramudaΔ. The exact meanings here of these semantically related terms are unclear; they are clearly intended to intensify the impact, in a way similar to kåmasya kåmåΔ,9 and we will encounter these terms used together in later literature. Van Buitenen (1979, 324) thinks that ånanda in these verses refers to “the joyous state of (drug-induced) ecstasy in which the ecstatic may hope for immortality”. Such a meaning is certainly possible; later we will see ånanda associated with surå (liquor). Elsewhere, however, ånanda is most frequently associated with sexual pleasure, and it is not farfetched to see a similar intent here. The soma sacrifice has clear connections to sexuality and fertility, and the soma juice itself is compared to semen.10 The process of crushing and squeezing out the juice from the stalks by means of the pressing stones is full of sexual imagery. It is this process that “generates ånanda”, which, as we will see, is closely associated in later literature with the ejaculation of the semen and the generation of offspring. We can assume that at least part of the meaning and imagery of ånanda in these R¢gvedic verses is sexual. In the AV(‡) ånanda and its derivatives are used six times. Twice we have the same phrase: ånandå modåΔ pramudo ’bhœmodamuda† ca ye.11 At AV(‡) 11.7.26 these experiences are among the various elements of the universe that are said to originate from the ucchiß™a, the diese versetze mich, o Pavamåna, in die unsterbliche, unvergängliche Welt! Fließe fr Indra ringsum ab, o Saft!” 8 Geldner translates: “Wo Wonnen, Freuden, Lste und Belustigungen wohnen, wo die Wnsche des Wunsches erlangt werden, dor mache usw [mich unsterblich! Fließe fr Indra ringsum ab, o Saft!].” 9 On this type of intensification, see Oertel 1937. 10 reto vai somaΔ (“Soma is indeed semen”) ‡B 1.9.2.9. See, Jamison 1996a, 127-46. 11 Whitney translates: “Delights, joys, enjoyments, and they that enjoy enjoyments.” The parallels at AV(P) 16.84.8 reads ånandå† ca pramodå† cåbhimodamuda† ca ye; and AV(P) 16.87.4 reads nandåΔ for modåΔ.

Orgasmic Rapture and Divine Ecstasy

79

sacrificial remains, while at AV(‡) 11.8.24 they are among the various powers that entered the human body. The contexts of these verses do not provide clues as to the precise meaning of ånanda. Såyaña here, as in other places where the three terms moda, pramud (or pramoda), and ånanda are listed together, explains the first as pleasure derived from seeing an object, the second as pleasure derived from obtaining an excellent object, and the third as the pleasure derived from enjoying the object.12 Although Såyaña’s interpretation does not tell us much about what the terms may have meant in their original contexts, I think his instinct in taking the three as a progressive intensifying of pleasure is correct. And his connection of ånanda with the actual enjoyment of the desired object is borne out by evidence from its usage elsewhere, especially within the context of sexual activity. In AV(‡) 10.2.9 [=AV(P) 60.1], a hymn that asks repeated questions about who created the various bodily parts and powers of Purußa, we have the terms ånandån and nandån. Here too the context provides few clues as to their specific meanings. A clue, however, may be found in the VS(M) 20, which also has a long list of bodily organs and powers; verse 9 reads in part: ånandanandåv åñ∂au me bhagaΔ saubhågya∫ pasaΔ.13 The connection between ånanda and the sexual organs made explicitly in this passage will become a recurrent theme in the middle and late vedic literature. The connection is also made in the AV(‡) itself, in a hymn (9.7) celebrating the bull. The various organs and activities of the bull are identified with different deities and cosmic entities. Verse 23 reads: mitra œkßamåña åvr¢tta ånandaΔ.14 The meaning of “looking” and “turning this way [or around]” is not altogether clear. Såyaña, however, may have had this passage in mind when he provided the gloss on AV(‡) 11.6.3 (see note 17) saying that a bull in a herd looks at the cows and, desiring to mount them, bellows. Such a scene provides a plausible explanation for this looking15 and turning around and for why the turning around is called ånanda. If we interpret the above terms as a prelude to mating, then the passage that immediately follows [AV(‡) 9.7.24=AV(P) 16.139.25] may also have sexual implications: yujyamåno vai†vadevo yuktaΔ prajåpatir 12 modåΔ = vißayadar†anajanyå harßåΔ; pramudaΔ = prakr¢ß™avißayalåbhajanyå harßåΔ; ånandåΔ = vißayopabhogajanitåΔ sukhavi†eßåΔ. Commenting on TB 2.4.6.5, Såyaña offers the following definitions: sukhavi†eßo modaΔ Ù dar†anajanyaΔ pramodaΔ Ù bhogajanya ånandaΔ; and at TB 2.4.6.5 he states explicitly that the three words imply a gradation: modådayas trayas tåra(ta)myenåvasthitåΔ sukhåvåntaravi†eßåΔ Ù vårtåjanyaΔ sukhavi†eßo modaΔ Ù dar†anajanyaΔ pramodaΔ Ù bogajanya ånandaΔ Ù For triadic intensifications of this type, see note 32. 13 Griffith (1917) was too embarassed to translate this section, but Såyaña explains it unabashedly and straightforwardly. Here the man prays that his two testicles will have ånanda and nanda, and that his penis will have sexual pleasure and success. This phrase occurs also in MS 3.11.8; KS 38.4; TB 2.6.5.6. 14 Whitney translates: “Mitra when looking, delight when turned this way.” The AV(P) 139.24 reads: åvr¢tta ånanda œkßamåño mitråvaruñau. 15 A similar sexual implication of looking is found at ‡B(M) 6.2.2.6 discussed below, p. 83, n. 24.

80

LANGUAGE, TEXTS AND SOCIETY

vimuktaΔ sarvam, which Whitney translates: “Belonging to all the gods when being yoked, Prajåpati when yoked, everything when released.” But the verb √yuj can also mean to unite sexually, and if that is the case, the meaning would be: “He is All-gods when about to couple, Prajåpati when coupled, and the Whole when uncoupled.” In this context the connection of “union” with Prajåpati, the creator/procreator god, also makes sense. The two final examples from the AV(‡) contain the feminine adjective ånandinœ, “one possessing ånanda”, i.e., joyful. In a charm to produce rain, there is the wish “may plants become joyful”–ånandinœr oßadhayo bhavantu [AV(‡) 4.15.16 = AV(P) 5.7.14]. The reference is to the joy that plants, parched after a long period of drought, feel (metaphorically) when the rains come. The sexual connotation, I think, is still there, because rain is often equated with semen, and at AV(‡) 11.4.3 the process of plants absorbing rain-water is explicitly compared to mating: yat pråña stanayitnunåbhikrandaty oßadhœΔ Ù pra vœyante garbhån dadhate ’tho bahvœr vi jåyante ÙÙ 16

Såyaña17 makes the sexual imagery explicit: thunder looks upon the plants and cries out, like a bull that bellows at seeing the cows; then the plants, by merely hearing that sound, become impregnated, carry the fetus, and bring forth progeny. AV(‡) 4.38 is a good-luck charm for victory in gambling. The first part of the charm is addressed to Apsarå, who is said to “rejoice in the dice” (akßeßu pramodante). She is characterized in verse 4 as ånandinœ∫ pramodinœm. The context appears to indicate that these terms are used with reference to the pleasure of gambling, but the Apsaras are closely associated with sexuality. Apart from passages parallel to the AV(‡), there are only two independent uses of the term in the AV(P). At AV(P) 12.6.8, in an eulogy of rain similar to AV(‡) 4.15.16, the sun is said to produce ånanda– sürya ånanda∫ janayan. At AV(P) 16.152.11 also the sun is said to bring ånanda. Turning now to the mantra sections of the Yajurveda Sa∫hitås, the TS uses ånanda only once (TS 5.7.19), and it is unclear whether this passage is a mantra or a bråhmaña.18 This section of the TS deals 16 Whitney translates: “When breath with thunder roars at the herbs, they are impregnated, they receive embryos, then they are born many.” 17 Såyaña glosses abhikrandati: abhilakßya †abdåyate Ù yathå goyüthamadhye dr¢pto vr¢ßabhaΔ garbham ådhitsus tå abhilakßya †abda∫ karoti tathety arthaΔ. He glosses pra vœyante: pråñåbhikrandanamåtråd eva garbha∫ gr¢hñanti [...] varßatuΔ sarvåsåm oßadhœnå∫ garbhagrahañakåla ity arthaΔ. In the next two verses the author uses the verb pramodati to refer to the joy of plants and animals at the coming of rain. 18 Såyaña appears to take it as a mantra, while Keith (1914, 479, n. 1) thinks it is a bråhmaña passage, although he is not completely sure because of the mantra-like ending of TS 5.7.20 with svåhå repeated three times.

Orgasmic Rapture and Divine Ecstasy

81

with how various parts and powers of the sacrificial horse are connected to gods and cosmic entities. The passage in question reads: ånanda∫ nandathunå kåma∫ pratyåsåbhyå∫ bhaya∫ †itœmabhyå∫ pra†ißa∫ pra†åsåbhyå∫ –“ånanda with the penis; love with the two Pratyåsas; fear with the two ‡itœmans; command with the two Pra†åsas.”19 Here we come across for the first time the clear connection of ånanda with a corresponding physical organ, the nandathu, literally “the delighter” but clearly referring to the penis that provides ånanda. The VS(M) uses the term four times. In the mantras to be recited at the Sautråmañœ sacrifice, the liquor (surå) is addressed: eßa te yonir modåya tvånandåya två mahase två –“This is your place of birth. You for delight! You for ånanda! You for joy!” [VS(M) 19.8; = KS 37.18; TB 2.6.1.5]. Here, as in ¥V 9.113, the context is a drink but here we see more clearly another aspect of ånanda, the hilarity and mirth of drinking. The term is used again in the mantras of the Sautråmañœ (VS(M) 20.9),20 a passage we have already examined and where the sexual connotation of ånanda is explicit. Similarly explicit are the two other occurrences, both in 30th Adhyåya containing mantras for the Purußamedha sacrifice. In listing the various types of men to be sacrificed for various aims, the text reads: ånandåya strœsakha∫ pramude kumårœputram –“for ånanda a women’s friend, for pleasure the son of an unmarried woman” [VS(M) 30.6=TB 3.4.2.1]. It is unclear what strœsakha precisely means; given the context of dance, song, and sex, it is at least a possibility that it refers to a libertine, a man all women run after. Further down the list we have vœñåvåda∫ påñighna∫ tüñavadhma∫ tån nr¢ ttåyånandåya talavam –“Lute player, hand clapper, flutist–these for dance; for ånanda a musician” [VS(M) 30.20]. In the VS, then, ånanda, besides its sexual meanings, is used with reference to the pleasure associated with drinking, dancing, and music. Taken together with the AV(‡) usage with regard to the Apsaras engaged in the game of dice, we see a pattern emerging in the early vedic literature of ånanda being associated with sex, gambling, drinking and dancing.

II.2 I turn now to the middle vedic texts represented by the Bråhmañas. With the exception of a single passage in the Kaußœtaki (=‡åõkhåyana 2.7), the term is used extensively only in two Bråhmañas, both belonging to the Yajurveda: the ‡atapatha and the 19 The meanings of these pairs are unclear. Såyaña merely states that they are pairs of organs near the sexual organ: guhyasamœpavartœny avayavayugalåñi. The A†vamedha section (13.9) of the KS reads nandathubhyåm. 20 The passage occurs also in MS 3.11.8; KS 38.4; TB 2.6.5.6.

82

LANGUAGE, TEXTS AND SOCIETY

Taittirœya. Although both are somewhat late texts, the ‡atapatha is probably the older of the two. Leaving aside for the moment the BU, which constitutes its final chapters, the ‡B(M) uses the term six times. The meaning of ånanda is most clear and explicit at ‡B(M) 10.5.2.11. This section deals with the connections between the sun, on the one hand, and the sacrifice and the body, on the other. With regard to the body (10.5.2.7-9), the golden person in the sun’s orb (mañ∂ale purußaΔ) and Indra are in turn identified with the person in the right eye (dakßiñe ’kßan purußaΔ), and the mate of the person in the sun’s orb and Indråñœ, the wife of Indra, are in turn identified with the person in the left eye.21 The male and the female persons in the right and left eyes remain out of each other’s sight by the partition created by the nose (10.5.2.9; cf. TS 2.3.8.2). During sleep, however, the two descend into the cavity or space within the heart and unite with each other; at the climax of this union the two become in some way unconscious (petit mort of orgasm) and in this rapture experiences the highest ånanda: tau hr¢dayasyåkå†a∫ praty avetya mithunœbhavatas tau yadå mithunasyånta∫ gacchato ’tha haitat purußaΔ svapiti tad yathå haiveda∫ månußasya mithunasyånta∫ gatvåsa∫vida iva bhavaty eva∫ haivaitad asa∫vida iva bhavati daiva∫ hy etan mithuna∫ paramo hy eßa ånandaΔ Ù ‡B(M) 10.5.2.11. The two descend into the space within the heart and engage in sexual intercourse. And when the two reach the climax22 of the sexual intercourse, the man here is then asleep. It is like this. As here when one reaches the climax of a human sexual intercourse one becomes in some way unconscious, so there he becomes in some way unconscious, for that is the divine sexual intercourse, for that is the highest ånanda.

Here ånanda refers clearly and explicitly to the orgasmic thrill that makes one lose one’s consciousness. In the very next paragraph (10.5.2.12), in fact, the text goes into further detail, stating that a) one should not awaken a sleeping man suddenly or violently, lest one disturb the sexual union of these two, and b) the mouth of people who have slept are slimy (†leßmaña) because these two have spilled their seed, thus comparing the slimy spit to the slimy semen. The connection between ånanda and orgasm is further established in ‡B(M) 6.2.2.6. In explaining why a white and hornless male goat is offered to Våyu Niyutvat, the text narrates the story of Prajåpati. After creating the creatures (prajåΔ, feminine), he looked (anuvyaikßata) at them and because of the excessive joy (atyånandena) he spilled his seed, which became the white goat. His orgasmic joy comes here not at the time of creation but afterwards when he looks These are identifications familiar also in the Upanißads: BU 4.2.2; 5.5.2 Eggeling translates anta as “end”, which misses the point. The anta of sex is not just the end but the climactic orgasm. He also misses the point when in a footnote he explains that “unconscious” means something like “indifferent, apathetic”. 21

22

Orgasmic Rapture and Divine Ecstasy

83

at his creatures. The use of the feminine prajå gives us a clue; he lusted after his own daughter(s). Prajåpati’s incest is frequently mentioned in the Vedas, and the story of Prajåpati lusting after his daughter is told repeatedly in the vedic literature.23 The “looking” by Prajåpati could also have sexual implications; looking at the melted butter by the wife of the sacrificial patron, for example, is viewed as sexual intercourse between the wife (woman) and butter (semen).24 These usages permit us to interpret the other four occurrences of ånanda in the ‡B(M), all in the single passage 10.3.5.12-14. tasya vå etasya yajußaΔ Ù rasa evopanißat tasmåd yåvanmåtreña yajußådh varyur graha∫ gr¢hñåti sa ubhe stuta†astre anuvibhavaty ubhe stuta†astre anuvya†nute tasmåd yåvanmåtra ivånnasya rasaΔ sarvam annam avati sarvam annam anuvyeti ÙÙ 12 ÙÙ tr¢ptir evåsya gatiΔ Ù tasmåd yadånnasya tr¢pyaty atha sa gata iva manyata ånanda evåsya vij∞ånam åtmånandåtmåno haiva sarve devåΔ så haißaiva devånåm addhåvidyå sa ha sa na manußyo ya eva∫vid devånå∫ haiva sa ekaΔ ÙÙ 13 ÙÙ etad dha sma vai tadvidvån priyavrato rauhiñåyana åha Ù våyu∫ våntam ånandas ta åtmeto vå våhœto veti sa ha sma tathaiva våti tasmåd yå∫ deveßv å†ißam icched etenaivopatiß™hetånando va åtmåsau me kåmaΔ sa me samr¢dhyatåm iti sa∫ haivåsmai sa kåma r¢dhyate yatkåmo bhavaty etå∫ ha vai tr¢ptim etå∫ gatim etam ånandam etam åtmånam abhisa∫bhavati ya eva∫ veda ÙÙ 14 ÙÙ 12. Now, of this yajus-formula, the hidden connection (upanißad) is flavor [or essence]. Therefore, when with ever so small a yajus-formula the Adhvaryu draws a cup of Soma, it [rasa] becomes equal to both the Stotra and the ‡astra, it measures up to both the Stotra and the ‡astra. The flavor of food, therefore, be it ever so small, enhances the entire food, pervades the entire food. 13. Its [of the yajus] completion is satiation. Therefore, when with food a man reaches satiation, then he considers himself in some way wiped out.25 Its body is ånanda—this is its true knowledge.26 For, indeed, all the gods have ånanda as their body. This, indeed, is the true knowledge of gods. 23 MS 4.2.12; AB 3.33; ‡B(M) 1.7.4. In the BU (1.4.3-4) there is the story of the first being (often identified with Prajåpati) who split himself in two, into pati (husband) and patnœ (wife). He copulated with her, producing the humans. But the woman thinks “After begetting me from his own body, how could he copulate with me?” She hid herself by becoming various animals, with all of whom he copulated, thus giving rise to the various kinds of animals. For an extensive discussion of Prajåpati’s incest, see O’Flaherty 1973, 111-40. 24 See ‡B 1.3.1.18. This topic is studied exhaustively by Jamison 1996a, 55-59. See also AV(‡) 9.7.23 discussed above. 25 The term gati (completion) indicates probably the progress and the completion of the progress of the yajus. In the earlier part of this section (‡B(M) 10.3.5.1-7) dealing with the etymology of yajus, it is repeatedly connected with motion. In the final image of a man who has eaten a lot, there appears to be a pun on gati. The man then feels as if he is gata, that is, “I’m gone” or “I’m wiped out”. 26 The nominal sentence ånanda evåsya vij∞ånam åtmå is problematic. In other nominal sentences with three nouns (A. B. C), such as CU 6.1.4-6, the most likely syntax is: B is A, (namely) C. Thus in the CU examples våcårambhaña∫ vikåro nåmadheyam is

84

LANGUAGE, TEXTS AND SOCIETY

And anyone who knows this is not a man; he is truly one of the gods. 14. Knowing this, indeed, Priyavrata Rauhiñåyana said to the wind as it was blowing: ‘Your body is ånanda. Blow this way or that way!’ And thus, indeed, does it blow. Therefore, a man who wishes to obtain a blessing from the gods should worship them within this: ‘Your body is ånanda. Here is my wish. May it be fulfilled for me!’ And whatever wish he may have, it will surely be fulfilled. A man who knows this will obtain this contentment, this fulfillment, this ånanda, and this body.

Here the yajus formula is compared to food. Of this yajus-food, the upanißad is the flavor (rasa),27 the completion (gati) is the satiation (tr¢pti), and the body (åtman) is ånanda. The text goes on to state that all the gods have ånanda as their åtman, providing the earliest evidence of ånanda being used as an essential attribute of gods. This knowledge of the essence of gods appears to be a secret, the knowledge of which gives a man power over the gods. Thus, if one tells the gods “Your åtman is ånanda,” one’s wishes will be fulfilled. The passage ends by saying that anyone who has this knowledge attains tr¢pti, gati, ånanda, and åtman; in this ascending hierarchy, the åtman that the man will obtain is clearly the åtman that consists of ånanda; that is, he becomes a god. This is a difficult passage. Eggeling’s translation compounds the difficulties, and I do not pretend to have solved all of them. Upanißad here means connection/equivalence, showing how the yajus is equal to the other ritual utterances, just as the flavor (rasa) permeates the food. The implication is that if it is the rasa, then even a small amount can surpass things that are much larger. The yajus is brief in comparison to the Stotra (Såmavedic chant) and the ‡astra (R¢gvedic recitation) that follow each other at a Soma sacrifice. The phrase ånanda evåsya vij∞ånam åtmå Eggeling translates as: “And joy, the knowledge thereof (viz. of the essence, the mystic import), is its soul (self)”. Clearly this is inaccurate. The subject is åtman, and vij∞ånam is probably a parenthetical statement. Then the åtman (which in this context probably means body) of the gods, just as the åtman of the yajus, is said to consist of ånanda, and this knowledge gives magical power to the knower.28 At the surface level translated: “The transformation is a verbal handle, a name.” In the present context, however, vij∞ånam appears as an intrusion both because the two parallel sentences in this structured series of identifications have only two nouns and because vij∞ånam is quickly dropped from the discussion; even in the final enumeration in § 14 it is omitted. I have followed Stephanie Jamison’s (private communication) suggestion that vij∞ånam may be a parenthetical comment and not part of the equation. The term then refers to this “knowledge” or “science”, i.e., the knowledge that “åtman is ånanda”, which the gods and Priyavrata Rauhiñåyana possessed. 27 Contrary to Eggeling, the subject of the nominal sentence is upanißad and the predicate is rasa; likewise, at the beginning of paragraph 14, gati is subject and tr¢pti the predicate (see Gren-Eklund 1978). Furthermore, I think that the phrase ånanda evåsya vij∞ånam åtmå is the third in the list, paralleling rasa and tr¢pti. 28 We have a similar connection between rasa. ånanda, and åtman in the TU 2.1-7.

Orgasmic Rapture and Divine Ecstasy

85

ånanda in this passage probably refers to the relishing of the flavor of food. But in this literature there are clear connections between food/eating and sex, and here the terms rasa and gati can have a double entendre, the former meaning semen (see TU 2.7 discussed in II.3) and the latter meaning “going”, i.e., sexual congress. And tr¢pti, as we will see in TB 2.4.6.5, also has sexual connotations. The author appears to be playing on this double meaning which permits him to say that gods have ånanda as their åtman, reminiscent of ‡B(M) 10.5.2.11 that speaks of “divine sexual intercourse” which is the highest ånanda. If we exclude the three passages (TB 2.6.1.5; 2.6.5.6; 3.4.2.1) occurring also in the VS, the TB uses ånanda four times. The sexual connotation is most explicit at TB 2.4.6.4-5, where the term is used twice: prajåpatiΔ striyå∫ ya†aΔ mußkayor adadhåt sapam Ù kåmasya tr¢ptim ånandam tasyågne bhåjayeha må ÙÙ modaΔ pramoda ånandaΔ mußkayor nihitaΔ sapaΔ Ù sr¢tveva kåmasya tr¢pyåñi dakßiñånå∫ pratigrahe ÙÙ Prajåpati put the penis in the vagina,29 the glory in the woman –the satisfaction of desire, the ånanda. O Fire, make me here partake of that! The penis is put in the vagina –the joy, the thrill, the ånanda, flowing somehow (with semen) toward the satisfactions of desire in accepting the sacrificial gifts.30

In this eulogy of the pride of masculinity, ånanda, as well as the two associated terms moda and pramoda,31 are identified with the penis placed within the vagina, the penis that brings the satisfaction (tr¢pti) of desire. Moda, pramoda, and ånanda32 appear as names of three of the fifteen muhürtas of a night at TB 3.10.1.1. The final example is from TB 2.4.5.7: indra† ca naΔ †unåsœrau ima∫ yaj∞a∫ mimikßatam Ù 29 Såyaña takes the dual mußkayor as referring to the testicles (añÛa), but the context, I think, suggests the labia majora (or minora) of the vagina. The term is used with the meaning of labia in VS(M) 23.28 (discussed by Jamison 1996, 71 and 276, n. 134); KßB 23.4; BU 6.3.3 (where the term appears to refer to the labia minora). 30 According to Såyaña, the sacrificial gift here refers to the practice of giving a virgin to the officiating priest as a dakßiñå. The exact meaning and syntax of sr¢två is unclear. 31 Both moda and pramoda are used with sexual connotations. Thus ¥V 10.30.5 says that Soma frolics with the waters (feminine) like a man with young girls: yåbhiΔ somo modate harßate ca kalyåñœbhir yuvatibhir na maryaΔ Ù At ¥V 10.10.12 Yama tells Yamœ, his sister, that she should not have sex with him but with some other man: anyena mat pramudaΔ kalpayasva. 32 This sequence clearly probably falls into the triadic intensifying device known as “Behaghel’s Law”, i.e., the law of increasing elements: Otto Behagel, “Beziehungen zwischen Umfang und Reihenfolge von Satzgliedern,” Indogerm. Forschungen 25 (1909-10), 110-42. This intensification consists in a progressively larger number of syllables or morae in the three words. An example from American English would be “life, liberty, and the pursuit of happiness”. In moda, pramoda, ånanda, we have a progressive increase in morae: 3, 4, and 5. I want to thank Mark Southern for pointing this out.

86

LANGUAGE, TEXTS AND SOCIETY

garbha∫ dhatta∫ svastaye ÙÙ yayor ida∫ vi†va∫ bhuvanam åvive†a yayor ånando nihito maha† ca Ù †unåsœråv r¢tubhiΔ sa∫vidånau indravantaethåm ÙÙ 33 Indra and you, O ‡una-Sœra, mix [prepare] this sacrifice for us and place the fetus for prosperity. Together with Indra and in agreement with the Seasons, O ‡una-Sœras, be pleased with this offering, you into whom this whole world has entered and in whom ånanda and exultation have been placed.

In this hymn to the obscure dual deity ‡una-Sœra,34 ånanda and mahas, a term often associated with ånanda, are said to be placed or deposited in these two deities. ‡una and Sœra are identified by Yåska with Våyu and Åditya, but originally they were probably agricultural deities, possibly personifications of the plow and the share. Although the passage is obscure, we can detect here the same type of agricultural metaphor that was found in AV(‡) 4.15.16, where plants are said to be joyful (ånandinaΔ) when it rains. There could also be a sexual imagery in the “union” of the plow and the share (or plow and the plowman), the union that produces ånanda leading to agricultural abundance. In a somewhat unclear passage of the Kaußœtaki Bråhmaña (7.2), ånanda is associated with three things, food, drink, and sexual intercourse: yaivaike cånandå anne påne mithune råtryå eva te sa∫tatå avyavacchinnåΔ kriyante Ù teßå∫ råtriΔ kårotaraΔ Ù ya u vaike cånandå annåd eva te sarve jåyante Ù “Whatever joys that are in food, drink, and sexual intercourse, all those are joined together without interruption through the night; for them the night is the sieve. Whatever joys there are, they are born from food.” Although here ånanda is said to be derived from food, the same passage goes on to state that the essence (rasa) of food gives rise to semen (retas) and the essence of semen 33 This is probably a Gåyatrœ verse, followed by a Triß™ubh. The meter of påda a of the Triß™ubh can be restored by dropping idam. This verse is an adaptation of the common verse-type beginning våyav indra† ca. Jamison (1988, 14) formulates succinctly the grammatical rule followed by this construction: “two vocatives may not be conjoined by ca; in place of the second expected vocative, another case will appear. In Vedic, this is always the nominative.” Our verse follows the “inverted construction” where the nominative is placed first (see Jamison 1988 for further bibliography and a detailed discussion of this construction). Theoretically, there should be a plural verb ending here, because three deities are addressed. The dual ending of the verb (mimikßatam) probably follows the stereotype of such constructions where two deities are addressed, generally Indra and another deity (usually Våyu). 34 The two are mentioned in ¥V 4.57.5, 8. The final Cåturmåsya sacrifice in the autumn is called ‡unåsœrœya and offered to these two deities, clearly indicating their association with agriculture and the bounty of the harvest: see ‡rauta Ko†a (Poona: Vaidika Sa∫†odhana Mañ∂Ûala, 1962), I.2: pp. 759-63, 895-98. In the ‡B (7.2.2.5) there is an explicit connection between plowing and sex: the plow (=penis) makes the furrow (=womb) and deposits the seed (=semen) in it, for “if one casts (seed) into unplowed land, it is the same as depositing semen in a place other than the womb.”

Orgasmic Rapture and Divine Ecstasy

87

gives rise to man. Here too, then, food and semen are closely associated with each other and with ånanda.

II.3 The literature of the late vedic period contains the most significant semantic developments of ånanda, developments that had a profound influence on later Indian theological vocabularies. The evidence for these developments come principally from the early Upanißads, the term occurring only once in the Årañyakas. Among a series of mantras for the Pitr¢medha, we read at TA 6.11.2: ånandåya pramodåya punar ågå∫ svån gr¢hån –“For bliss, for delight I have returned to my home.” The context offers no clues, but we have seen the two terms ånanda and pramoda frequently used together. Of the early Upanißads, ånanda is most prominent in the two Yajurvedic documents, the Br¢hadårañyaka and the Taittirœya, and to a somewhat lesser extent in the Kaußœtaki, which belongs to the R¢gveda.35 I will deal with the three together because all three reflect parallel semantic developments of the term. Continuing, and extending, the trend already noticed in the earlier literature, these Upanißads present ånanda as the faculty or power of the sexual organ parallel to the sensory and motor faculties associated with other organs, e.g., seeing with the eyes, hearing with the ears, and motion with the feet. Although the term upastha can refer to both male and female sexual organs, it is clear that these texts deal with the male rather than the female orgasm; upastha in these contexts undoubtedly refers to the penis. The BU 2.4.11 (=BU 4.5.12) presents the sexual organ as the point of convergence of all ånanda (eva∫ sarveßåm ånandånåm upastha ekåyanam), in the same way as the ocean is of all waters, skin of all sensations of touch, and sight of all visible appearances. A similar association is made in the TU 3.10.3: prajåtir amr¢tam ånanda ity upasthe. The meaning of the elliptical phrases in this passage is far from clear, but probably the sense is that one should “venerate”, that is, perceive the equivalence of, brahman in the sexual organ as “procreation, immortality, and ånanda.” Here we have an interesting coupling of immortality and ånanda, a connection that becomes important when brahman comes to be defined as ånanda. In the present context, it is procreation that links ånanda to immortality: ånanda, the ejaculatory bliss, precedes procreation, and sons are identified with immortality in the early vedic literature.36 35

The transmission of the KßU, however, has been much less faithful than that of the other early Upanißads. It is, therefore, difficult to make firm conclusions from the presence of the term in the KßU, which may have been influenced by the Yajurvedic documents. Such an assumption is supported by the fact that neither the parallel documents of the Aitareya †åkhå (AB, AA, and AU), nor the Kaußœtaki Bråhmaña contain the word ånanda. 36 See, for example, ¥V 5.4.10: “Through offspring, O Agni, may we attain immortality”–prajåbhir agne amr¢tatvam a†yåm. See Olivelle 1993, 41-46.

88

LANGUAGE, TEXTS AND SOCIETY

In explaining the pre-eminence of intelligence (praj∞å), the KßU shows how all human powers and the objects in the world corresponding to those powers are derived from intelligence. In this context KßU 3.5 states: upastha evåsyå ekam aõgam udü¬am Ù tasyånando ratiΔ prajåtiΔ paraståt prativihitå bhütamåtrå Ù The sexual organ is one part drawn from it [i.e., from intelligence], and ånanda, delight, and procreation constitute the particle of being that corresponds externally to it.

Here, instead of the amr¢ta of the TU, we have the more usual rati associated with ånanda and procreation. The subsequent paragraphs (KßU 3.6-8) make the same associations: praj∞ayopastha∫ samåruhyopasthenånanda∫ rati∫ prajåtim åpnoti Ù When someone mounts the sexual organ by means of intelligence, he grasps ånanda, delight, and procreation through his sexual organ. na hi prajåpeta upastha ånanda∫ rati∫ prajåti∫ kå∫cana praj∞åpayed anyatra me mano ’bhüd ity åha nåham etam ånanda∫ na rati∫ na prajåti∫ pråj∞åsißam iti Ù For without intelligence, the sexual organ would not make someone perceive any ånanda, delight, or procreation. So, one says: ‘My mind was elsewhere, I did not perceive that ånanda, delight, or procreation, nånanda∫ na rati∫ na prajåti∫ vijij∞åsœtånandasya rateΔ prajåter vij∞åtara∫ vidyåt Ù It is not the ånanda, delight, or procreation that a man should seek to apprehend; rather, he should get to know the one who apprehends ånanda, delight, or procreation.

In KßU 1.7 Brahman asks the man who has managed to arrive in the world of Brahman a variety of questions centering on how the man will grasp various objects. Thus, odors are grasped by the sense of smell, visible objects by sight, tastes by the tongue, actions by the hands, and so on. Brahman asks: kenånanda∫ rati∫ prajåtim iti –“(By what means do you grasp my) ånanda, delight, and procreation.” The man replies: upastheneti –“By my sexual organ.” Even though the sexual organ is not explicitly mentioned, the list37 of the father’s powers that he assigns to his son during the poignant ceremony of transmission when the father is about to die contains the same three powers: ånanda, rati, and prajåti (KßU 2.15). 37 The list of these powers is the same as the one repeated four times together with the corresponding organs at KßU 3.5-8. This long list is absent in the abbreviated ceremony recorded in BU 1.5.17.

Orgasmic Rapture and Divine Ecstasy

89

During this ceremony the son lies on top of the father, every limb and organ of his touching the corresponding organ of the father. Clearly, the transfer is from the organs of the father to those of the son, and though unstated the transfer of ånanda, rati, and prajåti must be from the father’s penis to that of the son. These Upanißads thus make a clear and explicit connection between ånanda and the penis on the one hand, and between ånanda and procreation, on the other. As opposed to the sexual organ, the BU 4.1.6 connects ånanda with the mind. The context is a long conversation between King Janaka and Yåj∞avalkya during which Janaka recounts what different teachers had told him about brahman. One of them had said that brahman is the mind. Yåj∞avalkya asks: “But did he tell you what its abode and foundation are?” He had not, and Janaka asks Yåj∞avalkya himself to tell him that. Yåj∞avalkya responds: mana evåyatanam åkå†aΔ pratiß™hånanda ity enad upåsœta –“The mind itself is its abode, and space is its foundation. One should venerate it as (i.e., take brahman to be) ånanda.” Janaka asks: kånandatå –“What is ånanda-ness (i.e., what constitutes ånanda)?” Yåj∞avalkya replies that is it the mind itself (mana eva) and gives the reason: manaså vai samrå™ striyam abhihåryate tasyå∫ pratirüpaΔ putro jåyate sa ånandaΔ Ù mano vai samrå™ parama∫ brahma Ù For surely, Your Majesty, it is with the mind that man takes a woman to himself and through her fathers a son who resembles him. And that is ånanda. So clearly, Your Majesty, the highest brahman is the mind.

Here we have a connection established between mind, space, and ånanda in a set of equations: between mind and ånanda, between mind and brahman, and, hence, between ånanda and brahman. Ånanda, moreover, is defined as the mind because it is through the mind that one takes a woman (or wife) and begets a son through (literally, in) her. Here it appears that the entire process of begetting a son is defined as ånanda. The connection between ånanda and brahman is established here through the son, in a way similar to TU 3.10.3 where procreation appears to be the link between ånanda and immortality. Further, the foundation of this brahman is space. The reference is probably to the space within the heart which, as we saw (‡B(M) 10.5.2.11), is associated with sexual activity in the context of sleep. And at BU 1.4.3 it is said that “the space here is completely filled by the woman” (ayam åkå†aΔ striyå püryata eva). This connection between ånanda and the mind throws light on the passage that follows the ceremony of transmission from a dying father to his son at BU 1.5.19. Among the divine faculties that enter the father after this ceremony is the daiva∫ manas, the divine mind. And this divine mind is defined as: tad vai daiva∫ mano yenånandy eva bhavaty atho na †ocati –“The divine mind is that by which he is just

90

LANGUAGE, TEXTS AND SOCIETY

(always) joyful (ånandin) and thereafter is never sorrowful.” We have seen in ‡B(M) 10.3.5.12-15 that the åtman of divine beings (deva) is said to be ånanda. The connection between ånanda and space in BU 4.1.6 permits us to interpret a rather difficult passage in TU 2.7 which immediately precedes the exegesis of ånanda discussed below. The author cites a verse: asad vå idam agra åsœt tato vå sad ajåyata Ù tad åtmåna∫ svayam akuruta tasmåt tat sukr¢tam ucyate Ù In the beginning this world was the non-existent, and from it arose the existent. By itself it made a body for itself; therefore it is called “well-made”.

Now, “well-made” (sukr¢ta) is an epithet that is used with reference to the human body; for example, in AU 1.2 the human body is distinguished from those of animals precisely because it is “well-made.” In ‡B (8.6.2.18) the body of Agni created in the Agnicayana ritual is called “well-made”, and the body of the sacrificer is likewise “wellmade”. But the TU appears to be playing on the two words svaya∫ √kr¢ and su-√kr¢; the body is “well-made” (sukr¢ta) because it is “self-made” (svaya∫kr¢ta). How does a man make his own body? In a similar context, the AU (2), speaking of the births of a man, declares that the semen (retas) is one’s very self in the form of an embryo (garbha); and a man carries this seminal self of his within himself and later deposits it in a woman. This depositing of semen is his first birth.38 When the semen has developed into a fetus and the woman gives birth, that is his second birth. It appears likely that the “self-made” nature of the body is connected to the fact that a man carries himself within himself in the form of his semen. This helps us understand the remainder of the TU (2.7) passage: yad vai tat sukr¢tam Ù raso vai saΔ Ù rasa∫ hy evåya∫ labdhvånandœ bhavati Ù ko hy evånyåt kaΔ pråñyåt Ù yad eßa åkå†a ånando na syåt Ù eßa hy evånandayåti Ù yadå hy evaißa etasminn adr¢†ye ’nåtmye ’nirukte ’nilayane ’bhaye39pratiß™hå∫ vindate Ù atha so ’bhaya∫ gato bhavati Ù yadå hy evaißa etasminn u daram40 antara∫ kurute ’tha tasya bhaya∫ bhavati Ù

The first problem is the meaning of rasa. It has been generally translated as “essence”, but that meaning does not make much sense within the context. Rasa has numerous meanings, including essence, seminal fluid, taste, and pleasure/desire. The author is probably playIn this sense, then, the connection between ånanda and prajåti can have another meaning. The ejaculation of the semen in ånanda is itself the prajåti or the self-procreation of the man. 39 So with Rau (1981). The vulgate reads ’bhaya∫. 40 So with Rau (1981). The vulgate reads udara∫. 38

Orgasmic Rapture and Divine Ecstasy

91

ing on the ambiguity of the term. At TU 2.1, for example, a man is said to be made of the essence of food (sa vå eßa purußo ’nnarasamayaΔ). The “well-made/self-made” condition of the human body consists in its being (or that it comes from) rasa; when one obtains this rasa one attains ånanda. The sexual implications of rasa here are unmistakable, and I think that here it means the “essence” of the human body, that is, the seminal fluid.41 The second problem concerns the somewhat mysterious statement about ånanda existing in space. The connection between mind, space, and ånanda in the above BU (4.1.6) passage, and the ånanda produced by the sleeping person when sexual intercourse takes place within the space in the heart (‡B 10.5.2.11-12), I think, permits us to see here a similar connection: ånanda is present, i.e., is able to be grasped, in the space of the heart. Indeed, the TU (1.6.1) itself locates the mind (manas) in the heart: sa ya eßo ’ntarhr¢daya åkå†aΔ tasminn aya∫ purußo manomayaΔ amr¢to hirañyayaΔ –“In this space here within the heart lies the immortal and golden person consisting of the mind”. Brahman, moreover, is said to reside within the cavity of the heart (CU 8.1). In this ånanda one loses the consciousness of being separate (see below BU 2.1.19; 4.3.21), one does not perceive even a smallest difference (daram antaram); this state of consciousness brings a man to the state of abhaya. We can then translate the TU 2.7 as follows: That which is well-made(=self-made) is nothing but semen, for when a man here obtains the semen, he comes to possess ånanda. Now, who would breathe in, who would breathe out, if this ånanda were not there in space; for that alone can grant ånanda. For when a man finds his support within that which is invisible, incorporeal, indistinct, supportless, and free from fear, then he becomes free from fear. For only when he creates even a small difference does fear come upon him.

We have seen how sleep was identified in the ‡B (10.5.2.11-12) with the persons in the right and left eyes having sexual intercourse within the space of the heart. The unconsciousness of sleep was there also compared to the loss of consciousness in orgasm. This theme is taken up again in the BU 2.1.19, where deep dreamless sleep is opposed to the state of dream when the person is in some way still conscious, except that the dream consciousness is creative, creating rather than perceiving its objects. In deep sleep, however, one is not aware of anything; during this time the self slips out of the space of the heart and rests in the pericardium: sa yathå kumåro vå mahåråjo vå mahåbråhmaño våtighnœm ånandasya gatvå †ayœta evam evaißa etac chete Ù 41 The AU (2.1), for example, calls semen (retas) “the radiance gathered from all the bodily parts” (tad etat sarvebhyo ‘õgebhyas tejaΔ sa∫bhütam). In BU 6.4.1 semen is called the

92

LANGUAGE, TEXTS AND SOCIETY

He rests there, just as a young man, a great king, or an eminent Brahmin rests after attaining the height of ånanda.

The term atighnœ is significant; in this context it must mean more than just the “summit of bliss” but to the apparent loss of awareness resulting from orgasmic bliss.42 Otherwise the comparison makes little sense, since the point the author wants to make is that in deep sleep a person enjoys bliss but is not conscious of anything. This meaning also corresponds to the way sleep is described in ‡B 10.5.2.11-12. The same metaphor is used to describe deep sleep also at BU 4.3.21 with a clearer statement about the loss of consciousness: tad yathå priyayå striyå sa∫parißvakto na båhya∫ ki∫cana veda nåntaram evam evåya∫ purußaΔ pråj∞enåtmanå sa∫parißvakto na båhya∫ ki∫cana veda nåntaram Ù It is like this. As a man embraced by a woman he loves is oblivious to everything within or without, so this person embraced by the self consisting of knowledge is oblivious to everything within or without.

The term ånanda is not used here and there is no direct reference to orgasm, but the reference is clearly to the oblivion created by the height of sexual bliss. The related term abhinanda43 is used for orgasm in the wellknown passage on the doctrine of five fires where the sexual organ of a woman and the sexual act performed in it are compared to a sacrifice (BU 6.2.13 = CU 5.8.1): yoßå vå agnir gautama Ù tasyå upastha eva samil lomåni dhümo yonir arcir yad antaΔ karoti te ’õgårå abhinandå visphuliõgåΔ Ù tasminn etasminn agnau devå reto juhvati Ù The fire is a woman, Gautama. Her firewood is the vulva; her smoke is the pubic hair; her flame is the vagina; when one penetrates her, that is her embers; and her sparks are the climax. In that very fire gods offer semen.

The dreaming state is connected with ånanda at BU 4.3.9: the dreaming person sees påpmana ånandå∫† ca, “both bad things and ånandas.” What these ånandas are the following passage explains (BU 4.3.10), describing how the dreamer creates his own dreamland: na tatrånandå mudaΔ pramudo bhavanti Ù athånandån mudaΔ pramudaΔ sr¢jate Ù essence of man: purußasya retaΔ (rasaΔ). Rau (1981) translated rasa as “der Lust gewährt”. 42 On the meaning and etymology of atighnœ, see M. A. Mehendale, “Some Lexicographical Notes on the Upanißads,” Indo-Iranian Journal 5 (1962), 184-86. 43 This term does not appear elsewhere in the early Upanißads. In the Bråhmañas the

Orgasmic Rapture and Divine Ecstasy

93

In that place there are no ånandas, pleasures, or delights; but he creates for himself ånandas, pleasures, and delights.

Later the text explains what these pleasures are (BU 4.3.13): svapnånta ucchåvacam œyamåno rüpåñi devaΔ kurute bahüni Ù uteva strœbhiΔ saha modamåno jakßad utevåpi bhayåni pa†yan ÙÙ Travelling in sleep to places high and low The god creates many a visible form– now dallying with women, now laughing, now seeing frightful things.

The påpmanaΔ and ånandåΔ of the first passage probably parallel the dallying with women (moda, we have seen, is used frequently with ånanda) and bhayåni of this passage. A long and interesting passage occurring in BU 4.3.32-33 identifies the world of brahman as one’s highest ånanda (eßo ’sya parama ånandaΔ) and goes on to explain how vast this ånanda is in comparison to other types of ånanda, beginning with what we are most familiar with: sa yo manußyåñå∫ råddhaΔ samr¢ddho bhavaty anyeßåm adhipatiΔ sarvair månußyakair bhogaiΔ sa∫pannatamaΔ sa manußyåñå∫ parama ånandaΔ Ù Among human beings, when someone is successful and rich, ruling over others and enjoying to the utmost all human pleasures–that is the highest ånanda of human beings.

The texts goes up the ladder of greater ånandas, each higher ånanda being a hundred times greater than the one below it. Here ånanda is not directly connected to sex but to the broader category of bhoga or pleasures. A similar gradation of ånanda is described in TU 2.8 in a passage entitle ånandasya mœmå∫så, analysis or exegesis of ånanda. Perhaps the most famous of the ånanda passages of the TU is the one on the five bodies (TU 2.2-5), which later literature identifies as sheaths (ko†a). A man has five bodies or selves (åtman) consisting of food (anna), breath (pråña), mind (manas), perception (vij∞åna), and finally ånanda, each surrounding the previous like layers of an onion. Of each such body, the text identifies the head, the two sides, trunk, and the bottom. In the case of the ånandamayåtman, the head is priya, the right side is moda, the left is pramoda, the trunk is ånanda, and the bottom is brahman. 44 Here again we come across the three terms moda, pramoda, and ånanda. The portion of the body I have term occurs only once, in JB 1.45 in the parallel description of the five fires. Verbal forms of the word occur only twice in the vedic literature, AV(‡) 9.2.2; 19.8.3. 44 These parts of the body derive from the image of a bird, the bottom being the tail. The bird image comes from the shape of the fire-altar built with bricks. For a comparison

94

LANGUAGE, TEXTS AND SOCIETY

translated as trunk (or torso) is called åtman; so here we find that åtman is ånanda, and the passage from åtman as the central part of the body to åtman as one’s essential self is easy. So, we find the two major concepts of the Upanißads, åtman and brahman identified as ånanda. At the end of the TU (3.10.5) these five åtmans are presented as the path that a person travels after death: he first reaches the åtman of food, then that of breath, then that of the mind, then that of perception, and finally the åtman of ånanda. The simple statement that brahman is ånanda is found in both the BU (3.9.28) and in the TU (3.6). When we turn to the other two early prose Upanißads, the Chåndogya and the Aitareya, it comes as a surprise to find the term ånanda almost absent from their vocabularies. It is completely absent not only in the AU but in the entire AA, within which the AU is embedded, and in the AB. Besides the term abhinandåΔ found in the passage on the five fires common to the CU and the BU that we have already examined, the term occurs only twice in the CU. When it rains the vital functions (pråñåΔ) are said to be full of ånanda (ånandinaΔ ) at the thought that there will be plenty of food (CU 7.10.1). We have already seen this usage of the term with reference to plants. In a theologically significant statement, the CU, speaking of a man who sees åtman everywhere, says: sa vå eßa eva∫ pa†yann eva∫ manvåna eva∫ vijånann åtmaratir åtmakrœ∂a åtmamithuna åtmånandaΔ sa svarå∂ bhavati – “A man who sees it this way, thinks about it this way, and perceives it this way; who finds pleasure in the self, who dallies with the self, who mates with the self, and finds ånanda in the self–he becomes completely his own master.” Here we have the two familiar terms rati and ånanda, together with two other terms (krœ∂a and mithuna) also with sexual connotations, in describing the activities of a man who has reached the ultimate state of oneness with his own åtman. Turning to the later verse Upanißads, it is even more suprising to find that ånanda is quite a rare word in their theological vocabularies. It is completely absent in the older group comprising Kena, Ka™ha,45 Œ†a, and ‡vetå†vatara. It occurs once in the Muñ∂aka (2.2.7b)46 where the åtman that the wise see in their heart is described as ånandarüpa, “having the form or appearance of bliss”. Only the Pra†na, an admittedly late work, returns to the theme of ånanda. In an eulogy of lifebreath (pråña), the PU (2.10) says that when it rains creatures becomes joyful (ånandarüpa). In enumerating the activities that a sleeping person does not engage in, the PU (4.2) says nånandayate, which echoes the similar enumerations in the earliof the TU passage to MtU 6.33, see van Buitenen 1979, 326-27. The MtU passage also concludes with the sacrificer becoming ånandin and modin. 45 A verse, variants of which occur both in KaU 1.3 and BU 4.4.11, contains the term anandåΔ (“joyless”). The BU(M), however, reads asuryå for anandåΔ. Charpentire (192829) on KaU 1.3 prefers to read ånandåΔ. 46 MuU (1.2.7) uses abhinandanti with reference to people who take delight in ritual activities.

Orgasmic Rapture and Divine Ecstasy

95

er literature and clearly means “he does not experience sexual pleasure”. Likewise, the PU (4.10) in enumerating the organs and their respective objects, lists upastha† cånandayitavya∫ ca –“the sexual organ and objects that can be sexually enjoyed”. The Måñ∂ükya (5), whose date is difficult to determine but is likely quite late, calls a person in deep dreamless sleep ånandamayaΔ and ånandabhuk, consisting of ånanda and enjoying ånanda. This is in line with the repeated use of ånanda by the older literature in the context of sleep. The transmission of the Maitråyañœya Upanißad has been extraordinarily bad, and the editor van Buitenen (1962) has shown that it is a composite text. The term ånanda occurs at MtU 6.7, 13, 23, 27; 7.3; and all these passages are placed within brackets by van Buitenen, indicating that they are editorial interpolations. The first passage has the term ånandayitå (“one who experiences ånanda”) in a list of agent nouns. The second (MtU 6.13) deals with the essence (rasa) of an ascending hierarchy of entities, from food to ånanda, which is said to be the essence of perception (vij∞åna). A person who knows these essences is said, among other things, to be ånandavån, “possessing ånanda”. At MtU 6.23 the top of O· is said to be Vißñu, who is characterized, among other things, as ånanda (occurring also at MtU 7.3). At MtU 6.27 the cup (ko†a) consisting of the space within the heart is said to be ånanda.

III Finally, I want to turn briefly to the literature of the period that followed the early Upanißads. This is a gray area because it is impossible to determine with any precision either this period or the texts belonging to it. Some of the late Upanißads may well be contemporary with some of the texts generally assigned to this period. In spite of these uncertainties, however, it is instructive to examine at least some of the literature from this period.47 Given the theological prominence of ånanda within the Brahmanical/Hindu religious vocabulary, it is surprising that the term is never used by the Buddhists or the Jains with reference to nirvåña or the ultimate state of liberation. Both traditions, nevertheless, claim 47 The term ånanda is rare in the ‡rauta- and the Gr¢hya-sütras. It occurs in the mantra eßa te yonir ånandåya två in both the Åpastamba ‡rautasütra (19.7.5) and the SatyåßÛå∂ha (Hirañyake†i) ‡rautasütra (23.1.26). This mantra is taken from VS(M) 19.8, which we have already examined. In the Baudhåyana ‡rautasütra (18.29) the term ånandinaΔ is found twice with reference to camasådhvaryavaΔ (assistant priests who carry the cups). Åpastamba ‡rautasütra (5.18.2) has the expression gr¢håñå∫ puß™im ånandam. The Kau†ikasütra (40.13) has ånandino modamånåΔ in a mantra to Agni. In the Gr¢hyasütras, the term occurs only in the Vaikhånasa (3.19) and the Ågnive†ya (1.4.1), both belonging to the early centuries C. E. The term is completely absent in the Dharmasütras and the Manu Smr¢ti. It does not occur in Påñini but is listed in the Gañapå™ha, 81.36. Pata∞jali uses ånanda twice, both in the identical phrase eti jœvantam ånandaΔ, which is a citation of Råm 5.32.6.

96

LANGUAGE, TEXTS AND SOCIETY

that the liberated state is one of bliss or happiness, but the term they use is sukha.48 The Påli Tipi™akaß Concordance lists 26 occurrences of ånanda, including both verbal (ånandati) and nominal forms.49 None has any religious or even an explicitly sexual significance; all refer to a normal sense of joy or happiness. The only usage even remotely connected to religion is the use of ånandajåte with reference to the gods (Suttanipåta 679), but the same expression is used a few verses later (687) when Asita became full of joy at seeing the young Bodhisattva. The same pattern holds true in the two epics, the Råmåyaña and the Mahåbhårata. There also, for the most part, ånanda means ordinary joy. The term occurs 45 times in the Råm,50 and 43 times in the MBh.51 None of the occurrences in the Råm and only a handful in the MBh have any religious/theological connotation. I give below a representative sample of the contexts in which the term is used in the epics. A son is often characterized as bringing or increasing the ånanda of his mother: Råma is kåusalyånandavardhana;52 Lakßmaña is sumitrånandavardhanaΔ;53 and Bharata is kaikeyyånandavardhanaΔ.54 The frequency of this usage especially in the Råm suggests that it had become a cliché. A large group of privative compounds is used to describe the state of grief and desolation of women who have lost their husbands, of men and women in exile, and even of towns at the death or exile of their king or favorite son.55 Only once is such a privative used with regard to a positive virtue,56 although there are frequent references to “tears of joy”.57 48 For a discussion of the Buddhist nirvåña as sukha, see Collins 1998, pp. 207-212. In Jainism also the liberated soul has infinite knowledge (anantaj∞åna) and infinite bliss (anantasukha): see Jaini 1979: 268. 49 Out of these 11 are from the Jåtakas, generally recognized as late texts. 50 Råm 1.1.16, 23; 10.28; 49.6; 50.1, 3, 10, 12; 64.21; 67.13, 15; 69.1, 7; 72.17. 2.5.19; 16.56; 39.13; 40.7; 46.76; 51.4; 53.13; 59.13; 66.33; 84.11; 105.24. 3.35.9. 4.20.9; 24.19. 5.11.29; 18.1; 32.33; 33.77; 34.11, 25. 6.24.31; 31.67; 39.7; 68.9, 12; 114.2, 36; 107.29; 115.1, 40. 7.87.4. 51 MBh 1.118.30; 163.16. 2.13.44; 59.1; 70.21. 3.118.29; 221.22; 261.13. 5.124.17; 136.17; 173.16. 6.2.18. 7.48.1; 50.9; 124.1; 159.42; 164.157. 8.46.9. 9.44.6; 45.11. 10.7.4; 16.24. 11.27.5. 12.31.36; 168.43; 187.33; 191.8; 212.2, 26; 236.21; 239.23; 267.26; 268.11; 301.3, 17. 13.16.55; 27.81; 135.33, 69, 79. 14.38.2; 45.8. 16.6.11. 52 Råm 1.1.16; 67.15; 72.17; 2.66.33; 84.11; 3.35.9; 6.31.67; 115.40. MBh 3.261.13. 53 Råm 1.1.23; 2.16.56; 46.76; 4.24.19; 5.32.33; 34.11; 34.25; 6.39.7; 107.29. 54 Råm 2.40.7. In the MBh (8.46.9) Karña is called suhr¢dånandavardhana, and in the Råm (2.5.19) a great festival of Ayodhå is called janasyånandarvardhana. 55 nirånanda: Ayodhyå after Råma’s exile (Råm 2.51.4; 105.24), Tårå when her husband is killed (Råm 4.20.9), Rumå at Sugrœva’s death (Råm 5.111.29), Sœtå in exile (Råm 6.68.9, 12; 114.36), people after a 12-year drought (MBh 1.163.16), the depraved world of the future (MBh 3.188.29), Ambå when rejected by ‡ålva (MBh 5.173.16), Draupadœ (MBh 10.16.24), the bank of the Ganges with widows of dead heroes (MBh 11.27.5), Dvårakå after Kr¢ßña’s death (MBh 16.6.11). nipatitånanda: Ayodhyå (Råm 2.53.13). anånanda: Hastinåpura after Påñ∂u’s death (MBh 1.118.30). vigatånanda: Påñ∂avas in exile (MBh 2.70.21. hatånanda: the Påñ∂ava camp (MBh 7.50.9). vigatånanda: a dead boy (MBh 12.31.36). 56 The sages Vålakhilyas and Vaikhånasas are said to be karmabhis te nirånandå dharmanityå jitendriyåΔ (MBh 12.236.21). 57 ånandå†ru: MBh 5.124.7; 136.17. 7. 124.1. a†rüñy ånandajåni: Råm 6.24.31. ånanda-

Orgasmic Rapture and Divine Ecstasy

97

It is only in the ‡åntiparvan, the book most subjected to later additions, that ånanda is used with religious or technical meanings. Thus, within a Så∫khya context, ånanda is listed among såttvika guñas and as a product of prakr¢ti (MBh 12.187.33; 212.26; 239.23; 267.26; 301.17; 14.45.8). Once ånanda is listed as a virtue (MBh 14.38.2), while elsewhere its absence is considered a virtue (MBh 12.191.8; 268.11). It is clear, however, that in the epics ånanda is the antonym of †oka.58 In a list of opposites that a person should abandon, we find the compounds satyånr¢te, †okånandau, priyåpriyau and bhayåbhaye (MBh 12.268.11). And in lists of synonyms or words with similar meanings, we find ånanda listed with praharßa, prœti, and sukha (MBh 12.187.33; 212.26; 239.23). The closest we come to a “religious” use of ånanda is in the list of the thousand names of Vißñu where we find surånanda, ånanda, nandana, nanda, and †atånanda (MBh 13.135.33, 69, 79). Only once have I found the term used with regard to the ultimate state to which people aspire, a state that is called paramam ånandam (MBh 13.16.55). In a passage reminiscent of the Bråhamañas and the Upanißads, the MBh (12.301.1) lists the bodily organs (adhyåtmam) and objects (adhibhütam) and divinities (adhidaivatam) that correspond to them. Thus we have: pådau, gantavyam, vißnuΔ; and hastau, kartavyam, indraΔ. In this list we have the group upasthaΔ, ånanda, prajåpatiΔ, connecting the sexual organ, ånanda, and Prajåpati, the procreative/creator god. It is significant that ånanda is completely absent in the Bhagavad Gœtå, a book that is inclusive both in doctrine and in vocabulary. In the sixth chapter that describes a true yogin, the Gœtå uses the term sukha repeatedly (BhG 6.21, 27, 28) to indicate the final bliss such a man attains. At 6.15 the Gœtå describes the bliss of a yogin as †ånti (peace) and nirvåña (possibly, “calm”).59 An examination of the non-Vedåntic literature shows that sukha was, in fact, the most common term for both ordinary and ultimate happiness. Besides the Gœtå and the Buddhist and Jain literature, sukha is the preferred term for the bliss of the liberated state in Så∫khyaYoga theology. In Så∫khya ånanda is said to be an attribute of prakr¢ti rather than of purußa.60 In Yoga, vitarka, vicåra, ånanda, and asmitå are four states of samådhi, but still short of the liberated consciousness.61 In the Nyåya-Vai†eßika the state of liberation lacks all pain or suffering but there is no positive feeling of bliss or ånanda.62 ja∫ payaΔ: Råm 2.39.13. ånandajam jalam: Råm 5.33.77. This expression is taken over by Kålidåsa: ånandottha∫ nayanasalilam (Meghadüta, uttaramegha, 4); ånandajaΔ [...] båßpaΔ (Raghuva∫†a, 14.3). 58 In the technical explanation of ånanda given in the Nyåyako†a (ed. B, Jhalakikar and V. S. Abhyankar; Poona, 1978), two meanings are given. The first is sukha, and the second is duΔkhåbhåva, the absence of suffering. 59 See, Dasgupta 1922, II: 450. 60 See, Dasgupta 1922: I: 238. 61 Yogasütras, 1.17. See, Dasgupta 1992, I: 366. 62 See, Dasgupta 1922, I: 366.

98

LANGUAGE, TEXTS AND SOCIETY

IV From the above survey of the use of ånanda in the early Indian literature we can draw the following conclusions: 1. In the early vedic literature ånanda is used in a variety of contexts, including the thrill of gambling, the convivial joy of drinking, and especially sexual pleasure. 2. The middle vedic literature of the Yajurveda emphasizes the sexual aspect of ånanda, using it almost as a technical term for orgasmic rapture. The absence of the term in non-Yajurvedic Bråhmañas, with the exception of a single passage in the KßB,63 indicates that this usage was by and large confined to the Yajurvedic schools. 3. In the late vedic literature also the term is most frequent in the two Yajurvedic Upanißads, the Br¢hadårañyaka and the Taittirœya, although the presence of the term with a sexual connotation in the R¢ gvedic Kaußœtaki Upanißad makes the picture somewhat less clear.64 The association of brahman/åtman with ånanda, however, takes place principally in the Yajurvedic Upanißads. This semantic development, I believe, took place specifically as an extension of the meaning of ånanda as orgasmic rapture, a meaning already found in the early Yajurvedic texts. The connection between these two meanings of ånanda, we saw, is made explicitly in BU 4.1.6. Two elements of orgasmic rapture are central in this extended meaning: 1) the connection of ånanda to procreation and, therefore, to Prajåpati, and 2) the loss of consciousness of individual identity associated with orgasm. TU (2.7) is the locus classicus for ånanda as the primary attribute of brahman/åtman. 4. The evidence of the Buddhist, Jain, and epic literature indicates that ånanda did not immediately enter the common religious vocabulary either as the joy of heaven or final release (mokßa) or as an attribute of the Ultimate Being or State. I think that after the composition of the BU and the TU ånanda as an attribute of brahman and as signifying the final state of bliss remained a technical usage confined to a somewhat narrow circle. There must have been a parallel semantic development of ånanda leading to its meaning as simple (not necessarily sexual) joy and happiness. This development took the term away from any specifically religious connotation. Unfortunately, we do not have the literary evidence to trace this development from the early vedic usage to the Buddhist and epic texts. We have, however, seen ånanda used with such a generic meanings in BU 4.3.32-33; CU 7.10.1; and TU 2.8. It is, however, clear that the religious usage of the term in the Brahmasütras and later literature is derived not from this 63 The KßB is probably younger than the Bråhmaña of its sister school, the AB, in which the term ånanda does not occur. 64 See above, note 35. It is interesting, however, that outside of the Yajurvedic documents, the term is used only in the KßB and the KßU.

Orgasmic Rapture and Divine Ecstasy

99

generic epic usage but from its specifically religious meaning that developed in the Upanißads. 5. The native tradition itself recognizes the connection between the Mœmå∫såsütras of Jaimini and the Brahmasütras. This association provides the basis for calling the former Pürva-Mœmå∫såsütra and the latter Uttara-Mœmå∫såsütra; as the former is the exegesis of the “earlier” part of the Veda, so the latter is the exegesis of the “later” part, namely the Vedånta or Upanißads. If we are to believe this association, then we should expect that the teachers of the two Mœmå∫sås belonged to the same religious/intellectual milieu. That may well be true, but at least in the case of the term ånanda, which is central to the UttaraMœmå∫så (both the Sütra itself and especially its commentators), the two appear not to have shared this religious vocabulary. Although the sütras of Jaimini themselves offer no clues, ‡abara’s commentary, written probably around the middle of the first millennium C. E., refers frequently to the “happiness” or “joy” of heaven (svarga) in his discussion of Mœmå∫såsütras 6.1.1-2. The most common word ‡abara uses for “joy” is prœti, although occasionally he uses sukha.65 6. Let us, finally, return to the Brahmasütra passage (1.1.12) that defines åtman/brahman as ånanda: ånandamayo ‘bhyåsåt –“The (self) consisting of ånanda is (the supreme self) because of repetition.”66 The reasons why the supreme self is defined as ånanda, according to Bådaråyaña, is because the Upanißads repeatedly say so. ‡a∫kara in his commentary expands on this terse statement, citing these repeated Upanißadic statements. Interestingly, however, all his citations except one are taken from the TU, and the one non-TU text is BU 3.9.28. So the repeated mention of brahman/åtman as ånanda is found only in the Yajurvedic Upanißads, and it is on the basis of these documents that the Brahmasütras assert the primacy of ånanda as the defining characteristic of brahman/åtman. We can thus see here the direct connection between the Yajurvedic Upanißads and the Brahmasütra definition of brahman as ånanda, at least if ‡a∫kara is right in identifying the sources that prompted Bådaråyaña’s statement about repetition. It is, furthermore, the extraordinary influence of the Brahmasütras on later theological discourse that made ånanda a central term and concept in the later Brahmanical/Hindu vocabulary.

65 See Mœmå∫såsütras (Ånandå†rama Sanskrit Series, 97), part 5: prœti, p. 175, l. 1; p. 176, l. 4; p. 177, l. 1-2; etc.; sukha, p. 177, 21; p. 179, l. 1. 66 The Brahmasütras define brahman as ånanda also at 3.3.11.

100

LANGUAGE, TEXTS AND SOCIETY

A

101

IV A∫r¢tå: Women and Indian Technologies of Immortality*

“I don’t want to become immortal through my works. I want to become immortal by not dying.” This wise quip, or something like that, is supposed to have been said by the American film director and comic Woody Allan. Apart from the intended humor, there is a deeper, perhaps unintended, meaning underlying this statement. Conceptions of immortality contain both an individual and a social dimension. People do live on after their death in the memories of friends and relatives, in the lives of people they have touched, and in the products they leave behind, be they films, books, or children.1 These two dimensions are inextricably interwoven, I believe, in debates on and technologies directed at achieving immortality. They become even more interwoven because, in spite of Woody Allan’s aspiration and in spite of ancient alchemy and modern medicine, all human beings die. So immortality is not about “not dying”, but about ways of postponing death and about coping with death at personal and ideological levels, coping which is, more often than not, intertwined with social memory. So we need not be surprised at recent news reports from Beijing announcing that the good old communist Deng Xiaoping had “joined the immortals”! Ideas and aspirations, moreover, do not exist in a vacuum; they are influenced by and in turn influence the social and economic conditions of the time and thus exist in a creative tension with their Originally published in the Journal of Indian Philosophy. 25 (1997): 427-49. The connection between social memory, in this case embodied in a literary product, and immortality was not alien even to ancient Indians. The god Brahmå tells Vålmiki, the author of the epic Råmåyaña: “As long as the mountains and rivers shall endure upon the earth, so long will the story of the Råmåyaña be told among men. And as long as the story of Råma you compose is told, so long will you live on in my worlds above and below” (Råmåyaña, 1.2.35-36; tr. Goldman). * 1

102

LANGUAGE, TEXTS AND SOCIETY

underlying socio-economic matrix. The aim of this paper is to explore the connections between ancient Indian ideas about immortality, debates regarding what it means to die and to survive death, and technologies for achieving a deathless state, on the one hand, and the changing social conditions of the time, on the other. My question is not what immortality is, although some attention needs to be paid to beliefs concerning immortality, but who can become immortal? My specific question relates to gender: are women part of the ancient Indian discussions about immortality, and if so, what sort of roles are they assigned? Are they agents or mere instruments in the emerging technologies of immortality? India is big, Indian history is long, and my own expertise is limited. So, I will limit my exploration of these topics to the time span from the middle to the late vedic period, that is somewhere between 700 and 300 BCE, give or take a few centuries, although I will occasionally cite evidence from later periods.

I. Meanings of Amr¢ta The Sanskrit term for “immortal” and “immortality” is amr¢ta. It consists of mr¢ta, which is the past participle of the verbal root √mr¢ = to die, with the negative prefix a. This is a somewhat curious formation, because theoretically it should mean “not dead” rather than “non-dieable” or “immortal”, which in Sanskrit should be amartya. Negated past participles, however, tend to have an “un-Xable” rather than just an “un-Xed” meaning,2 and the former is the primary meaning of amr¢ta. The term amr¢ta as “immortal”, however, is not a Sanskrit invention; it has solid Indo-European roots, with the Greek cognate ambrotos (from which is derived ambrosia) and the Avestan ameıa. Paul Thieme (1968b) has studied these terms and their meanings in detail; I summarize here his findings. Thieme has shown that in Indo-European the term had two distinct meanings: the first he calls “Lebenskraft spendend”, “giving vitality”; and the second, “unsterblich”, “immortal”. The spectrum of objects to which the term amr¢ta is applied in the vedic texts, objects from the clarified butter, gold, and the Soma drink to food, water, semen, son, and gods, supports Thieme’s conclusion.3 The term amr¢ta does not always mean immortal in the sense we usually attach to it; it often means vitality or vital energy (Thieme’s Lebenskraft)–it is a full and prosperous life and all things that sustain and promote such a life, including food, drink, cattle, and medi2 I am indebted to Stephanie Jamison for these two fine expressions and for elucidating several aspects of this term. There are other examples of negated past participles functioning as gerundives, for example, adhr¢ß™a “undareable-against”. 3 For a study of amr¢ta in the vedic literature, see Gonda 1965b, 38-70. 4 agnir amr¢tam, “the immortal is fire” (‡B 10.2.6.17); annam amr¢ta∫ vadanti, “they say that the immortal is food” (KS 70.6); amr¢ta∫ vai pråñåΔ, “the breaths are indeed the immortal” (TS 2.6.8.7); amr¢ta∫ vå åjyam, “clarified butter is indeed the immortal” (TA

A∫r¢tå

103

cine.4 Amr¢ta can thus indicate both life/immortality, as well as instruments that sustain life and ward off death. The two terms amr¢ta and åyus (long and full life) are often juxtaposed and form a single complex of meanings. Indeed, the ‡atapatha Bråhmaña (10.2.6.7-9) sees long life as a visible sign that a man is destined to become immortal: “The life of a hundred years makes for heaven [...]. He alone who lives a hundred years or more attains to that immortal life.” I disagree with Thieme, however, in taking these two meanings as somehow distinct and separate. Both because of the identity of the term and because life in all its forms confronts death as its opposite, I think what we have, at least in ancient India, is a spectrum of meanings that are never totally separate, each merging into and influencing the conception of the others. It is also this broad semantic range of the term that permitted its use simultaneously with regard to a wide variety of objects without contradiction. The term underwent further widening as the conceptions of death and the after-death state underwent drastic change within the ideology of rebirth (sa∫såra), coming to signify not just survival after death but the liberation from the cycle of rebirth (mokßa).

II. Conceptions of the Self The changing conceptions of immortality in ancient India were closely connected with the changing conception of “self”, of what it means to be a human individual. Michael Carrithers (1985), responding to Marcel Mauss’s attempt to trace the history of the ‘self’,5 makes a useful distinction between personne and moi, a distinction that provides a helpful heuristic tool to understand the connection between three things: conception of self, conception of immortality, and the position of women. Carrithers (1985, 235-36) defines personne as “a conception of the individual human being as a member of a (1) significant and (2) ordered collectivity,” and moi as “a conception of (1) the physical and mental individuality of human beings within (2) a natural or spiritual cosmos, and (3) interacting with each other as moral agents.” I want to extend Carrithers’ somewhat narrow definition of per2.17.2); amr¢ta∫ vai hirañyam, “gold is indeed the immortal” (TS 5.2.7.2); amr¢tam eva saptamœ citiΔ, “the seventh layer is truly the immortal” (‡B 8.7.4.18); amr¢to vai somaΔ, “Soma is indeed immortal”( ‡B(Kåñva) 4.3.4.12); ådityo amr¢tam, “the immortal is the sun” (‡B 10.2.6.16). These examples could be multiplied. I have translated amr¢ta uniformly as “immortal” for the sake of consistency, but as one can see in many of these examples the more accurate translation, following Thieme, would be something like “fire is what gives vitality”, “they say that what gives vitality is food”, etc. 5 Mauss’s article “A Category of the Human Mind: The Notion of Person; the Notion of Self,” originally published in 1938, is reprinted in the same volume (pp. 1-25) as Carrithers’s paper. 6 Carrithers appears to restrict personne to those societies that have legally defined cit-

104

LANGUAGE, TEXTS AND SOCIETY

sonne6 to include the selfhood of an individual, especially of a married male, within the society of ancient India hierarchically ordered according to class and gender. A personne is defined in terms of interlocking social relationships–to living and deceased members of one’s family, to members of one’s caste, to residents of one’s village, and so forth. When self is defined in terms of personne, the collectivity of which an individual is a member plays a determining role in the definition of a person’s self and vitally affects an individual’s choices and goals. Self as moi, on the other hand, sees the individual as a unique and self-contained entity transcending temporary social relationships. This conception of the self emerges in India within the context of the rebirth ideology which sees social relationships as fleeting and ephemeral, not affecting the inner core of one’s self. In distinguishing the two conceptions of self I am not suggesting that personne and moi are watertight and self-contained categories. These two conceptions of the self do exercise influence on each other especially in complex societies where an individual often belongs to several “collectivities”.7 In ancient India, such collectivities may have included family/lineage, caste, language group, city/village, kingdom, and sectarian religious affiliation. I like to see the two conceptions of self at two ideal-typical poles of a continuous line [see chart], in which the features of the one penetrate the other. Different social experiences, furthermore, must underlie these conceptions of self. I have argued elsewhere8 that the development of large bureaucratic states with complex economies and the rise of urbanization along the Gangetic plain in the middle of the first millennium BCE. were, at least in part, responsible for the rise of worldrenouncing ideologies and of conceptions of selfhood divorced from social relationships. The differing conceptions of the self, in turn, influenced differing conceptions of what it means to die, to survive death, and to become immortal. From among these conceptions I have selected four for comment [see chart]: son (together with the world of the fathers), a full life span, heaven, and liberation from the cycle of rebirth. The reason for selecting these is because they are the ones that are prominent in the technologies of immortality found in the extant literature of the period. It would be a mistake, however, to think that these izenship carrying rights and responsibilities. He denies that such a category is applicable to the Indian varña system (1985, 251). I think that the selfhood within the varña system is very much a personne, because it is defined by one’s belonging to a particular social group, a belonging that carries rights and responsibilities, even though that group is not the nation-state. 7 Carrithers (1985, 236) gives one example: “a German of the early nineteenth century might have found himself to be a citizen of a small principality, a member of the Catholic Church, and a member of some notionally powerful German nation.” 8 Olivelle 1993, 55-67. I am not suggesting here that changes in the perception of the self can be reduced simply to socio-economic changes, but that such changes form a significant part of the causal complex.

A∫r¢tå

105

were the only views regarding immortality in ancient India or that they represent a chronological history of the conceptions of immortality. The literature that has survived was produced for the most part by a male elite of the Brahmin class; their concerns dominate the discourse. We have no idea of what other classes and other peoples of ancient India may have thought about these matters. Even more significantly, we do not know the aspirations of ancient Indian women themselves or their thoughts about their own selves and their mortality/immortality. The women we encounter in the vedic literature are literary creations of men and women’s voices from ancient India are really the ventriloquial speech of men. Given the social prestige of Brahmanical writings, especially of the vedic texts, however, the “minority view” on self and immortality found in them did have a disproportionate influence on Indian society, an influence that Collins (1982, 32), borrowing an expression from Gramsci, has called the “culturally hegemonous role” of Bråhmañism vis-à-vis other groups and ideologies of India. Immortality SOCIAL EXPERIENCE WOMEN as instruments WOMEN as agents

+ + + Personne

Moi

SON/WORLD OF FATHERS

LONG LIFE (åyus)

HEAVEN (svarga)

LIBERATION (mokßa)

Semen Marriage, Ancestral Rites

Food Medicine

Sacrifice Ritual Purity/ Knowledge

Knowledge Renunciation/Celibacy

I will also argue, somewhat in the manner of Mary Douglas’s (1982) paradigm of grid and group, that A) social experiences and constraints are strong [represented by + in the chart] at the personne pole and become weaker [- in the chart] as we move toward the moi pole; and B) the perception of women as instruments is strong [ + ] at the personne pole, while the perception of women as agents and individuals with desires and roles independent of their connection to males is strong at the moi pole.

III. Sons and Immortality The instrumentality of the wife is most evident and the self as personne most central in the conception of the son as the physical and ritual continuation of the father, as the father’s immortality. This idea is very old. Already in the R¢gveda (5.4.10) we find the prayer: prajåbhir agne ’mr¢tatvam a†yåm, “Through offspring, O Agni, may we attain immortality.” The Taittirœya Bråhmaña (1.5.5.6) exhorts: prajåm anu prajåyase tad u te martyåmr¢tam, “In your offspring you are born again;

106

LANGUAGE, TEXTS AND SOCIETY

that, O mortal, is your immortality.”9 The wife is often said to be the completion of the husband; he becomes a complete “self” only when he is married: ardho ha vå eßa åtmano yaj jåyå tasmåd yåvaj jåyå∫ na vindate naiva tåvat prajåyate ’sarvo hi tåvad bhavaty atha yadaiva jåyå∫ vindate ’tha prajåyate tarhi hi sarvo bhavati, “A full half, surely, of one’s self is one’s wife. As long as one does not obtain a wife, therefore, one can never be reborn, for he then remains incomplete. As soon as he obtains a wife, however, he is reborn, for then he becomes complete” (‡B 5.2.1.10). Now, it is not altogether clear whether amr¢tatva and amr¢ta in passages such as these mean immortality or merely life/vital energy. Children, just like wives and cattle, can be seen as the expansion of the father’s life, a life that is defined here not just as biological existence but “living a full life”, the life of a rich, prosperous householder. But I think there is more to it than that; the child is seen here as the continuation of the father both when the father is alive and especially after his death.10 It is, however, not just any child that constitutes the continuation of the father; it is the son: ya u vai putraΔ sa pitå yaΔ pitå sa putraΔ, “The father is the same as the son, and the son is the same as the father” (‡B 12.4.3.1). At the birth of a son the father ritually takes him, saying: aõgåd aõgåt sa∫bhavasi hr¢dayåd adhijåyase, åtmå vai putranåmåsi sa jœva †aradaΔ †atam, “From my every limb you spring; out of my heart you are born. You are my self (åtman) called ‘son’; live a hundred autumns!” (Påraskara Gr¢hyasütra 2.3.2; cf. BU 6.4.9). The family line continues uninterrupted in the son despite the death of the father; the son inherits the paternal estate and replaces the father as the ritual and economic head of the family–the father’s personne continues in the son. As the son survives his father’s death, so the father in his son survives his own death. In a very moving song the Aitareya Bråhmaña (7.13) eulogizes the son as the new birth of the father: A debt he pays in him, and immortality he gains, the father who sees the face of his son born and alive. The husband enters the wife; becoming an embryo he enters the mother. Becoming in her a new man again, he is born in the tenth month. A wife is called ‘wife’ (jåyå), because in her he is born again (jåyate). The gods said to men: ‘She is your mother again.’ A sonless man has no world. All the beasts know this. Therefore a son mounts even his mother and sister.11

In this song that would have, had he but known it, brought joy to Freud’s heart, the instrumentality of the wife in accomplishing the immortality of her husband is brought out in starkest clarity. The 9

For a detailed discussion of this ideology, Olivelle 1993, 41-53. A text from a later period points this out clearly: “Through a son he conquers the worlds, through a grandson he obtains immortality, but through his son’s grandson he gains the world of the sun” (VaDh 17.5). This conception points to the significance of offspring for the continued felicity of deceased ancestors. 11 r¢ñam asmin sa∫nayaty amr¢tatva∫ ca gacchati Ù pitå putrasya jåtasya pa†yec cej jœvato 10

A∫r¢tå

107

same vision of wife and son is the focus of ‡akuntalå’s heart-wrenching outburst to DuΔßanta when he feigned not to remember his affair with her: “Because a husband enters his wife and is born (jåyate) again from her, the poets of old knew that this is the ‘wifehood’ (jåyåtva) of a wife (jåyå) [...]. A son, the wise say, is the man himself born from himself; therefore a man will look upon his wife, the mother of his son, as his own mother. The son born from his wife is like a man’s face in a mirror.”12 Mother and wife exchange roles and become fused together in their role of begetting male children to continue the male line unbroken, thus assuring male immortality.13 Ancient speculation regarding the nature of semen also contributed to the equation of father and son. A man’s sperm is viewed as his rasa or essence. In other words, a man replicates himself, creates a second self for himself, in his sperm. The Aitareya Upanißad (2.1; see note 37), for example, calls semen a man in embryonic form that he carries within himself; when he deposits it in a woman it becomes his first birth: “At the outset, this embryo comes into being within a man as semen. This radiance gathered from all the bodily parts he bears in himself (åtman) as himself (åtman). And when he deposits it in a woman, he gives birth to it. That is his first birth.” The Taittirœya Upanißad (3.10.3) locates procreation, immortality, and orgasmic bliss in the sexual organ (prajåtir amr¢tam ånanda ity upasthe), underscoring the connection between these three concepts.14 In this view of a man’s replication of himself through the ejaculation of semen and the accompanying bliss, the wife plays a passive and clearly instrumental role; 15 she is the fertile field, the soil, in which the seed is planted.16 The continuation of the father in the son is ritually and dramatically expressed in the ancient rite of transmission (BU 1.5.17-20; KßU mukham ÙÙ … patir jåyå∫ pravi†ati garbho bhütvå sa måtaram Ù tasyå∫ punar navo bhütvå da†ame måsi jåyate ÙÙ taj jåyå jåyå bhavati yad asyå∫ jåyate punaΔ ÙÙ … devå manußyån abruvann eßå vo jananœ punaΔ ÙÙ nåputrasya loko ‘stœti yat sarve pa†avo viduΔ Ù tasmåt putro måtara∫ svasåra∫ cådhirohati ÙÙ 12 bhåryå∫ patiΔ sa∫pravi†ya sa yasmåj jåyate punaΔ Ù jåyåyå iti jåyåtva∫ puråñåΔ kavayo viduΔ ÙÙ … åtmåtmanaiva janitaΔ putra ity ucyate budhaiΔ Ù tasmåd bhåryå∫ naraΔ pa†yen måtr¢vat putramåtaram ÙÙ bhåryåyå∫ janita∫ putram ådar†e svam ivånanam Ù MBh 1.68.36, 4748 (van Buitenen’s translation modified). 13 The view that the husband is born again in his wife finds expression even in later texts: see MBh 3.13.62; MDh 9.8; YDh 1.56. A passage in the epic Råmåyaña (2.35.8) illustrates this: the earth is viewed as the wife of a king; when the king dies she becomes a widow, but regains a husband when his sons becomes the new king. 14 For a more detailed study, see Olivelle 1997, 163-70; see above pp. 75-99. 15 The instrumentality of the woman is nicely illustrated in the bridal prayer (ņvalåyana Gr¢hyasütra 1.7.22): “May my husband live long, and may I have offspring.” And the husband pours ritually consecrated water on the bride, saying: “The evil substance which dwells in you that brings death to your husband, death to your children, death to cattle, destruction to the house, destruction to fame, that I change into one that brings death to thy paramour. Thus live with me to old age, N. N”. 16 For the idea of wife as the field (soil) for man’s seed, see MDh 10.70-73. This idea also

108

LANGUAGE, TEXTS AND SOCIETY

2.15). When the father is about to die, the son comes and lies on top of the father, each of his organs touching the respective organs of the father. The father consigns all his faculties to the son; entering the son he stands firm in the world even after death. In the son the father continues his personne, his role as paterfamilias. The reversal of roles is dramatic as it is permanent; if the father happens to recover he is expected either to leave home and live as an ascetic or to live at home under the authority of his son. The theological articulation of the identity between father and son and of this generational continuity as a form of male immortality leaves mothers, wives, and daughters out of the discourse except insofar as wife and mother. ‡akuntalå’s speech to her husband underscores the male expectation (articulated by the male author of the tale) that wives and mothers would unhesitatingly internalize this theology. Another conception of after-death connected with the centrality of a son is the “world of fathers”. The happiness of one’s ancestors depends on food and water offerings made by their male descendants, and this belief is often presented as a cornerstone of the imperative to marry and to father sons and as an argument against celibate modes of life.17 The world of fathers, however, appears as an archaic concept already in the Bråhmañas, and it is often presented as a counterpoint to immortality associated with the “world of gods”. Fathers are said to be mortal, while gods are immortal. But like many things in Indian religious history, the belief that the destiny of the deceased ancestors is tied to their continuing relationship to their descendants endured and still endures, ritually enshrined in the †råddha offerings. Even in death the self is very much a personne tied to enduring kinship relations. In the middle and late vedic literature we come across the interesting concept of punarmr¢tyu, re-death; people who die may be subject to death once again. Bodewitz (1996a) in his recent article has argued, correctly I believe, that “re-death” is more an argument than a belief; an argument by anti-ritualists that the worlds won by rites, for example, the world of fathers, are still subject to death. One reason why re-death appears to be a debating point is that it appears only in ritualist discourses (we must assume that they are responding to anti-ritualist arguments) and always with a corrective: those who do X will not die again, spawns the legal view that a son born in one’s field (i.e., one’s wife) belong to the owner of the field (i.e., the husband of the mother) and not to the owner of the seed (i.e., the biological father), just as a tree planted by someone in another man’s field belongs not to the man who planted the seed but to the man who owns the field (for this comparison see MDh 9.3641). On the controversy as to who owns the son begotten on someone’s wife, see MDh 9.4244. One of the several type of legitimate sons is called kßetraja, “born in one’s field”. 17 It is also the central element of the so-called theology of debt: a man is born with debts to gods, ancestors, and seers. The debt to ancestors is paid by fathering a son. On the question of debts as an argument against celibate modes of life, see Olivelle 1993, 41-53, 83-91.

A∫r¢tå

109

a conclusion that appears to be a preemptive answer to possible antiritualist claims. Nevertheless, the world of fathers becomes identified with death (and re-death) and appears again in later discussions about rebirth: the path of fathers entailing rebirth and the path of gods assuring immortality and liberation from rebirth. The “world of fathers” is also interesting in what it leaves out –the mothers, the women. As its name suggests, the world of fathers is clearly a male conception closely connected with social memory and the inheritance of property.18 Thus, only the three previous generations of ancestors –father, grandfather, and great-grandfather– are addressed by name in the ritual food offerings of †råddha. These are called the a†rumukha or tearful fathers, because they are still in the collective memory of the family. Prior generations are called nåndœmukha or joyful fathers, and they are anonymous. When one’s father dies, his grandfather departs from social memory and joins the anonymous group of fathers. In ancestral food offerings female ancestors enter into the picture only as wives of the fathers.19 Here, as we will see also in our discussion of heaven, women enter into the discourse and ritual only as appendages of their husbands. The very fact that “ancestors” are called pitaraΔ (“fathers”) indicates the male bias in the social/ritual memory of the dead.

IV. Long Life and Immortality The meaning of amr¢ta that comes closest to Thieme’s (1968) first meaning is åyus, a long and full life. Long life may seem unproblematic to us with our public health programs and medical technology; but in ancient times the probability of a new-born child living its full life span, viewed ideally in ancient India as 100 years, must have been extremely low. A recurring refrain at the conclusion of vedic descriptions of most rites is that the person “will live his full life span” sarvam åyur eti.20 So, the ‡atapatha Bråhmaña (2.1.4.4) says that a man who sets up his fires during the period when the sun moves north lives his full life span, because that period belongs to the gods, who are immortal; while the period when the sun moves south belongs to the fathers, who are mortal. 18 In Hindu jurisprudence the obligation to offer †råddhas to an ancestor is connected with inheritance: whoever inherits the estate of a deceased person has the obligation to make †råddha offerings to him: “Whoever inherits the estate has to offer the piñ∂Ûa (rice balls) to him” (Vißñu Dharmasütra, 15.40). The point appears to be that the owner of ancestral wealth has the obligation to feed those who depend on that wealth, the living, as well as the dead, members of the family. See Kane 1962-75, iii: 734-38. 19 For sources and discussion, see Kane 1962-75, iii: 474-76. Even when the female side of the ancestral line is represented, the focus is on the males of that line. Thus, for example, offerings are made to the mother (i.e., father’s wife) and then not to the mother’s mother but to the maternal grandfather etc. After death women appear to fall from the collective social memory–at least ritually enacted memory–much faster than men. 20 See ‡B 2.1.4.9; 2.3.3.6; BU 2.1.10, 12; 6.4.14-18.

110

LANGUAGE, TEXTS AND SOCIETY

As Thieme (1968) has pointed out, amr¢ta is not simply “not dying” or “un-dieble” but also that which stands apart from death, that is, life, vitality, health, and all that promotes life. The connection between åyus and amr¢ta in the vedic mind can be seen in the many substances that are called both åyus and amr¢ta.21 The two terms are also used together, as in amr¢tam åyur hirañyam.22 There were at least some who proposed that a long life was the only “immortality” that humans can aspire to. The ‡atapatha Bråhmaña (9.5.1.10) says: etad vai manußyasyåmr¢tatva∫ yat sarvam åyur eti, “For a man immortality is simply this–that he lives his full life span.”23 Likewise: tad dhaitad yåvac chata∫ sa∫vatsarås tåvad amr¢tam anantam aparyanta∫ sa yo haitad eva∫ vedaiva∫ haivåsyaitad amr¢tam anantam aparyanta∫ bhavati, “A hundred years is as much as immortality–unending and everlasting; and a man who knows it in this way will have immortality–unending and everlasting” (‡B 10.1.5.4). The ‡atapatha (‡B 2.2.2.14) also says that gods became immortal by establishing the fire in their inmost self; a man who does the same will get to live his full life span, because “there is for him no hope of immortality” (nåmr¢tatvasyå†åsti). Gods themselves were not naturally immortal; they achieved immortality through a variety of means consisting mostly of different ritual technologies. Likewise, humans cannot achieve their full life span naturally; it is the outcome of ritual activity and ritual knowledge. Living a long life may appear as a somewhat individualistic enterprise, closer to the moi conception of self than the personne. Yet, within the ancient Indian context of the ritual persona, living a long life is rooted within a ritual/familial context and is bound up with fame, riches, children, and social position. It would have seemed absurd to associate åyus or amr¢ta with a poor, low-class, and ignorant man.24 Åyus is not simply living but living well and living long. We have this 21 agnir åyuΔ, yaj∞a åyuΔ, “long life is fire”, “long life is sacrifice”, (MS 2.3.4; cf. KS 5.3; JB 1.70); annam u vå åyuΔ, “long life is indeed food” (‡B 9.2.3.16); åyus sa∫vatsaraΔ, “long life is the year” (MS 4.6.8); åyur ghr¢tam, “long life is ghee (TS 2.3.2.2); åyus is also identified with gold (‡B 3.8.2.27; 4.5.2.10; MS 1.7.5; 2.1.7; 4.4.2; KS 9.2; 11.8). Here the meaning is that these substances contain the means to securing a long life. Compare these statements with parallel ones regarding amr¢ta in note 4. The connection between åyus and amr¢ta is highlighted also in the TS 3.3.4.3: amr¢tam asi pråñåya tveti hirañyam abhi vyanity amr¢ta∫ vai hirañyam åyuΔ pråño ‘mr¢tenaivåyur åtman dhatte, “‘You are the immortal! You for breath!’ With these words he breathes over the gold. The gold is the immortal, breath is life. With the immortal indeed he places breath in himself.” 22 ‡B 3.7.2.27; 4.5.2.10; 4.6.1.6. Eggeling translates “gold is immortal life”, but the two terms may well qualify gold independently: gold is amr¢ta and gold is åyus. 23 A similar statement is made in the Vaikhånasa Gr¢hyasütra (3.21): “When eighty years and eight months, reckoned according to the solar year, have passed as a man practices this, he has seen a thousand moons. They call such a man Brahman’s body–a body that is endowed with rites and has performed meritorious deeds to the highest extent.” tad eva∫ vartamånasya yady aß™amåsådhik农tivarßåñi ravivarßeñådhigåny adhigaccheyuΔ sa dr¢ß™asahasracandro bhavati tam ena∫ kriyåyukta∫ puñyakr¢ttama∫ brahma†arœram ity åcakßate. 24 The importance of wealth to being a “religious man” in ancient India has often been ignored. Rau (1957, 32-34) has shown that the terms påpœyån (worse) and †reyån (better)

A∫r¢tå

111

refrain repeated ten times in the Chåndogya Upanißad (2.11-20): sa ya evam etad … veda … sarvam åyur eti jyog jœvati mahån prajayå pa†ubhir bhavati mahån kœrtyå, “When in this manner a man knows this [...], he lives his full life span; he lives a long life; he becomes a big man on account of offspring and livestock; and he becomes a big man on account of his fame.” As we will see in our discussion of heaven, it is not just any human being who is the subject of this discussion about åyus; it is a married male possessing the ritual fires as the head of a household. A wife is an absolute necessity for performing ritual functions; she is as much a sacrificial instrument as the priests, spoons, knives, and fires, an instrument in assuring her husband’s åyus.

V. Heaven and Immortality Human desire knows no bound, and the advice of some who asked humans to be content with a long life and children mostly fell on deaf ears. The Aitareya Årañyaka (2.3.2-3) astutely observes: martyenåmr¢tam œpsati … sa eßa purußaΔ samudraΔ sarva∫ lokam ati Ù yad dha ki∫cå†nute ‘ty ena∫ manyate yady antarikßalokam a†nute ‘ty ena∫ manyate yady amu∫ lokam a†nuvœtåty evaina∫ manyeta, “By means of the mortal he desires to obtain the immortal [...] This man is an ocean, beyond the entire universe. Whatever he obtains, he thinks beyond it. If he obtains the intermediate world, he thinks beyond it. If he were to obtain the world up there, he would surely think beyond it.” Men, in other words, want to be like the gods. They want to be immortal not merely in their sons, but in their own self-identity and self-consciousness. “To have the same world as the gods,” devånå∫ salokatå, is the refrain one hears with reference to people who perform sacrifices. We are moving here closer to a moi definition of self, a self that can survive and transcend death, a self that is the architect of its future, a self that does not require an unbroken ritual connection to its former kin for its existence and happiness. By the middle vedic period it was a common belief that gods themselves were not originally immortal. A universal principle applicable to all beings, gods and humans alike, appears to have been established: immortality is not a natural attribute of any being; it is something to be achieved. At first gods were on earth and they were mortal. It was through their full knowledge and correct performance of the sacrifice that they became immortal and reached heaven. yaj∞ena vai devå divam upodakråman, “It is by means of the sacrifice that the gods ascended to heaven” (‡B 1.7.3.1).25 The seers discovered this divine secret, the ritwithin the context of the class distinctions existing within ancient Indian society are not merely religious or ethical terms. The “worse” are people who are poor and powerless, while the “better” are the rich and the powerful. The two terms frequently refer specifically to the economic standing of a person, as at CU 4.16.3. 25 The ‡B (2.2.2.8-14) portrays gods and Asuras as mortal because they lacked a self,

112

LANGUAGE, TEXTS AND SOCIETY

ual technology of immortality, which the gods tried their level best to hide from men, and revealed it in the Vedas.26 Now men also possess the ritual technology to become immortal, to ascend to the world of the gods, to become like the gods.27 But this ritual technology for reaching heaven is a male prerogative. Only men are taught the vedic secrets; only men can sacrifice. But single men are not entitled to sacrifice, just as they are not entitled to father children. Only married men accompanied by their wives have the capacity to sacrifice and thus aspire to heavenly immortality. The woman is viewed as the completion of the male sacrificial persona. The ‡atapatha Bråhmaña (5.2.1.10), we have seen, proclaims the wife to be one half of the husband; it is only when they are together that they constitute a full ritual person. At one level this requirement elevates the position of a woman. Stephanie Jamison (1996) has demonstrated both the centrality and the complex nature of a wife’s participation in the vedic ritual, bringing a much needed corrective to the oft-repeated platitudes about the status of women in “Hindu” society. Note, however, that the entire discourse regarding the female participation in the ritual is carried out from a male perspective; one needs a wife to perform a sacrifice, just as one needs firewood and ghee. They are all instruments. Women, nevertheless, are different from other instruments; they are also ritual actors. Women, indeed, can aspire to heaven, but only as wives of their heaven-bound husbands. In a significant rite during the Soma sacrifice the sacrificer and his wife pretend to climb the sacrificial pole that connects heaven to earth. In the Taittirœya Sa∫hitå (1.7.9.1-2) the sacrificer tells his wife: jåya ehi suvo rohåva rohåva hi suvar aha∫ nåv ubhayoΔ suvo rokßyåmi [...] suvar devån aganmåmr¢tå abhüma prajåpateΔ prajå abhüma sam aha∫ prajayå sa∫ mayå prajå sam aha∫ råyaspoßeña sa∫ mayå råyaspoßaΔ, “Come here, wife, let us climb to heaven, let us, indeed, climb to heaven. I will climb to heaven for both of us [...]. To heaven, to the gods we have come! We have become immortal! We have become Prajåpati’s children! May I be joined with children; may children be joined with me! May I be joined with increase of wealth; may increase of wealth be joined with me!” Besides the union of husband and wife in heaven, this passage also highlights the connections among several things we have examined: sacrifice, heaven, children, wealth. Women, in this ritual theology, are not independent ritual actors.28 They can only hope to get to heaven as wives, hanging on to they were anåtmånaΔ. The sacred fire is the amr¢ta. The gods made it their self and became immortal. See ‡B 10.4.3.8-10; 11.1.2.12-13; TS 2.3.2.1. 26 See ‡atapatha Bråhmaña 2.3.4.4. 27 The world of gods in the Vedas has clear spatial dimensions. It is up above beyond the atmosphere–sometimes associated with the sun, sometimes with the moon which contains the divine ambrosia, and sometimes with the milky way, the divine river in the sky. See Witzel 1984. 28 This is, of course, not an invariable rule. Jamison (1996a, 36-38) has shown that

A∫r¢tå

113

their husbands’ ritual coat-tails, but at least they can get there. We have come some way from immortality defined as the son where the wife and women are instruments but not participants.

VI. Mokßa and Immortality All the technologies of immortality we have looked at thus far are located within the social context of family, rituals, and wealth. Indeed, the possession of wealth was regarded as a prerequisite for performing any rite, many of which were quite expensive to conduct, requiring the services of numerous ritual experts.29 To put it crassly, you have to be rich to become immortal. The middle of the first millennium BCE was a watershed in the cultural and religious history of India. Drastic social, political, and economic changes took place during this period principally in the Gangetic plain. A surplus and complex economy, the creation of larger political units coming close to state formation (Thapar 1984), facility of travel, trade, and urbanization–all contributed to the emergence of new religions such as Buddhism and Jainism based on ideologies and religious practices very different from the vedic,30 ideologies that underlie at least some of the literature of the late vedic period, including the Upanißads. Let me highlight two elements of the new religious culture which underscore the emerging centrality of self as moi. They are the belief in rebirth and the institution of world renunciation. Rebirth asserts the continuity of individual identity across life times both in the past and into the future. The same individual is born, dies, and is reborn, repeating this cycle indefinitely. The relationships and roles a person establishes within a given lifetime–wife/husband, children, and parents, as well as kinship, caste, professional, and political ties–, relationships that constitute a personne, are all fleeting and do not constitute one’s self, one’s moi. Indeed, over several lifetimes an individual will enter into and sever many such relationships, as presented graphically in a passage from the Råmåyaña (2.98.25-26): “As two pieces of wood might meet upon the open sea and, having met, drift apart after a few brief within very restricted parameters women can act as sole ritual agents, that is, perform rites unaccompanied by their husbands. In spite of these exceptions, however, the theology of ritual dependence of women remained firm. Manu, as usual is explicit and emphatic: “No sacrifice, no vow, no fast must be performed by women apart (from their husbands)” (MDh 5.155); “By a girl, by a young woman, or even by an aged one, nothing must be done independently, even in her own house” (MDh 5.147). 29 That only a sufficiently rich man was entitled to perform a Soma sacrifice is clearly stated even is as late a text as Manu: “A man who possesses a supply of food sufficient to maintain his dependents for three years or more is entitled to drink Soma [i.e., to perform a Soma sacrifice]. If a twice-born man possessing less wealth drinks Soma, he derives no benefit from it, even though he has drunk Soma.” MDh 11.7-8. The BU (1.4.17) expresses the reason for desiring wealth: “I wish I had wealth so I could perform rites.” 30 For a longer discussion with bibliography, see Olivelle 1993, 55-70.

114

LANGUAGE, TEXTS AND SOCIETY

moments, so too do your wives and children, your relatives and riches meet with you and hasten away” (tr. Pollock). The doctrine of karma, furthermore, proclaims that an individual is the architect of his or her own future. A renouncer does here and now what death does at the end of life; he severs the social relationships that constitute his personne. An ascetic leaves home and family, severs all kinship and economic ties, and lives as a homeless, wandering mendicant. He loses all title to property, his marriage is dissolved ipso facto, and he is often regarded as ritually dead.31 The new ideology puts into question many of the central elements of previous conceptions of immortality. Let us take son and semen, for example. In the rebirth ideology the son is a separate individual with a long series of prior births and deaths; he is not the continuation of the father and his relation to his present father is contingent at best. The new ideology also questions older ideas about the meaning of semen. In one of the earliest attempts to describe the process of rebirth (BU 6.2.9-16; CU 5.3-10; cf. JB 1.45-46), the self of the deceased person is said to go up as smoke to the sky. It finally reaches the moon and comes down as rain. The individual, now transformed into water, is absorbed into plants and finally becomes food. A man eats that food and transforms it into semen, which he deposits in a woman, giving rise to a new birth of the dead man. As opposed to the doctrine which sees the semen as the condensed self of the father, the new doctrine sees the semen as totally another person transformed into a new dimension. The father is a mere conduit, a reprocessing machine, for another being who is reborn by means of his semen. The relationship of the current generation to its forefathers and the imperative of begetting sons are also put into question. The Br¢hadårañyaka Upanißad (4.4.22) has this to say about people who have discovered their true self, their moi: etad dha sma vai tat pürve vidvå∫saΔ prajå∫ na kåmayante Ù ki∫ prajayå karißyåmo yeßå∫ no ‘yam åtmåya∫ loka iti Ù te ha sma putraißañåyå† ca vittaißañåyå† ca lokaißañåyå† ca vyutthåyåtha bhikßåcarya∫ caranti, “It was when they knew that that men of old did not desire offspring, reasoning: ‘Ours is this self, it is our world. What then is the use of offspring for us?’ So they gave up the desire for sons, the desire for wealth, and the desire for worlds, and undertook the mendicant life.” Note again how sons, wealth, and worlds are brought together here as the objects of desire; the desire for worlds demand the desire for the means of attaining them, namely, sons and wealth (= rites). Another passage responds to the older idea that a wife completes the self of the husband: tasmåd apy etarhy ekåkœ kåmayate jåyå me syåd atha prajåyeyåtha vitta∫ me syåd atha karma kurvœyeti Ù sa yåvad apy 31

For the legal consequences of renunciation within Hindu law, see Olivelle 1984.

A∫r¢tå

115

eteßåm ekaika∫ na pråpnoty akr¢tsna eva tåvan manyate Ù mana evåsyåtmå våg jåyå pråñaΔ prajå cakßur månußa∫ vitta∫ … †rotra∫ daiva∫.., “So even today when one is single, one has the desire: ‘I wish I had a wife so I could father offspring. I wish I had wealth so I could perform rites.’ As long as someone has not obtained either of these, he considers himself to be utterly incomplete. Now, this is his completeness –his mind is his self; his speech is his wife; his breath is his offspring; his sight is his human wealth […]; and his hearing is his divine wealth [...]” (BU 1.4.17). This is as clear a statement of self as moi as we can expect: one’s self is self-contained, it does not need external things or relationships to make it complete. In the new ideology, immortality is not seen as a form of survival after death; within the rebirth ideology survival is guaranteed to all. Neither is it some desirable location (e.g., heaven) after death, for now all those locations are regarded as way-stations in the unending cycle of births and deaths. Immortality is the liberation from that cycle, from being subject to repeated births and deaths, by means of some type of secret and powerful knowledge. How do women fare in this new ideological climate? An answer to this question emerges in the story about Yåj∞avalkya and his two wives related twice in the Br¢hadårañyaka Upanißad (2.4; 4.5): “Now, Yåj∞avalkya had two wives, Maitreyœ and Kåtyåyanœ. Of the two, Maitreyœ was a woman who took part in theological discussions, while Kåtyåyanœ’s understanding was limited to womanly matters. One day, as he was preparing to undertake a different mode of life, Yåj∞avalkya said: ‘Maitreyœ, I am about to go away from this place. So come, let me make a settlement between you and Kåtyåyanœ.’ Maitreyœ asked in reply: ‘If I were to posses the entire world filled with wealth, sir, would it, or would it not, make me immortal (amr¢tå).’ ‘No,’ said Yåj∞avalkya, ‘it will only permit you to live the life of a wealthy person. Through wealth one cannot expect immortality.’ ‘What is the point in getting something that will not make me immortal?’ retorted Maitreyœ. ‘Tell me instead, sir, all that you know.’”32 A couple of points about this story: first, this is possibly the first time in the whole of the vedic literature that the feminine adjective amr¢tå (“immortal”) is used with reference to an actual flesh-andblood woman, the adjective I have used as the title of this paper.33 Maitreyœ dares to ask –or, to be more precise, the author dares to put this question in Maitreyœ’s mouth–how she can become immortal, 32 atha ha yåj∞avalkyasya dve bhårye babhüvatur maitreyœ ca kåtyåyanœ ca Ù tayor ha maitreyœ brahmavådinœ babhüva Ù strœpraj∞aiva tarhi kåtyåyanœ Ù atha ha yåj∞avalkyo ‘nyad vr¢ttam upåkarißyan Ù maitreyœti hovåca yåj∞avalkyaΔ Ù pravrajißyan vå are ‘ham asmåt sthånåd asmi Ù hanta te ‘nayå kåtyåyanyånta∫ karavåñœti Ù så hovåca maitreyœ yan nu ma iya∫ bhagoΔ sarvå pr¢thivœ vittena pürñå syåt syå∫ nv aha∫ tenåmr¢tåho neti Ù neti hovåca yåj∞avalkyaΔ Ù yathaivopakarañavatå∫ jœvita∫ tathaiva te jœvita∫ syåt Ù amr¢tatvasya tu nå†åsti vitteneti Ù så hovåca maitreyœ yenåha∫ nåmr¢tå syå∫ kim aha∫ tena kuryåm Ù yad eva bhagavån veda tad eva me brühœti Ù BU 4.5.1-4 33 The feminine is used with regard to goddesses, as well as objects that are grammatically feminine: e.g., ¥V 1.113.13; AV 10.8.26.

116

LANGUAGE, TEXTS AND SOCIETY

and she does that as an individual in her own right and not just as the wife of Yåj∞avalkya. Indeed, if the wife accompanied the husband into immortality, as we saw within the context of the sacrifice, then Maitreyœ needn’t have worried; she would have become immortal automatically as part of her husband’s self. But now the secret to immortality is not the sacrifice which demands wealth but knowledge which cannot be obtained vicariously. To become immortal Maitreyœ did not need wealth; she had to know what Yåj∞avalkya knew. In an interesting reversal, the acknowledgment of female agency with respect to immortality is accompanied by the denial to women of an instrumental role in the acquisition of male immortality.34 The passage I cited above BU (1.4.17) tells a man not to consider himself incomplete if he lacks a wife or son; his completeness lies within himself, and his wife is his speech. In a curious but significant way male independence from women created independence for women as well, and vice versa. Women’s liberation obviously did not come to India with the rebirth/liberation ideology, but at least now women become part of the discourse on human aspiration for immortality. This is nicely demonstrated by the third and fourth books of the Br¢hadårañyaka Upanißad, a section that originally formed the conclusion of the great ‡atapatha Bråhmaña35 and constitutes what I would call “The Triumph of Yåj∞avalkya,” the individual responsible for the composition of the White Yajurveda. This triumphant conclusion of that Veda contains four episodes that establishes Yåj∞avalkya as the foremost theologian of the time. King Janaka appears as a main figure in the first three; the fourth is the conversation between Yåj∞avalkya and his wife Maitreyœ. Renowned for his knowledge,36 the presence of Janaka is a deliberate literary strategy of the author to highlight Yåj∞avalkya’s supremacy in knowledge. If he can teach Janaka, a listener will think, he must be the greatest! In the first of these episodes a group of distinguished theologians have assembled at the king’s court, and Janaka wants to find out who among them is the most learned. In this group of theologians is one woman, Gårgœ Våcaknavœ. Her mere presence would have raised many a Bråhmañical eyebrow, and I think that this was a deliberate literary strategy of the author. Within the literary structure of this story, she 34 At a much later time, in Buddhist and Hindu tantric traditions where women become the sexual partners and objects in Tantric rituals aimed at liberation, women emerge once again as instruments of male immortality. 35 In the extent BU this section is followed by another containing chapters 5 and 6 of the BU. But even the indigenous tradition recognized this as a supplement, calling it khilakåñÛ∂a; it was probably added at a later time to the original BU consisting of chapters 1-4. 36 The fame of Janaka’s wisdom and learning is revealed in a remark by King Ajåta†atru. When Bålåki tells him that he will teach him brahman, Ajåta†atru exclaims: “We’ll give you a thousand cows for such a speech! People are sure to rush here, crying, ‘Here’s a Janaka! Here’s a Janaka!’ ” (BU 2.1.1; KßU 4.1). To be called a second Janaka appears to have been a great compliment.

A∫r¢tå

117

plays a crucial role precisely because she is a woman. First, Gårgœ is the only member of the group who questions Yåj∞avalkya twice. On the second occasion she makes this boast to her male colleagues [BU 3.8.12]: bråhmañå bhagavanto hantåham ima∫ dvau pra†nau prakßyåmi Ù tau cen me vakßyati na vai jåtu yußmåkam ima∫ ka†cid brahmodya∫ jeteti, “Distinguished Brahmins! I am going to ask this man two questions. If he can give the answers to them, none of you will be able to defeat him in a theological debate.” Then she challenges Yåj∞avalkya: aha∫ vai två yåj∞avalkya yathå kå†yo vå vaideho vograputra ujjya∫ dhanur adhijya∫ kr¢två dvau båñavantau sapatnåtivyådhinau haste kr¢tvopottiß™hed evam evåha∫ två dvåbhyå∫ pra†nåbhyåm upodasthåm Ù tau me brühœti, “I rise to challenge you, Yåj∞avalkya, with two questions, much as a fierce warrior of Kå†i or Videha, stringing his unstrung bow and taking two deadly arrows in his hand, would rise to challenge an enemy. Give me the answers to them!” The military image is interesting; as Mrs. Thatcher is said to have been in her cabinet, Gårgœ is presented here as the only man in that male assembly. The image is also interesting because both of Gårgœ’s questions contain the metaphor of weaving, in all likelihood an occupation closely associated with women (see Rau 1970). The author of the tale mixes his metaphors, so to speak, to create the wonderful character of Gårgœ–woman, theologian, and warrior. After Yåj∞avalkya answers her questions, she turns to her male colleagues: bråhmañå bhagavantas tad eva bahu manyedhva∫ yad asmån namaskåreña mucyedhvam Ù na vai jåtu yußmåkam ima∫ ka†cid brahmodya∫ jeteti, “Distinguished Brahmins! You should consider yourself lucky if you escape from this man by merely paying him your respects. None of you will ever defeat him in a theological debate” [BU 3.8.12]. She, in other words, tells her male colleagues: “If I can’t beat him, none of you can!” The leader of the group, ‡åkalya, did not heed her warning, and lost his head after he was defeated by Yåj∞avalkya. The triumph of Yåj∞avalkya concludes with his instruction of another woman, his wife Maitreyœ. The prominence of women in this section, I think, is a literary strategy intended to show the triumph of the new ideology connected with Yåj∞avalkya: even women can understand the truth better than those old fogies! Going beyond the Bråhmañical tradition, we see in Buddhism and Jainism the world’s first voluntary organizations for women–the Buddhist and Jain orders of nuns. Here are women who leave their families, break their kinship ties, and voluntarily and possibly against the wishes of their families enter a community of celibate women. This is quite a departure from Manu’s dictum: “A girl, a young woman, or even an old woman should not do anything independently, even in her own house. In childhood a woman should be under her father’s control, in youth under her husband’s, and when her husband is dead, under her son’s. She should never be independent” (MDh 5.147-48). Here are Buddhist and Jain nuns exercising a daring freedom of choice, living lives in female communities outside direct

118

LANGUAGE, TEXTS AND SOCIETY

male control, and taking control of their own sexuality. What better example of the ultimate triumph of the self as moi.

VII. The More Things Change … The chart I have drawn and the way I have structured my comments may leave the impression that the path from personne to moi and the concomitant changes in the position of women vis-à-vis immortality chart a clear chronological line, each ideological change leading its predecessors in the dust. This, of course, is far from the truth. Although there is some detectable chronology to these changes, given the problems inherent in dating Indian texts with any degree of accuracy and certainty, building a chronology of these changes is very much like building a house of cards. More importantly, however, in India older ideologies did not change, yielding place to new; as Louis Dumont (1960) has accurately observed, Indian religious history by and large has moved by way of aggregation, putting new stones on old, rather than substitution. So, for example, rites and ideologies involving sons and the world of the fathers coexisted and continue to coexist in India side by side with the ideologies of rebirth and liberation. Attempts, however, were made to synthesize the differing conceptions of death and immortality. The Aitareya Upanißad (2.1-4), for example, attempts a synthesis in terms of the three births of a man: At the outset, this embryo comes into being within a man as semen. This radiance gathered from all the bodily parts he bears in himself (åtman) as himself (åtman). And when a man deposits [lit. pours] it in a woman, he gives birth to it. That is his first birth. It becomes one with the woman’s body (åtman), as if it were a part of her own body. As a result, it does not harm her. And she nourishes this self (åtman) of his that has entered her. As she nourishes him, so he should nourish her. The woman carries him as the embryo. At the beginning, he nourishes the child even before its birth. When he nourishes the child even before its birth, he thereby nourishes himself (åtman) for the continuance of these worlds, for it is in this way that these worlds continue. That is his second birth. And he–this self (åtman) of his–is appointed to carry out holy rites, while his other self, after it has done all it has to do, becomes old and dies. As soon as he departs from this world, he is born again. That is his third birth.37

Here we have a curious combination of “selfhoods” in terms of new births, that is, taking on new selves. The first self/birth is when one ejaculates oneself as one’s semen into a woman. Here the coming out of the semen from one’s body is viewed as a birth, just as the puruße ha vå ayam ådito garbho bhavati yad etad retaΔ Ù tad etat sarvebhyo 'õgebhyas tejaΔ sa∫bhütam åtmany evåtmåna∫ bibharti Ù tad yadå striyå∫ si∞caty athainaj janayati Ù tad asya 37

A∫r¢tå

119

coming out of the baby from the mother’s body. Both these hark back to the conception of self as son. The self of his that is the son continues his ritual tasks after his death, while his “other self” takes birth anew. This third birth is, of course, based on the rebirth ideology. Another attempt at synthesis is found in the Br¢hadårañyaka Upanißad (1.5.16) in terms of the different worlds that a man aspires to win: atha trayo våva lokå manußyalokaΔ pitr¢lokaΔ devaloka iti Ù so ‘ya∫ manußyalokaΔ putreñaiva jayyo nånyena karmañå Ù karmañå pitr¢lokaΔ Ù vidyayå devalokaΔ Ù devaloko vai lokånå∫ †reß™haΔ Ù tasmåd vidyå∫ pra†a∫santi, “ Now, there are only three worlds: the world of men, the world of ancestors, and the world of gods. One can win this world of men only through a son, and by no other rite,38 whereas one wins the world of ancestors through rites, and the world of gods through knowledge. The best of these, clearly, is the world of gods, and for this reason they praise knowledge.” The coexistence of different notions of self and of after-death states makes for a messy situation. Clarity is often achieved in such situations only at the expense of accuracy. The Hindu funeral rite is a good example of this “mess theory” of Indian religious history. Theoretically, this is the one rite that should bring to the foreground a culture’s ideas about death and afterlife. Given that the dominant ideology in Indian culture with respect to afterlife is rebirth, we should expect to find this belief clearly articulated in the funeral rite. The opposite is, in fact, the case. An observer looking solely at the Hindu funeral rite will have no idea that Indians believed in rebirth; this belief is completely ignored in favor of the ideology of the world of the fathers. The same is true of all rites connected with ancestral offerings. In the religions rooted in the ideologies of rebirth and liberation and wedded to the concept of self as moi, such as Buddhism and Jainism, furthermore, we should expect to find the claim of women to liberation and immortality articulated loud and clear. But that is not the case. There is great ambivalence and ambiguity in both traditions regarding the spiritual capabilities of women. The Buddha is supposed to have opened his monastic order to women with great reluctance and at the urging of his favorite disciple Ånanda; after giving his reluctant assent, he is said to have predicted that had women not been admitted the Buddhist way of life would have lasted 1000 years, but because women had been admitted it will last only 500 years (Lamotte 1958, 211). We do not know, of course, what the Buddha prathama∫ janma Ù tat striyå åtmabhüya∫ gacchati yathå svam aõga∫ tathå Ù tasmåd enå∫ na hinasti Ù såsyaitam åtmånam atra gata∫ bhåvayati Ù så bhåvayitrœ bhåvayitavyå bhavati Ù ta∫ strœ garbha∫ bibharti Ù so 'gra eva kumåra∫ janmano 'gre 'dhi bhåvayati Ù sa yat kumåra∫ janmano 'gre 'dhi bhåvayaty åtmånam eva tad bhåvayaty eßå∫ lokånå∫ sa∫tatyå eva∫ sa∫tatå hœme lokåΔ Ù tad asya dvitœya∫ janma Ù so 'syåyam åtmå puñyebhyaΔ karmabhyaΔ pratidhœyate Ù athåsyåyam itara åtmå kr¢takr¢tyo vayogataΔ praiti Ù sa itaΔ prayann eva punar jåyate Ù tad asya tr¢tœya∫ janma Ù 38 Note how the sexual act of procreation is implicitly regarded as a rite alongside "other rites". For passages that regard sex as a sacrifice, see BU 6.2.13; 6.4.1-4.

120

LANGUAGE, TEXTS AND SOCIETY

himself thought; but the story is direct evidence of the great ambivalence the Buddhist male monastic community (we have no idea as to what the nuns themselves thought about this) must have felt about the order of nuns and the spiritual capabilities of women. Diana Paul (1985) details the debates that raged within Mahåyåna Buddhism regarding the possibility of women becoming a Buddha. And more recently Padmanabh Jaini (1991) has chronicled the debates between the Digambara and ‡vetåmbara Jains about the possibility of women attaining liberation as women, that is, without being reborn as men. The same ideologies that brought a modicum of agency to women also created ascetic traditions that considered women as temptresses and created a truly misogynic literature (Olivelle 1995a). Within the Brahmanical/Hindu traditions there has been an ongoing controversy whether it was legitimate for women to become ascetics (Olivelle 1993, 183-90). And as we have seen, the legal codes, composed many centuries after the emergence of the rebirth ideology, repeatedly insist that women are never to be independent agents either in the legal/economic or in the ritual/religious field (MDh 5.147-56; see note 28). Nevertheless, within the rebirth ideology women have at least had the opportunity to choose not to marry, to control their sexuality, to live independently of male authority, to pursue individual aspirations, and even to become religious élite. They do not have to be mere instruments of their husbands’ aspirations and have the opportunity to become agents of their own destiny.

A

121

V Power of Words: The Ascetic Appropriation and the Semantic Evolution of dharma*

The closest Sanskrit equivalent to the modern concept of “asceticism” is probably tapas, associated both in etymology and theological discourse with fire and heat, and belonging to a group of terms and concepts, the most significant of which is tejas, that reveals an ancient Indian ontology of individualized power–a source of energy located within an individual, whether it is a god, a warrior, or an ascetic.1 It is this energy that makes the wrath of a tapasvin, the ascetic in possession of tapas, so potent and so feared. Ascetics as “powerful men” is an image so ingrained in Indian culture and discourse as to require little comment. Recall the story of the Buddha overpowering the fire of the snake with his own, thus dumbfounding and then converting his Brahmin hosts. At a more scholastic level, there is the list of siddhis, the superhuman powers such as levitation, multi-location, and the ability to kill, believed to result from the practice of Yoga. My paper, however, is not directly concerned with these modes of ascetic power or with the ontologies that underlie them and the technologies that generate them. My concern here is with another type of power–power as a social reality, power that defines and is defined by social relationships. These two kinds of power, however, are related; we know that the projection of ascetic power, carried out especially in the form of miracles, is a fruitful way of gaining followers, resources, and even political patronage, although ancient Indian politicians, including Kau™ilya the author of the Artha†åstra, the most famous Indian work on polity and politics, show both a healthy skepticism of * To be published in Asceticism and Power in the Asian Context edited by Peter Fluegel and Gustaaf Houtman. London: Curzon. 1 For a discussion, see Whitaker 2000.

122

LANGUAGE, TEXTS AND SOCIETY

such miracles and a willingness to use popular credulity in such powers to further their political aims. Ascetics individually, but ascetic institutions especially, needed and coveted power within society–the power of numbers by recruiting lay and religious followings, the power of physical resources by soliciting grants and donations, and the power of political patronage by gaining influence among kings and rulers. The methods used by various ascetic institutions for these purposes were undoubtedly many and varied. I want here to focus on one: the employment of royal symbols and vocabulary by early ascetic leaders and institutions to lay claim to a new and different type of royal authority. Buddhism and Jainism offer good examples. The founders of these ascetic groups are called jina, “conqueror”; they are called cakravartin, “roller of the wheels” or universal emperor. The Buddha’s doctrine is compared to a wheel, a metonym for the war chariot and conquest; and his first sermon is the dharmacakrapravartanasütra, “the Sütra that set the wheel of Dharma rolling”. The Buddha’s teaching is †åsana, the counterpart of a royal edict. These are all clearly royal symbols used, deliberately I think, to define new ascetic groups and new religious ideologies. Siddhårtha Gautama and Mahåvœra are, of course, given royal pedigrees. As I have shown elsewhere (Olivelle 1993), claims to royal heritage were not limited to these new religions. The Brahmanical tradition places several new teachings of the Upanißads in the mouths of kings; and, of course, Råma and Kr¢ßña, divine incarnations and central figures of new devotional religions, are kings, not Brahmins. Such appropriations of royal symbols are, of course, not unique to India; consider the figure of Christ the King within the Christian tradition. The thesis that I want to propose here is that one such item of the royal lexicon co-opted and redefined within ascetic traditions, and then absorbed back into the royal and public domains, was dharma, clearly the most central and pivotal term and concept in the whole of Indian culture. If my thesis is correct, we can interpret this co-optation of a royal term as one more indicator of the use of royal symbols by ascetic institutions to garner power, prestige, and influence by appealing to a new generation of élite created by state and imperial formations in northern India around the middle of the first millennium before the common era.

Dharma in Vedic Literature It may come as a surprise to many that dharma is not a central term in the vedic lexicon during its 1000 years or so of history, from the earliest strata of the R¢gveda to the early Upanißads, as we are generally led to assume. It is not possible to present a detailed study of dharma in the vedic literature within the confines of this paper. The semantic range of the term is outlined in the excellent study by Paul

Power of Words

123

Horsch (1967). I have dealt with its semantic history and its possible socio-political underpinnings in a Olivelle 2004b (below, pp. 137-154). Here I can present only a thumbnail sketch. We encounter the term for the first time in the R¢gveda, always in the neuter form dharman. It occurs 67 times in 55 separate hymns, 59 times as an independent word and 9 time within a compound.2 The distribution of the term within the Mañ∂alas [Books] is instructive: it occurs 50 times in Books 1, 8, 9, and 10, which are the non-family Books, generally viewed as belonging to the later strata of the R¢gveda, and only 18 times within the family Books. This distribution, plus the fact that no cognate term exists in Avestan or other Indo-European languages, raise the possibility that dharman entered the R¢gvedic vocabulary somewhat late and possibly as a neologism. The term is used with reference to at least 11 gods, but its association with Mitra and Varuña, often within their dual personality as Mitrå-Varuña, is unmistakable. The precise meaning of dharman in its various contexts is difficult to determine, but in general it has the meaning of “institute” in the sense of constitution or foundation in a ritual, legal, and moral sense, thus containing the meanings of statute, command, rule, and law.3 This meaning, as Horsch (1967) has suggested, originated from its etymological connection to the verb dhr¢, “bearing or holding,” especially in the sense of holding apart (vi-dhr¢)4 of heaven and earth: dyåvåpr¢thivœ varuñasya dharmañå vißkabhite ajare bhüriretaså: “The heaven and earth, unaging and with abundant seed, were propped apart by the institute of Varuña” (R¢gveda 6.70.1). Horsch (1967) has rightly pointed out that the cosmogonic dimension of dharma is its basic and possibly earliest meaning, directly connected with it etymological derivation. He has also noted the closely related ritual meaning of the term, and the somewhat extended moral and social dimensions. The moral dimension of the term can be detected within the R¢gveda itself. This should come as no surprise, because Varuña, besides being the sovereign, is also the god of moral law, and his companion, Mitra, is the god of compact and contract. In R¢gveda 7.89.5 we read: acittœ yat tava dharmå yuyopima må nas tasmåd enaso deva rœrißaΔ –“When without thought we have transgressed your institutes, O god, may you not punish us because of that evil.” This is echoed in the Atharvaveda 6.51.3: yat ki∫ ceda∫ varuña daivye jane ‘bhidroha∫ manußyå† caranti Ù acittyå cet tava dharmå yuyopima må nas tasmåd enaso deva rœrißa –“Whatever, O Varuña, that is hateful to the people of the gods human beings practice here, if without intention we have 2 Six of these nine compounds consist of the repeated refrain satyadharman. The three other compounds occur just once each: svadharman, dharmakr¢t, and dharmavat. 3 For discussions of the term dharma, see Halbfass 1988; Horsch 1967; Renou 1964, 1978; Wezler 1999; Brereton 2004. This article was written before the appearance of Brereton’s study of dharman in the ¥V. 4 From this verb is derived vidharman, occurring 14 time in the R¢gveda, which has the more specific meaning of holding apart.

124

LANGUAGE, TEXTS AND SOCIETY

obstructed your institutes, do not, O god, harm us for that sin” [Tr. Whitney]. Within this context, however, the moral dimension is not divorced from the legal: law and morality, crime and sin, punishment and penance are all closely linked in ancient Indian discourse. One would have expected that this term, which made such a promising start in the R¢gveda, would grow in significance and centrality in the vocabulary of the middle and late vedic periods. This, however, was not to be the case. Except for changing its gender from the neuter dharman to the masculine dharmaΔ, it became somewhat of a marginal term in the later vedic lexicon. Indeed, its use within the other vedic Sa∫hitås, the Bråhmañas, and even the Upanißads, which together form an enormous corpus of literature, is infrequent. Parallel with its relative infrequency, we can detect also a gradual semantic restriction of the term. It is used most frequently with reference to Varuña, the heavenly king par excellence, and to his earthly counterpart. If we exclude verses cited from the R¢gveda and repetitions, the Atharvaveda, which is next in age to the R¢gveda, uses the term in both its masculine and its neuter forms 13 times; all the Yajurvedic Sa∫hitås combined use it 28 times; in the enormous corpus of the Bråhmañas, it occurs just 13 times. Even in the four ancient prose Upanißads, texts where one would expect dharma to play a central role, it is found in only 9 passages.5 A startling statistical fact emerges from this: the term dharma occurs 68 times within the relatively short R¢gveda, whereas it is used in just 63 independent passages within the rest of the vedic literature. I think we need to rethink our assumptions regarding the centrality of dharma within the cultural history of India. The clear direction of the semantic development of the term within the later vedic texts, moreover, relates to the earthly king. His connection to Varuña, and the dual Mitra-Varuña, is well established. Gonda (1959, 113) notes: “råjånau [two kings] has in connection with Mitra and Varuña assumed, to a certain extent, the character of a fixed phrase.” It is Varuña, and not Indra, who is associated with the royal duties of the king and especially with his royal consecration. Heesterman (1957, 85-86) observes: “The råjasüya [royal consecration] is also known as the Varuñasava. This latter name indicates that the royal sacrificer being anointed impersonates [or better, assumes the persona of] the god Varuña: ‘It is Varuña whom they anoint’, as ‡åõkhåyana ‡rautasütra 15.13.4 asserts.”6 5 At Chåndogya 7.2.1 [repeated at 7.7.1], Br¢hadårañyaka 2.5.11; 4.4.5, the term occurs within lists of categories and is not singled out for comment. More substantive discussions of the term occurs in Br¢hadårañyaka 1.4.11-14; 1.5.23; Only at Chåndogya 2.23.1 and Taittirœya 1.11.1, 4 does dharma appear with meanings close to the classical. The uncertainty as to the chronology of these two Upanißads makes it difficult to assess the significance of these usages. My own belief is that these Upanißads are later than and influenced by the ascetic appropriation of the term. 6 Even as late a text as the Månava Dharma†åstra (9.244-245) considers Varuña to be the lord of kings and judicial punishment.

Power of Words

125

Within the ritual for the royal consecration, for example, Varuña is addressed as dharmapati, “the lord/protector of dharma”; the new king is explicitly said to be Varuña with respect to dharmas: varuña eva dharmapatir dharmasya pati∫ karoti paramatå vai så yo dharmasya patir asad –“Varuña himself, the lord of dharma, makes him [the king] the lord of dharma. This, clearly, is supremacy, that he is the lord of dharma” (‡atapatha Bråhmaña 5.3.3.9). Dharma is thus closely linked to the king, to the exercise of his power, and to his duties as the guardian of social and moral order. An interesting passage in the ‡atapatha Bråhmaña (11.1.6.24) links dharma to the northern quarter and water, indicating that social order depends especially on the bounty of the soil; it is the king’s special duty to ensure the timely arrival of rain: athodœcœ∫ di†am apa†yat Ù tåm apo ‘kurvatopainåm itaΔ kurvœmahœti ta∫ dharmam akurvata dharmo vå åpas tasmåd yadema∫ lokam åpa ågacchanti sarvam eveda∫ yathådharma∫ bhavaty atha yadåvr¢ß™ir bhavati balœyån eva tarhy abalœyasa ådatte dharmo hy åpaΔ Ù They then saw the northern quarter. They made it the waters, saying: “Let us assist it from here!”. They made it dharma; the waters are dharma. Therefore, when the waters come to this world everything here comes to be in accordance with dharma; but when there is drought, then the stronger seize the weaker; for the waters are dharma.

One of the most suggestive passages is Br¢hadårañyaka Upanißad 1.4.11-14, which confirms the close connection between dharma and kingship. This passage views dharma as the power that stands above the king, a view that comes close to the notion of absolute justice, which even the king is obliged to follow; it constitutes the essence of kingship. The Upanißadic story deals with creation. Brahman created all beings one by one, including the four social classes: Bråhmaña, Kßatriya, Vai†ya, and ‡üdra. After the creation of each item, we hear the refrain: “It still did not become fully developed” –sa naiva vyabhavat. Finally Brahman created dharma as the power superior to the ruling power (kßatra): tacchreyo rüpam atyasr¢jata dharmam Ù tad etat kßatrasya kßatra∫ yad dharmaΔ Ù tasmåd dharmåt para∫ nåsti Ù atho abalœyån balœyå∫sam å†a∫sate dharmeña Ù yathå råj∞aivam Ù yo vai sa dharmo satya∫ vai tat Ù tasmåt satya∫ vadantam åhur dharma∫ vadatœti Ù dharma∫ vå vadantœ∫ satya∫ vadatœti Ù It created dharma, a form superior to and surpassing itself. And dharma is here the ruling power [kßatra] of the ruling power. Hence there is nothing higher than dharma Therefore, a weaker man makes demands of a stronger man by appealing to dharma just as one does by appealing to a king. Now, dharma is nothing but the truth. Therefore, when a man speaks the truth, people say that he speaks dharma; and when a man speaks dharma, people say that he speaks the truth.

126

LANGUAGE, TEXTS AND SOCIETY

The expression kßatrasya kßatram (lit. “ruling power of the ruling power”) is noteworthy. The identical formula is used frequently in the phrase satyasya satyam (“the real of the real”, Br¢hadårañyaka 2.1.20)7 Such expressions appear to refer to the essential core of something. In the latter formula, it refers to the “truly real” or the quintessence of what is normally perceived as the real. In other expressions, such as “the sight of sight” or “the hearing of hearing” (see Br¢hadårañyaka 4.4.18; Kena Upanißad 1.2), the meaning appears to include the power or the entity that is behind the external act of seeing or hearing, i.e., that without which hearing would not take place. In this sense food is said to be the “breath of breath” in Taittirœya Årañyaka 3.7.3. In the expression kßatrasya kßatram, therefore, dharma is viewed as both constituting the very essence of kingship and the transcendent power that lies behind the visible power and authority of the king. Even though dharma here stands above the king and gives legitimacy to his rule, it is still intimately connected with the discharge of royal functions, with the internal regulation of society, and with the judicial process. The connection of dharma to the public and judicial sphere is nicely illustrated by one ritual detail. In the Purußamedha, the real or symbolic human sacrifice, different types of men are sacrificed to various deities. The kinds of men sacrificed to dharma and its opposite adharma are instructive: to dharma, a sabhåga (a man who participates at the royal audience hall where judicial proceedings are carried out), and to adharma, a deaf man (badhira), quite the opposite of the debater in the assembly hall (Taittirœya Bråhmaña, 3.4.1). Besides the vedic literature, Brahmanical texts that may have some claim to being pre-Buddhist or at least pre-A†okan are the ‡rauta-sütras, which give the ritual procedures of vedic or solemn sacrifices, and the Gr¢hya-sütras, which deal with domestic rites. Dharma occurs just 38 times in the 10 ‡rauta-sütras I have examined, and in all but one of its occurrences8 the term has a clear technical meaning: dharma is a specific ritual detail or a ritual unit. The normal way that these texts proceed is to describe fully one basic rite and then show what ritual details [dharmas] of this are transferred to other similar rite. For example: sarveßv iß™ipa†ubandheßu dår†apaurñamåsikå dharmå anuyanti –“The dharmas of the New-moon and the Full-moon sacrifices are carried over into all Iß™is and Animal sacrifices” (Bhåradvåja ‡rautasütra, 6.15.5). 7 For an extensive study of this formula, see H. Oertel, Zum altindischen Ausdrucksverstärkungstypus satyasya satyam “das Wahre des Wahren” = “die Quintessenz des Wahren.” Sitzungsberichte der Bayerischen Akademie der Wissenschaften, Philosophischhistorische Abteilung, Jahrgang 1937, Heft 3. Munich: Verlag der Bayerischen Akademie der Wissenschaften. 8 This is Bharadvåja ‡rautasütra 9.18.3: yad åryåñå∫ dharmaj∞ånå∫ dharmakåmånåm abhojanœya∫ na tena devån yajeta –“He should not offer to gods what Åryas who know the dharma and who love the dharma would consider as unfit to be eaten.” The use of dharma here is so classical that it casts doubt about the antiquity of this section of the sütra.

Power of Words

127

The Gr¢hya-sütras deal with domestic rites such as initiation and marriage and contain a lot of material found also in the Dharmasütras. They are one class of texts that would be expected to make full use of the term dharma. But that is not the case. In all of the more ancient sütras, the term occurs just 6 times, and only three of these carry anything close to the classical meaning.9 To recapitulate, I think there is an unmistakable link in the vedic usage between dharma and the king. The term is especially connected to social order, to the rules that govern society, and to the king’s duty to maintain order and to administer the judicial process.

Ascetic Appropriations and A†okan Edicts Given the rather peripheral role dharma plays in the vedic vocabulary and therefore in vedic theology, it comes as somewhat of a surprise that there should emerge within the Brahmanical tradition an entire genre of literature devoted to dharma –that is the Dharmasütras and later the Dharma†åstras. Even more surprisingly, these texts deal principally with how a Brahmin (and to a lesser degree other Årya males) should live his life, and not with socio-legal matters around which the late-vedic usage of the term revolved.10 What precipitated this move? Two things, I believe, happened in the intervening period. First, dharma came to occupy a central position in the religious, political, and perhaps even in the popular lexicon. Second, dharma underwent pari passu a semantic evolution that was able to underpin its new and central role within the Indian religious lexicon. Both these changes, I believe, were facilitated by the ascetic usage of dharma within its broader appropriation of royal symbols and vocabulary, on the one hand, and by the re-appropriation and redefinition of dharma by A†oka to articulate a new religio-moral foundation for the unprecedented imperial formation of which he was the principal architect.11 The frequency with which A†oka uses the term dharma in his brief edicts and inscriptions stands in sharp contrast to the vedic texts. By my count he uses it well over one hundred times, excluding repetitions in the same edict published in different locations. In all but 16 of these, dharma is used within a compound; and except in the openly Buddhist inscription (Minor Rock Edict 3) where we have the compound saddharma, dharma stands as the first member of a com9 ‡åõkhåyana 2.16.2; Påraskara 3.4.18; Kå™haka 54.6. Dharma in the classical meaning is found in ‡åõkhåyana 3.3.7; ņvalåyana 1.7; Baudhåyana 3.3.31. 10 Royal duties (råjadharma) and legal procedures and disputes (vyavahåra) take up very little space in the early Dharmasütras. 11 It is, of course, unclear whether A†oka’s understanding and use of dharma was unprecedented, or whether he built upon similar usages of his father and grandfather, the founders of the Maurya dynasty.

128

LANGUAGE, TEXTS AND SOCIETY

pound.12 This usage is quite unusual; dharma in these compounds appear to function almost as an adjective to qualify the final member of the compound–for A†oka everything is “dharmic”. A†oka used dharma as the central concept to construct a theology that gave legitimacy to his rule and a religious/moral foundation to his empire. Unlike the vedic texts, however, A†oka gives us quite a clear picture of what he meant by dharma. In Rock Edict 3 he instructs his ministers to teach the dharma in the following words: “Obedience to mother and father is good (sådhu). Giving (dåna) to friends, acquaintances, and relatives, and to Bråhmañas and ‡ramañas is good. Not killing is good. Spending little and possessing little is good.” Other lists add the proper treatment of slaves. Similar sentiments are expressed frequently in the edicts, which he refers to as dharmalipi –“inscriptions relating to dharma”–, or perhaps “dharmic inscriptions”. In the long Pillar Edict 7, where he uses dharma 30 times, he defines the practice of dharma as consisting of compassion, liberality, truthfulness, purity, gentleness, and goodness. A†oka boasts that previous kings had no officials called dha∫mamahåmåtå [dharmamahåmåtra]. These were officials or ministers specially commissioned to foster dharma among his people, as also to look into the affairs of ascetics sects (påßañ∂as). He also instructs his various officials to go on circuit every five years; they were to, among other duties, preach the dharma in the words I have cited above. A†oka also refers to many of his own activities as “dharmic”. Ten years after his consecration, he went to the site of Buddha’s enlightenment and calls this journey dharmayåtrå –clearly contrasting this with the normal form of royal yåtrå, the military expedition. In the past there was killing and suffering, but now things are different because A†oka is practicing dharma. The sound of drums [warfare] 12

PE = Pillar Edict; RE = Rock Edict; MRE = Minor rock edict. dha∫ma: 4 in MRE 1; 1 in MRE 3; 2 in RE4; 1 in RE12; 1 in RE13; 1 in RE15; 4 in PE1; 2 in PE2 = Total16; dhammakåmatå: 1 in RE13; 2 in PE1 = Total 3; dha∫maguña: 1 in MRE 2 = Total 1; dha∫maghoßa: 1 in RE 4 = Total 1; dha∫macaraña (calana): 5 in RE 4; 1 in RE15; 1 in PE4 = Total 7; dha∫mattha∫ba: 1 in PE7 = Total 1; dha∫madåna: 1 in RE9; 2 in RE 11 = Total 3; dhammådhithåna: 1 in RE5 = Total 1; dha∫maniyama: 4 in PE7 = Total 4; dha∫månugaha: 1 in RE9 = Total 1; dha∫månupa™œpatti: 3 in PE7 = 3; dha∫månusathi: 1 in RE3; 1 in RE4; 1 in RE8; 3 in RE13; 2 in PE7 = Total 8; dha∫månusåsanå: 1 in RE4 = Total 1; dhammanisita: 1 in RE5 = Total 1; dha∫måpadåna: 2 in PE7 = Total 2; dha∫mapalipuccha: 1 in RE8 = Total 1; dha∫mapaliyåyåni: 1 in MRE 3 = Total 1; dha∫måpekkhå: 1 in PE1 = Total 1; dha∫mamaõgala: 3 in RE9 = Total 3; dha∫mamahåmåtå: 3 in RE5; 1 in RE12; 3 in PE7 = Total 7; dhammayåtå: 1 in RE8 = Total 1; dha∫mayuta: 3 in RE5; 1 in PE4; 1 in PE7 = Totl 5; dha∫malipi: 2 in RE1; 1 in RE5; 1 in RE6; 1 in RE13; 1 in RE14; 1 in PE1; 1 in PE2; 1 in PE4; 2 in PE6; 2 in PE7 = Total 13; dha∫mavåya/sœlana: 1 in RE13 = Total 1; dha∫mavijaya: 3 in RE13 = Total 3; dha∫mavi∂∂hi: 1 in RE5; 1 in PE6; 9 in PE7 = Total 11; dha∫mavuta: 1 in RE10; 1 in RE13 = Total 2; dha∫masa∫ba∫dha: 1 in RE11 = Total 1; da∫masa∫vibhåga: 1 in RE11 = Total 1; dha∫masa∫stava: 1 in RE11 = Total 1; dha∫masåvana: 3 in PE7 = Total 3; dhammasusußå: 3 in RE10 = Total 3; saddha∫ma 1 in MRE 3 = Total 1

Power of Words

129

has now been converted into the sound of dharma [dharmaghoßa; Rock Edict 4]. Now A†oka seeks not the conquest of war but the conquest of/for dharma [dharmavijaya – Rock Edict 13]. A†oka’s dharma reminds me of the rhetoric centered on “family values” that was in vogue in the United States several years ago. This was first articulated by the right wing of the American political establishment. It was a masterful political move. Like “motherhood and apple pie”, who can stand against “family values” and hope to win an election in the American heartland? Soon, the theme was co-opted by every segment of the American political spectrum. But what precisely are “family values”? It was an attractive container into which anyone could pour any content. For A†oka and his political operatives, dharma was the “family-value” cliché of the 3rd century BCE It was a masterly political move. A†oka may have been sincere; for him, talk of dharma was probably not simply rhetoric. He claims to have institutionalized dharma within his bureaucracy, even appointing dharmaspies against members of his own family. Nevertheless, the political underpinnings of this usage is evident; dharma was the linchpin of A†oka’s new imperial theology. A†oka’s use of dharma, on the one hand, follows a long-standing tradition of the close connection between dharma and the king, a connection we have traced within the vedic literature. There is also, on the other hand, a departure from that tradition: A†oka makes dharma almost entirely an ethical concept. Dharma is the ethical principal that must guide the lives of all people–Brahmin and ascetic, king and subject. Dharma in A†oka is the central concept of a common public ethic; dharma is the right way of living. These semantic changes and the central position accorded to dharma suggest that A†oka’s use of dharma was mediated by its appropriation sometime before A†oka by some of the emerging ascetic communities. I believe that this appropriation was part of the broader adoption of royal symbols by ascetic leaders and communities. Within its royal usage, we saw that in the middle and late vedic literature dharma defined royal authority; governing was implementing the requirements of dharma. We saw in the Br¢hadårañyaka passage that even a weak man could challenge a stronger man by recourse to dharma, as he would by recourse to the king or a court of law. Within the new ascetic religions that emerged in Gangetic planes of northern India towards the middle of the first millennium BCE, and especially within Buddhism to which A†oka was converted, dharma stands at the very heart of their theologies.13 Dharma defines the very essence of what the Buddha, the Universal Emperor of a new moral universe, was thought to have discovered. Dharma here defines the truth discovered by the Buddha, the truth that conferred authority on him. In this transfer from king to ascetic, however, dharma 13

For a detailed study of the Buddhist usages of dharma, see Geiger 1920.

130

LANGUAGE, TEXTS AND SOCIETY

came to be redefined in terms of doctrine, especially moral doctrine, and it was given a centrality within religious discourse that it lacked in the vedic lexicon. The literature of the earliest ascetic communities in the 5-4th centuries before the common era, if such a literature existed, is now lost. The manner in which dharma was first appropriated within ascetic discourse, therefore, has to be reconstructed from the tidbits of evidence scattered in the extant sources. One such source is the A™™akavagga of the Suttanipåta, by common consensus one of the earliest texts of the Buddhist canon and possibly the earliest extant example of ascetic literature. Recently Louis Gomez (1976) and Tilmann Vetter (1990) have independently investigated the A™™akavagga using different criteria. Vetter concludes that seven suttas in the A™™akavagga14 form its oldest stratum, and Gomez’s study broadly confirms Vetter’s conclusion. The word dhamma is used in 34 of the 210 verses comprising the A™™akavagga. In 5 or possibly 6 of these instances, dhamma carries a positive meaning, as when it speaks of the correct dhamma, evidently with a strictly Buddhist meaning (933-4, 947). All these, however, occur in passages identified by Vetter as belonging to a later stratum. In all other instances –and uniformly in the seven older suttas– dhamma carries a negative connotation. Dhamma is something that other ascetics were involved in; it is something that a good ascetic should shun. One should not take part in debates (våda) relating to dhammas (dhammesu in the plural: 787). Some ascetics claim “here alone is purity (suddhi) not in other dhammas”(837). The authors of these texts especially condemn those who claim their dhamma to be true and all others to be false: They call their own dhamma the highest; the dhamma of others the lowest (904). They call someone from another sect a fool with impure dhamma (893). If you do not know my dhamma you become an animal; this is how they speak (880). Some call a dhamma the highest; others call the same the lowest (903).

The authors claim that when one passes beyond such disputes, one does not consider one dhamma to be the best: “nothing has been grasped by me from among the doctrines (dhammesu), after consideration”, which appears to be a stock phrase appearing frequently (785, 801, 837, 907). This early usage of dhamma within the early stratum of the A™™akavagga indicates a period when dhamma was not a universally accepted term. It was, however, gaining popularity among the ascetic groups as defining what they proposed as doctrine and way of life. Frequently, the meaning of dhamma in these texts come close to that 14 Suddha™™aka (4), Parama™™aka (5), Pasüra (8), Mågandiya (9), Kalahavivåda (11), Cü¬aviyüha (12), and Mahåviyüha (13).

Power of Words

131

of di™™hi, viewpoint or doctrine. It also parallels other terms such as sacca (truth), sammuti (opinion), and yåna (vehicle/path). In the Mahåvagga (23, etc.) we have the stock question asked by one renouncer of another: ka∫ ‘si tva∫ åvuso uddissa pabbajito, ko vå te satthå, kassa vå tva∫ dhamma∫ rocesi –“Dependent on whom, sir, did you go forth; who is your teacher; to whose dhamma do you adhere.” Here also dhamma is taken as a special teaching of some sectarian leader, very similar to the way it is used in the A™™hakavagga. This association of dharma with a single charismatic and authoritative leader stands in sharp contrast to its usage in the vedic texts; I will revisit this issue later in dealing with the Dharmasütras. Vetter believes that the group or groups that produced this early literature of the A™™akavagga were pre- or non-Buddhist: “The A™™aka contains texts of a group that existed before or alongside the first Buddhist community. After some time this circle was integrated into the Buddhist sangha. It then produced more texts. In these it tried to combine the old teaching with the teaching of the Buddha.” I think in the A™™akavagga we catch a glimpse of dharma as it was emerging into the ascetic vocabulary, assuming a central role in some and subjected to censure in others.

Dharma in Later Brahmanical Discourse A†oka was what we would call today “a Buddhist”; he calls himself an upåsaka, a lay follower. He was also partial towards ascetic communities in general and embraced the ascetic ethic of non-injury (ahi∫så). He boasts about the reduction in the number of animals killed in the royal kitchens, and promises to reduce that number even further. We saw that abstention from killing was one of the elements of the dharma he preached. Given that honoring Bråhmañas and ‡ramañas was also a key element of his dharma, it is unlikely that A†oka was hostile to the Brahmanical community. Yet, something had changed radically from the viewpoint of the Brahmanical theology relating to the special relationship between king and Brahmin. This special relationship is well articulated in the vedic texts: the royal power (kßatra) and the Brahman power (brahma) can thrive only when they are united. Royal power and legitimacy were sustained and enhanced by Brahmanical intervention through ritual and sacrifice, especially the imperial horse-sacrifice. A†okan reforms clearly displaced the Brahmin from the unique position within society which he is accorded in vedic theology; now there were two models of religious authority and power, the Brahmin and the ascetic. Another, and perhaps a more severe, blow to Brahmañical theology and privilege was the A†okan insistence on not killing. This referred in a special way to sacrificial killing: no living being should be killed for sacrificial purposes (hidå nå kichi jive ålabhitu pajohitaviye;

132

LANGUAGE, TEXTS AND SOCIETY

Sanskrit: iha na ka†cit jœvaΔ ålabhya prahotavyaΔ—Rock Edict 1). This downgrading, if not the abolishing, of animal sacrifice cut at the very heart of the Brahmanical self-definition as the guardians of sacrifice. It was the theology of the sacrifice that granted Brahmins their special and central role in society, for it was the proper offering of sacrifice that assured rain, prosperity, and social harmony. At the very least, the A†okan social reforms had placed the uniqueness and centrality of the Brahmin in jeopardy. A†oka uses the compound †ramaña-bråhmaña to indicate the entire class of religiosi, placing the two on an equal footing. A century or so later, the grammarian Pata∞jali (on Påñini 2.4.12) would use the same compound to illustrate the class of co-ordinative (dvandva) compounds where the two members are in eternal opposition (yeßå∫ virodhaΔ †å†vatikaΔ). Let us review what had happened by at least the middle of the third century before the common era. The word dharma was used by ascetic sects, especially by Buddhists, that were gaining in popularity and influence to define their doctrine and way of life. In the process, dharma has begun to be ethicized and to be given a general application as the central term within theological discourse. All should practice dharma, and being a good person meant abiding by dharma. A†oka made dharma the centerpiece of his political philosophy and moral theology. He published this in numerous edicts placed across the Indian subcontinent; he instructed his officials to teach the dharma in a formal way; and he encouraged the members of his court and his subjects to practice dharma. We can assume with some degree of confidence that all this must have penetrated the common vocabulary of the people. This is confirmed by a statement in the Br¢hadårañyaka Upanißad [1.4.14] that I have already cited: tasmåt satya∫ vadantam åhur dharma∫ vadatœti Ù dharma∫ vå vadanta∫ satya∫ vadatœti –“Therefore, when a man speaks the truth they say ‘He is speaking dharma’, or when a man speaks dharma, they say ‘He is speaking the truth’.’ In the Upanißads “they” stands for what is commonly seen among people, and the authors use this to support their arguments especially for the various verbal equivalencies with which the Upanisads are replete. Within this state of affairs, it is not surprising that a genre of literature devoted to dharma should arise within the Brahmanical scholastic circles. They attached it to the most obvious literary structure already in place: the Kalpasütras, literature in the sütra or aphoristic style that dealt with matters relating to the solemn and the domestic rituals and comprising the ‡rauta-sütras and the Gr¢hya-sœtras. The ritual strand of the semantics of dharma already found in the Vedas and in the ‡rauta-sütras must have facilitated this. I propose that the emergence of the new genre of Dharma-sütras is a direct Brahmanical response to the events beyond their control that I have outlined above. Dharma had become as a central concept of theology and public discourse; this new reality could not be ignored. The Brahmanical

Power of Words

133

response in the form of the Dharma-sütras re-appropriated and redefined dharma in ways that would underpin not the new theologies opposed to Brahmanical ritualism but the traditional values of the Brahmanical tradition rooted in the Vedas.15 Let me highlight one aspect of the emergent Brahmanical discourse on dharma that may have some relationship to the ascetic appropriations of the term. It is dharmamüla (the root or source of dharma) or dharmapramåña (the authoritative basis of, or more epistemologically, the means of knowing dharma). This is the very first topic treated in all the dharma texts. The Åpastamba Dharmasütra [1.1.1-3], which I consider the oldest extant Brahmanical work on dharma, begins: athåtaΔ såmayåcårikån dharmån vyåkhyåsyåmaΔ Ù dharmaj∞asamayaΔ pramåñam Ù vedå† ca –“And now we shall explain the dharmas based on agreed-upon customs. The authority [means of knowledge] is the agreement [or acceptance] by those who know the dharma, and the Vedas.” Here the Vedas are relegated to the end, almost as an afterthought. The central pramåña of dharmas [note Åpastamba’s use of the plural] is the samaya or agreement among experts in dharma.16 The Gautama Dharmasütra [1.1-2] gives a more developed answer that becomes standard in later literature: vedo dharmamülam Ù tadvidå∫ ca smr¢ti†œle –“The source of dharma is the Veda, as well as the tradition [smr¢ti] and the practice of those who know it (the Veda).” Åpastamba in particular, but also Gautama, places the pramåña for dharma squarely on the consensus of a community, not of a generic community or the general public, but of a community of experts, who, by definition, are Brahmins. It is their agreement on any point of dharma, or their habitual practice, that makes it authoritative. This stands in sharp, and possibly deliberate, contrast to the charismatic personality and individual authority of the leaders of the ascetic sects, the tœrtha∫karas, on which the dharma preached by the ascetic groups depended. We need only to note the theology of the buddhavacana, the ipsissima verba of the Buddha, that made a text, or more generally a point of doctrine or practice, authoritative. Here we have a contrast between community standards, at least the standards of a cultured élite, and the charismatic authority of an individual. By the middle of the second century before the common era, that is, about a century after A†oka, we find further developments within the Brahmanical theology of dharma that set even stricter emphasis on the community. The Baudhåyana Dharmasütra [1.1.4-5] makes this sig15 I do not mean to suggest that this re-appropriation was carried out solely within the Dharma-sütras. We find similar but less systematic appropriations in other literature, including the Upanißads. The Brahmanical usage of the term is evidenced in the grammatical literature as well: see Påñini 4.4.41, 91, 92; 5.1.124; 5.2.132; 5.4.124; and Yåska’s Nirukta, 2.10; 3.4; 6.19; 8.10; 12.13; 14.6; 14.30. 16 For a detailed discussion of this and other texts dealing with the basis of dharma, see Wezler 1999. For the Dharmasütras and their dating, see Olivelle 1999 and 2004d.

134

LANGUAGE, TEXTS AND SOCIETY

nificant addition to the pramåñas of dharma: tr¢tœyaΔ †iß™ågamaΔ –“The third [means of knowing dharma] is the conventions/traditions of the cultured élite.” The same statement is made in the other ancient work on dharma, the Vasiß™ha Dharmasütra [1.5]. The introduction of the concept of †iß™a is noteworthy, because the same cultural élite is the subject of discussion by the grammarian Pata∞jali at about the same time. For Pata∞jali the question is how one learns good Sanskrit and good grammar; he tells, just listen to how †iß™as speak. These cultured people set the standard for good dharma as much as for good Sanskrit.17 A second criterion also enters into the discussion about how we can know dharma, and it is geography. A totally new concept and word enters the vocabulary, the Åryåvarta, “the land of Åryas” located broadly in north-central India, and interestingly excluding what is today Bihar, the heartland of the new religions. All three authors, Pata∞jali, Baudhåyana, and Vasiß™ha refer in almost exactly the same words to this region as the place where †iß™as live.18 The language and the customs of this region provide us a sure way of knowing correct linguistic and dharmic usage. We have come a full circle. A brand new term invented by ancient Brahmanical poets of the R¢gveda has become the central and defining term for the Brahmanical religion and way of life with considerable help from their rivals for religious authority and influence, the ascetic communities, and from an emperor with dreams of conquering the world through and for dharma. If my hypothesis regarding the adoption of the term by ascetic groups with the intention of defining their message in royal terms and thus gaining power and prestige is correct, their gambit, at least as far as the term itself is concerned, proved to be far more successful than they could ever have imagined. The changes we have examined during two or three centuries in the second half of the millennium before the common era catapulted dharma into the forefront of Indian cultural and religious discourse. It will remain so through the next two millennia with the Buddhist dharma spreading to almost all the countries of Asia and more recently to every corner of the globe and the Brahmanical dharma producing the longest-lasting literary tradi17

Pata∞jali, Mahåbhåßya on Påñini 6.3.109 (ed. Kielhorn, III, p. 174). On †iß™a and its usage by grammarians, see Deshpande 1985. 18 prågådar†åt pratyak kålakavanåd dakßiñena himavantam udak påriyåtram etad åryåvartam– “The region to the east of Ådar†a, west of Kålaka forest, south of the Himalayas, and north of Påriyåtra mountains is the land of the Åryas. The practices of that land alone are authoritative.” BDh 1.2.9. pråg ådar†åt pratyak kålakavanåd udak påriyåtråd dakßiñena himavataΔ uttareña vindhyasya [...] etad åryåvartam ity åcakßate– “... the region east of Ådar†a, west of Kålaka forest, north of Påriyåtra mountains, and south of the Himalayas –(or) north of the Vindhya mountain– ... is called the land of the Åryas.” VaDh 1.8-12. kaΔ punar åryåvartaΔ Ù pråg ådar†åt pratyak kålakavanåd dakßiñena himavantam uttareña påriyåtram– “Now, what is the Åryåvarta? The region east of Ådar†a, west of the Kålaka forest, south of the Himalayas, and north of Påriyåtra mountains.” Pata∞jali on Påñini 2.4.10; 6.3.109.

Power of Words

135

tion of India, the Dharma†åstras, impacting deeply modern Indian personal law both under the British and under independent rule, and influencing even the modern nationalistic political discourse of Hindutva and Hindu dharma.

136

LANGUAGE, TEXTS AND SOCIETY

A

137

VI Semantic History of Dharma: The Middle and Late Vedic Periods*

The studies of the term dharma in the early vedic period by Paul Horsch (1967)/(2004) and Joel Brereton in this volume1 have shown that the term is used with relative frequency in the R¢g Veda (67 times) with a somewhat broad semantic range including the cosmological, ritual, and ethical spheres. The frequency drops drastically in the other text from this period, the Atharva Veda (13 times), although maintaining a semantic range similar to the R¢g Veda. Given the centrality of dharma in later religious literature of India, both Brahmanical and Buddhist, one would expect that the term gradually began to assume a central role in the religious vocabulary of the middle and late vedic periods represented by the Bråhmañas, Årañyakas, the Upanißads, and the ‡rauta- and Gr¢hya-sütras, which together constitute a vast body of literature both in extent and in the variety of their topics and concerns. A close study of the use of dharma in these texts, however, demonstrates the opposite: dharma was at best a marginal term and concept within the vocabulary of these texts, and it did not play a central role in the religious world depicted in them. On the one hand, the frequency of its usage drops rather dramatically and, on the other, its semantic range becomes narrower, being restricted for the most part to Varuña and his earthly counterpart, the king. In the concluding postscript I will attempt to show how this specialized meaning of dharma may have contributed to its further semantic development in its adoption and adaptation by Buddhism, by A†oka, and in the later Brahmanical literature. * Originally published in the Journal of Indian Philosophy 32 (2004): 491-511. 1

See also Willman-Grabowska 1934.

138

LANGUAGE, TEXTS AND SOCIETY

The Yajurveda Sa∫hitås The Yajurveda Sa∫hitås are the oldest texts we have after the R¢g Veda and the Atharva Veda. I have examined all four extant Sa∫hitås: Maitråyañœ, Kå™haka, Taittirœya, and Våjasaneyi. The study of dharma2 in these texts is complicated by two factors: first, they often cite verses from the R¢g Veda; and second, because they are parallel recensions of the same ritual material, they frequently present the same passages. To determine the frequency of dharma in the Yajurvedic vocabulary, therefore, I have discounted citations from the R¢g Veda and counted only once the passages that are repeated almost verbatim in several of these Sa∫hitås. In the four Yajurveda Sa∫hitås the term dharma is found in 22 passages.3 Even this may be an overestimate because several of these occurrences are in stock phrases and epithets. Many of these refer to Varuña or Mitra-Varuña. Thus we have the recurrent phrase dhruvéña dhármañå (“with firm/enduring dharma”) with reference to Mitra-Varuña: dyutånás två måruto minotu mitrÌvaruñayor dhruvéña dhármañå –“May Dyutåna Måruta establish you in accordance with the enduring dharma of Mitra and Varuña (at the planting of the sacrificial post; TS.1.3.1.2; MS 1.2.11; VS 5.27). The same phrase is repeated in a different context: váruñas två dhr¢távrato dh√payatu mitrÌváruñau dhruvéña dhármañå –“May Varuña, whose commandment is upheld, perfume you with incense, may Mitra and Varuña, in accordance with their enduring dharma” (at the perfuming of the pan; MS 4.9.1).4 And again: mitrÌváruñau tvottarataΔ pári dhattå∫ dhruvéña dhármañå –“May Mitra and Varuña lay you around in the north in accordance with their enduring dharma” (at the laying of the enclosing sticks [paridhi]; KS 1.11; TS 1.1.11.2; VS 2.3; ‡B 1.3.4.4). The use of dharman here closely agrees with its principal usage in the R¢g Veda; dharman is the institute or commandment of Varuña, an institute that is here said to be dhruva, firm and enduring. The MS (3.8.9) provides a commentary on the phrase: mitrÌváruñau dhruvéña dhármañeti Ù mitrám evaínå∫ dådhÌra, váruñaΔ kalpayati, vídhr¢tyai ca khálu vÌ eßÌ∫ prajÌnå∫ kl¢ptyai ca mœyate, mitrÌváruñau vaí devÌnå∫ dhármadhÌrayau, daivá∫ vÌ etád dhármam adœdharatå∫ Ù Mitra and Varuña in accordance with their enduring dharma. Mitra, 2 In this study of the middle vedic texts, I have discounted the difference between the old neuter form dharman and the newer thematic stem dharma. In texts of this period there does not appear to be any semantic difference between the two forms. 3 MS 1.1.8; 1,2,11; 1.5.4, 11; 1.11.1; 2.6.6, 8; 2.8.8; 3.16.5 (75, 77); 4.12.6; KS 1.11; 8.7; 35.7; TS 1.8.14.1; 1.8.16.2; 3.2.8.1; VS 10.29; 20.9, 17; 30.6; 38.19. 4 Other texts with parallel passages omit dhruveña dharmañå: TS 4.1.6.1; KS 16.5; VS 11.60; ‡B 6.5.3.10. I have adopted the translation from Brereton (1981: 55), as also his translation of dhruveña dharmañå in other places.

Semantic History of Dharma

139

indeed, holds it fast, and Varuña establishes it. For the upholding of these creatures and for their establishment, it is set up. Mitra and Varuña are clearly the ones who uphold dharma among the gods. They have upheld here the divine dharma.

Note that the explanation clearly connects dharma with its root dhr¢, to bear, to support, to uphold. The connection between dhruva and dharma is evident also in a passage of the KS (35.7): dhruvå dyaur dhruvå pr¢thivœ dhruva∫ vi†vam ida∫ jagat Ù devå ha dharmañå dhruvå yajamånaΔ pa†ubhir dhruvaΔ –“The sky is enduring; the earth is enduring; this whole world is enduring. The gods are enduring through dharma, and the sacrificer is enduring through the sacrificial animals”.5 The close connection between dharma and Varuña within the Yajurveda Sa∫hitås is most evident within the context of two rituals associated with kingship, the royal consecration (råjasüya) and the horse sacrifice (a†vamedha) performed by a king to enhance and proclaim his sovereignty.6 At the royal consecration the king is placed in close relationship with both Varuña and dharma. Several cakes (caru) are offered to deities who are identified with various characteristics (TS 1.8.10.1): Agni is the lord of the house (gr¢hapati), Soma is the lord of the forest (vanaspati), Rudra is the lord of cattle (pa†upati), and Br¢haspati is the lord of speech (våcaspati). The last two are Mitra, characterized as satya, and Varuña, who is dharmapati, the lord of dharma. In the parallel passage at MS 2.6.6 the wording is mitrÌya satyásya pátaye ... váruñåya dhármasya pátaye –“To Mitra the lord of truth; to Varuña the lord of dharma”.7 Within the same context we have a mantra asking Savitr¢ and other gods to stimulate the new king (savitÌ två ... suvatåm). Once again the deities are given their distinctive characteristics (TS 1.8.10.2): mítraΔ satyÌnå∫ váruño dhármapatœnåm (or in MS dhármåñåm).8 After this the new king is presented with the words: eßa vo bharatå råjå –“This, O Bharatas, is your king!” In the MS (2.6.8; 4.4.2), gods are invoked to make the new king an upholder of dharma: sómå índro váruño mitro agnís té devÌ dharmadhr¢to dhárma∫ dhårayantu –“Soma, Indra, Varuña, Mitra, Agni –may these gods, the upholders of dharma, uphold dharma.” The TS (1.8.16.2) addresses the king, identifying him with Varuña: váruño ’si satyadhármå –“You are Varuña, whose dharma is true/real.” Here again we have the implicit statement that the epithet of Varuña as satyadharmå is applicable to the king as well. Within the A†vamedha rite also the same epithet is used with reference to the reins of Mitra and Varuña’s chariot: yó vå∫ rátha r¢j¶ra†miΔ sátyadharmå mithucárantam upayÌti düßáyan –“You, whose chariot with straight reins and 5

The first half-verse is R¢V 10.173.4. The second half-verse is not found elsewhere. For an examination of kingship, see Rau 1957, 90-95. 7 See also KS 15.6-7; VS 10.4, 6-9, 16, 18; TB 1.7.6; ‡B 5.3.4.27; 5.3.5.16-37; 5.4.1.15. 8 On these passages where dharma is connected to the king, see Rau 1957, 90-96. 6

140

LANGUAGE, TEXTS AND SOCIETY

true/real dharma draws near, frustrating him who acts falsely” (MS 3.16.5; KS 22.15; TS 4.7.15.2). In these and in other passages we will examine in the Bråhmañas and the Upanißads, the kingly power or kßatra is intimately connected with dharma. In the same way as Varuña, the heavenly sovereign, his earthly counterpart is also dharmapati; the two are lords and upholders of dharma in the cosmos and within society. The connection of dharma to the royal and judicial sphere is nicely illustrated by one ritual detail. In the Purußamedha, the real or symbolic human sacrifice, different types of men are sacrificed to various deities. The kinds of men sacrificed to dharma and its opposite adharma are instructive: to dharma, a sabhåga (presumably, a man who participates at the royal audience hall where judicial proceedings are carried out), and to adharma,9 a deaf man (badhira), quite the opposite of the debater in the assembly hall (VS 30.6, 10; TB 3.4.1).

The Bråhmañas The corpus of the Bråhmañas is even more vast than the Yajurveda Sa∫hitås. For this study I have chosen three: Aitareya, Taittirœya, and ‡atapatha. Although paralleling the Aitareya, the Kaußœtaki does not use the term dharma at all.10 In the three Bråhmañas that are the focus of this study, the term dharma occurs independently a total of 11 times.11 This is striking, especially because the Bråhmañas, unlike the Sa∫hitås, contain a wide spectrum of material and address numerous issues, including ritual and ethical ones, pertaining to Brahmanical life. One would assume, a priori, that if dharma was a key concept that defined the Brahmanical view of the world and of human life, it would have been used more frequently and within a variety of contexts. The term is used only once in the Aitareya Bråhmaña, and, significantly, here also it occurs within the context of the royal consecration (råjasüya). At AB 8.12-14 there is a description of the great anointing (mahåbhißeka) of Indra as king by the gods, which is immediately followed at AB 8.15-23 by the anointing of the king. Clearly, the anointing of Indra provides the model for the anointing of the king. In both we have the following public proclamation of the sovereignty of Indra and the king, the former proclaimed by the Vi†vedevas and the latter by the “king-makers” (råjakartåraΔ). The wording is identical in the two contexts with the exception of the audience addressed, gods (devåΔ) in the case of Indra and men (janåΔ) in the case of the king: This is the earliest occurrence of the negative adharma in the vedic literature. According to Vishva Bandhu’s Vedic Word-Concordance (Hoshiarpur, 1973), dharma occurs in 11 passages of the Jaiminœya Bråhmaña and once in the Gopatha Bråhmaña. 11 Here also I have eliminated citations from the Sa∫hitås, which are for the most part mantras for ritual recitations. 9

10

Semantic History of Dharma

141

ima∫ janå [devå] abhyutkro†ata samråja∫ såmråjya∫ bhoja∫ bhojapitara∫ svaråja∫ svåråjya∫ viråja∫ vairåjya∫ parameß™hina∫ pårameß™hya∫ råjåna∫ råjapitara∫ kßatram ajani kßatriyo ‘jani vi†vasya bhütasyådhipatir ajani vi†åm attåjany amitråñå∫ hantåjani bråhmañånå∫ goptåjani dharmasya goptåjanœti. (AB 7.12, 17) Do ye proclaim him, O men [O gods], as overlord and overlordship, as paramount ruler and father of paramount rulers, as self ruler and self rule, as sovereign and sovereignty, as supreme lord and supreme lordship, as king and father of kings. The kßatra (royal power) has been born, the kßatriya has been born, the suzerain of all creation has been born, the eater of the commoners (vi†) has been born, the slayer of foes has been born, the guardian of bråhmañas has been born, the guardian of dharma has been born. (Tr. Keith with modification)

The proclamation reaches its climax with the announcement that Indra/king has been born through the right of consecration as the guardian of dharma. Here Indra has replaced Varuña, but the connection of the king to dharma is brought to the forefront. In the Taittirœya Bråhmaña dharma occurs four times. In three of these dharma is closely associated with Varuña and with the overlord (adhipati). At TB 3.9.16.2, which is a commentary on TS 7.4.16 dealing with the horse sacrifice, the mantra of the TS namó ’dhipataye (“Homage to the overlord”) is interpreted by the TB as: dhármo vÌ ádhipatiΔ, dhármam evÌvarundhe (“The overlord is dharma. He does, indeed, obtain dharma”). The person who is here identified with ådhipatya and with dharma is the king performing the horse sacrifice. In the TS, significantly, the mantra commented on by the TB comes at the end and climax of a longer mantra: “Homage to the king! Homage to Varuña! Homage to Prajåpati! Homage to the overlord!” Here dharma as the adhipati is placed last as the highest. At TB 3.11.4.1 we have the common designation of Varuña as “Lord of dharma”. Indeed the juxtaposition of indrapatnœ and dharmapatnœ (“wife of Indra” and “wife of dharma”) at TB 3.11.1.20 makes it likely that dharma here stands for Varuña, whose epithet is dharmapati. In the voluminous ‡atapatha Bråhmaña, if we exclude the last section consisting of the Br¢hadårañyaka Upanißad, the term dharma occurs in just six passages. In one of the most significant statements regarding the relation between dharma, Varuña, and the king, the ‡atapatha (5.3.3.9), within the context of the royal consecration (råjasüya), states: athá váruñåya dhármapataye Ù våruñá∫ yavamáya∫ caru∫ nírvapatíi tád ena∫ váruña eva dhármapatir dhármasya páti∫ karoti paramátå vai så yó dhármasya pátir ásad yó há paramátå∫ gácchati ta∫ hí dhárma upayánti tásmåd váruñåya dhármapataye ÙÙ Then to Varuña the lord of dharma he offers a cake made with barley. Thereby Varuña himself, the lord of dharma, makes him [the king] the

142

LANGUAGE, TEXTS AND SOCIETY

lord of dharma. That, surely, is the highest state when one becomes the lord of dharma. For when someone attains the highest state, (people) come to him (in matters relating) to dharma. Therefore, to Varuña, the lord of dharma.

Here we get a clearer picture as to what the author of the ‡atapatha means by dharma, certainly clearer than the texts we have examined thus far. It has to do with matters regarding which people come to the king and must refer principally to legal disputes. Dharma is thus placed squarely within the public realm of law and social norms that must be overseen by the king. We can now understand why the king is the dharmapati, in the same way as Varuña, the sovereign who oversees the cosmic dharma. Perhaps this was the meaning generally attached to dharma during the middle vedic period and underlies the statements on dharma in other texts that we have examined. Another significant passage occurs within a creation story, a passage that throws further light on the meaning of dharma for the author of the ‡B. After Prajåpati was born, he created four gods: Agni, Indra, Soma, and Parameß™hin. These five deities then observe the four directions. Our passage deals with the north (‡B 11.1.6.24): áthódœcœ∫ dí†am apa†yan Ù tÌm ápo ’kurvatópainåm itáΔ kurvœmahÎti ta∫ dhármam akurvata dhármo vå Ìpas tásmåd yádemá∫ lokám Ìpa ågácchati sárvam évedá∫ yathådharmá∫ bhavaty átha yadÌvr¢ß™ir bhávati bálœyån eva tarhy ábalœyasa Ìdatte dhármo hy ÌpaΔ ÙÙ Then they saw the northern direction. They made it waters. “Let us improve it from here,” they said. They made it dharma. The waters, surely, are dharma. Therefore, when waters come to this world, all this becomes in accordance with dharma. But when there is no rain, then the stronger indeed seizes the weaker ones, for the waters are dharma.

This passage further supports the meaning of dharma as social order founded on law. The argument here corresponds to the “law of the fish” (matsyanyåya) of later Dharma†åstras. Here, however, there is no mention of a king or the lack of a king; it is the natural phenomena of abundance created by water and famine created by drought that bring about adherence to and deviation from dharma. The assurance of timely rain, however, is very much a function of kingship, and Varuña is closely associated with water. Indeed, as Insler has recently pointed out, the royal consecration (råjasüya) initially may have been an annual ritual conducted during the winter solstice to assure the timely progress of the seasons and the timely arrival of rain.12 12 Stanly Insler’s paper “The Function of the Varuña Hymns of the R¢gveda” at the Third International Vedic Workshop, Leiden, May 29–June2, 2002. See the story of Devåpi and ‡antanu in the Br¢haddevatå (8.2): when a transgression against dharma occurred (here the younger brother becoming king in place of the older) the realm got no rain for twelve years.

Semantic History of Dharma

143

At ‡B 13.4.3.14 it is Indra who is identified with dharma, but even here note the connection between dharma and Indra as king of the gods: dhárma índro rÌjety åha tásya devå ví†aΔ –“Indra the king is dharma, he says; his people are the gods.” The ‡atapatha also uses the negative adharma twice in the sense of something done contrary to law or the natural order of things. At ‡B 3.7.3.7 the author is refuting an opponent’s view about how to bind the sacrificial animal. He says that doing it the opponent’s way “would be as if he intended to do an adharma stealthily” (yathÌdharma∫ tira†cáthå cíkœrßet). Likewise at ‡B 5.2.4.13 with reference to the Dakßiñå cow that draws a cart (ana∂uhœ), it says: yát strÎ satœ váhati ádharmeña tád asyai våruñá∫ rüpám –“In that, while she is a female she draws contrary to dharma, that is her Våruña form”.13 In both these instances, adharma appears to indicate an activity that is either against the rule, whether it is ritual or natural. An unexpected and possibly a new connotation of dharma as either a specific attribute or a right/obligation is found at ‡B 11.5.7.1, a passage in praise of svådhyåya, vedic study, through which a man becomes independent and wealthy, and by which his intelligence grows (praj∞åvr¢ddhi). Such a Brahmin is said to acquire four dharmas and people serve him with four other dharmas: práj∞å várdhamånå cat¶ro dhármån bråhmañám abhiníßpådayati brÌhmañya∫ pratirüpacaryÌ∫ yá†o lokapaktí∫, lokaΔ pácyamåna† cat¶rbhir dhármair bråhmañá∫ bhunakty arcáyå ca dÌnena cåjyeyátayå cåvadhyátayå ca Ù The growing intelligence brings to the Brahmin four dharmas: Brahmanical stature, fitting deportment, fame, and ‘cooking’ the world.14 The world, as it is being ‘cooked’, gratifies the Brahmin with four dharmas–with veneration, with gifts, with the condition of not being oppressed and of not being subject to capital punishment.

There appears to be a semantic development here from “law” or “the way things are or should be” to specific attributes that characterize a particular entity.15 The four dharmas of the world vis-à-vis the Brahmin, indeed, can be seen as legal privileges granted by society to the Brahmin class. Except for this last passage, the Bråhmañas do not expand the semantic range of dharma in a significant way. It remains closely associated with Varuña and with the royal power of the king. We detect, 13 It is unclear why acting in an adharmic way makes her connected with Varuña; one would have expected the opposite. However, even here we have the correlation between Varuña and dharma. 14 Eggeling, following Såyaña, takes this expression to mean “perfecting of the people”. 15 We can see how such an extension may have influenced also the ritual meaning of dharma as a peculiarity or ritual detail of a particular rite and especially the later Buddhist Abhidharma meaning of dharma as particles of the ultimately real, that which makes something that thing (svabhåva), which parallels the ritual use of svadharma in the ‡rautasütras that we will examine below.

144

LANGUAGE, TEXTS AND SOCIETY

however, more clearly that dharma has acquired the primary meaning of law and order within society, a law that is hypostatized into an abstract entity as dharma that stands above and gives legitimacy to kßatra, the ruling power of the king.

The Årañyakas The term dharma occurs only three times in the Årañyakas, twice in the Taittirœya and once in the Aitareya. They do not add much to the semantics of dharma. At AÅ 2.1.7, within a section that describes the creation of the world through the organs of the Purußa, we have a statement that again connects dharma with Varuña and with water: varuño ‘sya prajå∫ dharmeña dådhåra –“Varuña supported his offspring through dharma”. The Taittirœya passages are brief. At TÅ 2.19.1 various cosmic entities or categories are homologized with various parts of a ‡i†umåra (alligator or porpoise). Here dharma is said to be the crown of its head (mürdhånam). At TÅ 4.42.5 we have a list of divine categories that begins with †rœ and end with dharma. The incorporation of dharma into lists of cosmic categories becomes a common feature in late vedic texts, and in these lists dharma is most often placed last as the highest of the categories.

The Upanißads The Upanißads are the texts where we would expect to find a sustained treatment of dharma, given that a central theme in these documents is human activity and knowledge leading to an ultimate state beyond death. That, however, is not the case. In the four early prose Upanißads16–Br¢h adårañyaka, Chåndogya, Taittirœya, and Aitareya (in which the term does not occurs at all)–the term occurs in just nine passages. Even more importantly, there is no sustained focus on the term as it applies to living a righteous life, except perhaps in CU 2.23.1 and TU 1.11.1.

The Br¢hadårañyaka In the Br¢hadårañyaka, which constitutes the last portion of the ‡B, the term occurs in four passages. Significantly, it occurs only once in the central Yåj∞avalkya-kåñ∂a (BU 3-4) , which forms the oldest core of the Upanißad. At 4.4.5, in a passage demonstrating that the åtman is made up of everything, there is a long list of categories. In this list, åtman is said to be made up of various categories and their opposites: 16 I am not including the later Upanißads in this study because they are probably later than the rise of Buddhism and the early Dharmasütras. The term is found in Kaußœtaki Upanißad 2.1-2; Ka™ha Upanißad 1.21; 2.13, 14; 4.14; ‡vetå†vatara Upanißad 6.6.

Semantic History of Dharma

145

tejomayo ‘tejomayaΔ kåmamayo ‘kåmamayaΔ krodhamayo ‘krodhamayaΔ dharmamayo ‘dharmamayaΔ sarvamayaΔ –“made of light and the lightless, made of desire and the desireless, made of anger and the angerless, made of dharma and adharma, made of everything”. We are not told what dharma means here, except that it must have been viewed as occupying the highest position within the list, because after it the author simply says sarvamayaΔ. The most sustained treatment of dharma is found in BU 1.4.14 within the context of a creation myth. In the beginning this world was only brahman. Because it was single, the author says, it “did not become fully developed” (na vyabhavat). Brahman then goes about creating “the ruling power” (kßatra), including the gods, then the Vai†ya class, and finally the ‡üdra class. After each creation, the author repeats the refrain na vyabhavat, “it still did not become fully developed.” As the final act of creation that made brahman fully develop and reach its full potential, it created dharma: tacchreyo rüpam atyasr¢jata dharmam Ù tad etat kßatrasya kßatra∫ yad dharmaΔ Ù tasmåd dharmåt para∫ nåsti Ù atho abalœyån balœyå∫sam å†a∫sate dharmeña Ù yathå råj∞aivam Ù yo vai sa dharmaΔ satya∫ vai tat Ù tasmåt satya∫ vadantam åhur dharma∫ vadatœti Ù dharma∫ vå vadanta∫ satya∫ vadatœti Ù etad dhy evaitad ubhaya∫ bhavati Ù So it created dharma, a form superior to and surpassing itself. And dharma is here the ruling power standing above the ruling power. Hence there is nothing higher than dharma. Therefore, a weaker man makes demands of a stronger man by appealing to dharma, just as one does by appealing to a king. Now, dharma is nothing but the truth. Therefore, when a man speaks the truth, people say that he speaks dharma; and when a man speaks dharma, people say that he speaks the truth. They are really the same thing.

This passage echoes two other significant passages of the ‡B that we have examined: 5.3.3.9, which presents dharma as the highest, something to which people go to settle disputes, and 11.1.6.24, which speaks of the stronger seizing the weak when there is the absence of dharma. In the BU passage, however, dharma is made the very essence of kßatra, the ruling power. A weaker man can take on even a stronger opponent by resorting to dharma in exactly the same way as he can by resorting to the king. I think the subtext here is litigation. A weaker man can drag a stronger man to the king’s court. As in many other vedic texts, dharma here is not only placed side by side with satya, truth, but is said to be identical with it.17 The significant point in this passage for our study is that, as in the Bråhmañas, here also dharma is associated with the legal and regal spheres. See the earlier connection of Mitra with satya and Varuña with dharma. The connection between dharma and satya continues into the classical period: see MDh 1.81. Their connection is also apparent within the context of judicial proceedings, especially in the context of witnesses who speak the truth: see MDh 8.14, 74, 80-8. 17

146

LANGUAGE, TEXTS AND SOCIETY

The other occurrence of dharma in the first chapter (BU 1.5.23) is in a †loka cited in support of the pre-eminence of breath over other faculties: yata† codeti süryaΔ asta∫ yatra ca gacchati Ù ta∫ devå† cakrire dharma∫ sa evådya sa u †va ÙÙ From which the sun rises, and into which it sets; the gods make it dharma. It is the same today and tomorrow.

The author of the BU comments after the first half-verse: pråñåd vå eßa udeti pråñe ‘stam eti –“From breath, indeed, does it rise, and into breath it sets”. The meaning of dharma here is unclear and the commentarial section of the text does not deal with this term. It appears likely that for the author of the BU dharma was the highest principle, and he equates it with breath/wind, which is here presented as the highest faculty, breath with respect to adhyåtma and wind with respect to adhidaivata. In the original setting of this verse, however, dharma may have been viewed as the ultimate institute/statute/commandment that is responsible for the regular rising and setting of the sun, a meaning familiar from the R¢g Veda. Finally, in the Madhukåñ∂a (BU 2.5.11) where all cosmic categories are said to be honey, dharma occurs just before satya. The connection between these two terms is the only thing remarkable about this occurrence; nothing is said that would permit us to understand what dharma meant for the author.

The Chåndogya The Chåndogya Upanißad, in which the term occurs in three passages, does not have a single extended discussion of dharma. The first time the term occurs is at 2.1.4. The context is the veneration of såman as sådhu (good). The section concludes: sa ya etad eva∫ vidvån sådhu såmety upåste ‘bhyå†o ha yad ena∫ sådhavo dharmå å ca gaccheyur upa ca nameyur –“When someone knows this and venerates the Såman chant as good, he can certainly expect that good dharmas will come his way and fall to his share”. The meaning of dharma in this context is quite unclear, but it may have the meaning of qualities, attributes, or simply things, somewhat similar to its usage we examined in ‡B 11.5.7.1. In an eulogy of speech (våc) at 7.2.1 we have another occurrence of dharma in a long list of things that speech makes known from the R¢g Veda down to worms, moths, and ants: dharma∫ cådharma∫ ca satya∫ cånr¢ta∫ ca sådhu cåsådhu ca –“dharma and adharma, truth and untruth, good and non-good.” Here dharma is juxtaposed with satya and sådhu. The first pair we have seen elsewhere; and I think that for the author of the CU satya and sådhu may be concepts that are closely related. Here again we get the negative adharma.

Semantic History of Dharma

147

The final and the most significant occurrence of dharma in the Chåndogya is in the famous passage on dharmaskandhas (2.23.1):18 trayo dharmaskandhåΔ Ù yaj∞o ‘dhyayana∫ dånam iti prathamaΔ Ù tapa eva dvitœyaΔ Ù brahmacåry åcåryakulavåsœ tr¢tœyaΔ Ù There are three types of persons whose torso is dharma.19 The first is the one who pursues sacrifice, vedic recitation, and gift-giving. The second is the one who is devoted to austerity. Third is a celibate student of the Veda living at his teacher’s house.

It is clear that in this passage dharma specifically refers to modes of religious life, probably to the life of a Brahmanical householder, an ascetic, and a vedic student. Here for the first time in vedic literature we have an unambiguous passage that uses dharma in a way very close to its usage in the later Dharma†åstras.

The Taittirœya In the Taittirœya Upanißad dharma occurs in a passage at the end of the first chapter (†ikßåvallœ), a passage that appears very much like a remnant of an old Gr¢hyasütra dealing with the instruction of a vedic student and his return home after his residence at his teacher’s house. In his parting words to the student (TU 1.11.1), the teacher tells him: satya∫ vada Ù dharma∫ cara (“Speak the truth. Practice dharma”). This is the first time in the vedic literature that the verb √car is used with reference to dharma, indicating that here the term is used in its classical Dharma†åstric meaning. Note, however, that satya is closely associated with dharma. In the same passage (TU 1.11.4) authoritative Brahmins, whose behavior is presented as a model to the young student, are characterized as dharmakåmaΔ (devoted to or loving dharma).

The ‡rautasüstras The dates of the numerous ‡rautasütras are uncertain, but at least some of them are probably from a pre-Buddhist period and overlap with some of the later vedic texts. Although they are not viewed as part of the vedic canon, they continue the tradition of the Bråhmañas in providing rules and explanations of the vedic rituals. In the 10 ‡rautasütras I have examined, the term dharma occurs in 39 passages.20 In all but a handful of them, however, dharma does 18

For a detailed study of this passage, see Olivelle (1996b); see above pp.53-74. As I have explained in my previous study (Olivelle, 1996b), I take the compound dharmaskandhåΔ as a Bahuvrœhi. Traditionally, it has been taken as a Tatpurußa and translated “divisions of dharma”. 20 These exclude the occurrence of the term in citations of mantras from the 19

148

LANGUAGE, TEXTS AND SOCIETY

not have the meaning found in either the earlier vedic texts or the later Dharma†åstras. It appears that the expert scholastic tradition focusing on the ritual developed a very special meaning of dharma that was restricted to this tradition. Within this specialized meaning, dharma refers to the specific ritual rules or ritual details of a rite. Many of these passages deal with how dharmas, taken as ritual details, are extended from ritual archetypes, such as the Dar†apürñamåsa (New- and Full-moon) sacrifice to others that are modeled after it. Thus, for example, the Bhar‡r (1.1.9) states the general principle: tatraißo ‘tyantaprade†o ye kecana paurñamåsœm amåvåsyå∫ vå dharmå anårabhyåmnåyanta ubhayatraiva te kriyante –“In this connection, this is the general rule. The characteristics (dharmåΔ) which have been prescribed in connection with the fullmoon day or the new-moon day without specification hold good with reference to both” (Tr. Kashikar). In a more specific case, the Bhar‡r (6.15.5) states: sarveßv iß™ipa†ubandheßu dår†apaurñamåsikå dharmå anuyanti –“The dharmas of the New-moon and the Fullmoon sacrifices are carried over into the iß™i and animal sacrifices.” It is significant that this meaning of dharma is absent in the vocabulary of the vedic texts, including the Bråhmañas and the Upanißads. This specialized and technical meaning probably developed within the expert tradition devoted to ritual that produced both the ‡rautasütras and the later Mœmå∫så texts. The compound svadharma, interestingly, is also used with a similar meaning; here it refers to the fact that a particular rite has its own ritual details (dharmas) specific and limited to it and not taken over from or extended to other rites.21 This meaning is quite similar to the use of dharma we saw at ‡B 11.5.7.1 and CU 2.1.4. The only passage where the term appears to have a Dharma†åstric meaning is Bhar‡r 9.18.4, where the question is how to tell whether an oblation is defiled: katha∫ duß™a∫ havir vidyåt Ù yad åryåñå∫ dharmaj∞ånå∫ dharmakåmånåm abhojanœya∫ na tena devån yajeta (“How does one know a defiled oblation? What Åryas who know dharma and who love dharma consider unfit to be eaten, with that he should not make an offering to the gods”). As in the TU, here also we have the term dharmakåma with the additional phrase dharmaj∞a, compounds that become commonplace in the Dharma†åstras. Apart from the specialized meaning of dharma as ritual details or characteristics22 that may be transferred between rites, the term does not play a central role in these ritual texts. Sa∫hitås. Åpastamba, 19.21.4, 16; 21.3.3, 10; ņvalåyana, 10.5.15; 10.7.9; Baudhåyana, 24.37; 27.24; Bhåradvåja, 1.1.9; 6.15.5, 7, 12; 7.6.7; 9.18.3; Hirañyake†in, 3.8.24, 32, 40, 41, 44, 46, 47; 13.4.27; 16.1.23; 22.4.2, 6; Kåtyåyana, 1.8.7; 4.3.19; 5.8.17; 9.5.10; Lå™yåyana, 7.7.19; ‡åõkhåyana, 4.5.13; 7.15.16-7; 9.26.2; 16.18.5; Vådhüla, 11.6.7-8, 24; and Våråha 1.1.1.68, 70; 1.2.4.9. 21 See Åp‡r 19.21.16; Hir‡r 3.8.40; 22.4.6; Bar‡r 6.15.12. 22 This meaning did not disappear after the ‡rautasütras. We find it used with precise-

Semantic History of Dharma

149

The Gr¢hyasütras Even though many of the Gr¢hyasütras date from a period after the rise of Buddhism, some of them may be earlier. It is in these texts devoted to the life and rituals of a Brahmin householder, however, that we would expect the term dharma to play a central role, as it does, for example, in the parallel texts of the Dharmasütras. That, however, is not the case. In the Gr¢hyasütras that can claim some antiquity23 the term occurs only six times, and two of these (‡åõGr¢ 3.3.7; PårGr¢ 3.4.18) deal with a ritual connected with the building of a house. There we have a mantra that identifies dharma with the main post of the house: dharma sthünaråjaΔ. In another ritual context, this time the offering of the Vai†vadeva oblation, the Kå™aka Gr¢hyasütra (54.6) enjoins an offering to dharma and adharma at the door: dharmådharmayor dvåre. Here too positive and the negative forms are taken as deities or cosmic principles, just as in the Upanißads. Only in three passages do we have a usage similar to that found in the Dharma†å†tras. At ‡åõGr¢ 2.16.2 we have the refrain iti dharmo vidhœyate in a verse that closely resembles MDh 3.110. The six †lokas in this khañ∂a appear to be taken from a Dharma text. In the BauGr¢ (3.3.31) we have the term dhårmika with a very classical meaning used in the instruction to the student who has completed his studentship: vratasamåptau vedasamåptau vå gurudakßiñåm åhared dhårmiko yathå†akti –“At the end of the observances or at the end of (the study of) the Veda, the dhårmika (student) should bring a gift to the teacher according to his ability.” And, finally, in ņGr¢ (1.7.1) we have a passage very reminiscent of Dharma†åstras: atha khalüccåvacå janapadadharmå gråmadharmå† ca tån vivåhe pratœyåt –“Now, manifold are the dharmas of regions and the dharmas of villages. One should observe these at a marriage.” Here dharma clearly refers to customs and local norms, a meaning that is identical to the Dharma†åstric concept of dharma.

Concluding postscript This somewhat brief though comprehensive survey of the use of dharma in texts roughly belonging to the middle and late vedic period (around 800-400 BCE, although some of the individual texts and passages may be from a later period), shows that in the early texts of this period, especially the Bråhmañas and the early Upanißads, the term is used most frequently with reference to Varuña and the king. It is likely that dharma was part of the specialized vocabulary associatly the same meaning in the Pürva-Mœmå∫så Sütra 2.1.9-10. 23 I have omitted the Vaikhånasa and the Ågnive†aya, which clearly belong to a much later date.

150

LANGUAGE, TEXTS AND SOCIETY

ed with royalty, especially because of its frequent use within the royal consecration (råjasüya). In all likelihood, dharma referred to social order and the laws of society that the king was obligated to enforce. Dharma thus becomes an abstract concept and entity, a cosmic force that stands above the king; it is called kßatrasya kßatram, the power behind the royal power.24 This hypostatization of dharma is continued on in other texts, as when dharma and adharma are considered deities or cosmic categories. There is no doubt that the classical meaning of dharma is encountered in a few passages of the Upanißads, ‡rautasütras, and Gr¢hyasütras. But the rarity of these passages makes one suspect that they reflect the ideology of a later period when dharma had come to define the very essence of the Brahmanical way of life. The question, then, is how did this term become transformed from a somewhat marginal and specialized context of the middle vedic texts to becoming the central concept within Brahmanical religious vocabulary, generating an entire genre of expert treatises, the Dharma†åstras? How did the meaning of dharma change from its earlier limited applications to defining the proper religious life both in Brahmanism and in Buddhism? Now, it is entirely possible that the evolution was gradual and internal to the Brahmanical tradition. Within this scenario, it was the Buddhists who borrowed the fully developed concept from the Brahmanical sources. There are several reasons why this is unlikely, the most significant of which is the fact that in texts that can be dated with some confidence to a pre-Buddhist period dharma does not occupy a central role within the religious vocabulary; the term does not appear still to have acquired the meaning that would enable theologians to define the good religious life as dharma. I will here only summarize the conclusions I have drawn in an earlier paper on the subject (Olivelle, see above pp 121-135). My hypothesis is that the Buddha25 borrowed dharma as he did many other royal symbols to locate and articulate his new religion. Buddha’s own ascribed pedigree makes him an heir to the throne. Prognosticators predicted that he will be a world-conquering (cakravartin) king or an enlightened being. Both the Buddha and other founders of new religions during this period are called jina, the conqueror. The Buddha’s doctrine is compared to a wheel, a metonym for the war chariot and conquest; and his first sermon is the dharmacakrapravartanasütra, 24 On the use of such expressions, also found in satyasya satyam, see H. Oertel, Zum altindischen Ausdrucksverstärkungstypus satyasya satyam “das Wahre des Wahren” = “die Quintessenz des Wahren.” Sitzungsberichte der Bayerischen Akademie der Wissenschaften, Philosophisch-historische Abteilung, Jahrgang 1937, Heft 3. Munich: Verlag der Bayerischen Akademie der Wissenschaften, 1937. 25 I use “Buddha” as a shorthand for the early Buddhists. What the historical Buddha actually taught is impossible to determine, because we have only one lens to view it and that is the Buddhist texts written, revised, edited over centuries.

Semantic History of Dharma

151

“the Sütra that set the wheel of dharma rolling”. The Buddha’s teaching is †åsana, the counterpart of a royal edit. These are all clearly royal symbols used, deliberately I think, to define a new ascetic group and a new religious ideology. Within this context, I think, we can see how dharma, which was called kßatrasya kßatram, the divine principle that gave legitimacy and meaning to a worldly ruler, the term associated with the divine sovereign Varuña, would be a natural choice to define the new dispensation, the new truth (satya) that the enlightened one had discovered. If this hypothesis is correct, then it was within early Buddhism that dharma changed from being a peripheral concept to becoming a central and key theological concept defining the Buddhist religion. Within this transformation, there must have been a semantic development; dharma becomes increasingly ethicized within the primarily ethical religion of Buddhism. It came to define the good and righteous life and the truth (satya) the Buddha discovered which made such a life possible. Another factor probably influenced the term’s ascendancy in the religious vocabulary of India. The use of dharma within a single new religion could be ignored. It’s use as the central concept in defining a new imperial ideology, the ethical ideology of the Maurya empire articulated by A†oka in his edicts, could not be ignored even by the scholastic Brahmins working within the vedic †åkhås. In his brief edicts, A†oka uses the term about 111 times (excluding the repetitions found in the multiple versions of the same edit).26 This number stands in sharp contrast to the numbers we have examined in the much more vast literature of the middle and late vedic periods. It is very likely that A†oka’s use of dharma is mediated by its use within Buddhism. A†oka not only gives pride of place to dharma, he also defines it in completely ethical terms. In Rock Edict 3, for example, he instructs his ministers to teach the dharma in the following words: “Obedience to mother and father is good (sådhu). Giving (dåna) to friends, acquaintances, and relatives, 26 Number of occurrences are given within parentheses: Rock edicts I (2); III (1); IV (10); V (10); VI (1); VIII (3); IX (5); X (2); XI (5); XII (2); XIII (11); XIV (1); XV (2); Pillar edicts I (8); II (3); IV (3); VI (3); VII (30); Minor rock edict I (4 only in Gujarra version); II (1); III (3); IV (1). The last is in the trilingual edict with Aramaic and Greek translations. The Aramaic equivalent of dharma given by A†oka is “qsyt” and the Greek is “eusebia”. John Huehnergard (personal communication) takes the root qst means “(be) true, truth”. The form qsyt is a substantivized adjective, “what is true”. This term thus corresponds to satya with which, as we have seen, dharma is often paired and sometimes identified. The Greek term may mean something like piety and reverence towards gods. According to Paul Woodruff (personal communication) there may be a connection between eusebia and kingship. Augustus is Latin for Sebastos, which is used in connection with Hellenistic kingship. In classical usage, sebein is used for respecting monarchs and eusebein for relations with the gods. As the line between kings and gods was blurred in the Hellenistic era, so was this rather soft distinction. A tantalizing possibility is that A†oka’s use of dharma may, in fact, have been influenced by Hellenistic or Persian royal vocabularies. 27 According to the Girnar version: sådhu måtari ca pitari ca susrüså mitrasa∫stuta∞åtœna∫ båmhañasamañåna∫ sådhu dåna∫ pråñåna∫ sådhu anåra∫bho appavyayatå appabhå∫Ûatå sådhu.

152

LANGUAGE, TEXTS AND SOCIETY

and to Bråhmañas and ‡ramañas is good. Not killing is good. Spending little and possessing little is good.”27 He uses dharma most frequently as the first member of compounds; for him everything is dharmic: so he calls his edits dharmalipi, and his pilgrimage dharmayåtrå. I have counted 31 such compounds used by A†oka.28 How did all this affect the Brahmanical community? Clearly the A†okan reforms displaced the Brahmin’s “special relationship” with the royal power; now there are two equal religious groups worthy of honor and vying for influence and patronage: †ramaña-bråhmaña, in the compound used and possibly coined by A†oka. The influence of the historical memory of this period is possibly reflected in Brahmanical literature of the following centuries.29 The term and the concept dharma acquired a prominence and centrality through both Buddhism and the A†okan imperial theology that was impossible to ignore. My hypothesis is that the emergence of the Dharma†åstric literature, first in the form of prose sütras and then in metrical treatises beginning with Manu, was a direct consequence of Buddhist and A†okan reforms. That a †åstra, an expert tradition of knowledge, be devoted to dharma would seem improbable from its marginal use within the theologies expressed in the middle and late vedic texts. Further, as Wezler (2004) and Lariviere (2004) have demonstrated, there is a deep divide between what Wezler calls “vedic dharma” and “Dharma†åstric dharma”. The latter is based squarely on “custom”, the traditional customs, usages, and practices viewed as authoritative within the Brahmanical communities. Historically, I think, this happened because the Brahmanical experts reflecting on this new dharma that defined one’s religious, ethical, and social life had to find sources that would provide access to such a dharma; in other words, they had to find a pramåña, an authoritative means of knowing dharma. The Buddhists had already formulated just such a pramåña, and that was the “words of the Buddha” (buddhavacana). Only the ipsissima verba of the Enlightened One could give us access to the truth, to 28 Here is the list of these compounds: dhammakåmatå, dha∫maguña, dha∫maghoßa, dha∫macaraña (calana), dha∫mattha∫ba, dha∫madåna, dhammådhithåna, dha∫maniyama, dha∫månugaha, dha∫månupa™œpatti, dha∫månusathi, dha∫månusåsanå, dhammanisita, dha∫måpadåna, dha∫mapalipuccha, dha∫mapaliyåyåni, dha∫måpekkhå, dha∫mamaõgala, dha∫mamahåmåtå, dhammayåtå, dha∫mayuta, dha∫malipi, dha∫mavåya/sœlana, dha∫mavijaya, dha∫maviÛÛ∂∂hi, dha∫mavuta, dha∫masa∫ba∫dha, da∫masa∫vibhåga, dha∫masa∫stava, dha∫masåvana, dhammasusußå. For further details about A†oka’s use of dharma, see Olivelle above pp. 121-135 and Thapar 1997. 29 Historical memory is not the same as historical accuracy. The historical memory of A†oka was different among different groups; the Buddhists generally have a very positive view of A†oka and his work, while the press is not entirely favorable to A†oka in the Brahmanical tradition. With regard to the possible influence of such historical memory in shaping later Brahmanical literature, see James Fitzgerald, The Mahåbhårata: 11. The Book of Women, 12. The Book of Piece. Vol. 7. Chicago: Chicago University Press, 2004; and Alf Hiltebeitel, Rethinking the Mahåbhårata: A Reader’s Guide to the Education of the Dharma King. Chicago: University of Chicago Press, 2001.

Semantic History of Dharma

153

dharma. The Brahmanical tradition already had such an infallible pramåña, the Veda. It was natural, therefore, for the expert tradition of Dharma†åstra to proclaim that the Veda is the pramåña or the müla (root) of dharma. This, however, was abstract theory and theology; it was Mœmå∫så. In reality, however, the Veda has little to say about dharma in its new and expanded meaning. The “real” source of dharma, as Wezler and Lariviere have pointed out, was custom: the smr¢ti (memory) and åcåra (practice) of cultured, learned, and virtuous Brahmins, who came to be designated as †iß™a. This brings up the issue of the dates of the earliest †åstras on dharma, the Dharmasütras. The earliest reference to such texts is in Kåtyåyana’s Vårttika 39 on Påñini 1.2.38, which uses the term dharma†åstra. Pata∞jali refers to “authors of Dharmasütras” (dharmasütrakåråΔ) in the plural twice in his Mahåbhåßya on Påñini 1.1.47 (p. 115), and 5.1.119 (p. 365). This would correspond to my estimate that the earliest Dharmasütras do not date earlier than the 3rd century BCE. If this historical reconstruction is accepted, then we will have another tool to date early Indian texts: those that reflected a developed meaning of dharma and give it a central role in their theology and vocabulary must be assigned to a date at least after the development of Buddhist doctrine, if not after the proclamation of A†okan imperial ideology. We have already seen some such passages in the Taittirœya Upanißad and the Chåndogya Upanißad, as well as in some of the Gr¢hyasütras. The dates of these texts, or at least the time when these passages were incorporated into them, may have to be rethought. The same may be said for the grammarians Yåska and Påñini, who also use dharma in ways that indicate a highly developed and expanded concept. In Yåska’s Nirukta, for example, we have the use of the verb √car with reference to dharma in the story of Devapœ and ‡antanu: adharmas tvayå caritaΔ –“You have done an adharma” (2.10).30 Then there is a verse that looks very much like a Dharma†åstric verse asserting the right of daughters to inherit property (3.4): avi†eßeña putråñå∫ dåyo bhavati dharmataΔ Ù mithunånå∫ visargådau manuΔ svåya∫bhuvo ‘bravœt ÙÙ –“In accordance with dharma the partition of inheritance is done among male and female children no differently from the sons; so has Manu Svåya∫bhu declaired”.31 Given that Yåska has been subject to later redactoral activities, it is unclear whether these usages are original. Påñini also, who is generally assigned to around 400 BCE, in several sütras gives derivatives of dharma that show the word to have a classical meaning. Thus in sütra 4.4.41 he indicates the derivative dhår30 The story is also found in the Br¢haddevatå, where we find the expression dharmavyatikrama, “transgression of dharma” (8.3). 31 See also Nirukta, 6.19; 8.10; 12.13. In chapter 14, generally considered to be a later addition, we also have similar use of dharma: 14.6, 14.

154

LANGUAGE, TEXTS AND SOCIETY

mika in the sense of dharma∫ carati. In sütra 4.4.91, he shows that the affix yat comes after dharma (i.e. dharmya) in the sense of dharmeña pråpyam (obtainable by dharma) and in the very next sütra shows that the same affix can be used with dharma to mean dharmåd anapetam (not deviating from dharma).32 These terms and usages must have been derived from or at least mediated by the spoken Sanskrit of the time (bhåßå) rather than directly from the vedic vocabulary. If with von Hinber and Falk33 we date Påñini 350-300 BCE, his use of dharma would be quite explicable within the hypothesis I have presented. The period we have examined was probably a watershed in the semantic development of dharma. Once the term became the central concept defining a civilization, this development accelerated. Other studies in this volume (Olivelle 2004b) discuss aspects of its semantic range. It is, however, important to recognize that words do not exist in a vacuum; they are used by individuals and groups that have their own histories and interests and that change the meanings of words as they use them. Philology must not simply look at the web but at the spider also. The analysis of words thus can give significant and unique insights into not just into human language but into human history. Given its centrality in Indian discourse, a close study of dharma has the potential to reveal interesting and hitherto unknown contours of ancient Indian history.

32 See also Påñini 4.4.47 and 6.2.65 for dharmyam, and 4.2.46 for dharmavat. See also 5.2.132; 5.4.124. 33 Oskar von Hinber, Der Beginn der Schrift und frhe Schriftlichkeit in Indien (Wiesbaden: Franz Steiner, 1989), p.34; Harry Falk, Schrift im alten Indien: ein Forschungsbericht mit Anmerkungen (Tbingen: Gunter Narr Verlag, 1992), p. 304.

A

155

VII Explorations in the Early History of Dharma†åstra*

The absence of the term †åstra in the vedic texts, including the Bråhmañas and the Upanißads,1 and the relative frequency of its use in the literature falling broadly within the period “Between the Empires” raise a set of questions relating to the emergence of the †åstra genre in the 4-5th centuries BCE. In this study I focus on a single sub-genre within †åstra, namely the Dharma†åstra. Why did a class of technical literature (†åstra) devoted to dharma arise in the second half of the first millennium BCE? What were the social, political, and religious contexts within which this genre of literature was created? Was its creation simply a natural development of Brahmanical scholasticism, or was it a response to external challenges? And, finally, how did the expert tradition created by this genre respond to the new challenges and changing circumstances during the first six or seven centuries of its existence, centuries that saw the rise and demise of two empires, the Maurya and the Gupta, and the rise of foreign dynasties including the Greek-Bactrian, the ‡aka, and the Ku†åna? The term †åstra is found for the first time in Yåska’s Nirukta (1.2, 14), where the reference is probably to the science or a text of etymology (nirukta). Likewise, the R¢gvedapråti†åkhya (11.36; 14.30) uses the term to refer to the pråti†åkya tradition. Kåtyåyana and Pata∞jali use it with reference to Påñini’s Aß™ådhyåyœ.2 Likewise, the * To be published in Between the Empires: Historical and Cultural Studies 4th Century BCE — 5th Century CE. Edited by Patrick Olivelle. New York: Oxford University Press . 1 The term occurs only in a single passage of the R¢gveda 8.33.16, where the meaning of the term is far from clear. 2 See especially Kåtyåyana on Påñini 1.1.1 (1, p. 8); 1.1.1 (10); .62 (1); 1.1.62 (1); 2.1.1 (16); 6.1.84 (4).

156

LANGUAGE, TEXTS AND SOCIETY

Vedåõgajyotißa uses the term to refer to astronomical treatises.3 Significantly, however, the latter text uses the expression vedåõga†åstråñåm, indicating that the †åstra may have been used as a generic term covering treatises dealing with the Vedåõgas. Yet, it is notable that †åstra is never used with reference to the specifically ritual manuals, the ‡rautasütras and the Gr¢hyasütras. The term occurs only in the Kåtyåyana ‡rautasütra (1.6.21), and there it appears to refer to the Veda. We find †åstra used with reference to the literature on dharma, however, as early as the vårttika of Kåtyåyana, who uses the expression “Dharma†åstra”.4 This brief sketch indicates that †åstra may have been used first with reference to manuals of instruction that were useful in understanding the Veda and practicing the Vedic rituals, but significantly not with reference to specifically ritual texts.5 Even though the tradition considers the earliest texts of dharma, the Dharmasütras, to form part of the Kalpasütras that included also the ‡rautasütras and the Gr¢hyasütras, it was only the texts dealing with dharma to which the label †åstra came to be attached. This may indicate that in reality the tradition considered texts on dharma to constitute a special category and that its connection to the other two categories of ritual sütras was not original.6 The hypothesis I put forward here is that Dharma†åstra did not develop as an integral part of the ritual literature produced within the Vedic †åkhås; it developed instead as an autonomous genre. This brings up the crucial question: how and why did this genre of literature come into being? This is a question that has not been raised in any of the extant “histories” of Dharma†åstra (Kane 1962-75 and Lingat 1973), a question that needs to be answered at least hypothetically if we are to understand the early history of this †åstric literature.7

Dharma and the Beginnings of Dharma†åstra One possible reason for the neglect to examine the origin of the Dharma†åstra genre is the assumption by most scholars that the term and concept of dharma has always been central to the Brahmanical understanding of religion, society, and the cosmos, and that scholarly discourse on this topic must have been an ongoing activity among 3

The R¢gvedic recension, 25, 36. dharma†åstra∫ ca tathå. Kåtyåyana on Påñini 1.2.3, vårttika 39 (I: 242). 5 Etymologically derived from the verb √†ås, to instruct, the term †åstra refers to an instrument of such instruction, probably a manual used for instructional purposes. 6 I do not mean that Vedic †åkhås did not produce literature on dharma. They evidently did, as seen in the works ascribed to Åpastamba, Hirañyake†in, and Baudhåyana. My point is that the expert tradition on dharma probably did not arise as an integral part of the ritual tradition of scholarship. 7 A notable exception is Wezler (2004), who raises significant points regarding the origin and codification of the early Dharmasütras. 4

Explorations in the Early History of Dharma†åstra

157

Brahmanical schools constituting the Vedic branches (†åkhå).8 Another reason may have been the conviction that dharma had a primarily ritual dimension and was ultimately based on the Veda, and exegetical and interpretive works dealing with the Veda, therefore, must necessarily involve discussions of dharma.9 Both these assumptions, I believe, are wrong. The term dharma was probably a neologism invented by the poets of the R¢gvedic hymns; it has no cognates in other Indo-European languages, including Avestan.10 Although the term occurs 67 times in the R¢gveda, the frequency drops dramatically in the middle and late vedic literature. As I have shown elsewhere,11 dharma is a marginal concept in the theology expressed in the Bråhmañas and the early Upanißads; it is used principally within the royal rather than the strictly ritual vocabulary.12 The likely reason for its rise to prominence within the religious discourse of India between the fourth and fifth centuries BCE is its assumption, along with other terms and symbols of royalty, by the newly emergent ascetic religions, especially Buddhism, and its use for an imperial theology by A†oka in the middle of the third century BCE. Given the marginality of dharma in the vedic vocabulary, it is unlikely that the term would have been the subject of intense scholarly or exegetical inquiry during the late vedic period. To claim that during the Vedic period “dharma was par excellence the sacrificial act which maintains and even conditions the cosmic order” (Lingat 1973, 3) simply ignores the facts and projects later Mœmå∫så views onto the vedic discourse. The fact is that the major 8 Kane (1962-75, I: 19) says: “It seems that originally many, though not all, of the dharmasütras formed part of the Kalpasütras and were studied in distinct sütracarañas,” and views dharmasütras as “closely connected with gr¢hyasütras in subjects and topics” (I: 20). Lingat (1973: 18) concurs: “Originally, it seems, most of these dharma-sütras, if not all of them, belonged to a collection of kalpa-sütras and were attached to a particular Vedic school.” 9 Lingat (1973: 3), for example, asserts: “During the Vedic period the fundamental laws of the universe were identified with the laws of the sacrifice. Consequently dharma was par excellence the sacrificial act which maintains and even conditions the cosmic order.” 10 For detailed studies of the early history of the term dharma, see Brereton 2004 and Horsch 1967. 11 I have dealt with this topic at length in two articles. In Olivelle 2004d (above, pp. 137-154). I have dealt with the semantic history of the term in the middle and late Vedic period, continuing thus the work of Brereton 2004. In “forthcoming-a” (above, pp. 121135) I have dealt with the use of the term in A†oka and in early Buddhism and proposed possible ways in which the term may have entered the mainstream of the theological vocabulary of Brahmanism. I will not argue this point at length here, therefore, referring the reader to these two earlier works. 12 The most frequent occurrence of the term is in connection with the royal consecration (råjasüya), but even though the consecration is a ritual act the term is used principally to point out the functions of the king. Here I disagree both with Horsch (2004 [1967]) and with my friend and mentor Albrecht Wezler (2004, 633), who asserts: “I assume that the Mœmå∫så was stimulated to apply this term [dharma] to the content of the Vedic prescriptions only by the Dharma†åstra, even though it has already been a well-known term in the sacrificial context, there denoting the cosmos-sustaining and life-preserving power.”

158

LANGUAGE, TEXTS AND SOCIETY

scholastic works on the ritual, the ‡rautasütras and Gr¢hyasütras, hardly ever use the term dharma either with reference to ritual activity or with reference to the ritual and other duties of a Brahmin (Olivelle 2004d). Why then do we see the proliferation of Dharmasütras during the last three or four centuries BCE? The hypothesis I want to propose is that once dharma had become a central concept in the religious discourse of Buddhism and once it had penetrated the general vocabulary of ethics especially through its adoption by the Maurya emperors, certainly by A†oka and possibly also by his predecessors, in developing an imperial theology, Brahmanical theologians had little option but to define their own religion, ethics, and way of life in term of dharma. Indeed, the scrutiny of the early meaning of dharma within its Dharma†åstric use suggests that it was not the Veda but the “community standards” prevalent in different regions and communities that were taken to constitute dharma. The early texts on dharma speak of de†adharma, jåtidharma, kuladharma –the dharma of regions, castes, and families/lineages. Clearly, these texts regard dharma as multiple and varied; each of these kinds of dharma can hardly be expected to be based on the Veda.13 The tradition of Vedic exegesis and hermeneutics known as Mœmå∫så, however, exerted a strong influence on the Dharma†åstric tradition, and gradually that influence led to the dominance of the Veda as the principle if not the single source of dharma within the theological understanding of the term, something that Wezler (2004, 643) has called the “vedamülatva concept”, a concept that was first articulated by Gautama (GDh 1.1). As we will see, “community standards” came to be restricted to the standards of Brahmanical elite, the †iß™as, living within a theologically defined sacred geography known as the åryåvarta. In spite of this theologizing, dharma of the Dharma†åstras remained rooted in the normative practices of the various communities. Wezler (2004, 648) remarks that religiosity based on dharma in contradistinction to the elitist sacrifice, led to “a democratization regarding the access to salvation”. I think this is noteworthy and important, because if we place the origin of Dharma†åstra within the context of Brahmanical response to the “democratic” ethics and religion preached by Buddhism and the new ascetic religions, we can better appreciate both the sociological context of the rise of this genre of literature and the significant role it played in the new Brahmanical religiosity and soteriology.14 My studies of the Vedic vocabulary has shown that at best dharma is a marginal concept in the vocabulary of the middle and later Vedic texts. 13 Yåj∞avalkya, in fact, explicitly states that when a king conquers a land he should make sure that the customs of that land are protected; he should not impose the customs of his own region on the newly conquered land: yasmin de†e ya åcåro vyavahåraΔ kulasthitiΔ Ù tathaiva paripålyo ‘sau yadå va†am upågataΔ ÙÙ YDh 1.343. 14 A similar hypothesis regarding the composition of the great epic, Mahåbhårata, has been proposed by Hiltebeitel (2001, 2004) and by Fitzgerald (2004).

Explorations in the Early History of Dharma†åstra

159

Lariviere (2004) and Wezler (2004) have argued, convincingly I believe, that the historical source of dharma in the Dharma†åstras is not the Veda but “custom” (åcåra), that is, the normative behavior and practices of various and varied historical communities. Lariviere (2004, 612) presents his view of the nature of Dharma†åstra clearly: “Let me begin by giving my view of the nature of the dharma†åstra literature. I believe that the dharma†åstra literature represents a peculiarly Indian record of local social norms and traditional standards of behavior.” Wezler (2004, 642) agrees completely with this new view of the source of dharma in Dharma†åstra: “The dharma of the Dharma†åstra [...] is, in its essential parts, a record or codification of custom and convention.” In seeing the Veda or some transcendent tradition as the source of dharma, historians of Dharma†åstra have bought into the theological position enunciated in most of the Dharma†åstras themselves as to the provenance of the dharma that they are teaching thus confusing history with theology.15 All extant texts on dharma begin with a discussion on the means of knowing (pramåña) or the sources of dharma. Although there are some differences in the texts, by and large, they point to three sources: Veda (or †ruti), smr¢ti, and normative custom (åcåra). The GDh (1.1-2) provides the most explicit statement: “The source [or root] of dharma is the Veda, as well as the tradition and practice of those who know it [the Veda]” –vedo dharmamülam Ù tadvidå∫ ca smr¢ti†œle.16 The category of smr¢ti is somewhat unclear; it may refer to the living memory of the Brahmanical community or to written sources of such immemorial customs. Clearly by the time of Manu smr¢ti meant texts, because he equates the term with Dharma†åstra. The Mœmå∫så tradition of vedic exegesis, which exerted a strong influence on the Dharma†åstric tradition from its very inception, began to interpret the multiple sources of dharma as having their origin in a single source, the Veda. This is stated explicitly by Manu (2.7): “Whatever dharma Manu has proclaimed with respect to anyone, all that has been taught in the Veda, for it contains all knowledge” –yaΔ ka†cit kasyacit dharmo manunå parikœrtitaΔ Ù sa sarvo ‘bhihito vede sarvaj∞ånamayo hi saΔ ÙÙ Veda contains all knowledge and thus, a priori, should contain all dharma. This position is already hinted at in the above statement of the Gautama Dharmasütra when it qualifies that only the tradition and practice “of those who know the Veda” are authoritative. The authority of tradition and practice 15 Even the later Brahmanical tradition is not unanimous in seeing the Veda as theologically the single source of dharma. Medhåtithi, the author of the earliest extant commentary on the MDh, in the context of the duties of the king states explicitly that not all of dharma is based on the Veda: pramåñåntaramülå hy atra dharmå ucyante na sarve vedamülåΔ (on MDh 7.1). 16 For an examination of the meanings of smr¢ti and †œla, see Wezler 2004. The process whereby the Veda came to be considered as the only or primary source of dharma, is called by Wezler (2004, 648) as “vedification”.

160

LANGUAGE, TEXTS AND SOCIETY

are here implicitly connected with the Veda. Åpastamba (ÅpDh 1.12.10-12) provides the earliest evidence of the hermeneutical argument for the position when he claims that all rules were originally found in the Bråhmañas; but some sections of these were lost over time, and they can be recovered by observing the practice: “All rules are described in the Bråhmañas. The lost Bråhmaña passages relating to some of them are inferred from usage” – atha bråhmañoktå vidhayas teßåm utsannåΔ på™håΔ prayogåd anumœyante. Here we have the Mœmå∫så concept of anumita†ruti, that is, Vedic passages that are inferred to have existed on the basis of either smr¢ti or practice.17 Given the other statement of Åpastamba on the sources of dharma that I will discuss below, however, this particular statement may indeed refer only to matters relating to Vedic ritual (vidhi) and not to dharma as such, especially because it is made within the context of explaining the daily recitation of the Veda. The “lost Veda” argument will be used later, however, to underpin the authority of other sources of dharma within the theological fiction that the Veda is the sole source of dharma. The Mœmå∫så view of dharma, then, is that the Veda is the sole means of knowing it; when a specific vedic text is wanting with regard to a particular aspect of ritual or behavior, then one can use supplementary sources, such as smr¢ti and åcåra, on the basis of which one can infer the existence of a vedic text. This theological claim that camouflages the historical sources of dharma is pointed out by Lariviere (2004, 612): “What I mean is that the whole of the dharma corpus can be viewed as a record of custom. It is not always a clear record because of the idiom and the fictions that came to be the mode of expression of the dharma literature. That the dharma is a record of custom is obfuscated by the fact that the idiom of all the dharma literature is one of eternality and timelessness. This means that there are no contemporaneous references which can help us to establish the chronology of these ideas nor is there admission that custom and practice changed and evolved over time. It is further obfuscated by the fact that the dharma literature clings to the claim that all of its provisions can be traced directly or indirectly to the Veda, the very root of dharma.” Indeed, as Pollock (1985, 1990) has shown, the reason for the “idiom of eternality and timelessness” is precisely the theological imperative that to be based on the Veda means to transcend time and historical context and change.18 The historical reality is very different from this theological position. The dharma taught in the Dharma†åstras has little to do with the Veda but reflects the actual practices of local groups; the 17

This doctrine was known to Åpastamba (1.4.8), who says that “a vedic text has greater force than a practice from which the existence of a corresponding vedic text has to be inferred” –†rutir hi balœyasy ånumånikåd åcåråt. 18 For a more extensive treatment of the connection between dharma and Veda, see Heesterman 1978.

Explorations in the Early History of Dharma†åstra

161

Dharma†åstras themselves are nothing but the textualization of such practice. Evidence from texts belonging roughly to the last three centuries before the common era indicates that this is not merely a historical conclusion of modern scholarship; it appears to have been the view of at least three major authors belonging to the early period of Dharma†åstric textual production: the grammarians Kåtyåyana and Pata∞jali, and Åpastamba, the author of both a Gr¢hyasütra and a Dharmasütra. Kåtyåyana, in his Vårttikas on Påñini’s Aß™ådhyåyœ, makes a distinction between what is found in the Veda (vede and vaidika) and what is prevalent in the world (loke and laukika). The major use of these two categories is for grammatical purposes, giving examples from the Vedic texts and common speech, the two areas of language encompassed by the early Sanskrit grammars.19 At least in Pata∞jali’s understanding of Kåtyåyana, the category of laukika (worldly) does not apply simply to what ordinary people say and do but to norms of behavior encoded in textualized form that are certainly Dharma†åstric injunctions. The clearest example of such injunctions is found in Pata∞jali’s commentary on Kåtyåyana’s Vårttika 5 (on Påñini 6.1.84; III: 57-8), which contains the expression yathå laukikavaidikeßu. Here the examples given by Pata∞jali support the view that in teaching dharma (dharmopade†a) the injunctions refer not to individuals (avayava) but to categories or classes (anavayava).20 As laukika examples, Pata∞jali gives: bråhmaño na hantavyaΔ (“A Brahmin should not be killed”), surå na peyå (“Liquor should not be drunk”), and pürvavayå bråhmañaΔ pratyuttheyaΔ (“An older Brahmin should be greeted by standing up.”). If these injunctions are taken as referring to an individual rather than to a class, then someone could fulfill the obligations by not killing a particular individual Brahmin, by not drinking liquor once, and by rising up to greet a single older Brahmin on a single occasion, after which time he is free to kill Brahmins, to drink liquor, and not to rise up to greet older Brahmins. This is obviously erroneous; so we must assume that the injunctions refer to classes rather than to individuals. The Vedic example given in this context is: vasante bråhmaño ‘gniß™omådibhiΔ kratubhir yajeta (“In the spring a Brahmin should offer sacrifices such as the Agniß™oma”). Here also the obligation is a continuous one and cannot be fulfilled by doing this act once. In support of the laukika injunction that a younger person should rise when approached by an older person, Pata∞jali cites a verse, which is found also in Manu (2.120): 19 See Kåtyåyana’s Vårttika 2 on Påñini 1.2.45 (I: 217); 15 on 6.1.1 (III: 3); 5 on 6.1.83 (III: 55); 2 on 6.2.36 (III: 125). 20 The terms used here are avayava and anavayava, literally “part” and “non-part”. The terms are somewhat obscure, but they appear to refer to individuals (dravya or vyakti) and classes (jåti or åkr¢ti).

162

LANGUAGE, TEXTS AND SOCIETY

ürdhva∫ pråñå hy utkråmanti yünaΔ sthavira åyati Ù pratyutthånåbhivådåbhyå∫ punas tån pratipadyate ÙÙ For when an older person comes near, the life breaths of a younger person rise up, and as he rises up and greets him, he retrieves them.

Other laukika examples also show that they are actually Dharma†åstric in nature. Thus on Påñini 1.1.1 (I: 5, 8) Pata∞jali repeats the following maxims twice: abhakßyo gråmyakukku™o ‘bhakßyo gråmya†ükaraΔ (“It is forbidden to eat a village fowl; it is forbidden to eat a village pig”),21 injunctions that are common in the Dharma†åstras. He also gives the well-known maxim (on Påñini 1.1.1; I: 5) pa∞ca pa∞canakhå bhakßyåΔ (“The five five-nailed [animals] may be eaten”). What is significant here is that for grammarians both the Veda and the loka are pramåña, authoritative with respect to correct speech. This authoritative nature of loka is carried over into the Dharma†åstric framework when Pata∞jali cites injunctions. Clearly, not every thing that is said or done in the world is so authoritative. Thus loka for Pata∞jali and most likely also for Kåtyåyana referred to Dharma†åstra. We have confirmation of this conclusion. The two examples on not killing Brahmins and not drinking liquor cited above, that Pata∞jali refers to loka at Påñini 6.1.84; III: 57-8, is cited by him again in his comments on Vårttika 39 on Påñini 1.2.64. The Vårttika read dharma†åstra∫ ca tathå (“And so also Dharma†åstra”), and as an example of “Dharma†åstra” Pata∞jali gives: bråhmaño na hantavyaΔ surå na peyå (“A Brahmin should not be killed; liquor should not be drunk”). Clearly, for Pata∞jali loka and dharma†åstra are, if not synonyms, at least equivalents with respect to authoritative injunctions. The question that still remains is whether “dharma†åstra” as used by Pata∞jali refers to texts as treatises such as the extant Dharmasütras or to the individual maxims that he cites. Wezler (2004, 642) has drawn attention to this issue in examining the meaning of “text” within the history of early Indian law: “Methodologically speaking, however, I consider it very important in this context to expound the problems of the notion of ‘text’ [...] because the consideration of textual genesis – and textual history – in India advises caution in using the term ‘text’. Thus in the present case, one may consider the possibility that certain ‘legal phrases’ which constitute the contents of the Dharma†åstra, originated initially without any connection to similar or dissimilar phrases as elements of ‘tradition’ (smr¢ti). In other words, the textualization might have begun with single elements of custom and legal tradition.” Clearly “dharma†åstra” in Pata∞jali may well refer to such individual legal or moral pronouncements. I think, however, that evidence points to the fact that 21

Also repeated on Påñini 7.3.14 (III: 320)

Explorations in the Early History of Dharma†åstra

163

Pata∞jali views the rules he cites as derived from “texts” in the sense of a complete treatise or †åstra. Pata∞jali uses the term “dharma†åstra” with reference to such texts and cites individual norms from such treatises. The general pattern of use of the term †åstra during and prior to this period shows that it was used with reference to treatises, such as the Veda or the Aß™ådhyåyœ, rather than single pronouncements. In his comments on Påñini 1.1.47 (I: 115),22 Pata∞jali’s statement: naive†vara åj∞åpayati nåpi dharmasütrakåråΔ pa™hanty apavådair utsargå bådhyantåm iti (“Neither does the Lord command nor do the authors of Dharmasütras declare: ‘General rules should be set aside by special rules’”) also points in this direction. Not only does this show that Pata∞jali was familiar with the genre of literature called dharmasütra that had authors, it also shows that in his mind the authors of Dharmasütras paralleled œ†vara. Now, it is unclear what or whom Pata∞jali refers to by this term. The term œ†vara (Lord) here, in the understanding of the later commentator Någojibha™™a, refers to the Veda. According to this interpretation, for Pata∞jali the Dharmasütras were as much texts as the Veda. It is more probable, however, that œ†vara refers to the king,23 in which case the authors of Dharmasütras parallel the king in authority to pronounce on matters of public importance. It is extremely unlikely that singular injuctions are referred to here by dharmasütra, and that the authors are merely stating individual injuctions. Even the proponents of “floating legal maxims” do not contend that they had actual and identifiable authors. In the early literature, furthermore, the titles Dharma†åstra and Dharmasütra were synonyms, the former referring to their substance and the latter to their linguistic form. Even in classical texts we do not find the distinction that modern scholars make between Dharmasütras and Dharma†åstras, the latter referring to texts composed in verse.24 There is another class of ancient texts, the Gr¢hyasütras, that are similarly divorced from the Veda. Wezler (2004) summarizes his assessment of the two classes of texts: “Gr¢hyasütras as well as Dharmasütras are verbalizations of certain regional or tribal-specific aspects of traditional social ‘practice’ of the Åryas. Although historically consecutive, they surely have factual points of contact. These verbalizations are textual ‘coagulations’ of the late Vedic period that 22 The identical wording naive†vara åj∞åpayati nåpi dharmasütrakåråΔ pa™hanti is also found in Pata∞jali’s comments on Påñini 5.1.119 (II: 365). 23 This is confirmed by Pata∞jali’s statement on Påñini 6.1.2 (Kåtyåna’s vårttika 9; III: 7): loka œ†vara åj∞åpayati where the command of the Lord refers to worlds (laukika) matters and not Vedic. Here, clearly, the œ†vara is the king and not the Veda or god. I thank Madhav Deshpande for his help in resolving the meaning of œ†vara in Pata∞jali. Further, in his comments on Påñini 1.1.38 (I: 177) Pata∞jali clearly states that œ†vara is a synonym (paryåya) of råjan (king). See also his comments on Påñini 2.3.9 (I: 447). 24 Even Yåj∞avalkya, composed around 4-5th century CE, lists under dharma†åstraprayojakåΔ the authors of the ancient Dharmasütras.

164

LANGUAGE, TEXTS AND SOCIETY

were apparently regarded by the authors as fundamentally different from older parts of the tradition.” Even though it is clear that the domestic (gr¢hya) rites are different from the vedic (†rauta) rites and are based on customary practice of the community, the authors of the Gr¢hyasütra texts themselves do not feel the need to make this distinction explicit or inquire about the sources of these rites. Only the Åpastamba Gr¢hyasütra (1.1.1) has this very brief comment: “Next, the rites that are obtained from customary practice” – atha karmåñy åcårad yåni gr¢hyante.25 Again the rules of Gr¢hya are taken to be based on customary practices rather than on the Veda. Indeed, no other †åstra of the early period saw the need to explicitly address the question of how one comes to know (pramåña) its subject matter. The issue of sources is passed over in silence.26 Why did the authors of the earliest texts on dharma feel the need to address this issue explicitly? I want to propose the hypothesis that the authors of early Dharma†åstras were working both within the model provided by the Buddhist texts and in response to the Buddhist appropriation of dharma. The Buddhist theory of dharmapramåña, that is, the way one knows whether a particular oral text is authoritative, is articulated in the doctrine of buddhavacana (“Buddha’s word”). Either proximately or ultimately all pronouncements on dharma must go back to the words of the Buddha himself for those pronouncements to have any authority or validity. This conviction is encapsulated in the opening words of every Buddhist text: eva∫ mayå †rutam (Påli: evam me suta∫): “Thus have I heard.”27 It is probable that the Brahmanical scholars writing on dharma, a term that we have noted did not have a central role within the previous Brahmanical discourse, were consciously responding to this Buddhist theory by proposing a different pramåña, a different authoritative source of dharma. This source they found at first not in the Veda, which has little to say on the topic, but in the customary norms and practices (åcåra) of living communities, among which, we must suppose, the practices of the Brahmanical community were considered the model and yardstick. The conclusions of Wezler and the use of the term såmayåcårika by Åpastamba as the first source of dharma support such an hypothesis. We also saw that for 25 The commentator Haradatta, however, is explicit: dviprakåråñi karmåñi †rutilakßañåny åcåralakßañåni ca Ù “Rites are of two kinds, those characterized by the Veda and those characterized by customary practice.” See also Åpastamba Gr¢hyasütra 3.7.23. It is very likely that the author of the Dharmasütra and of the Gr¢hyasütra ascribed to Åpastamba was the same individual (Olivelle 2000, 4, n. 5). On the customary nature of the Gr¢hya material, see also Gobhila Gr¢hyasütra 3.3.29; ņvalåyana Gr¢hyasütra 1.7.1. 26 Later texts, such as the Månava-Dharma†åstra, the Nå™ya†åstra, and the medical †åstras, do deal with its provenance, but not in terms of authoritative sources but in terms of their mythical origin from the creator god. 27 What connection the evolution of the term †ruti to mean the Veda has to this usage of the Buddhists is unclear. No historical-linguistic work has been undertaken thus far on the crucial term †ruti. For an attempt at relating †ruti and smr¢ti, see Pollock 1997.

Explorations in the Early History of Dharma†åstra

165

Pata∞jali Dharma†åstric prescriptions were located principally in loka, that is, the actual practices, behavior patterns, and pronouncements of living communities rather than in the Veda. The problem for the authors of the †åstras was how to limit and control the enormous diversity with respect to norms of conduct prevalent in the different regions, castes, and communities across that vast land. The early history of the Dharma†åstras testifies to the continuing efforts to define and limit the universally authoritative practices to those prevalent in Brahmanical communities, and even there to draw boundaries, both geographical and ideological, around especially authoritative Brahmins.

The Early History When the earliest Dharmasütras were written (or orally composed and transmitted) has been a matter of scholarly conjecture. If my argument for the history of the term dharma and its incorporation into an expert tradition within Brahmanism is accepted, however, then the earliest writings on dharma cannot be earlier than second half of the fourth century BCE, perhaps even somewhat later. As we have seen, the earliest reference to Dharmasütra is by Pata∞jali, generally assigned to the middle of the 2nd century BCE. The term Dharma†åstra is used by his predecessor Kåtyåyana (kårikå 39 on Påñini 1.2.64; I: 242). There are four extant Dharma†åstric texts written in the sütra mode and ascribed to Åpastamba, Gautama, Baudhåyana, and Vasiß™ha.28 It is evident, however, that a much larger body of literature on dharma did exist in the three centuries prior to the common era. The extant Dharmasütras cite or refer to opinions of seventeen authorities.29 Although it is unclear whether the statements ascribed to them are single memorable quotes or derived from larger compositions authored by or ascribed to those individuals, it is probable that at least some of these authors did write Dharmasütras that have not survived. This is clearly the case with Hårœta, who is 28 For the dating of these texts, see Olivelle 2001, 4-10. I take Åpastamba to be the oldest, followed by Gautama. Wezler (2004, 650, n. 11) calls Kangle’s argument for placing Gautama late “not entirely convincing”. Besides the arguments I have spelt out in the above work, it is very clear that Gautama represents a much more advanced stage in the development of thinking regarding judicial procedure (vyavahåra), as can be seen by comparing the four Dharmasütras on this subject in my forthcoming Dharmasütra Parallels (Delhi: Motilal Banarsidass). 29 Aupajaõghani (BDh 2.3.33); Bhållavins (BDh 1.2.11; VaDh 1.14); Eka (ÅpDh 1.19.7); Hårœta (ÅpDh 1.13.11; 1.18.2; 1.19.12; 1.28.1, 5, 16; 1.29.12, 16; BDh 2.2.21; VaDh 2.6); Kañva (ÅpDh 1.19.3; 1.28.1); Kåñva (ÅpDh 1.19.2, 7); Kapila (BDh 2.11.28); Ka†yapa (BDh 1.21.2); Kåtya (BDh 1.3.46); Kautsa (ÅpDh 1.19.4; 1.28.1); Kuñika (ÅpDh 1.19.7; Kutsa (1.19.7); Mahåjaj∞u (BDh 3.9.21); Manu (ÅpDh 2.14.11; 2.16.1; GDh 21.7; 23.28; BDh 2.3.2; 4.1.13; 4.2.15; VaDh 1.17; 3.2; 11.23; 12.16; 13.16; 19.37; 23.43); Maudgalya (BDh 2.4.8); Pußkarasådi (ÅpDh 1.19.7; 1.28.1); Vårßyåyañi (ÅpDh 1.19.5, 8; 1.28.1).

166

LANGUAGE, TEXTS AND SOCIETY

cited eight times by Åpastamba, and once each by Baudhåyana and Vasiß™ha. Hårœta’s positions were conservative, clearly to the right of what Åpastamba considered the mainstream. Both Åpastamba and Baudhåyana cite him as saying simply and imperiously “That is false” (mithyå etad) with reference to opinions of others: in Baudhåyana (2.2.21) against those who say that the children of those excommunicated from caste do not become outcastes themselves; and in Åpastamba (1.28.16) against suicide as a penance for those who commit incest. At another point Hårœta claims that a person who covets another’s property is a thief, an opinion shared by two other Dharmasütrakåras, Kañva and Pußkarasådi, whereas Vårßyåyañi thinks that there are exceptions, such as fodder for an ox. Hårœta, however, is unconvinced and asserts that in all cases the owner’s permission must be obtained (ÅpDh 1.28.1-5). All the citations of ancient authors in the extant Dharmasütras are in the context of diverse opinions and controversies on various points of proper conduct and ritual procedure. Clearly we have here a vibrant scholarly debate on a variety of issues, very different from the later tradition which sought to present a singular point of view eliminating or interpreting away divergent voices. The Brahmanical community in ancient India, although it shared some common interests especially in preserving its privileges, was not monolithic. As I have shown elsewhere (Olivelle 1993), Brahmins represented diverse interests and points of view, especially in the realm of religious ideas and institutions. Some clearly favored the ascetic movements and ideologies that emerged in the middle of the first millennium BCE and were the architects of the new and somewhat revolutionary theory of the four å†ramas. One question that emerges from our discussion of the creation of the new genre of literature dedicated to dharma is this: from what segment of the Brahmanical community did the authors of Dharma†åstric texts come? There is, of course, no clear and definite answer to this, but a clue may be found in the manner the early Dharmasütras are organized. All the ancient Gr¢hyasütras begin with marriage and the establishment of a household with a new ritual fire. This is only to be expected, as the entire Brahmanical ritual system and way of life are centered on the married householder. The Gr¢hyasütras then follow the married householder through his life, providing instruction about a variety of subjects, especially the procreation of children, their initiation and education, their eventual marriage, and finally death and funerary ceremonies. One would have expected the Dharmasütras to follow this model, given that they are also principally concerned with the married householder. That, however, is not the case. After a brief discussion of the sources of dharma, the authors of Dharmasütras begin their works with a long discussion of Vedic initiation (upanayana) and the period of Vedic study that a

Explorations in the Early History of Dharma†åstra

167

student embarks upon his initiation (brahmacarya).30 What is the reason for this structural change? The model of the original å†rama system and the organization of the Gautama Dharmasütra, I think, provide some clues. As I have demonstrated elsewhere (Olivelle 1993), the original system of four å†ramas envisaged four permanent states of life chosen by a young adult upon the completion of his vedic studentship. Four å†ramas are open to that young adult: to remain a permanent student, to get married and become a householder, or to enter either of the two ascetic modes of life, forest hermit or wandering mendicant. It must be remembered that in the original system the preparatory period of study after vedic initiation is not viewed as an å†rama. Rather it was a kind of novitiate that prepares the young man to choose one of the permanent modes of adult life envisaged in the å†rama system. If the early composers of Dharma†åstric texts were following or were under the influence of the å†rama model in envisioning human life, then we can see their logic in beginning the treatises on dharma with initiation and the temporary studentship that opened the way to assuming one of the four adult modes of life. Indeed, this is precisely the structure that is found in the Gautama Dharmasütra. It begins with the sources of law followed by a discussion of initiation and the duties of a temporary student (Ch. 1-2). Chapter 3 is devoted to a discussion of the å†rama system and of three of the four å†ramas, permanent student, hermit, and mendicant. As Gautama rejects the validity of this system in the last sütra of Ch. 3, he is able to pass on to a discussion of marriage and the householder in Ch. 4, the only å†rama that he considers valid. As the examples of Gautama and Baudhåyana show, several authors of Dharmasütras were clearly antagonistic towards the å†rama system in general and ascetic and celibate lifestyles in particular. What the structural differences between the Gr¢hyasütras and the Dharmasütras show is that the creators of this new genre of texts were either partial to or influenced by the new and revolutionary system of the å†ramas, so much so that even those writers who did not agree with that system followed the structural model created on the basis of that system. Within the new structure of the Dharmasütras, the first topic of discussion is the sources or the means of knowing dharma. As I have already noted, it is clear that the historical sources from which the authors derived their rules and norms were the actual customs, practices, conventional rules, moral precepts, and the like prevalent in different communities. These sources are given different terms that 30 The exception is the Vasiß™ha Dharmasütra. This text, unfortunately, has had a problematic transmission and it is very unclear whether the structure of the extant text is original. The sequence of topics does not follow any logical plan and initiation and studentship are dealt with in the middle of chapter 11.

168

LANGUAGE, TEXTS AND SOCIETY

have similar and overlapping semantic ranges: åcåra (ÅpDh 1.1.1; VaDh 1.5), †œla (GDh 1.2), and ågama (BDh 1.1.4). The authoritative nature of the customary practices of a family, caste, village, or region is recognized by the authors. This recognition undercuts the theological argument that dharma is derived solely from the Veda. Nevertheless, the vedic exegetical tradition (mœmå∫så) influenced Dharma†åstric thinking from the very beginning. What we see in the early Dharmsütras is a sustained attempt to limit the authoritative source of dharma so that it would be in the hands of Brahmins. We see a very broad-minded and liberal attitude regarding who can teach dharma in the work of Åpastamba. He says that the knowledge found among women and ‡üdras is a valid source of dharma, and that, according to some, aspects of dharma not taught in Dharma†åstras can be learnt from women and people of all classes (ÅpDh 2.29.11, 15). Surely such knowledge cannot be viewed as deriving from the Vedas or any other scriptural source. This liberal attitude changed over time. For Gautama the only source of dharma is the Veda and the memory and conduct (smr¢ti and †œla) of those who know the Veda, thus firmly restricting its teaching to Brahmins. The major shift, however, occurs perhaps in the middle of the 2nd century BCE with the parallel introduction of two novel concepts: the notion of †iß™a and the definition of Åryåvarta as a sacred and authoritative region. With the notion of †iß™a we find the introduction of a restricted community of Brahmins who are viewed as both learned and virtuous; it is their conduct and memory that should be trusted in matters of dharma. Thus Baudhåyana (BDh 1.1.4) gives †iß™ågama (the conventions of †iß™as) and Vasiß™ha (VaDh 1.5) †iß™åcåra (conduct of †iß™as) as a third source of dharma, after the Veda and smr¢ti, which by this time had come to mean a textualized form of memory. We have a parallel development in the grammatical tradition, with Pata∞jali looking to the †iß™as as the source of correct Sanskrit. Now, the term †iß™a in its usage as a substantive to refer to certain kinds of Brahmins is not very old. It is not found with this meaning in the Vedic Sa∫hitås, the Bråhmañas, the early Upanißads, the ‡rautasütras, or the Gr¢hyasütras. It is absent in the vocabulary of Yåska’s Nirukta and Påñini’s Aß™ådhyåyœ. The earliest usage I can trace is in Kåtyåyana’s Vårttika (3 on ‡ivasütra 2: I: 20) and the Åpastamba Dharmasütra (2.24.3). It is remarkable that even though both Åpastamba and Gautama31 know and use the term, they do not connect it with the sources or pramåña of dharma. We see the connection between †iß™a and dharmapramåña for the first time in Baudhåyana (1.1.5-6) and Vasiß™ha (1.6). It is at about the same time that Pata∞jali uses the term with reference to individuals who have a natural ability to speak correct Sanskrit. Both in grammar and in dharma, then, †iß™as come to be 31

See GDh 19.3; 28.48, 50.

Explorations in the Early History of Dharma†åstra

169

viewed as individuals setting the standard and whom others should look up to if they want to learn correct Sanskrit and proper dharma. Pata∞jali’s (on Påñini 6.3.109; III: 174) definition of †iß™a includes both the place of their residence, the Åryåvarta, and their conduct (åcåra) and livelihood, a conduct that, according to him is found only in Åryåvarta: “(Brahmins) who possess just a jarful of grain, who are without greed, who have no detectable motive (for their behavior), and who have reached the highest expertise in some branch of learning within a short period of time” (alolupå agr¢hyamåñakårañåΔ ki∫cid antareña kasyå†cid vidyåyåΔ påragåΔ). Baudhåyana (1.1.5-6) gives a similar but longer definition: “Now, †iß™as are those who are free from envy and pride, who possess just a jarful of grain, who are without greed, and who are free from hypocrisy, arrogance, greed, folly, and anger.”32 We find a parallel appeal to the living tradition of authoritative Brahmins with reference to a sacred geography in both the dharma and the grammatical traditions of around the 2nd century BCE. For Pata∞jali the living tradition of correct speakers of Sanskrit exists only in the area he calls “Åryåvarta”.33 For Baudhåyana and Vasiß™ha, writing during or shortly after the time of Pata∞jali, Brahmins living in Åryåvarta provide the paradigm for correct dharma. The earlier authors, Åpastamba and Gautama, are ignorant of this geographical concept. Pata∞jali, Baudhåyana, and Vasiß™ha define this geographical area using what appears to have been a set formula; their descriptions agree almost verbatim: “The region to the east of where the Sarasvatœ disappears, west of Kålaka forest, south of the Himalayas, and north of Påriyåtra mountains is the land of the Åryas.” kaΔ punar åryåvartaΔ Ù pråg ådar†åt pratyak kålakavanåd dakßiñena himavantam uttareña påriyåtram. Pata∞jali on Påñini 2.4.10; 6.3.109. pråg ådar†åt pratyak kålakavanåd dakßiñena himavantam udak påriyåtram etad åryåvartam. BDh 1.2.9. pråg ådar†åt pratyak kålakavanåd udak påriyåtråd dakßiñena himavataΔ uttareña vindhyasya [...] etad åryåvartam ity åcakßate. VaDh 1.8-12.

What the socio-political impetus for developing the theory of Åryåvarta around the 2nd century BCE remains unclear. It is, however, a common imperative in any kind of demarcation, whether it is geographical as in this case or the formation of canons in the case of religious texts, that drawing boundaries is not merely an act of affirmation 32 †iß™åΔ khalu vigatamatsarå nirahaõkårå kumbhœdhånyå alolupå darbhadarpalobhamohakrodhavivarjitåΔ. Baudhåyana also cited this verse: “Cultured people are those who have studied the Veda together with its supplements in accordance with the Law, know how to draw inferences from them, and are able to adduce as proofs express vedic texts.” See also VaDh 6.43. 33 For an extensive discussion of this, see Deshpande 1993b, 1999.

170

LANGUAGE, TEXTS AND SOCIETY

(these are in) but also of exclusion (these are out). A canon is intended as much to exclude other texts from being considered religiously authoritative as to assert the authority of the texts included. In most cases canon formation is a political move to exclude groups that favored other texts with different doctrines and practices. Likewise, in demarcating the Åryåvarta as the religiously and linguistically authoritative region the authors must have been attempting to exclude from their authoritative community individuals and groups living in other regions.34 In this context, we can think of the foreign rulers in the north-western regions during the last few centuries BCE, or even the Buddhist expansion into those regions. Were there Brahmanical groups in those regions working for or under the influence of foreign rulers or of Buddhist doctrines? What about the expansion into the Deccan? Was the Åryåvarta intended to exercise the hegemony of Brahmins living broadly within the Ganga-Yamuna Doab? We have then two interlocking authoritative sources for dharma within the developing scheme presented in the Dharma†åstras, sources that are not textual but live individuals and communities. These individuals should first pass the test of †iß™a and second live in the sacred and authoritative region of Åryåvarta. The appeal to the live testimony of qualified Brahmins as a living source of dharma is already found in one of the oldest pieces of Dharma†åstra preserved in the Taittirœya Upanißad (1.11), which advises the young student who has just graduated: “should there be experienced, qualified, and gentle Brahmins devoted to dharma who are able to make a judgment in that matter, you should observe how they act in that regard and behave likewise.” Deshpande35 has shown how in the grammatical tradition the source of authority for correct Sanskrit shifted from the living community of †iß™as living in Åryåvarta during the time of Pata∞jali to the dead letter of grammatical texts during later times. Pata∞jali had direct access to spoken Sanskrit, whereas later grammarians focused on the grammatical texts to derive correct usage. Using the concepts of lakßya (linguistic usage) and lakßaña (grammatical rules), Deshpande shows the shift in the gaze of the grammarians from the first (found in Pata∞jali) to the second (by the time of Bhartr¢hari) as linguistic usage independent of grammar became extinct: “What is suggested is that the individuals described as being 34 On the connection between årya, åryåvarta, and the politics of exclusion, as well as the demarcation of sacred geographies by other religious traditions, such as Buddhism and Jainism, see Deshpande 1999 and Forthcoming-c. 35 See Deshpande 1993, forthcoming-a, and forthcoming-b. In a personal communication, Deshpande elaborates that even in the grammatical tradition, in spite of the extreme ideological stance found in authors such as Någe†a, post-Påñinian grammarians found ways to accommodate later linguistic changes into their grammars through a variety of strategies. In this, their work is similar to that of the later Dharma†åstric commentators, who had to both maintain the immutable authority of the early Dharma†åstras and incorporate the changing practices and mores into their presentations.

Explorations in the Early History of Dharma†åstra

171

lakßyaikacakßußka place the final authority in the usage that they know independently of the inherited grammatical system, and are for that reason deemed to be entitled to adjust the rules of grammar to fit to that usage. The second term, lakßañaikacakßußka, is taken to refer to those individuals who place the final authority in the inherited rules of grammar, and not in the known usage of the language, which it is felt is no longer independent of the received rules of grammar” (Deshpande, forthcoming-b, p. 7). A similar, though not as radical, a shift happened in the Dharma†åstric tradition especially with regard to the category of smr¢ti. The term is used by Gautama : tadvidå∫ ca smr¢ti†œle (“And the tradition/memory and practice of those who know it [the Veda]; GDh 1.2), where the reference is clearly to the memory and tradition of those who know the Veda, the kind of community later called †iß™a. In simple terms, one can go to a person from this community and ask “How do you perform the Upanayana?” And he would reply “This is how we do it.” That would be dharma, correct behavior pattern and ritual procedure, for other people. One could also watch their behavior (†œla), and imitate their example (see the passage of the Taittirœya Upanißad cited above). This living tradition encompassed by the term smr¢ti is gradually replaced by textualized versions containing rules of correct behavior and ritual practice. When Baudhåyana (BDh 1.1.3)says smårto dvitœyaΔ (“What is given in tradition is the second”), smårta most likely means norms found in textual sources, given that, unlike Gautama, he fails to state whose smr¢ti it is, in sharp contrast to his third source †iß™ågama (BDh 1.1.4), where the source is not just any ågama but that of the †iß™as. Likewise, the last of the Dharmasütrakåras, Vasiß™ha (VaDh 1.4), makes the simple statement †rutismr¢tivihito dharmaΔ (“Dharma is set forth in †ruti and smr¢ti”), thus making smr¢ti parallel to †ruti or the Veda, both textualized forms of religious norms. By the time of Manu smr¢ti had come to be equated with Dharma†åstra as such, so that Manu can simply declare: dharma†åstra∫ tu vai smr¢tiΔ (“Smr¢ti, however, is Dharma†åstra; MDh 2.10). Whereas the earlier authors see smr¢ti as external to the †åstra on dharma, indeed an external source from which the †åstra derives its norms, for Manu the †åstra itself is smr¢ti and thus the authoritative font of dharma. Even though Manu speaks of other sources (müla) of dharma, yet the two textual forms represented by Veda and Dharma†åstra are at the forefront. Recourse to other sources are legitimate only when these two fail to address an issue, as explicitly stated by Vasiß™ha (1.4-5): “The Law is set forth in †ruti and smr¢ti. When these do not address an issue, the practice of †iß™as becomes authoritative” (†rutismr¢tivihito dharmaΔ Ù tadalåbhe †iß™åcåraΔ pramåñam). Unlike Sanskrit grammar, however, Dharma†åstra could not simply dismiss the living tradition, because dharma as the rules of correct living, correct social interaction, and correct governance was a living and, therefore, changing reality. The reality of dharma was akin not

172

LANGUAGE, TEXTS AND SOCIETY

to the book-learnt Sanskrit but to the natural, living, and evolving languages of ancient India represented by the Pråkr¢ts. Dharma†åstras continued to maintain the authority of the dharma of families, castes, villages, regions, and groups for members of these communities. Kings and judges are advised to make careful note of the divergent norms of various groups before passing judgment. The broad nature of dharma even in later Dharma†åstric discourse is evidenced by the fact that royal edicts are also considered dharma. So, Manu (7.13) advises people: “When the king issues a dharma favorable to those he favors or unfavorable to those out of favor, therefore, no one should transgress that dharma.” Kåtyåyana (670) is even more explicit about the right of the king to promulgate dharma: “The man who does not follow the dharmas promulgated by the king should be censured and punished, as he has violated the king’s edict.” So, in spite of the textualization of the tradition and the theologizing of the sources of dharma under the influence of Mœmå∫så, Dharma†åstra to some degree remained rooted in the lived reality of communities. In this sense, the judgment of Lariviere and Wezler that dharma is nothing but a codification of custom remained true throughout the history of Dharma†åstra, even though this rootedness in living communities is often camouflaged by the theological rhetoric.36

Developments in ‡åstric Production ‡åstric production continued during the last three centuries of the last millennium BCE and the early centuries of the common era. How extensive this production was is impossible to estimate, because, as is usual in the case of ancient India, only a few of the Dharma†åstric texts produced during that period have survived. But the citations of other works in the extant literature and the vast body of Dharma†åstric material incorporated in the great epic Mahåbhårata clearly point to a vibrant scholarly tradition of textual production. My intention here is not to discuss in detail the history of that textual production but merely to point out certain central elements, both formal and substantive, within that history. The most significant formal change in the textual production is the gradual change from sütra-style prose to metrical composition principally in †loka verse. As I have shown elsewhere (Olivelle 2000, 67), the early texts, although written in prose, contain numerous verse citations given as proof texts in support of positions advocated in the prose exposition37. In the eyes of the authors, these verses appear to 36 For a more detailed and nuanced discussion of the relationship of †åstra to actual practice (prayoga) in a variety of intellectual domains, see Pollock 1989 and Olivelle 2004a, 62-66. 37 The only text not to have such verses is the Gautama Dharmasütra. This is probably due to the fact that the author has attempted to write a texts in an artificial manner completely in the sütra style, rather than to its early date. On this see Olivelle 2000: 8.

Explorations in the Early History of Dharma†åstra

173

have greater authority than their own composition. This is an old practice already recorded in the early Upanißads. The Br¢hadårañyaka Upanißad, for example, frequently cites supporting verses with the introduction tad eßa †loko bhavati (“In this connection there is this verse”).38 The significance of these verses in the eyes of the authors is indicated by the fact that they write commentaries on some of them, as in Br¢hadårañyaka 1.5. In the later Dharmasütras, however, we find increasing use of verses not simply as citations but integrated into the composition. This strategy is used with increasing frequency by Vasiß™ha, the author of the latest Dharmasütra; chapters 25-27, for example, are entirely in verse. The following table gives the percentages of verses (both cited and as part of the composition) in the three Dharmasütras: TABLE 1

Åpastamba Baudhåyana Vasiß™ha

Total sütras

Total verses

Non-quote verses

No. 1364 1236 1038

No. 30 279 288

No. 5 176 155

% 2 22 28

% 0.4 14 15

It appears that during the last few centuries prior to the common era †lokas had assumed an aura of authority, and proverbial wisdom was transmitted as memorable verses.39 The logical outcome of this tendency was for authoritative texts themselves to be composed in verse, lending authority to the text by its very literary genre. We see this already in some of the earliest Buddhist texts, such as the anthologies of the Suttanipåta and the Dhammapada and in the verses of the Jåtakas. The same process was probably responsible for the fact that the early prose Upanißads, such as the Br¢hadårañyaka and the Chåndogya, are followed by a series of Upanißads composed entirely in verse, such as the Ka™ha, the Muñ∂aka, and the ‡vetå†vatara. The parallel between the older and the later Upanißads is true of the dharma literature as well. Whereas the earlier texts are in prose with verse citations, the later ones are composed entirely in verse. The first such text was that of Manu. His use of verse for the composition of his Dharma†åstra, therefore, must have been part of a deliberate plan to lend the kind of authority to his text that would come only through this literary genre. We have, of course, the parallel examples of the epics Mahåbhårata and Råmåyaña composed in verse and claiming religious authority. This move away from prose to verse continues especially in religious compositions such as the Puråñas. In See 2.2.3; 4.3.11; 4.4.6, 7, 8. See also ChU, 3.11.2; 5.2.9; 5.10.9; 5.24.5; 7.26.2; 8.6.6. We find this also in Pata∞jali as exemplified in the verse cited above on the obligation of a younger person to rise when an older person approaches. 38

39

174

LANGUAGE, TEXTS AND SOCIETY

what could be regarded as expert traditions, however, the picture is mixed. The artha and kåma traditions continued to produce prose works, as did the ritual, philosophical, and grammatical traditions. The dharma tradition followed the trail blazed by Manu; all later Dharma†åstras are written in verse, prose entering the tradition only in commentaries and medieval digests (nibandha). At the substantive level, the greatest change in the content of the Dharma†åstras was the incorporation of matters relating to the king, the state, and the judiciary (an area I will call artha for the sake of convenience). These matters, we must assume, were the object of a separate expert tradition that also produced †åstras. Unfortunately, the early textual production of the artha tradition has been almost completely lost, with the single exception of the Artha†åstra ascribed to Kau™ilya.40 It is likely that the early exponents of Dharma†åstra already conceived of dharma broadly enough to include the running of society –the duties of the king and administrative officers, as well as the judiciary– as part of dharma. Indeed, as I have shown elsewhere (Olivelle 2004d and forthcoming-a), in the middle Vedic period the term dharma was most frequently associated with the duties of the king to maintain social order and justice. Yet, the treatment of artha in the early Dharmasütras is spotty at best, indicating that the authors were not particularly concerned about this area. In their treatment of artha, moreover, they focus principally on the duties of the king, such as the protection of the subjects, suppression of crime, and equitable taxation, rather than on the technical matters relating to the affairs of the courts of law and to judicial procedure. The earliest Dharmasütra, that of Åpastamba, devotes a total of 73 sütras to the duties of the king, under which he also includes crimes and punishment, topics that are later included within judicial procedure (vyavahåra).41 Åpastamba has little to say, however, about judicial procedure, devoting just six sütras: two dealing with the judges and four with witnesses. Even though Baudhåyana is considerably later than Åpastamba, he also has little to say on artha, devoting just 26 sütras to the king and 10 to witnesses. Baudhåyana’s text has undergone repeated emendation and corruption, so it is impossible to estimate what the original state of the text would have been. The texts of both Åpastamba and Baudhåyana, however, have come down as part of a larger complex, the Kalpasütra, containing both the 40 Indeed, it may well have been that the very incorporation of artha material in the Dharma†åstras and the ensuing demise of a separate expert tradition devoted solely to artha, may have contributed to the disappearance of artha texts. Even the extant Artha†åstra was discovered only in the 20th century and very few manuscripts of this work are in existence. 41 By contrast, Åpastamba devotes 119 sütras to the single and, from a broader societal point of view, narrower subject of a Brahmin’s annual course of Vedic study and 99 sütras to the times when such recitation has to be suspended.

Explorations in the Early History of Dharma†åstra

175

‡rauta- and the Gr¢yha-sütras of these schools. Their close connection to the respective Vedic †åkhås may have had something to do with their attitude towards matters of artha. The text of Gautama is the earliest to have been composed as an independent treatise. It is also the text that shows greatest connection to the artha tradition. Gautama devotes 52 sütras to the king and state and a further 109 to judicial procedure. As can be seen in the attached table, Gautama devotes a comparatively larger segment of his book to affairs of state, a tradition that will be continued in later Dharma†åstric literature. Although not as large as the section on artha of Gautama, Vasiß™ha also incorporates artha material far more extensively than his predecessors. He devotes 56 sütras to the king and state and 37 to judicial procedure. TABLE 2 On King Åpastamba Baudhåyana Vasiß™ha Gautama Manu

73 26 56 52 300

On Court 6 10 37 109 251

Total on artha 79 36 93 161 551

% of text 5.8 2.9 9 16.1 20.5

Gautama’s discussion of judicial procedure (vyavahåra) is not simply long; it also covers a larger number of topics: theft, abuse, assault, property damage, rates of interest, ownership, debts, and witnesses. I have dealt elsewhere (Olivelle, forthcoming-b) with the close connection between the Gautama Dharmasütra and the Månava Dharma†åstra. This connection extends also to their treatment of the king and judicial procedure. Manu, however, represents a watershed in the development of Dharma†åstra not only because of his exclusive use of verse and his placing the entire discourse in the mouth of the creator god himself, but also because he integrated to an extent not seen before material from the artha tradition bearing on king, state, and judicial procedure.42 The most obvious borrowing is Manu’s adoption of the 18 vyavahårapadas, or Grounds for Litigation (often called Titles of Law). Some of these grounds for bringing a lawsuit are given in the earlier literature, but never in such a systematic manner; neither are they called by the technical term vyavahårapada. According to Manu, a plaintiff when bringing a case before a tribunal has to indicate the exact vyavahårapada under which he is filing the suit. The significance of this topic for Manu is indicated by the fact that he devotes 251 verses or 9.3% of his entire book to it, and it covers two entire chapters (8-9) of the twelve-chapter work. 42 On Manu’s connection to the Artha†åstra of Kau™ilya, see Olivelle 2004c(below pp. 277-287).

176

LANGUAGE, TEXTS AND SOCIETY

Another telling example is the technical term såhasa. This term is used by Manu, the Artha†åstra, and all later authors of Dharma†åstras with reference to a set of three levels of fines imposed on most common crimes. The lowest is called pürvasåhasa (or prathamasåhasa), generally 250 Pañas; the middle (madhyamasåhasa) is 500 Pañas; and the highest (uttamasåhasa) is 1000 Pañas.43 This system was, in all likelihood, created within the artha tradition; both for †åstric purposes and for the actual administration of justice the experts in this tradition must have found it easy to settle on three levels of punishment for crimes. This meaning of såhasa, or any system of fines, is absent in the Dharmasütras.44 The incorporation of judicial procedure into the framework of Dharma†åstra culminates in Yåj∞avalkya, who divides his treatise into three sections: åcårakåñ∂a, vyavahårakåñ∂å, and pråya†cittakåñ∂a. Thus vyavahåra becomes a self-contained section of the literature on dharma, so much so that works dealing solely with vyavahåra (works that today we would call “monographs”) come to be composed toward the middle of the first millennium CE. The only extant work of this type is that of Nårada, although, judging by the extant citation from their works, the †åstras of Kåtyåyana and Br¢haspati also appear to have been devoted solely to vyavahåra.

Conclusion The missing ingredient in all the “histories” of Dharma†åstra is the historical, social, political, and economic contexts of the early textual production in this tradition. As Pollock (1985, 1990) has shown, Brahmanical writings, especially those, such as Dharma†åstras, produced under the influence of Mœmå∫så, deliberately exclude all references to the lived reality of their authors, to the social, religious, political, and economic conditions in which they lived and wrote. It is a mistake, however, for the modern scholar and historian to assume that such lack of reference means that these texts were produced in a social vacuum; the lack of reference should be recognized for what it is, a theological viewpoint or ploy to make their work seem eternal and transcending social and historical contexts. The task of the historian is to raise this theological veil to see under and behind it and to uncover the real historical conditions in which these texts were produced. Now, this is a difficult enterprise, given the notoriously difficult task of locating these texts in space and time and the incomplete picture we have of Indian history “Between the Empires”, between 43 See MDh 8.138. According to YDh 1.366, the highest fine is 1080, the middle half that, and the lowest half the middle. The Artha†åstra (3.17 from 8-10) gives a range from 48 to 96 for the lowest, 200 to 500 for the middle, and 500 to 1000 for the highest. 44 The term såhasa is found once in the ÅpDh (2.13.7) and once in the GDh (1.3) with the simple meaning of violence.

Explorations in the Early History of Dharma†åstra

177

the Mauryas and the Guptas. Yet, this is the task we as cultural historians must undertake in attempting to understand the rich tradition of Dharma†åstra. This essay is but a modest contribution to that enterprise. I think the influence of Buddhism on the Brahmanical textual production during this period has long been underestimated. Recently, the work of Hiltebeitel and Fitzgerald, both contributors to this volume, have thrown light on the historical underpinnings of the great epic, Mahåbhårata, seeing the epic as one answer to the challenges posed by the new religions and the A†okan reforms. A fresh look at the early history of Dharma†åstric texts will show similar historical underpinnings, as I have tried to show in my recent critical edition of the Månava Dharma†åstra.45

45

See Olivelle 2004a, pp. 18-25, 37-41.

178

LANGUAGE, TEXTS AND SOCIETY

A

179

VIII Structure and Composition of the Månava Dharma†åstra*

I have been engaged in preparing a critical edition of the Månava Dharma†åstra (MDh) for the past several years on the basis of over fifty manuscripts, nine commentaries, and citations in several medieval texts. As the editorial task nears its completion, I want to present before the scholarly audience some preliminary results of my close reading of and engagement with this text. These observations relate to the deep structure of the text, its authorship, and possible intervention of one or more redactors between the time of its composition and the period when we obtain manuscript and other evidence for the text. Scholars traditionally have regarded the composition of the MDh as a gradual process at the hands of anonymous and successive compilers, editors, and copyists lasting for several centuries, the same sort of agent-less process that many have thought lay behind the composition of the great epic Mahåbhårata. These compilers and editors, we are told, did nothing more than gather together proverbial sayings, moral maxims, and legal axioms that were floating in the mouths of people and handed down from generation to generation. The composition of the text is thus divorced from authorial intent and agency and from social, political, and economic context. The first to propose such a hypothesis was E. Washburn Hopkins:1 I draw the conclusion that the çåstram [MDh] was in great part collated between the time when the bulk of the epic [MBh] was composed and its final completion, that previous to its collation there had existed a vast number of sententious remarks, proverbial wisdom, rules of Originally published in the Journal of Indian Philosophy 30 (2002): 535-74. “On the Professed Quotations from Manu Found in the Mahåbhårata,” Journal of the American Oriental Society 11(1885), 268; cited by Lariviere 1989, II: x-xi. * 1

180

LANGUAGE, TEXTS AND SOCIETY

morality etc. which were ascribed, not to this treatise of Manu at all, but to the ancient hero Manu as a type of godly wisdom. These I conceive to have floated about in the mouths of the people, not brought together but all loosely quoted as laws or saying of Manu and these sayings were afterwards welded into one with the laws of a particular text [sect?] called the Månavas – a union natural enough, as the two bodies of law would then bear the same title, although the sect had no connection with Manu except in name [...]. According to my theory, these Manu-verses found in the Månava treatise were simply caught up and drawn from the hearsay of the whole Brahman world, keeping their form after incorporation with the Månavas’ text.

In the introduction to his influential translation of the MDh, Bhler agreed substantially with the Hopkins hypothesis. Bhler (1886a, xc) thought that the composers of both the MDh and the Mahåbhårata drew on a common stock of Spruchweisheit that at the hands of the teachers of specialized schools had spread to all legal topics. This general view regarding the creation of the MDh was accepted by Lariviere (1989, xi) and extended to other metrical smr¢tis, such as that of Nårada: “I doubt whether such texts as the Nåradasmr¢ti or the Manusmr¢ti were composed by a single individual.” I want to challenge this view regarding the composition of ancient texts in general and of the MDh in particular. This vision of composition in the case of the Mahåbhårata has recently been rejected, rightly I believe, by several leading epic scholars.2 That there were proverbs and legal maxims, principally composed in †loka verses, outside of texts is beyond doubt. Indeed, it is probably such verse maxims that are cited by the authors of Dharmasütras to support their judgments rendered in aphoristic prose, often with the introductory remark: athåpy udåharanti –“Now they also quote”. The term udåharanti probably means that these verses were recited by experts when questions about some point arose or when circumstances warranted.3 It would have been natural for authors of texts in almost any given field, but most especially those, such as the Dharma†åstras, dealing with morality and human relationships, to draw upon these maxims. Indeed, the example of the Dharmasütras indicates that they clearly did so. These verse maxims, however, are easily detectable in the Dharmasütras, because they are surrounded by the author’s own prose. In the metrical †åstras it is more difficult to separate the cited 2 See Hiltebeitel 2001; Madeleine Biardeau, cited by Hiltebeitel, p. 165 (“I prefer to suppose the creation of a sole Brahman of genius”); and James Fitzgerald, cited by Hiltebeitel, p. 25-26. 3 Indeed, as one of my students, David Brick, has pointed out, the term smr¢ti in its earliest usage may have referred precisely to such memorable maxims to which attention (smr¢ti) is drawn in particular circumstances and, of course, reside in the collective memory of the community at large or, in the case of law or grammar, in the memory of a community of experts. The citation of a maxim (nyåya), now mostly in prose, is also a feature of later medieval texts. See Appendix E of V. S. Apte’s The Practical Sanskrit-English Dictionary for a collection such maxims.

Structure and Composition of the Månava Dharma†åstra

181

maxims from the author’s own composition. Let me offer a couple of examples of such maxims in the MDh: “When an older person comes near, the life breaths of a younger person start to rise up, and by rising up and greeting him, he retrieves them” (2.120). This must have been a proverbial saying concerning respect for older persons; it is cited by Pata∞jali and given twice in the Mahåbhårata.4 The verse at MDh 4.57 is likewise clearly a proverbial saying about inappropriate actions: “He must not sleep alone in an abandoned house; awaken a sleeping superior; speak with a menstruating woman; or go to a sacrifice uninvited.”5 The authors clearly drew upon such maxims and, indeed, on previous scholarship in composing their texts. My point, however, is that the composition of these texts did not happen as an unconscious and gradual accumulation at different hands and at different times and places; these texts were authored by individuals with clear authorial intent. They gave their texts a particular structure; they argued for particular positions in law and morality; they disagreed with other experts, both their contemporaries and their predecessors; and they had particular social, economic, and political axes to grind. In all this they are not much different from modern authors. The unitary authorship of the MDh was proposed over a century ago by Bhler (1886a, xcii), who answered the objections of the proponents of a gradual textual evolution, objections based on such criteria as the contradictions in the extant text. My argument for the unitary authorship of the MDh is based primarily on the structure of the text, a structure that has thus far gone unnoticed perhaps because it was obscured by the chapter division to which the text was subjected probably through redactorial activities after its initial composition. I do not propose that the original text of the author, whom I will call “Manu” for convenience, has remained unaltered through the ages. Through a form of higher textual criticism, I will propose that certain sections are later additions. Indeed, when these additions are removed, the structure I have uncovered becomes more transparent. My argument, then, is that such a unique and symmetrical structure could not have been given to this text except by a conscious plan created by a single gifted individual. A deep structure that runs through the entire book –a structure that is not apparent at first glance and that remained undetected even by the commentators– could not have simply happened over time as the text was being put together by different individuals separated by centuries. If not an individual, then it must have been composed by a “strong chairman of a committee” with the help of research assistants who carried out his plan. 4

Pata∞jali’s Mahåbhåßya on Påñini 6.1.84 (Kielhorn, III: 58); MBh 5.38.1; 13.107.32. Especially in the case of Pata∞jali, it is more likely that he would cite a well-known saying to illustrate a grammatical rule than a verse from a specialized text. 5 For a historical analysis of this verse and its vedic precedents, see Jamison 2000.

182

LANGUAGE, TEXTS AND SOCIETY

I. The Structure The manuscript tradition of the MDh divides the text into 12 adhyåyas (lessons or chapters). This appears to be an old division; it is followed by all the commentators. I believe, however, that this division is not original. It was probably imposed on the text when it was subjected to a revision that added several sections (see IV below), most notably the table of contents given at the conclusion of the first chapter.6 Although several of the chapters follow the natural sequence of topics, a close reading of the text shows that they are artificial divisions. The chapters also contain different topics that the author, as I will demonstrate, intended to be separate: Ch. 2 contains the sources of dharma, rites of passage, and the duties of a student; the duties of a king is spread over chapters 7, 8, and 9; the single topic of judicial procedure and the grounds of litigation is spread over chapters 8 and 9; and Ch. 9 contains the final discussion of the king’s dharma and the dharma of Vai†yas and ‡üdras. More importantly, however, the division into chapters obscures the latent and deeper structure of the text, a structure that spans the entire corpus and must go back to the author himself. Manu uses the unique technique of “transitional verse” to mark the conclusion of one subject and the beginning of another. Here is an example (2.25): eßå dharmasya vo yoniΔ samåsena prakœrtitå Ù sa∫bhava† cåsya sarvasya varñadharmån nibodhata ÙÙ I have described to you above succinctly this source of the Law, as also the origin of this whole world. Learn now the Laws of the social classes.

This verse marks the transition from the two introductory topics, creation and the sources of dharma, to the main body of the text, the dharma of the four varñas. Such a technique is unique to Manu; it is not used in the Dharmasütras and sparingly, if at all, in the later Dharma†åstras. Note also the use of the verb nibodhata in most transitional verses; this manner of expression becomes a signature of Manu. This device was, I believe, an innovation conceived by Manu and provides an insight into the plan he had for his book. By following the trail of these transitional verses, we can uncover the overall plan and structure of the MDh. The chart below presents schematically the structure that emerges through this method together with the transitional verses at the beginning and/or end of topics that provide the clues to uncovering that structure: 6 Note the parallel imposition of chapters onto the Artha†åstra noted by Scharfe (1993).

Structure and Composition of the Månava Dharma†åstra

183

1. SARVASYA SA·BHAVA¸ [Creation of World] 1.1-119 2. DHARMASYA YONI¸ [Sources of Dharma] 2.1-24 eßå dharmasya vo yoniΔ samåsena prakœrtitå Ù sa∫bhava† cåsya sarvasya varñadharmån nibodhata ÙÙ 2.25 I have described to you above succinctly this source of the Law, as also the origin of this whole world. Learn now the Laws of the social classes. 3. CÅTURVAR˜YASYA DHARMA¸ [Dharma of the Four Varñas] 2.25—11.266 3.1. DHARMAVIDHI¸ [Rule of Dharma] 2.25—10.131 3.1.1. Anåpadi KarmavidhiΔ [Rules of Action in Normal Times] 2.26—9.336 3.1.1.1. Bråhmañasya CaturvidhaΔ DharmaΔ [Fourfold Dharma of a Brahmin] 2.26—6.97 eßa vo ‘bhihito dharmo bråhmañasya caturvidhaΔ Ù puñyo ‘kßayaphalaΔ pretya råj∞å∫ dharma∫ nibodhata ÙÙ 6.97 I have explained to you above the fourfold Law of Brahmins, a Law that is holy and brings imperishable rewards after death. Listen now to the Law of kings. 3.1.1.2. Råj∞aΔ KarmavidhiΔ [Rules of Action for King] 7.1—9.325 eßo ‘khilaΔ karmavidhir ukto råj∞aΔ sanåtanaΔ Ù ima∫ karmavidhi∫ vidyåt krama†o vai†ya†üdrayoΔ ÙÙ 9.325 I have described above in its entirety the eternal rules of action for the king. What follows, one should understand, are the rules of action of the Vai†ya and the ‡üdra in their proper order. 3.1.1.3. Vai†ya-‡üdrayoΔ KarmavidhiΔ [Rules of Action for Vai†yas & ‡üdras] 9.325-336 eßo ‘nåpadi varñånåm uktaΔ karmavidhiΔ †ubhaΔ Ù åpady api hi yas teßå∫ krama†as tan nibodhata ÙÙ 9.336 I have described above the splendid rules of action for the social classes outside times of adversity. Listen now to the rules for them in the proper order for times of adversity.

184

LANGUAGE, TEXTS AND SOCIETY

3.1.2. Åpadi KarmavidhiΔ [Rules of Action in Times of Adversity] 10.1-129 eßa dharmavidhiΔ kr¢tsna† cåturvarñyasya kœrtitaΔ Ù ataΔ para∫ pravakßyåmi pråya†cittavidhi∫ †ubham ÙÙ 10.131 I have described above the entire set of rules per taining to the Law of the four classes. Next, I will explain the splendid rules pertaining to penance. 3.2) PRÅYA‡CITTAVIDHI¸ [Rule of Penance] 11.1-265 cåturvarñyasya kr¢tsno ‘yam ukto dharmas tvayånagha Ù karmañå∫ phalanirvr¢tti∫ †a∫sa nas tattvataΔ paråm ÙÙ 12.1 “You have described this Law for the four classes in its entire ty, O Sinless One! Teach us accurately the ultimate consum mation of the fruits of actions.” 4. KARMAYOGASYA NIR˜AYA¸ [Determination of Karmayoga] 12.3-116 sa tån uvåca dharmåtmå maharßœn månavo bhr¢guΔ Ù asya sarvasya †r¢ñuta karmayogasya nirñayam ÙÙ 12.2 Br¢ghu, the son of Manu and the very embodiment of the Law, said to those great seers: “Listen to the determination regarding engagement in action.” 4.1. KARMA˜Å· PHALODAYA¸ [Fruits of Action] 12.3-81 eßa sarvaΔ samuddiß™aΔ karmañå∫ vaΔ phalodayaΔ Ù naiΔ†reyasa∫ karmavidhi∫ viprasyeda∫ nibodhata ÙÙ 12.82 I have declared to you above the fruits arising from actions. Listen now to these rules of action for a Brahmin that secures the supreme good. 4.2. NAI¸‡REYASA¸ KARMAVIDHI¸ [Rules of Action for Supreme Good] 12.83-115 etad vo ‘bhihita∫ sarva∫ niΔ†reyasakara∫ param Ù asmåd apracyuto vipraΔ pråpnoti paramå∫ gatim ÙÙ 12.116 I have explained to you above all the best means of securing the supreme good. A Brahmin who does not deviate from it obtains the highest state.

The structure that emerges from tracing the transitional verses consists of four major divisions of uneven length and importance: 1) Creation of the world; 2) Sources of dharma; 3) The dharma of the four social classes; and 4) Law of karma, rebirth, and final libera-

Structure and Composition of the Månava Dharma†åstra

185

tion. Obviously, the main section in terms of both length and importance is the third dealing with the four varñas. The other three are presented as a preamble, an introduction, and a concluding postscript. The preamble and the introduction are mentioned at the end of the second (2.25) in the transitional verse that also introduces the central third section on the four varñas. The latter section is mentioned also at its conclusion (12.1) in the transitional verse that also introduces the final section on karma. The central third section has two major subdivisions: the first is called dharmavidhi (rules relating to dharma), and the second, pråya†cittavidhi (rules relating to penances). These two sections, 3.1 and 3.2 in the above chart, are mentioned only once, at the conclusion of the first (10.131): “I have described above the entire set of rules pertaining to the Law of the four classes. Next, I will explain the splendid rules pertaining to penance.”7 The first subsection (3.1) of dharmavidhi is the longest in the entire book and is further subdivided into two: rules of action in normal times (anåpadi karmavidhi) and in times of adversity (åpadi karmavidhi). These two subdivisions, 3.1.1 and 3.1.2 in the above chart, are also introduced just once in the transitional verse at the conclusions of the first (9.336): “I have described above the splendid rules of action for the social classes outside times of adversity. Listen now to the rules for them in the proper order for times of adversity.” There is a fourth level of division in section 3.1.1 on rules for normal times. This section has three further divisions. The first, 3.1.1.1 in the chart, is called bråhmañasya caturvidhaΔ dharmaΔ (“The Fourfold Dharma of a Brahmin) and deals with the four å†ramas. This subsection is mentioned in the transitional verse at its conclusion (6.97), which also introduces the next subsection, 3.1.1.2 in the chart, dealing with the king: “I have explained to you above the fourfold Law of Brahmins [...]. Listen now to the Law of kings.” The third subdivision, 3.1.1.3 in the chart, deals with the remaining two varñas, the Vai†ya and the ‡üdra; it is introduced at the conclusion of the section on kings (9.325): “I have described above in its entirety the eternal rules of action for the king. What follows, one should understand, are the rules of action of the Vai†ya and the ‡üdra in their proper order.” The final postscript dealing with karma, rebirth, and liberation, which was introduced in 12.1, also has two subdivisions: the first (12.382) is on the fruits of actions (karmañå∫ phalodayaΔ) and the second (12.83-115) is on achieving the highest bliss (naiΔ†reyasaΔ karmavi7 Manu may have found a precedent for this division of the book in one of his primary sources, the Gautama Dharmasütra. Gautama (19.1) also begins his section on penance with the statement that he has completed his discussion of the varñas: “The Law pertaining to the social classes and the Law pertaining to the orders of life have been stated” (ukto varñadharma† cå†ramadharma† ca).

186

LANGUAGE, TEXTS AND SOCIETY

dhiΔ). These two are introduced in the transitional verse at the end of the first subsection (12.82): “I have declared to you above the fruits arising from actions. Listen now to these rules of action for a Brahmin that secures the supreme good.” An objection may perhaps be raised to my analysis, because these are not the only verses that introduce a topic. This is no doubt true, but all such verses simply signal the passage to a new topic within the broad structure I have outlined. In these verses, Manu does not say that he has finished one topic and is about to begin another; rather, with a few exceptions I will consider below, they simply indicate the new topic. Here is an example (2.89): ekåda†endriyåñy åhur yåni pürve manœßiñaΔ Ù tåni samyak pravakßyåmi yathåvad anupürva†aΔ ÙÙ I will explain precisely and in their proper order the eleven organs described by wise men of old.

This is part of a long list that uses the word pravakßyåmi to introduce a new topic.8 There are other verses using this term that both introduces a new topic and marks the end of the previous topic, in a manner similar to the transitional verses I have listed within the structure. In each of these cases, however, the topics are not broad themes but specific sub-themes within the structure I have identified. Verse 5.26 is an example: etad ukta∫ dvijåtœnå∫ bhakßyåbhakßyam a†eßataΔ Ù må∫sasyåtaΔ pravakßyåmi vidhi∫ bhakßañavarjane ÙÙ I have described above completely foods that are forbidden and permitted to the twice-born. I will now explain the rule on eating and on avoiding meat.

Here the author introduces the minor topic of meat-eating after his long disquisition on permitted and forbidden foods. Most such verses occur in the long section dealing with the eighteen grounds for litigation (vyavahårapada). At 8.214 the passage is from the non-delivery of gifts to the non-payment of wages; at 8.218, from the non-payment of wages to breach of contract; at 8.266, from boundary disputes to verbal assault; at 8.278, from verbal assault to physical assault; and at 8.301, from physical assault to theft. In the section on inheritance, at 9.56 there is a transition from the discussion of the relative importance of the seed (man) and the womb (woman) in procreation to the dharma of women in a time of adversity. At 11.99 there is a transition from penances for drinking liquor to those for stealing gold. 8 Variants of the verb include vakßyåmi and sa∫pravakßyåmi. These introductory verses are found at: 2.89; 3.22; 3.124, 169, 266; 5.57; 7.1, 36; 8.61, 119, 131, 229; 9.1; 10.25; 11.211; 12.30, 39.

Structure and Composition of the Månava Dharma†åstra

187

There is another group of verses that uses the verb nibodhata. This is the verb of choice in the transitional verses I have analyzed within the structure. Beyond those, however, Manu uses similar transitional verses to mark the passage from one minor topic to another. Thus at 2.68 the transition is from the rite of vedic initiation to the duties of an initiated student; at 5.100, from the purification following a death for those of the same ancestry (sapiñ∂a) to the purification for other individuals (asapiñ∂a); at 5.146 from purification to the dharma of women; at 6.86, from the discussion of ascetics (yati) to that of holy retirees (vedasa∫nyåsika); at 9.25, from the duties of husband and wife to a discussion of children; at 9.103, from the duties of husband and wife to partition of inheritance; at 9.148 from partition among children by wives of the same caste to that among children by wives of different castes; at 9.220 from partition of inheritance to gambling, the last ground for litigation; at 11.71 from the list of sins to the penances for their expiation; at 11.248 from penances for public sins to those for secret sins.9 The verb √†ru is used a few times in introductory verses: at 3.286 the transition is from the five sacrifices to the livelihood of Brahmins; at 5.110, from bodily purification to purification of articles; and at 11.180, from penances for sinners to penances for those who associate with them. Taken collectively, all these other uses of transitional verses merely indicate smaller subdivisions of the text. They uniformly refer only to the topics dealt with just before and just after the verse. With regard to such transitions Manu is not consistent in his use of verses; sometimes he uses them, but most often he does not. Such usages, however, does not impinge on the broad structure I have outlined above. Those verses stand out from the rest both because of their consistency and because they refer back not to the topic immediately preceding it but often to a broad theme introduced hundreds of verses before.

I.1 The dharma of a Brahmin The largest portion of the central section on the four varñas is devoted to the four-fold dharma of a Brahmin encompassing much of chapter 2 and all of chapters 3-6. This section is explicitly organized around the four å†ramas.10 The traditional material, however, could not be contained within the scheme of the four å†ramas, especially the sections on the childhood rites of passage, rules of a bath-graduate (snåtaka), and holy life styles falling outside the å†ramas of forest hermit and wandering mendicant. Manu, however, attempts to include these within his overall structure. 9 At the following places, nibodhata simply introduces a minor topic or is an invitation to the audience to be attentive: 1.68, 119; 2.1; 3.20, 183, 193; 9.31; 12.53. The verb √†ru is also used in similar contexts: 1.4, 60. 10 For the history of this institution, see Olivelle 1993.

188

LANGUAGE, TEXTS AND SOCIETY

Chapter 4 on the snåtaka is sandwiched between chapters 3 and 5 dealing with various aspects of an householder’s life. We see the difficulty Manu had with blending the snåtaka into the å†rama framework when we look at the beginnings of chapters 3 and 4. Chapter 3 begins quite naturally with the return home of a student who has completed his vedic studies. The author deals with the selection of a bride and marriage, with a long disquisition on the various kinds of marriage. Then at the beginning of Chapter 4 Manu has to repeat this within the context of the å†rama system: after dwelling at the teacher’s house during the first part of his life, a man should return home, get married, and lead a householder’s life during the second part. The fifth chapter is introduced with a question from the seers to Bhr¢gu about how a Brahmin could be subject to death. This opens the way to a discussion of permitted and forbidden foods and means of bodily purification. The theme of the four å†ramas is taken up again at the beginning of chapter 6: after living as a householder a man may become a hermit and live in a forest. The same is repeated at 6.33: after living the third part of his life as a forest hermit, he should become a wandering mendicant during the fourth and final period of his life. Even though this section is explicitly said to deal with the dharma of Brahmins, a close examination shows that Manu is here following a practice common in ritual texts. They describe fully the ritual procedure only for the archetypal rite of a group of related rites; the description of the other rites (ectypes) consists of pointing out only those ritual elements unique to each and different from the archetype. For Manu, the dharma of Brahmins constitutes the archetype, and he describes it fully. Mutatis mutandis these rules are applicable to all varñas. Indeed, within this section itself Manu often points out how the dharma is modified for other varñas. For example, under initiation he points out the different times for the different varñas, the different ways of manufacturing the girdle, different kinds of staffs, and the like (2.41-47). Likewise, he enumerates the kinds of marriages and the number of wives permitted for the different varñas (3.13). This principle of descriptive parsimony permits Manu to deal with the other varñas, especially Vai†yas and ‡üdras, briefly. Only the dharmas specific to them are discussed.

I.2 The Rules for a King The section devoted to the king, statecraft, and law in the MDh is disproportionately large compared to his predecessors in the expert tradition of dharma.11 The disproportion becomes even more striking 11 The MDh allots 971 verses (36%) of its 2680 verses to statecraft and law, only slightly smaller than the section devoted to the Brahmin, which consists of 1034 verses (38.6%). This stands in sharp contrast to the earlier works on dharma. The Åpastamba Dharmasütra, for example, devotes 83 (6 %) of its 1364 sütras to statecraft and law, and the Gautama Dharmasütra 115 (11.8%) of its 973 sütras.

Structure and Composition of the Månava Dharma†åstra

189

when we take into account the fact that this section deals with matters specific to the king and the kßatriya class, whereas the section on the Brahmin includes issues common to all varñas. The section on the king shows that Manu organized his material around a simple structure in three parts. The first part spanning 7.1-142 deals with the origin of the king, the organization of the state machinery including the appointment of officials, the construction of the fort, the king’s marriage, the conduct of foreign policy including war, and finally taxation. It appears that Manu’s narrative scheme here envisages a new king occupying a virgin territory. He is unmarried; he has to settle the land and build a capital; and he has to organize the state apparatus. This structure suited Manu’s purpose well, because it enabled him to discuss all the points associated with statecraft. Real life, however, is quite another matter; most kings would gain a kingdom either through inheritance or conquest. In either case there would be pre-existing cities, forts, and a state bureaucracy. In the second part, Manu changes his narrative scheme to span a single day, from the morning when the king awakens until nightfall when he goes to bed. Manu squeezes into this brief time period all the duties of a king. The morning routine extends from 7.145 to 7.215; the afternoon routine from 7.216 to 7.222; and the evening routine from 7.223 to 7.226. This part concludes with the king going “to bed at the proper time and rise up refreshed.” The third part deals with the justice system and comprises the 18 grounds for litigation (vyavahårapada, often translated “Titles of Law”). After dealing briefly with the organization of the court, Manu organizes his material on law and the dispensation of justice under the 18 titles. The issues relating to evidence and the interrogation of witnesses are dealt with not separately but under the first ground for litigation, the non-payment of debts. This appears to have been a convention borrowed from the artha tradition, to which Manu is indebted for the material relating to the king.12 Manu’s organization of the 18 vyavahårapadas is based on a few clear principles and, I believe, is superior to the structure give to them in any other text. Manu’s structure is significantly different from that of the extant Artha†åstra, as well as from the other two major Dharma†åstras, Nårada and Yåj∞avalkya. I give below a chart giving the organization of the vyavahårapadas in the four texts:13 12 I have dealt with this topic extensively, including word studies between the MDh and the Artha†åstra, in the introduction to my critical edition. 13 I give below a translation of the terms in Manu: non-payment of debt; deposits sale without ownership, partnerships; non-delivery of gifts; non-payment of wages; breach of contract; cancellation of a sale or purchase; disputes between owners and herdsmen; the Law on boundary disputes; verbal assault; physical assault; theft; violence; sexual crimes against women; Law concerning husband and wife; partition; and gambling and betting. The others have the same titles in different orders. The major category of the others not found in the MDh is prakœrñaka (miscellaneous).

190

LANGUAGE, TEXTS AND SOCIETY

Manu

Artha†åstra

1. r¢ñådåna 2. nikßepa 3. asvåmivikraya 4. sa∫bhüyasamutthåna 5. dattasyånapåkarma 6. vetanådåna 7. sa∫vidvyatikrama 8. krayavikrayånu†aya 9. svåmipålavivåda 10. sœmåvivåda 11. våkpårußya 12. dañ∂apårußya 13. steya 14. såhasa 15. strœsa∫grahaña 16. strœpu∫dharma 17. vibhåga 18. dyütasamåhvaya

14

*strœpu∫dharma dåyavibhåga sœmåvivåda samayasyånapåkarma r¢ñådåna upanidhi karmakarakalpa sa∫bhüyasamutthåna vikrœtakrœtånu†aya dattasyånapåkarma asvåmivikraya såhasa våkpårußya dañ∂apårußya dyütasamåhvaya prakœrñaka

Yåj∞avalkya

Nårada

r¢ñådåna upanidhi dåyavibhåga sœmåvivåda svåmipålavivåda asvåmivikraya dattåpradånika krœtånu†aya abhyupetyå†u†rüßå sa∫vidvyatikrama vetanådåna dyütasamåhvaya våkpårußya dañ∂apårußya såhasa vikrœyåsa∫pradåna sa∫bhüyasamutthåna steya strœsa∫grahaña prakœrñaka

r¢ñådåna nikßepa sa∫bhüyasamutthåna dattåpradånika abhyupetyå†u†rüßå vetanasyånapåkarma asvåmivikraya vikrœyåsa∫pradåna krœtånu†aya samayasyånapåkarma kßetrajavivåda strœpu∫sa∫yoga dåyabhåga såhasa våkpårußya dañ∂apårußya dyütasamåhvaya prakœrñaka

The three dharma†åstras have r¢ñådåna (“non-pament of debt”) as the first. This is only to be expected, because disputes regarding debts both personal and commercial must have been the most common reason for litigants to come before a court. It is also within the context of this first ground for litigation that the authors, including Manu, deal with judicial procedure, including rules of evidence and the examination of witnesses. Only the Artha†åstra departs from this practice; it begins the discussion with marriage and the partition of the paternal estate. The reason for this appears to be stated in the opening sütra: “All legal transactions begin with marriage” (vivåhapürvo vyavahåraΔ; A‡ 3.2.1). The other convergence in these lists is våkpårußya (verbal assault) and dañ∂apårußya (physical assault), which always go together, with såhasa (violence) coming very close. Other than these, the order of the lists diverge remarkably, making it clear that there was no traditionally fixed order for the vyavahårapadas. The order of enumeration in the MDh, therefore, was probably the creation of Manu himself, and we get a glimpse into his systematic way of thinking also in his arrangement of these topics. His arrangement, I think, is far superior and more systematic than any of the others and can be presented schematically: A. B. C. D. 14

Disputes between individuals and groups (= 1-10) Criminal law (= 11-15) Personal law (= 16-17) Public order and safety (= 18)

This term is not given in the A‡ but the topic is treated at the very outset.

Structure and Composition of the Månava Dharma†åstra

191

Manu begins with disputes between individuals and groups. Such disputes must have been the most common reason for litigation and cover the first ten grounds. The first nine for the most part deal with individual disputes, with the possible exception of the fourth on partnerships, where the dispute is between an individual and a partnership of which he is a member. Likewise, the seventh on breach of contract may happen between individuals and between an individual and a corporate body. The tenth, disputes over boundaries, can happen between individual land owners; but the typical dispute discussed by Manu concerns boundaries between villages. The next category is criminal law, involving verbal and physical assault, theft, robbery, and sexual crimes. It is interesting that såhasa, which I have translated “robbery”, is clearly viewed by Manu as belonging broadly to the area of theft; except that robbery involves violence in the act of theft. The other authors, including the Artha†åstra, appear to view såhasa as simply violence; they have a hard time, therefore, in distinguishing this from physical assault. The third category is personal or family law. The first ground for litigation under this rubric is disputes between a husband and a wife, although much of what is discussed is more general matters relating to laws and conventions governing marital relationships. The second and clearly the more significant is the partition of inheritance. It is in these two topics that there is often an overlap with material covered under marriage especially in chapter 5. This was probably inevitable when the dharma tradition incorporated strictly legal matters and, therefore, had to deal with marriage and family in two places, under proper conduct (åcåra) and law (vyavahåra). The final category is gambling and betting. One would have expected Manu to present rules for the orderly conduct of these practices, as is done in other texts.15 Manu, however, was strictly opposed to gambling and betting. For him these areas of social practice should be suppressed rather than regulated. It is, therefore, natural for him to follow his brief discussion of gambling with the important topic of the “eradication of thorns” (kañ™aka†odhana), that is, the suppression of criminal activities, especially theft, in the kingdom. This is a topic found in all artha and dharma texts, but it falls outside the grounds for litigation. Litigation, according to ancient Indian jurisprudence, is initiated by private individuals; the king and his officials are explicitly barred from initiating law suits. The eradication of thorns, on the other hand, is one of the principle duties of a king; it is a police activity and falls outside the judicial process. Nevertheless, Manu sees the eradication of thorns and the suppression of gambling as part of the same administrative process. The section on the duties of the king concludes with this pithy statement typical of Manu (9.324): “Conducting himself in this man15

See, for example the Åpastamba Dharmasütra, 2.25.12-14.

192

LANGUAGE, TEXTS AND SOCIETY

ner and always devoted to the Laws pertaining to kings, the king should direct all his servants to work for the good of his people.”

I.3 The Rules for Vai†yas and ‡üdras Manu’s discussion of Vai†yas and ‡üdras, the last two of the varñas, is extraordinarily brief. Six verses are devoted to the Vai†ya and just one to the ‡üdra. Even granting that, according to the ritual principle of parsimony I have discussed earlier, much of the material for these two classes was included in the discussion of the Brahmin, yet one would have expected something more than just seven verses. The reason for this brevity is unclear, but I think it must be understood within the context of the socio-political motives behind Manu’s composition, a topic beyond the scope of this paper. Simply put, Manu’s interest lay not in the lower classes of society, which he considered to be an ever-present threat to the dominance of the upper classes, but in the interaction between the political power and Brahmanical priestly interests, interests that were under constant threat ranging from the A†okan imperial polity to the foreign invasions toward the turn of the millennium.16

I.4 On Sin and Penance The methodical approach demonstrated in the sections on Brahmin and king is evident also in the chapter on sin and penance.17 Manu begins the topic with a discussion on whether penance can actually remove sins. After justifying the need for penance, he divides his inquiry into two sections: public sins (11.55-189), which occupies much of the discussion, and private or secret sins (11.227). Manu first presents the major classifications of sins: 1) the five grievous sins that causes the loss of caste (mahåpåtaka: 11.55-59); 2) a large group of secondary sins that also causes the loss of caste called upapåtaka: (11.60-67); and 3) four further classes of sins (11.68-71) causing a man a) to be excluded from caste (jåtibhra∫†akara), b) to become mixed caste (sa∫kœrñakara), c) to be unworthy of receiving gifts (apåtrœkaraña), and d) to be impure (malåvaha). He concludes the classification of sins with this transitional verse: “Listen now attentively to the specific penances by which all these sins individually enumerated above may be removed” (11.72) Manu then goes on to discuss the appropriate penances for each of the categories of sins: 1) the first four of the grievous sins (11.73108); 2) secondary sins (11.109-124); and 3) the four further classes of 16 I have dealt with the political and social environment of the MDh more extensively in my introduction to the critical edition. 17 There are, of course, some inconsistencies and extraneous material in this chapter. I will deal with them later in section IV.

Structure and Composition of the Månava Dharma†åstra

193

sins (11.125-26). Finally, he turns from sins personally committed to association with sinners who have become outcastes as a result of their sins, a category that form the fifth grievous sin (11.181-190). Manu introduces the last discussion with the transitional verse: “I have described above the expiation for all four kinds of sinners. Listen now to the following expiations for those who associate with outcastes” (11.180). The mention of the four kinds of sinners has caused some confusion. Grievous and secondary sins make two. The third category consists of four sins, but the penances for the four are dealt within two verses. I think Manu viewed the first (jåtibhra∫†akara) as one class and presented the penances for it in a single verse (11.125). He appears to have viewed the other three as forming a single class, dealing with their penances in a single verse (11.126). So, we have four categories of penances relating to the sins listed previously. The attempt to come up with four sinners as indicated in verse 11.180 may have led editors or scribes to insert the four offenses listed in 11.127-79 (see below section IV). The section on penance for publicly known sins concludes with two crisp statements. First: “No one should transact any business with uncleansed sinners; and under no circumstances should anyone abhor those who have been cleansed” (11.190). There follows an excursus containing miscellaneous items on sins and penances, which is clearly an interpolation. Manu concludes the section on penances for public sins with the transitional verse (248 omitted in my critical edition), which also introduces the section on penances for private sins: “Thus I have described to you above penances for sins in accordance with the rules. Listen next to the penances for secret sins.” When we take out the accretions in this chapter, the clear and impressive structure of the original composition emerges. That this concludes the central portion of the treatise dealing with the dharma of the four varñas, number 3 in the structure I have outlined above, is evident in the opening verse of the last chapter: “You have described this Law for the four classes in its entirety, O Sinless One! Teach us accurately the ultimate consummation of the fruits of actions” (12.1).

I.5 Karma Chapter 12 dealing with actions and their consequences, as well as with the attainment of ultimate happiness beyond the realm of rebirth, is quite different from the style and contents of the rest of the book. I am not willing to call this chapter a later addition; sufficient evidence does not exist to draw that conclusion. The entire chapter is taken up with the theme of action (karma), both the consequences of good and bad actions (karmavipåka) and the final triumph over action and the attainment of the ultimate goal beyond the process of rebirth. Broadly this discussion falls into two section, the one dealing with the fruits of action (3-81) and the other dealing with actions

194

LANGUAGE, TEXTS AND SOCIETY

leading to the supreme good (83-106). These two sections are divided by Manu’s signature transitional verse: eßa sarvaΔ samuddiß™aΔ karmañå∫ vaΔ phalodayaΔ Ù naiΔ†reyasakara∫ karma viprasyeda∫ nibodhata ÙÙ I have declared to you above the fruits arising from actions. Listen now to these rules of action for a Brahmin that secures the supreme good.

This chapter has also undergone redactorial intervention which I will address in section IV.

II. Narrative Structure and Composition Manu introduces two major innovations in comparison to the previous literature of the dharma tradition. First, he composed his text entirely in †loka verse. Second, he set his text within a narrative structure that consists of a dialogue between an exalted being in the role of teacher and others desiring to learn from him. Late vedic texts, especially the early prose Upanißads, regularly cite verses in support of statements and viewpoints.18 It appears that these verses were somehow viewed as having greater authority and, therefore, able to lend greater support to the author’s views, much like citations from scripture. The Br¢hadårañyaka Upanißad, for example, frequently cites supporting verses with the introduction tad eßa †loko bhavati (“In this connection there is this verse”).19 The significance of these verses in the eyes of the authors is indicated by the fact that they write commentaries on some of them, as in Br¢hadårañyaka 1.5. We see this practice continued by the authors of the Dharmasütras.20 They also present verses as providing support for or confirmation of views they have already presented in prose. They introduce these with athåpy udåharanti (“Now, they also quote”), indicating that these verses were well-known sayings that people generally cite.21 In the later Dharmasütras, however, we find increasing use of verses not simply as citations but integrated into the composition.22 This strategy is used 18 For a study of verses in the vedic corpus, see Paul Horsch, Die vedische Gåthå- und ‡loka-Literature. Bern: Francke Verlag, 1966. 19 See 2.2.3; 4.3.11; 4.4.6, 7, 8. See also Chåndogya Upanißad, 3.11.2; 5.2.9; 5.10.9; 5.24.5; 7.26.2; 8.6.6. 20 The exception is Gautama. For a discussion, see Olivelle 2000. 21 See Åpastamba Dharmasütra 1.19.15; 1.25.9; 1.31.23; 1.32.23; 2.9.13; 2.13.6; 2.17.7. At 2.23.3 two verses are cited with the introduction atha puråñe †lokåv udåharanti (“Now, they quote a couple of verses in a Puråña”) indicating that such verses may have been found in the genre Puråña, although this probably refers not to any extant Puråña. See Baudhåyana Dharmasütra 1.1.8; 1.2.11, 15, 17; 1.7.1; 1.8.23, 25, 53; 1.10.6, 23; 1.11.16, 14; 1.21.2; 2.1.6, 17, 21; 2.2.26; 2.3.14, 16, 19, 31, 45; 2.4.1, 10, 14, 18; 2.5.4, 7, 9; Vasiß™ha Dharmasütra 1.22, 38; 2.6, 27, 30, 31, 41, 48; etc. 22 For a discussion see Olivelle 2000, 6-7. For such verses, see BDh 1.10.26; 1.19.8; 2.6.32-42.

Structure and Composition of the Månava Dharma†åstra

195

with increasing frequency by Vasiß™ha, the authors of the latest Dharmasütra; chapters 25-27, for example, are completely in verse. It appears that during the last few centuries prior to the common era †lokas had assumed an aura of authority and that proverbial wisdom was transmitted as memorable verses. The logical outcome of this tendency was for authoritative texts themselves to be composed in verse, lending authority to the text by its very literary genre. We see this already in some of the earliest Buddhist texts, such as the anthologies of the Suttanipåta and the Dhammapada and in the verses of the Jåtakas. The same process was probably responsible for the fact that the later Upanißads are composed entirely in verse. The parallel between the older and the later Upanißads is true of the dharma literature as well. Whereas the earlier texts are in prose with verse citations, the later ones are composed entirely in verse. The first such text was that of Manu. His use of verse for the composition of his Dharma†åstra, therefore, must have been a deliberate plan to lend the kind of authority to his text that would come only through this literary genre. We have, of course, the parallel examples of the epics Mahåbhårata and Råmåyaña composed in verse and claiming religious authority. This move away from prose to verse continues especially in religious compositions such as the Puråñas. In what could be regarded as expert traditions, however, the picture is mixed. The artha and kåma traditions continued to produce prose works, as did the ritual, philosophical, medical and grammatical traditions. The dharma tradition followed the trail blazed by Manu; all later Dharma†åstras are written in verse, prose entering the tradition only in commentaries and medieval digests (nibandha). The second innovation in the composition of the MDh is its narrative structure. The Dharmasütras are not only written in prose, but they are also presented as nothing more than scholarly works. There is no literary introduction; the author gets right down to business. He presents his material in a straightforward manner, and on points of controversy and debate he presents opposing viewpoints. All this is eliminated by Manu. Here the real author is presented not as a scholar but as the primeval lawgiver, the Creator Svaya∫bhü, and his intermediaries, his son Manu and the latter’s disciple Bhr¢gu. The law is promulgated authoritatively; there cannot be any debate, dissension, or scholarly give and take. An anonymous group of seers approach Manu and asks him to teach them dharma. Manu accedes to their wishes. He narrates the creation of the world up to the emergence of human society hierarchically arranged into the four varñas. Then he asks his pupil Bhr¢gu to teach them the rest (1.59), reminding me of a busy professor letting his graduate assistant do the dirty work of teaching an undergraduate class. Bhr¢gu takes up the task in earnest; the rest of the book is the oral teaching of Bhr¢gu. The seers reappear only twice—once at the beginning of Chapter 5 when they ask how Brahmins can be subject to death, a ques-

196

LANGUAGE, TEXTS AND SOCIETY

tion that leads to a discussion of food practices, purification, and duties of women; and a second time at the beginning of Chapter 12 when they ask Bhr¢gu to teach them the effects of actions (karma). The narrative structure given prominence at the opening of the text fizzles out; there is no conclusion to the narrative. A similar structure is found in the Pa∞catantra, where the original setting –Vißñu†arman’s instruction of princes in statecraft and policy under the guise of animal stories– is lost sight of in the conclusion. We have no way of knowing all the reasons for Manu’s strategy of departing from the tradition of textual composition found in the earlier dharma tradition. The tradition of dialogue where a teacher instructs a pupil, a son, or a king goes back to the Bråhmañas and the Upanißads. The literary structure of these dialogues, however, places these individuals within human history. The transition into divine instruction is found already in the Chåndogya Upanißad where we have the instruction of Nårada by Sanatkumåra (7.1) and of Indra and Virocana by Prajåpati (8.7). Nevertheless, I think the example of the Buddhist texts was probably a factor. Hiltebeitel (2001, 167) has argued that the Mahåyåna literature, especially the Lotus Sütra, offers a parallel to the narrative structure of the Mahåbhårata. For the first time in India, the words of a single charismatic individual were taken as the sole fountain of authority in a religious tradition. The doctrine of buddhavacana, that the sole form of textual authority is the words of the Buddha, governed the production of texts both in the early forms of Buddhism and in the Mahåyåna.23 All texts begin with the preamble “Thus have I heard”, placing the text in the mouth of the Buddha and making the function of the “author” merely that of a transcriber or reteller of what he had heard. The narrator narrating what he had heard and placing his narrative in the distant past is also at the heart of the Mahåbhårata structure.24 Although its narrative structure is much simpler, the same is true of the MDh as well. We have here five layers of “telling”, “hearing”, and re-telling. At the most remote level, we have the creator himself soon after his creative activity composing a treatise and reciting it to his son Manu (1.58). Manu is the first “hearer”. He transmits it to Marœci and the other sages (1.58), who form the second tier of “hearers”. At Manu’s command, one of these sages, Bhr¢gu, teaches the seers who had come to Manu with the mission of learning dharma. Bhr¢gu’s first word (1.60), significantly, is “Listen” (†rüyatåm). This group of seers, still placed in illo tempore, constitutes the third tier of “hearers”. The narra23 For an interesting study on ancient instructions on how to “produce” a new Buddhist text, see Gregory Schopen, “If You Can’t Remember, How to Make it Up: Some Monastic Rules for Redacting Canonical Texts,” in BuddhavidyåsudhåkaraΔ: Studies in Honour of Heinz Bechert on the Occasion of his 65th Birthday, ed. P. Kieffer-Plz and Jens-Uwe Hartmann, Indica et Tibetica 30. Swisttal-Odendorf, 1997. 24 C. Minkowski, “Snakes, Sattras, and the Mahåbhårata,” in A. Sharma, ed., Essays on the Mahåbhårata (Leiden: Brill, 1991), 384-400; Hiltebeitel 2001.

Structure and Composition of the Månava Dharma†åstra

197

tor of the entire text makes only a fleeting and implicit appearance in the very first verse of the text: “As Manu was seated, absorbed in contemplation, the great seers came up to him, paid him homage in the proper manner, and said to him.” Here we have the voice of the narrator introducing the first group of characters; then he becomes silent. Evidently the narrator himself, who at one level can be identified with the historical author of the text, heard the text presumably from the seers; or he has been eavesdropping on Bhr¢gu’s instruction of the seers. This narrator is the fourth “hearer”. There is then the implied fifth “hearer”, that is, all those who listen to or read this text, including modern scholars. Although mediated by a series of tellers and hearers, the ultimate authority of the text lies in its original promulgator, the Creator himself. Paralleling the Buddhist doctrine of buddhavacana and doing one better than that, the MDh grounds its authority (pramåña) on the svayambhüvacana, the words of the Self-existent One, the very ground of creation. This appeal to a single source of authority stands in sharp contrast to the traditional source of authority for and means of knowing (pramåña) dharma, namely the Veda supplemented by traditional texts (smr¢ti) and the conduct of the virtuous (åcåra).25 Indeed, the MDh itself presents this doctrine when it discusses the sources of dharma in Chapter 2. There is thus a disjuncture between the narrative structure of Chapter 1 and the body of the text. The author is a traditional pandit, and his habitual methods of reasoning, argumentation, and public presentation take over in the substantive parts of the text.26 One may ignore the references to “that is the teaching of Manu” or “so said Manu”,27 which are peculiar in a text that was composed by Manu’s father and perhaps edited by Manu. The whole text, after all, constitutes the “sayings of Manu”. But such self-referential statements occur also in other Sanskrit texts. More revealing are the following. “Whatever Law Manu has proclaimed with respect to anyone, all that has been taught in the Veda, for it contains all knowledge” (2.7). This assertion is strange within a text that is Manu’s own composition. It appears that the author is trying here to reconcile the authority of Manu with the authority of the Veda as the source of dharma. “Tradition holds that the various groups of ancestors are the sons of all the seers headed by Marœci, seers who are the children of Manu, the son of Hirañyagarbha” (3.194): why ascribe to tradition (smr¢tåΔ) a view when the creator himself is the speaker. “Because of discipline, on the other hand, Pr¢thu, as well as Manu, obtained a kingdom; Kubera, lordship over wealth; and the son of Gådhi, the rank of a Brahmin (7.42). 25 This contrasts even more with the “community standards” (såmayåcårika) espoused in the Åpastamba Dharmasütra (1.1.1-2). 26 We find similar forgetfulness in other texts as well. See, for example, Br¢haspati’s reference to the contradictions in the MDh cited below in section III. 27 iti manor anu†åsanam: 8.139, 279; 9.239; abravœn manuΔ, manur åha: 3.150, 222; 4.103; 5.41, 131; 6.54; 8.124, 168, 204, 242, 292, 339; 9.158, 182; 10.63, 78.

198

LANGUAGE, TEXTS AND SOCIETY

Here Manu, along with other traditional heroes, are treated in the third person. “Bed, seat, ornaments, lust, hatred, behavior unworthy of an Årya, malice, and bad conduct –Manu assigned these to women” (9.17). Again, Manu appears here in the third person. The author of the MDh does not, indeed cannot, openly present diverse opinions of scholars both due to metrical reasons, as Bhler (1886a, xciii) has already pointed out, and because it would violate his narrative structure; how can the creator present diverse opinions on points of law? Yet, the pandit mentality is hard to suppress, and our author repeatedly forgets his narrative and engages in ordinary scholarly give and take. The most obvious is the expression iti cet (“if you argue thus”) which introduces an objection or a doubt occurring at 9.122; 10.66, 82; 12.108. At 8.140 we have a particular interest rate set by Vasiß™ha, and at 8.110 an appeal to the practice of former sages and gods with reference to the legitimacy of oaths. The author sometimes refers to the opinions of others with the common kecit. At 3.53 the opinion of some that a bull and a cow are given as a bride price at a seer’s type of marriage is refuted; at 3.261 he refers to different customs regarding the disposal of ancestral offerings; and at 9.32 he introduces two opinions about the person to whom a son belongs, the biological father or the husband of the mother.

III. Contradictions The text of the MDh as it has come down to us contains numerous contradictory statement. This feature has drawn the attention of not only modern scholars but also ancient commentators. The author of the Br¢haspati Smr¢ti, a text written a few centuries after the MDh, had the same problem with Manu. In his section on niyoga (levirate), Br¢haspati comments (1.25.16): uktvå niyogo manunå nißiddhaΔ svayam eva tu Ù yugahråsåd a†akyo ‘ya∫ kartu∫ sarvair vidhånataΔ ÙÙ Manu has prescribed the leviratic union, and then he himself has forbidden it. Because of the shortening of each age, no one can carry it out in accordance with the prescriptions.28

Br¢haspati provides a traditional solution to the problem; the contradictory rules apply to different ages. So there is no true contradiction. Modern scholars cannot accept the traditional hermeneutical solution. Many have seen these contradictions as proof that the text had multiple authors over a long period of time; the contradictions 28 The “shortening” has multiple meanings: the ages themselves become shorter in duration, the life span of humans become correspondingly shorter, and their proclivity to virtue also becomes weaker (cf. MDh 1.83-84).

Structure and Composition of the Månava Dharma†åstra

199

represent changing customs and norms. Bhler (1886a, xcii-xciii), the first scholar to posit the unitary authorship of the MDh, sought to answer these critics. His answer was two-fold: Thus in weighing the value of the argument drawn from the occurrence of contradictory passages, two circumstances, which mostly have been left out of account, must be kept in mind: first, that it is a common habit of Indian authors to place conflicting opinions, supported by authorities of equal weight, side by side, and to allow an option, or to mention time-honoured rules, legal customs, and social institutions, and afterwards to disapprove of them; and secondly, that, as our Smr¢ti is in any case a recast of an earlier Sütra, that fact alone is sufficient to account for contradictions.

The second point is based on Bhler’s assumption of a pre-existing Månava Dharmasütra and must be ignored. In my view, the author’s work cannot be limited to that of an editor; even though he used older sources, as all authors do, he created a new work, and one would have to assume that a good author would seek to avoid contradictions, especially contradictions that sit side by side. Bhler’s first point, however, is well taken. As I have pointed out above, Manu was unable to demarcate various views with the traditional iti followed by the name of the authority both because the treatise was composed by the creator and because it was difficult to incorporate such attributions into a verse composition. The various views are here woven into the very fabric of Manu’s narrative. The fact the Manu does cite conflicting opinions is demonstrated by the fact that sometimes he does ascribe conflicting views to different factions. So, for example, the conflicting opinions about the relative superiority of the seed and the field in determining the person to whom a son belongs are stated clearly (9.32-44). Although stated less clearly, two views on primogeniture are presented at MDh 9.105-110 and 9.111f. Here the presence of the particle vå (“or”, “or rather”), indicating an alternative that the author himself prefers, at the beginning of 9.111 shows that we are dealing with two opinions. The clearest attribution of opinions to authorities is found at 3.16: According to Atri and the son of Utathya, a man falls from his caste by marrying a ‡üdra woman; according to ‡aunaka, by fathering a son through her; and according to Bhr¢gu, by producing all his offspring through her.

A clear example of an apparent contradiction where two viewpoints are juxtaposed is found at 9.97-100: 97

If, after the bride-price has been paid for the girl, the man who paid the price dies, she should be given to the brother-in-law, if she consents to it.

200

LANGUAGE, TEXTS AND SOCIETY

98

Even a ‡üdra should not take a bride-price when he gives his daughter; for by accepting a bride-price, he is engaging in a covert sale of his daughter. 99 That after promising her to one man, she is then given to another— such a deed was never done by good people of ancient or recent times. 100 The covert sale of a daughter for a payment under the name “brideprice”– we have never heard of such a thing even in former generations.

We see here that verse 97 assumes the payment of a bride-price, a view rejected in the next verse. In verse 99, likewise, the bride-price is assumed; the father takes money from one man and then gives the girl to another, thus getting two payments. The next verse again inveighs against the practice of bride-price. One cannot but assume that Manu is here deliberately pairing the two views.29 Contradictory positions on marrying a ‡üdra woman are, likewise, place side by side at 3.13-14: 13

A ‡üdra may take only a ‡üdra woman as wife; a Vai†ya, the latter and a woman of his own class; a Kßatriya, the latter two and a woman of his own class; and a Brahmin, the latter three and a woman of his own class. 14 Not a single story alludes to a Brahmin or a Kßatriya taking a ‡üdra wife even when they are going through a time of adversity.

Here, we may have not just a juxtaposition of views but Manu actually siding with the latter view. From the verses that follow, it becomes clear that Manu did not approve a Brahmin marrying a ‡üdra, a view that is backed by ancient authorities who permit hypergamous marriage to all classes. We may be able here and in other instances to distinguish two voices of Manu, the legal and the moral, which may also have contributed to apparent contradictions. As a lawyer he has to deal with the reality of social life and accepted custom; he has to decide whether certain actions are legal or not. It is apparent that he accepted upper classes marrying lower class women, even ‡üdra women, as a social fact. At 9.149-157, for example, in dealing with the practical matter of dividing the paternal estate, Manu states what share the son born of a ‡üdra wife should receive. Nevertheless, dharma is not just law but also right living, and when he uses the moral voice Manu often contradicts what he may have said within a legal context. Even today, for example, opposition to birth out of wedlock (the moral position) can exist side by side with legal and social provisions for the welfare and legal status of illegitimate children. We see such a bifurcation in Manu’s discussion of permitted foods (5.4-26) where meat eating is taken for granted. His discussion of food at an ancestral offering (3.267-72) also presupposes meat. His discussion of permit29 We see a similar juxtaposition of views at 9.122-26. The first view permits seniority among sons born to wives of equal status, whereas the second view denies this and treats all of them equally.

Structure and Composition of the Månava Dharma†åstra

201

ted foods, however, is followed by a long statement about the evils of meat (5.27-55). Clearly, Manu here is not so much contradicting himself as providing two sets of guidelines, the one traditional that includes meat, within which he provides rules on permitted and prohibited animals, and the other reflecting the new morality of vegetarianism. The moral tone is clearly brought out at the end of Manu’s long discussion on food (5.56): It is not a sin to eat meat, to drink liquor, or to have sex; that is the natural activity of creatures. Abstaining from such activity, however, brings great rewards.

The most famous instance of contradiction is Manu’s position on leviratic union, that is, the fathering of a son by a brother upon his dead brother’s widow. This had been a long-established practice in India recorded in all the Dharmasütras. Manu himself presupposes this practice, for example, in his discussion of partition (9.143-47). At 9.57-63, moreover, he not only admits this practice but also lays down rules as to how it should be carried out. In the very next breath, however, he opposes this practice vehemently (9.64-68). This contradiction does not lend itself to an easy resolution, but I think here also Manu is talking in two voices. He acknowledges the traditional practice, provides rules for its proper implementation, and deals with the legal rights of children born from such a union. But he does not like this custom and calls it a bestial practice. He also preaches chastity to young widows; they do not need children to attain happiness here or in the hereafter (5.158-60): Aspiring to that unsurpassed Law of women devoted to a single husband, she should remain patient, controlled, and celibate until her death. Untold thousands of Brahmins who have remained celibate from their youth have gone to heaven without producing offspring to continue their family line. Just like these celibates, a good woman, though she be sonless, will go to heaven when she steadfastly adheres to the celibate life after her husband’s death.

One other reason for apparent contradictions is the practice of using the best argument to win a debate. We see Kr¢ßña using this sort of argumentation in the Bhagavad Gœtå; you should do this for reason X, and if not for that then for reason Y, even if X and Y and incompatible. Thus, with reference to the question “To whom does a son belong?”, Manu clearly sides with those who argue that a son belongs to the husband of the mother, just as the crop belongs not to the owner of the seed but the owner of the field (9.41-55). Manu, however, dislikes obtaining children through surrogate fathers, as permitted within the doctrine of the twelve kind of sons (9.158-60), as demonstrated by his denunciation of leviratic union. All except the natural-born son, that is, one’s own biological son, are viewed by

202

LANGUAGE, TEXTS AND SOCIETY

Manu as surrogates. Within this context, the argument that a son belongs to (or takes after) the biological father, that the seed is stronger than the field because different plants and trees grow in the same field from different seeds, is a convenient argument, and Manu uses this at 9.181: “Those born from another man’s seed who have been designated here as sons belong only to the one from whose seed they were born and to no one else.” A final reason for apparent contradictions may be found in the frequent use of hyperbole in didactic literature. Early scholars of the Vedas encountered what they perceived as a problem: the vedic hymns take different gods to be the highest. This went against their monotheistic presuppositions, as also against common logic. How could several gods in a pantheon be supreme at the same time. Max Mller coined the word “henotheism” to account for this phenomenon. A quick glance at the royal panegyrics (pra†astis) of a millennium later is instructive. All kings, from petty chieftains to Gupta emperors, are eulogized as conquerors of the whole world. One would hardly expect the panegyrist to be accurate –he is a petty king, a tributary to king X, who managed a couple of small-scale victories! That would hardly do. Neither can you go to god X and say “You are in the third rank below Y and Z. And by the way, will you give me a thousand cows?” Even in the Catholic faith, where theological orthodoxy takes center stage, devotees of the Virgin or St. Jude do not give theologically accurate descriptions of the powers of their favorite saint; hyperbole is the norm. The vedic poets were no exception. This form of rhetoric was recognized as an alaõkåra, a poetic ornament, in Sanskrit aesthetics under name ati†ayokti, the hyperbole. As Gerow observes, a hyperbole contains a falsity but is so framed as to open the door to a deeper truth: A “skyscraper is so tall they had to put hinges on the two top stories so to let the moon go by” [Carl Sandburg]. But of course it is not a falsity that is capable of being detected in truth tables; for the falsity, the ‘exaggeration’ is only apparent, only for effect. For the lie in the hyperbole is so framed as to conceal a greater truth, and a truth urgently required by the context; as these skyscrapers are the tallest things that man ever built.30

The use of hyperbole in religious and didactic literature is clearly different from its use in poetry. There is, however, one thing in common; neither can be taken at face value or read literally without distorting the language and producing interpretive monstrosities like “henotheism”. In religious literature we have statements that giving a cow to a learned Brahmin is worth one thousand horse sacrifices, or bathing in a particular river is better than bathing ten thousand times in the Ganges at Kå†i. In attempting to inculcate the virtue of abstention from meat, Manu (5.53) says: “A man who 30

E. Gerow, Indian Poetics (Wiesbaden: Harrassowitz, 1977), p. 242.

Structure and Composition of the Månava Dharma†åstra

203

abstains from meat and a man who offers the horse sacrifice every year for a hundred years–the reward for their meritorious acts is the same.” An ascetic text wishes to promote the importance of giving food to ascetics: “Even if a man gives the entire earth, it would not equal the merit of preparing almsfood and giving it to a mendicant.”31 No one expects such statements to be taken literally; they are meant to eulogize certain lesser known practices and to recommend their performance. Likewise, in didactic and legal literature a sin or crime or a virtue or good act is said to be the worst or the best. “Lack of generosity is the gravest sin; generosity is the highest austerity,” says the Baudhåyana Dharmasütra (1.10.6). And Manu (4.224-25): The gods once evaluated the food of a miserly vedic scholar and that of a generous usurer and pronounced the two to be equal. Prajåpati came up to them and said, “Don’t make equal what is unequal. The food of the generous man is cleansed by the spirit of generosity, whereas the other food is defiled by the lack of generosity.”

When the topic is generosity, even a usurer is said to be better than a miser. But when the topic is usury, it is a sin worse than even abortion: “Usury and abortion were once weighed in a balance. The abortionist rose to the top, while the usurer trembled” (BDh 1.10.23). It is impossible to think that these authors intended their statements to be taken as literally true, just as it is not possible to think that there is a device in the skyscrapers to let the moon pass. As in literature and poetry, so in religious, didactic, and legal literature hyperbole is simply a literary device. Failure to recognize this can only cause serious misinterpretation of texts. So, it is not a contradiction when Manu (9.14-16), in warning husbands to guard their wives, waxes eloquent on the evil tendencies inherent in women: They pay no attention to beauty, they pay no heed to age; whether he is good looking or ugly, they make love to him with the single thought, “He’s a man!” Because of the lechery, fickleness of mind, and hardheartedness that are innate in them, even when they are carefully guarded in this world, they become hostile towards their husbands. Recognizing thus the nature produced in them at creation by Prajåpati, a man should make the utmost effort at guarding them.

and in urging men to respect women, he eulogizes them (9.26-28): On account of children, the wife is the bearer of many blessings, worthy of honor, and the light within a home; indeed, in a home there is no distinction at all between the wife (strœ) and ‡rœ, the Goddess of Fortune. She begets children, and once they are born, she brings them up—in daily domestic affairs, the wife is obviously the linchpin. 31 Cited in Yådava Prakå†a, Yatidharmasamuccaya, ed. P. Olivelle (Albany, NY: State University of New York Press, 1995), 6.314.

204

LANGUAGE, TEXTS AND SOCIETY

Offspring, rites prescribed by Law, obedient service, the highest sensuous delights, and procuring heaven for oneself and one’s forefathers—all this depends on the wife.

and warns against abusing them (3.56-58): Gods take delight in a family where the womenfolk are revered; but where they are not, no rite bears any fruit. Where the womenfolk are sad, that family soon comes to ruin; but where they are not sad, that family always prospers. When womenfolk, not receiving due reverence, curse any house, it comes to total ruin, as if struck down by witchcraft.

I do not propose that all apparent contradictions in the MDh can be resolved in these ways. Authors, even modern ones, do contradict themselves occasionally. Further, the work of redactors down the centuries, to which I now turn, cannot be completely detected. Changing norms and mores may have prompted some of them to introduce opinions at variance with those expressed in the original text.

IV. The Work of Redactors After it leaves the hand of the author, every text assumes an independent life; it continues its life as it is read, studied, interpreted, commented on, and copied by succeeding generations of readers, scholars, and scribes. It is this after-life of a text that a critical edition in its critical apparatus uncovers and presents to the reader. This aspect of a critical edition is as important as its better known feature of attempting to reconstruct the text as composed by the author. Both these aspects –the original text and the after-life– laid out in the text and the critical apparatus of the edition presuppose that changes are introduced into the author’s text by those responsible for its after-life. Some of these changes are inadvertent, such as scribal errors and misreadings; others are deliberate, such as the different but equally cogent and intelligible readings found in different recensions and the additional verses found in numerous manuscripts. These changes introduced into the text and detectable through the examination of the extant manuscripts and commentaries can be identified and moved to the critical apparatus, thus restoring the text. The manuscript and other evidence we possess, however, often do not cover the entire period from today to the time of the author. In the case of the MDh, we pick up the textual history midstream, at least several centuries after its composition. If the later tradition of readers and copyists introduced changes, it is fair to assume that earlier generations did so too. Any such changes that were taken over by the extant manuscript tradition cannot be identified by the normal methods employed in the critical constitution of texts. These methods are thus called “lower criticism”; they are dependent on extant manuscripts, citations, commentaries,

Structure and Composition of the Månava Dharma†åstra

205

and the like. The text constituted by these methods can only be an approximation to the original text as written by the author. I agree with Lariviere’s (1989, xii) hypothesis that the Dharma†åstras continued to expand with addition of new materials “until a commentary on the collection was composed. A commentary would have served to fix the text, and the expansion of the text would have been more difficult after that.” Because I consider the MDh to have a single author, I take these emendations as produced by redactors working on the original text. Such activity ceased for the most part after the text was “fixed” by early commentators such as Bhåruci and Medhåtithi; changes after that period were limited to the addition of individual verses and minor changes in the wording of verses. Methods of identifying changes undetectable by manuscript evidence fall into the category of “higher criticism” and are not as surefooted as methods employed in “lower criticism”. There some amount of conjecture is inevitable. In Biblical studies higher criticism has been used profitably especially with regard to the Pentateuch, but there the criticism was used not to construct the original text of the author but to uncover the sources he used in constructing his text. Hence, it has been called source criticism. Bhler’s (1886a, lxv-lxxiv) attempt to separate the ancient portions from later additions within the MDh was an exercise in source criticism. His aim was to distinguish the passages going back to his hypothetical Månava Dharmasütra from those added by the editor of the versified version. His criteria were somewhat crude. He accepted sections that had parallels in other Dharmasütras as part of the ancient sütra, and took sections without such parallels to be the work of the editor. My aim is different. Mine is not a source criticism but a criticism aimed at identifying possible additions and interpolations into the text composed by Manu between the time of its composition and the earliest manuscript and other evidence that we possess. The criterion I use is based on Manu’s overall plan for and organization of his work, which I have outlined in detail. Manu is careful to let his reader know that he has completed one subject and is moving onto another through “transitional verses”. The structure of his work shows that Manu had a very methodical and systematic mind. It is extremely unlikely that he would have introduced extraneous material right in the middle of his carefully crafted plan thus vitiating the work’s organization. In the following examination of the entire text, I will follow Manu’s organizational scheme and identify sections that fall outside that scheme. Most, if not all, extraneous material are found at the interstices of his plan, at the junctures between two topics. Clearly, it was easier to introduce new material at these fault lines than within the discussion of a given topic. I have identified these passages as “Excursus”; even if someone does not accept my contention that these are interpolations, they must at best be viewed as parenthetical statements. Even

206

LANGUAGE, TEXTS AND SOCIETY

though I think that the chapter division is a later innovation, I will follow the chapter sequence for easy reference to the text.

Chapter One Why would a treatise on dharma begin with the story of creation? Bhler (1886a, lxvi) in fact remarks: “The whole first chapter must be considered as a later addition. No Dharma-sütra begins with a description of its own origin, much less with an account of creation.” Long before Bhler, the 9th century commentator Medåtithi expressed a similar concern: Where did we start? And where have we ended? He [Manu] was asked the dharmas prescribed in the †åstras, and he indeed promised to explain them. To then describe the world in its unmanifest state is both irrelevant and serves no human purpose. This truly exemplifies the common saying: “Asked about the mango trees, he talked about the Kovidåra trees.” With regard to this matter, there is neither an authoritative basis nor does it serve any purpose. Therefore, this entire chapter should not be studied.32

There is, however, a cogent defense of Manu’s introductory statements found in the first chapter. I have already dealt with why Manu begins with the origin of the text thereby investing it with supreme authority. Why he gives an account of creation is less obvious. I think the clue is found in the manner the first account of creation ends (MDh 1.31): “For the growth of these worlds, however, he produced from his mouth, arms, thighs, and feet, the Brahmin, the Kßatriya, the Vai†ya, and the ‡üdra,” evoking the conclusion of the creation story in the Purußa Hymn (R¢gveda 10.90). The seers questioned Manu not simply about dharma, but about the dharmas of all social classes (1.2: sarvavarñånå∫ dharmån). I think the creation account is intended to show how the four varñas came into being; they are not contingent and temporary social phenomena but part of the very fabric of creation. The clarity of Manu’s presentation is obscured at this point by three “excursus” containing a quite superfluous second account of creation (1.32-41), a classification of fauna and flora (1.42-50), and an account of cosmic cycles (1.51-57). I believe these are interpolations introduced at the interstice between creation of the varñas and the composition of the †åstra (1.58). If we eliminate the intervening 26 verses, we see an elegant transition from the creation account culminating in the production of the four varñas to the composition of the treatise by the Creator and its transmission to Manu, culminating in 32 Medhåtithi on MDh 1.5: kva asthåΔ kva nipatitåΔ Ù †åstroktanipatitadharmån pr¢ß™as tån eva vaktavyatayå pratij∞åya jagato ‘vyåkr¢tåvasthåvarñanam aprakr¢tam apurußårtha∫ ca Ù so ‘ya∫ satyo janapravådaΔ “åmrån pr¢ß™aΔ kovidårån åcaß™a” iti Ù na cåsmin vastuni pramåña∫ na ca prayojanam astœty ataΔ sarva evåyam adhyåyo nådhyetavyaΔ Ù

Structure and Composition of the Månava Dharma†åstra

207

Manu’s instruction to Bhr¢gu to transmit it to the sages thereby publishing it to the world: For the growth of these worlds, however, he produced from his mouth, arms, thighs, and feet, the Brahmin, the Kßatriya, the Vai†ya, and the ‡üdra. (1.31) After composing this treatise,33 he himself imparted it first only to me according to rule; and I, in turn, taught it to Marœci and the other sages. Bhr¢gu here will recite that treatise to you in its entirety, for this sage has learnt the whole of it completely from me. (1.58-59)

As the initial dialogue between the seers and Manu ends (1.4) with Manu’s command “Listen!” (†rüyatåm), so the entire preliminary narrative ends (1.60) with Bhr¢gu’s command “Listen!”. At this point we should have expected Bhr¢gu to get on with his task and to expound the dharmas of the varñas. That, however, is not he case. Here at this interstice between the introductory material and the body of the text, there are five “excursus” containing an account of time and cosmology (61-86), the occupations of the varñas (87-91), and the excellence of Brahmins (92-101); an eulogy of Manu’s treatise (102-110); and a table of contents (111-18). This material takes up the rest of Chapter One. Most scholars have taken the table of contents to be a later addition. I think this entire section represents redactorial interventions and is quite out of place here. Even though the origin of the †åstra and its transmission to Manu and Bhr¢gu had already been stated, the eulogy of the treatise ascribes its composition not to the creator but to Manu himself (1.102). This section ends with the instruction of Bhr¢gu (1.119): “Just as, upon my request, Manu formerly taught me this treatise, so you too must learn it from me today.” This duplicates Bhr¢gu’s command “Listen!” and does not have the same force or elegance. I think the authentic voice of the author is heard in this imperious “Listen!”, which is repeated in the two other answers to the seers’ queries at 5.3 and 12.2 later in the text. If we place the first verse of Chapter Two immediately after verse 60 of Chapter One, we see the smooth transition from the preliminary narrative to the body of the text that the author accomplished: tatas tathå sa tenokto maharßir manunå bhr¢guΔ Ù tån abravœd r¢ßœn sarvån prœtåtmå †rüyatåm iti ÙÙ 1.60 vidvadbhiΔ sevitaΔ sadbhir nityam adveßarågibhiΔ Ù hr¢dayenåbhyanuj∞åto yo dharmas ta∫ nibodhata ÙÙ 1.61 When Manu had spoken to him in this manner, the great sage Bhr¢gu was delighted; he then said to all those seers: “Listen!” 33 The introduction of this treatise appears to be abrupt, but it was actually introduced at the very beginning 1.3: “For you alone, Master, know the true meaning of the duties contained in this entire ordinance of the Self-existent One, an ordinance beyond the powers of thought or cognition.”

208

LANGUAGE, TEXTS AND SOCIETY

Learn the Law that people who are learned, virtuous, and free from love and hate always adhere to, the Law that they assent to in their hearts.

Here we have two signature expressions of Manu: †rüyatåm (“listen”) concluding the first verse, and nibodhata (“learn”) concluding the second. I think these two verses followed each other in the original composition of Manu, the first concluding the preamble and the second opening the main body of the work.

Chapters Two to Seven I find few if any identifiable interpolations in the central chapters of the book, 2-7. It may well be that the opportunity and/or the impulse to add new material were present at the beginning and in the concluding chapters. Suspicion was already raised by Bhler (1886a, lxvii) regarding the first eleven verses of Chapter Two. I think his doubts about verses 6-11 are unfounded; he considers these to be repetitions. We would have to eliminate a lot of verses throughout the text if we were to eliminate all duplications and repetitions; an author surely has the right to repeat and reiterate. So, for example, Manu deals with women both in his treatment of marriage and household life (Chapters 3-5) and under grounds for litigation (Chapter 9). The only doubt I have focuses on verses 2-5 of Chapter 2 that deal with desire. This section stands outside the flow of the discussion in the rest of the chapter. The kind of certainty I have with regard to the interpolations I have identified in the first chapter, however, is lacking here. The second passage identified by Bhler is 2.88-100. This deals with the control of organs and includes an enumeration eleven organs. I agree with Bhler’s (1886a, lxvii) assessment that this passage “interrupts the continuity of the text very needlessly, and has nothing whatsoever to do with the matter treated of.” Although such assessments often contain a heavy dose of subjectivity, in this particular case it is backed by textual evidence. The Bhavißya Puråña contains much of the early chapter of the MDh (Laszlo 1971; Sternback 1974). Bhavißya 1.2.5-27a-b reproduces the entire section on the student MDh 2.69-87. Immediately after this section at 1.2.27c-d, the Bhavißya gives the verse MDh 2.101, thus omitting the section 2.88-100, precisely the passage that we have suspected of being an interpolation (Sternback 1974, 7). It is a possibility that the author of the Bhavißya had before him a copy of the MDh in which this section was missing. Furthermore, the very wording of verses 2.87 and 2.101 indicates that they probably followed each other in the original text; the transition here is smooth, with verse 101 picking up the word japa from verse 87:

Structure and Composition of the Månava Dharma†åstra

209

japyenaiva tu sa∫sidhyed bråhmaño nåtra sa∫†ayaΔ Ù kuryåd anyan na vå kuryån maitro bråhmaña ucyate ÙÙ 2.87 pürvå∫ sandhyå∫ japa∫s tiß™het såvitrœm årkadar†anåt Ù pa†cimå∫ tu samåsœta samyag r¢kßavibhåvanåt ÙÙ 2.101 Only by soft recitation does a Brahmin achieve success; on this there is no doubt. Whether he does anything else or not, a benevolent man, they say, is the true Brahmin. At the morning twilight, he should stand reciting softly the Såvitrœ verse until the sun comes into view; but at the evening twilight, he should always remain seated until the Big Dipper becomes clearly visible.

Medhåtithi also, commenting on 2.88, appears to consider this section an appendix (pari†iß™a) to the section on twilight worship and lacking injunctive force (arthavåda). On the whole, however, these central chapters are remarkably free of tampering or interpolations. One may quibble about this verse or that, but there is no sure way of determining the authenticity of individual verses.

Chapter Eight I think there are several identifiable interpolations within this chapter. I will list them first and offer explanations for my decisions: 2022 (‡üdras as legal interpreters), 27-29 (property of minors and women), 30-40 (lost and stolen property), 386-420 (miscellanea). Chapter 8 begins with the king entering the court to adjudicate a law suit. This is in keeping with Manu’s penchant for placing his discourse in concrete situations, as, for example, discussing the king’s duties by following him through a regular day from getting up in the morning to going to bed at night. So here Manu places his discussion of the system of justice within the context of the king coming into the court. After enumerating the 18 grounds for litigation (vyavahårapada; 4-8), he opens his discussion on legal proceedings. In this section he deals with a) the judges, who substitute for the king (9-11), b) the demands of Justice that the law be administered impartially (12-19), and c) the beginning of the trial proper with a brief summary of judicial conduct and reasoning (23-26, 41-46). This admirable structure is marred by three “excursus”. The first (20-22) is a diatribe against a king employing a ‡üdra to interpret the law, a topic having little to do with matters at hand, coming between sections a) and b). Likewise, the section c) on judicial conduct and reasoning is cut in two by the intervention of some totally extraneous material having to do with property: the property of minors and women (27-29) and stolen and lost property (30-40). These matters have little to do with litigation or court proceedings. Indeed, once we remove these two “excursus” the section on judicial conduct reads smoothly: the judge should infer the truthfulness of litigants and witnesses by their external demeanor (25-26), find

210

LANGUAGE, TEXTS AND SOCIETY

out the special laws of the region, caste, and family of the litigants (4142), never initiate a law suit suo moto or try to suppress an action brought before him (43), apply correct judicial reasoning (44-45), and stick to the norms recognized by the cultured elite but only if they are not in conflict with those of particular regions, castes, and families (46). I think it is most probable that the three sections I have identified, verses 20-22, 27-29, and 30-40, are the work of later redactors. The rest of the chapter moves smoothly and there are no obvious interpolations that interfere with the flow of the text. The next interpolation comes right at the end of the chapter after the section of sexual crimes against women (386-420). This “excursus” contains a motley group of topics including the control of trade, ferries and tolls, and the occupations of the varñas, topics that have nothing to do with the administration of justice or law suits and belongs to Chapter 7 that deals with state administration. This section is so out of place and so obviously the work of redactors that it is unnecessary to argue the point at length. It is also instructive that this long section of 35 verses comes at the end of the Chapter, indicating that the addition was made after the MDh was subjected to the chapter division.

Chapter Nine This chapter addresses the last three grounds for litigation: marital law, inheritance, and gambling. As I have already noted, the section on gambling and betting does not regulate this practice but rather seeks to abolish it. For Manu gambling and betting is a social curse and the source of many social ills. Although, following tradition, Manu lists it under grounds for litigation, in his eyes it is properly a police function. This naturally leads to the final topic under the duties of a king, namely the eradication of thorns, that is, the elimination of social parasites. Unlike litigation which are brought before the king by private litigants, the eradication of thorns is to be initiated by the king himself and his officials. Immediately after the section on gambling (221-28), we should have expected some concluding statement wrapping up Manu’s discussion of the 18 grounds for litigation. And we find precisely such a statement in verse 250: “I have described above in great detail how lawsuits brought by litigants and falling within the eighteen avenues of litigation are to be decided.” After this Manu turns his attention back to the king and his duties with the statement (251): “Carrying out properly in this manner his duties flowing from the Law, the king should both seek to acquire territories not yet acquired and protect well those that have been acquired,” after which he proceeds with the subject of the eradication of thorns (252): “After properly settling the country and building a fort according to textual norms, he should direct his maximum effort constantly at the eradication of thorns.” This admirable and methodical discussion is marred and inter-

Structure and Composition of the Månava Dharma†åstra

211

rupted by two section that have little to do with the subject and that have all the marks of redactorial activity. The first comes between the conclusion of the section on gambling (228) and its closing statement (250). This section (229-49) of 21 verses contains various materials dealing with punishments and grievous sins, topics that are dealt with elsewhere in the book. The natural flow of Manu’s discourse is restored when we eliminate this section. The next “excursus” (294-311) deals with various subjects including the constituents of a kingdom and the activities of the king, topics belonging to Chapter 7. This interpolation is inserted between the section on the eradication of thorns (252-93) and the statement that closes this section (312). This statement finds its natural place immediately after verse 293; the intervening verses makes the closing statement out of place. The final “excursus” (313-23) instructs the king never to offend a Brahmin and comes between the closing statement on the eradication of thorns (312) and the concluding statement that brings the entire section on the king to a close (324-25). Again this interpolated section stands out like a sore thumb next to the finely flowing narrative. See how the flow is restored once this section is eliminated: In this and other ways should the king, always alert and tireless, suppress thieves within his own realm and even in others. (312) Conducting himself in this manner and always devoted to the Laws pertaining to kings, the king should direct all his servants to work for the good of his people. (324) I have described above in its entirety the eternal rules of action for the king. What follows, one should understand, are the rules of action of the Vai†ya and the ‡üdra in their proper order. (325)

Chapter Ten Chapter 9 concludes with the transitional verse that introduces the section on åpaddharma (the Law in times of adversity): eßo ‘nåpadi varñånåm uktaΔ karmavidhiΔ †ubhaΔ Ù åpady api hi yas teßå∫ krama†as tan nibodhata ÙÙ I have described above the splendid rules of action for the social classes outside times of adversity. Listen now to the rules for them in the proper order for times of adversity.

But Chapter 10 begins with a long discourse on mixed varñas (173) that at first sight appears to have little to do with how one must act in a time of adversity. At first sight, this section also appears to have resulted from the work of later redactors who were unable to understand how Manu could have omitted a discussion on mixed classes. Nevertheless, I think this section is part of the original treatise. For Manu, I think, a

212

LANGUAGE, TEXTS AND SOCIETY

time of adversity was not just a temporary emergency but also a permanent state of affairs, given the decadent state of society in the age of Kali. This permanent time of adversity is signaled by the intermixture of the varñas giving rise to several intermediate and lower castes (jåti). This was probably the reason why Manu deals with the mixture of varñas at the start of his discussion of åpaddharma. Other reasons also support this conclusion. The initial request (1.2) of the seers that prompted the narration of the text included the dharmas of not just the four varñas but also of those that are in between: bhagavan sarvavarñånå∫ yathåvad anupürva†aΔ Ù antaraprabhavånå∫ ca dharmån no vaktum arhasi ÙÙ Please, Lord, tell us precisely and in the proper order the Laws of all the social classes, as well as those of born in between.

It would be unlikely that a methodical writer like Manu, having introduced this issue at the very beginning, would fail to address it in the body of his text. There is no other place in the text that deals with mixed varñas. Furthermore, the Gautama Dharmasütra, which was one of the sources Manu used,34 has a similar section on mixed varñas (4.16-28). Likewise, the Artha†åstra (3.7.20-37) treats this topic in the course of his discussion of sons. With these examples in his own sources, it is unlikely that Manu would have neglected to treat this topic. The rest of the chapter flows smoothly, and I see no section that raises suspicions of redactorial intervention.

Chapter Eleven Chapter 10 concludes with the transitional verse that introduces the new topic of penance: eßa dharmavidhiΔ kr¢tsna† cåturvarñyasya kœrtitaΔ Ù ataΔ para∫ pravakßyåmi pråya†cittavidhi∫ †ubham ÙÙ 10.131 I have described above the entire set of rules pertaining to the Law of the four classes. Next, I will explain the splendid rules pertaining to penance.

One would expect Manu to open the topic of penance immediately, that is, at the beginning of Chapter 11. That, however, is not the case. We have to wait until verse 44 for the introduction of penance. When he finally gets to penance, Manu is as usual lucid and methodical. He begins with a clear and succinct introductory verse 34 See Olivelle forthcoming -b; below, pp. 261-274. I also deal with the sources of the MDh, including the Artha†åstra, in my introduction to the critical edition.

Structure and Composition of the Månava Dharma†åstra

213

(44): “When a man fails to carry out prescribed acts, performs disapproved acts, and is attached to the sensory objects, he is subject to a penance.” Then he discusses the significant issue of whether penance does any good. Can a person erase sins through penance? He devotes the next eight verses to this discussion, and concludes in the affirmative (54): “Therefore, one should always do penances to purify oneself; for individuals whose sins have not been expiated are born with detestable characteristics.”35 After this Manu proceeds to a methodical treatment of sins and their respective penances (see above I.4). Before we come to this lucid presentation of penance, however, there is a long (43 verses) section dealing with a gamut of topics unrelated either to penance or to each other. First is a section on occasions for giving and begging (1-26) followed by verses dealing with times of adversity (27-30), the power of Brahmins (31-35), and sacrifices (36-43). I am convinced that this entire section added at the beginning of the chapter represents not the original work of Manu but the supervening activities of redactors. The discussion of sins and the appropriate penances for them takes up verses 55-126. I have already dealt with this section (I.4) and its impressive structure. Manu then turns from personally committed sins to association with outcaste sinners, within which section he deals with both excommunication from caste and re-admission to it (18189). The entire section on penance concludes with the nice and pithy statement (190): enasvibhir anirñiktair nårtha∫ ki∫cit samåcaret Ù kr¢tanirñejanå∫† caiva na jugupseta karhi cit ÙÙ No one should transact any business with uncleansed sinners; and under no circumstances should anyone abhor those who have been cleansed.

At the interstice between the sections on penances for sins (73-126) and association with outcaste sinners (181-89), however, we have a long interruption that is attributable to redactorial activity. This “excursus” (127-79) deals with penances for a) injury to living beings, b) eating forbidden food, c) theft, and d) sexual offenses. As I have already noted (I.4), the reason or opportunity for this interpolation is probably to make up the four sins that Manu mentions in his transitional verse introducing the topic of association with sinners (180): “I have described above the expiation for all four kinds of sinners. Listen now to the following expiations for those who associate with outcastes.” These four kinds of sinners, however, were already presented in the previous authentic segment of the text on the classification of sins. 35 In this case also, Manu is probably following the example set by Gautama (19.310), who introduces his discussion with the explicit statement: “With regard to this (i.e. penance) people raise the question: Should (a sinner) perform a penance or not” (tatra pråya†citta∫ kuryån na kuryåd iti mœmå∫sante).

214

LANGUAGE, TEXTS AND SOCIETY

There follows a long section of 57 verses (191-247) containing miscellaneous expiations for a motley list of sins and infractions. The extraneous nature of this section is highlighted by verse 227: “By these observances should twice-born persons cleanse themselves of public sins; they may cleanse themselves of secret sins, however, through ritual formulas and burnt offerings.” This is quite out of place here, because Manu introduces his section of penances for secret sins many verses later at the end of this excruciating “excursus” (248) with the signature expression of Manu in transitional verses:36 ity etad enasåm ukta∫ pråya†citta∫ yathåvidhi Ù ata ürdhva∫ rahasyånå∫ pråya†citta∫ nibodhata ÙÙ Thus I have described to you above penances for sins in accordance with the rules. Listen next to the penances for secret sins.

The transition from the discussion of public sins to private sins would have occurred smoothly if this verse came immediately after verse 190 cited above without the intervening 57 verses that are clearly the work of redactors.

Chapter Twelve Chapter 12 poses unique problems because it is so very different from the rest of the work. It begins with the seers making one final request of Bhr¢gu to teach them the law of karma. One is tempted to see this entire chapter as deriving from the work of redactors. There is, however, no clear evidence that it did not belong to the original work of Manu; we cannot detect the breaks in the line of discussion that we detected in other interpolated passages or the violation of the structure Manu has laid out. Broadly this discussion on karma falls into two section, the one dealing with the fruits of action (3-81) and the other dealing with actions leading to the supreme good (83-106). These two sections are divided by one of Manu’s signature transitional verses (82). It is after these two central sections of the chapter that one begins to suspect redactorial intervention. At the conclusion of the section on actions leading to the supreme good there is what appears to be the usual transitional verse (107): naiΔ†reyasam ida∫ karma yathoditam a†eßataΔ Ù månavasyåsya †åstrasya rahasyam upadi†yate ÙÙ 36 Jolly (1887) expressed some reservation regarding the authenticity of this verse. I have removed it from my critical edition; it is not found in the southern manuscripts in Grantha, Telugu, and Malayalam scripts. Medhåtithi also doubted this verse, saying that this verse is recited by some; he apparently decided to accept it. Even if we discount this particular verse, a section on secret penances is appropriate here, especially because the Gautama Dharmasütra (24) also contains a section on them.

Structure and Composition of the Månava Dharma†åstra

215

This is the totality of activities leading to the supreme good as prescribed. The secret doctrine of this Treatise of Manu will now be taught.

There follows a section (108-115) on how to resolve matters relating to dharma that are not covered by Manu’s treatise. The language of this verse is quite different from all the other transitional verses; the passive verb upadi†yate is not found elsewhere. Manu prefers the active nibodhata and pravakßyåmi. The expression yathoditam in the first line is also absent in other transitional verses. The expression “månava†åstra” is also uncharacteristic. Although the issue of how to resolve matters not covered by the text is addressed at the conclusion of the Dharmasütras as well,37 nowhere are these rules called rahasya. My best guess is that the text of Manu concluded with the closing verse 116: etad vo ‘bhihita∫ sarva∫ niΔ†reyasakara∫ param Ù asmåd apracyuto vipraΔ pråpnoti paramå∫ gatim ÙÙ I have explained to you above all the best means of securing the supreme good. A Brahmin who does not deviate from it obtains the highest state.

The “excursus” on secret teaching contain material such as the legal assembly (parißad), however, that are found also in the Dharmasütras and thus has some claim for authenticity. One possibility is that this section formed the concluding statement of Manu and that the introductory verse was put there by a not very competent redactor, especially because the word rahasya does not occur in the body of this section. Nevertheless, the concluding verse 116 appears out of place after this section, because it refers back to the major theme of this chapter, namely, the attainment of the highest good (niΔ†reyasa), which was the topic covered by verses 83-106. If verse 116 came directly after 106 it would have provided a fitting and elegant conclusion to the entire book. As it is, this verse is further followed by another “excursus”. It appears that a redactor wanted to bring back the voice of Bhr¢gu, because this section begins with Bhr¢gu reflecting on the origin of the text: (117) “In this manner, the blessed god, desiring to do what is beneficial for the people, revealed to me in its entirety this highest secret of the Law.” There follows several verses recommending the contemplation of the Self with a strong Advaita tinge. The final verse in the extant text is an eulogy of the text itself where the voice is not that of Bhr¢gu but of a third person (126): “When a twice-born recites this Treatise of Manu proclaimed by Bhr¢gu, he will always follow the proper conduct and obtains whatever state he desires.” This entire section is in all likelihood a later addition to bring the 37

See Åpastamba Dharmasütra 2.29.13-15; GDh 28.48-51.

216

LANGUAGE, TEXTS AND SOCIETY

text to a close. I for one think that Manu’s original closure is much better; it is strong and pithy, without unnecessary emotion. In total, then, the sections I have identified as possible interpolations through the activities on one or more redactors contain 329 verses. Out of a total of 2680 verses in my critical edition, those verses account for just 12% of the text.

Conclusion In the current scholarly atmosphere of anti-orientalism there is a strong distrust of critical editions, textual criticism, and the work of philology in general. Much of this critique stems from ignorance of the process of critically editing texts and of the nature and purpose of a critical edition. This critique, unfortunately, has been fueled by the work of early philologists enamored with the “ancient” whose single ambition appeared to be the uncovering of an ur-text and in the process discarding everything else in the traditionally received texts. This enterprise appeared very much like separating the wheat from the chaff and throwing the chaff into the dustbin. A better comparison, however, is an archeological dig. No one faults an archeologist for digging deep and uncovering the oldest stratum. But good archeologists do not take all the evidence of later strata to the dump. The entire story of the village they are digging is told not just in the oldest layer but at every layer; the oldest is only the beginning of the story, but not the whole story. The editor, likewise, attempts to uncover the oldest detectable form of the text. All the supervening additions and emendations, however, are important for the life of the text, just as the succeeding layers are for the history of the village, even though the text scholar is often unable to date these additions with the kind of precision that the archeologist can. These additions and emendations are given below the line in the critical apparatus; the story of the text is told both above and below the line. The life of this particular text begins with its composition, I believe, by a single author –“Manu” for convenience– sometime around the beginning of the common era. It’s extraordinary life spans about 2000 years. Only about 1200 of these is represented by the evidence we have in the form of commentaries, manuscripts, and citations. I have attempted above to uncover its life during the first 800 years or so of its existence. Given the scope of this paper, I have limited myself to the text itself. A broader study, which will include how this text was used and remodeled repeatedly in the later Dharma†åstras such as those of Yåj∞avalkya, Nårada, Br¢haspati and Kåtyåyana, will form part of the introduction to my critical edition of the MDh.

A

217

IX Caste and Purity: A Study in the Language of the Dharma Literature*

With this volume on the theme of tradition, pluralism, and identity, we celebrate the life-long achievements of Professor T. N. Madan both as scholar extraordinaire and wonderful human being, a man who is as secure in his own identity when he is talking with a villager in Kashmir as when he is addressing a scholarly audience in Texas. The Indian tradition down the centuries, however, has managed pluralism primarily within the context of interlocking group identities, the most basic of which is caste. And the caste system, according to the currently prevalent view, is based on purity, each caste being located on a hierarchical gradation of purity. The higher the caste the greater the degree of purity. This thesis was laid out most boldly and most compellingly by Louis Dumont in his seminal work Homo Hierarchicus first published in 1966. Purity, according to Dumont, is the basis of hierarchy in traditional India and is, therefore, the ideological principle behind the caste system. Quigley (1993: 1), in his critique of the Dumontian thesis, acknowledges that the prevalent view among both Hindus and outside observers considers castes to be “hierarchically ranked on a purity-pollution scale”. Madan (1989: 365) himself states that “according to traditional caste ideology, which is obviously the brain-child of Brahmins, the key to the rank order lies in the notion of ritual purity.” Dumont’s thesis has not gone unchallenged. In a “Review Symposium” published just four years after the appearance of Homo Hierarchicus and organized by Madan with contributions from ten * I want to thank James L. Fitzgerald, who read this paper carefully and gave valuable suggestions and criticisms. Originally published in Contributions to Indian Sociology 32 (1998): 190-216.

218

LANGUAGE, TEXTS AND SOCIETY

scholars,1 several critiqued the central point of Dumont’s thesis that caste is based on purity. McKim Marriott’s (1990) efforts to construct an “Indian ethnosociology” using what he terms “Hindu categories” are in large part directed against Dumont. In the same volume Nicholas Dirks undertakes a frontal attack on Dumont’s thesis, claiming that in his own ethnographic work he has “found that purity and pollution are not the primary relational coordinates which endow hierarchy with its meaning and substance” (Dirks 1990, 61). In his recent book Quigley (1993) likewise challenges the premise that caste is based on the gradation of ritual purity. The connection between caste and purity that is at the heart of Dumont’s thesis, furthermore, appears not to be based on ethnographic data. This is a point asserted repeatedly by reviewers and is the basis for Dirks’ (1990) critique. Madan (1971: 9) notes the ethnologists’ “complain of his [Dumont’s] attitude to empirical evidence” and acknowledges Dumont’s “devaluation of the ethnographic datum,” but he notes: What distinguishes this work from the usual social anthropological discussions of caste is that it does not proceed from fieldwork to a model of how the system works. Instead it begins with a cardinal explanatory principle—hierarchy—and boldly sets out to build a model thereon, throughout maintaining the position that theory or ideology overrides and encompasses ethnography [...]. Models, therefore are not to be judged as true or false but as possessing more or less explanatory power.

Yet, Dumont cannot invoke the principle of hierarchy and its basis in purity totally a priori; they must be derived in some way from the evidence of how the Indian society functions or from the native social ideology. So, if not ethnography then what is the source of such evidence? Some think that his souce is Sanskrit texts. Berreman (1971: 22-23) explicitly states that Dumont “relies heavily on some classical Sanskrit texts while ignoring others” and concludes that his thesis “conforms well to the theory of caste purveyed in learned Brahmanical tracts. But it bears little relationship to the experience of caste in the lives of the many millions who live it in India.” Von Frer-Haimendorf (1971: 24) notes Dumont’s “expertise in classical Indology.” Are the classical texts of Brahmanism, then, the inspiration behind Dumont’s “model” of the caste system? These texts, especially the technical literature on Dharma (dharma†åstra), however, recognize only the division of human society into four varñas and their social ideology is based on varña and not on caste (jåti), castes being subsumed under varña ideology as hybrid forms. Leach (1971: 15) 1 Madan et al. (Contribution to Indian Sociology5: 1971). Besides Madan, the reviewers were E. Leach, G. D. Berreman, C. von Frer-Haimendorf, R. S. Khare, V. Das, J. S. Uberoi, J. C. Heesterman, D. Kantowsky, and M. Singer, with a response by Dumont. Especially critical of Dumont’s thesis are Berreman, Veena Das, and Uberoi. For a recent defense of purity as the basis of caste from a psychoanalytic perspective, see Dundes (1997).

Caste and Purity

219

appears to hint that Dumont’s ideology is borrowed from the Brahmanical theory of varña when he remarks on Dumont’s insistence that the structure of jati organization which is ‘out there’, external to the observer, is integral with the structure of varna hierarchy [...]. As anthropological outsiders we need to pay close attention to the varna system because it helps to make sense of the facts on the ground and, in turn, continually forces the facts on the ground into a coherent pattern.

Brahmanical scholarship is consistent on this point, and the technical literature on Dharma continues to focus on varña as the basis of Indian society throughout the medieval period and right up to modern times. In Dumont’s (1980: 66-91) own treatment of varña, however, he states clearly that although varña and jåti are both based on hierarchy divorced from “power”, the varña hierarchy is not based on the pure and the impure (1980: 66, 73). So, according to Dumont, on this crucial point the two systems of social hierarchy in India diverge. And yet he acknowledges that much of the caste ideology, which Madan (1989: 365) calls “ the brain-child of Brahmins,” is derived from the varña ideology. We are then faced with not just a theoretical or ideological problem but with a historical question: when and how did caste ideology of purity/impurity emerge out of the varña hierarchy?2 The historical question is raised by Madan (1971: 12): “It is not clear, however, at what time Dumont believes the crucial structural elements of the caste system, as presented by him, to have crystallized.” Is it plausible, furthermore, that the same Brahmin intelligentsia that kept purity out of the varña hierarchy would have made it the cornerstone of the caste hierarchy? Even though, as Dumont himself acknowledges, varña hierarchy is not based on a gradation of purity/impurity, nevertheless the same Brahmanical social thinking that developed the varña ideology also placed great emphasis on purity, creating intricate and minute rules on pollution and purification. These rules and the ideology underlying them have not been subjected to thorough scholarly scrutiny. An important way to get a handle on that ideology, I think, is to study the vocabulary of purity/impurity in these texts. Scholars regularly use Sanskrit equivalents in dealing with purity under a tacit assumption that this will somehow take us closer to the reality on the 2 Another question is the ideological basis of the varña hierarchy, a question Dumont never fully answers except to draw a distinction between “dignity” and “power” and the separation of status from power. The mythical legitimization of the varña division is, of course, contained in the Purußa Hymn of the R¢g Veda (10.90) which portrays the creation of the four varñas from four parts of the creator’s body and posits thereby a biological (racial/genetic?) basis for varñas. The early vedic texts, however, indicate that raw power was a central ingredient in varña hierarchy, illustrated by the metaphor of food and eater, the lower varñas being the ‘food’ of the upper varñas (see Smith 1990).

220

LANGUAGE, TEXTS AND SOCIETY

ground.3 Yet, there is no single term in Sanskrit for either the substantives “purity/impurity”4 or the adjectives “pure/impure”. The existence of a large number of terms in a language for a broad area of human experience is prima facie evidence for that area being central to that culture and for its nuanced and often technical treatment by that culture. It is interesting to note, however, that, despite the enormous amount of writing on the concept of purity in India, there has been no sustained and detailed study of the Sanskrit (or other vernacular) terms for “pure/impure”. That is what this paper attempts to do within the confines of the ancient Dharma texts in an attempt to uncover the complex ways in which they deal with the category today subsumed under “purity.” This is an exercise in philology in the best sense of the word, a philological study that is context-sensitive and therefore refuses to reify anything,5 least of all words, and examines the changing patterns of word usage that opens a window into the changing reality of the social world lying behind the language. Such a study may also throw some valuable light on the ongoing scholarly debate about the caste system and contribute to the cross-cultural study of purity and pollution, to which I will return in the concluding section of this paper.

1. The Vocabulary of Pure/Impure The principal terms for pure/impure in the Dharma literature belong to seven families,6 some of which contain several individual terms with their own nuances and technical meanings. For convenience I introduce each family under its most prominent representative in the order I treat them.7 1. †uci — †auca, a†uci, å†auca 2. †udhyati (†odhayati, †odhana) — †uddhi, †uddha, a†uddha 3. prayata — aprayata 3 See, for example, Marriott (1990: 33). Madan himself is concerned about the use of the modern English term “purity” to describe the lived reality of Indian life: “I use deliberately, but only as far as seems reasonable, the Sanskrit words †ubha and †uddha instead of ‘auspiciousness’ and ‘purity’. The former two words or derivatives from the same are in use in most languages of India. My hesitation in using the two English words throughout the chapter arises from the fact that they have become omnibus words and conceal more than they reveal” (Madan 1988: 49-50). 4 The abstract term “purity”, as we will see, is absent in the vocabulary of Dharma texts. 5 Marriott’s (1990) use of the term “Hindu” is misleading, pointing as it does to a reified and unchanging substance or category. There is a similar reified use of “Hindu” with reference to food in Khare (1992), about which I have written elsewhere (Olivelle 1995b). On the context-sensitive nature of Indian culture, see Ramanujan (1990). 6 I have ignored a few terms, such as kalmaßa, which in the Dharma literature always refers to moral turpitude or sin: A 1.22.4; 1.24.25; 1.28.18; 1.29.1 (the last three repeat the same expression); B 2.4.5; Va 28.6 (same as B); M 4.260; 12.18; 12.22; Vi 1.36; 23.60; 52.14; Y 3.218. 7 So as not to make matters more complex than they already are, I have not treated separately compound terms, such as vi†uddhi and pari†uddhi (and other compounds with

Caste and Purity

4. 5. 6. 7.

221

püyate (püta, apüta) — pavitra, påvana ucchiß™a medhya — amedhya mala (amala/nirmala)

Let me make a few general comments before proceeding to a more detailed analysis of each term. First, a grammatically negative term (e.g., a†uci or amedhya) is not always simply the opposite of the positive; negatives often develop their own semantic overlay, creating highly technical meanings. This is a common phenomenon in Sanskrit as witnessed by the well-known term ahi∫så, whose meaning goes well beyond what may be expected from the simple negative of hi∫så. Since there is considerable semantic overlap between several of these terms, the use of one in a particular text or context may not always be of special significance. Often the selection may be based on the exigencies of meter (in verses), or on alliteration and other “sound effects”. There is the danger of “reading too much” into a word. In this study, I have tried to survey a broad sampling of cases before determining whether nuances and technical meanings are attached to particular terms in particular contexts or by particular authors. In my discussion of individual terms, I will focus on several significant questions. Does a term refer: (a) to a person or to a thing; (b) to a condition (that is, a person or thing being pure, or more generally to the abstract “purity”) or to a transition (that is, a person or thing becoming pure, the recovery of lost purity); (c) to what we have come to call “ritual purity”, to common cleanliness, or to the areas of morality and criminal law (four areas that are not always as cleanly distinguished in our sources as in our own minds); or (d) to an individual in his own existential being or to a class or group (or to an individual as part of a group)? The resolution of these questions, I believe, is important not only to our understanding of ancient Indian world views and ritual practices, but also to the ongoing debate on the role of purity in caste identity and hierarchy. For a diachronic study of the terms for pure/impure we need both an absolute and a relative chronology of the Dharma texts. Unfortunately, as is the case with most ancient Indian literature, we cannot place certain or accurate dates against any of these texts. In footnote 8 I give a tentative chronology that may be of some help for those unfamiliar with this literature.8 the prefixes vi-, pari-, etc.). These, as far as I can tell, do not add new meanings or nuances, except to make the term more intensive. So pari†uddha or vi†uddha may mean “purified to an extraordinary degree”. 8 Kane (1962-75) gives the following chronology: Gautama 600-400 BCE; Åpastamba 450-350 BCE; Baudhåyana 500-200 BCE; Vasiß™ha 300-100 BCE; Manu 2nd cent. BCE. to 2nd cent. CE (I would place him closer to the later than the former date); Vißñu 300 BCE100 CE (the current inflated text 400-600 CE); Yåj∞avalkya first two centuries CE; Nårada

222

LANGUAGE, TEXTS AND SOCIETY

Within the confines of this paper it is not possible to discuss fully and in detail every occurrence of these terms in the Dharma literature. I will confine myself to presenting illustrative examples for each major meaning or nuance, relegating to the charts attached to each family of terms the exhaustive listing of the evidence. 1. ‡uci The term †uci is an adjective derived from the verbal root †uc and in its earliest usage meant “shining, bright, white”. By the time of the early Dharma texts the term had acquired its traditional meaning of “pure”, although more literal meanings continued to coexist.9 As its primitive meaning indicates, †uci alludes to a positive quality (bright, pure) in a subject, and its dynamic meaning refers to the regaining of this lost quality. In contrast, we will see that †uddhi refers directly to the getting rid of impurities, the positive quality being indicated indirectly as the result of such purification. For the sake of clarity, I give below four significant areas in the semantics of †uci. 1. An important distinction our texts make is between the purification of persons and of things, such as vessels, seats, or a piece of ground.10 ‡uci is the most common term for “pure” with reference to persons.11 When a text wants to make this distinction clear, it invariably uses †uci (or its derivative †auca: see 1.2) for persons and †uddhi for things. The use of separate terms points to an ideological distinction between persons and non-persons in the area of purity. Gautama is the first author to use the expression dravya†uddhiΔ (“purification of things”). He introduces the section (G 1.29-34) on the 100-300 CE; Vaikhånasa 300-400 CE This sort of chronology is a mere house of cards without too much supporting evidence. My own opinion is that, whatever may be said about the original versions, with regard to the texts as we have them now, Åpastamba is the oldest, followed by Gautama, Baudhåyana (although this text has undergone extensive additions), and Vasiß™ha, all probably composed between the fourth and first centuries B.C.E. Then comes Manu, Nårada, Yåj∞avalkya, Vißñu, and Vaikhånasa in no particular order (see Lariviere’s comments: N, vol. 2, pp. xix-xxii; Manu is probably the oldest of this group), all of them composed well into the common era, although Vißñu contains sections belonging to an old sütra text. 9 Kålidåsa, in his R¢tusa∫håra (1.2), for example, calls “summer”, the hot season, †uci. References in the Dharma texts to god as †uci probably has the meaning of bright or shining: A 1.22.7; B 1.104; 1.13.2. Interestingly, the word “pure” itself is derived from the IndoEuropean root for fire, from which is derived the Greek pyra and English “pyre” and “fire”, as well as the Sanskrit püyate (see section 4 on püyate). 10 In this study I make the broader distinction between persons and “non-persons”, including animals, objects, activities (e.g., rites), bodily discharges, and detached parts of the human body, such as hair and nails. Within the category of “person” I include “pure” used with reference to the human body (†arœra) or bodily parts (aõga), such as hands and mouth. 11 Åpastamba, possibly the earliest Dharma text, uses prayata (see section 3) in preference to †uci with regard to personal purity, using the latter mostly in its second meaning. This may indicate a development of the purity vocabulary within the Dharma tradition. It may, however, be simply a matter of personal preference or idiosyncrasy, or even a regional difference in the vocabulary.

Caste and Purity

223

different methods for cleansing articles made of metal, clay, wood, and so forth with this expression (G 1.29) and distinguishes that from personal purification, †auca (G 1.35), a topic he deals with in the very next section (1.35-45). This distinction is brought out most clearly by Manu. In the first chapter he gives a table of contents which lists †aucam and dravyåñå∫ †uddhiΔ (M 1.113) as two separate entries, the former referring to personal purification and the latter to the cleansing of things. In the body of the text he maintains this distinction with a verse that concludes his discussion of personal purification and introduces the next topic, the purification of things: eßa †aucasya vaΔ proktaΔ †årœrasya vinirñayaΔ / nånåvidhånå∫ dravyåñå∫ †uddheΔ †r¢ñuta nirñayam // (“I have explained above the settled practice regarding the purification of the body. Listen now to the settled practice regarding the cleansing of things”) [M 5.110]. This distinction is also evident in a rule that prohibits vedic recitation when a person finds that “the ground where the recitation takes places is unclean or he himself is impure”: svådhyåyabhümi∫ a†uddhåm åtmåna∫ cå†ucim (M 4.127). Throughout the Dharma literature, the term for “pure” within the sections on personal purification is invariably †uci. Even outside these sections, †uci most frequently refers to the purity of persons rather than of things (see Fig. 1: A.1.a-b). ‡uci is not used exclusively with reference to people, however, although this is its primary sense. It has a broader application; animals, clothes, mines, and water are said to be †uci. The most frequent use of †uci outside the context of persons, however, occurs in the set phrase †ucau de†e (“in a clean [or cleansed] place”) with reference to the area where a ritual act is to take place (see Fig. 1: A.2.b). 2. The connection of †uci with persons carries over into the area of morality and personal character. Thus, in many contexts, especially in the appointment of ministers and other public officials, †uci is used with the meaning of “upright”, “honest”, “loyal”, “trustworthy”– that is, a man of character and integrity. The king “should appoint Åryas who are upright and honest to protect his subjects” (årya∞ chucœn satya†œlån prajåguptaye nidadhyåt) [A 2.26.4]. The king himself, as well as his ministers, officials, supervisors of mines and gambling houses, ambassadors, judges, and witnesses in a court should be †uci, upright (Fig. 1: A.1.c) 3. In the area of criminal justice †uci means “innocent”, although this meaning is recorded only in the somewhat late texts of Manu and Nårada. A man who has undergone an ordeal successfully is said to be †uci (M 8.115). A king’s power is demonstrated by that fact that his word can make a guilty person (a†uci) innocent (†uci) and vice versa (N 18.49). 4. When †uci is used with the first meaning, it most frequently indicates “purification” or “becoming pure” rather than “being pure”; that is, †uci indicates that a person has become pure through some purificatory activity (see Fig. 1: A.1.a). This usage, together with its counter-

224

LANGUAGE, TEXTS AND SOCIETY

part in †uddhi (see section 2.1), shows that the concern of Dharma authors was with the constant struggle to recover lost purity rather than with some abstract notion of “purity” that may attach to persons or groups. This dynamic meaning is most evident in the expression †ucir bhavet (“he becomes pure”) coming at the end of the description of purificatory rituals. Thus, after he has bathed in and drunk a mixture of cow’s urine, cowdung, milk, curd, ghee, and a decoction of Ku†a grass, “a man who has been bitten by a worm is purified” (kr¢midaß™aΔ †ucir bhavet: B 1.11.38). More commonly, however, the verb bhavet is dispensed with and the prescription of the rite concludes with just †uciΔ –“after bathing he becomes pure” (snåtvå †uciΔ: M 2.176); “after bathing the father becomes pure” (upaspr¢†ya pitå †uciΔ: M 5.62); and with regard to food particles stuck to the teeth, “a man becomes pure by simply swallowing them” (nirgirann eva tac chuciΔ: G 1.40; B 1.8.25; Va 3.41). The dynamic meaning is also prevalent whenever †uci is used adjectivally. Thus, when a text states that a “†uci Brahmin” should do something, it does not mean that a “pure Brahmin” (static meaning) should do it but that a Brahmin “having become or made himself pure” should do it, and refers to a purificatory rite such as bathing or sipping water that would precede any ritual act. Thus, when Manu (M 4.35) describes what a vedic student should do when he recites the Veda: k¬ptake†anakha†ma†rur dåntaΔ †uklåmbaraΔ †uciΔ (“having cut his hair, beard and nails, keeping himself subdued, wearing white clothes, pure), the term pure as an attribute of the student while studying indicates that he should make himself pure before study, just as he should cut his hair, make an effort to subdue his senses, and wear white clothes. There are occasions, however, when †uci means being pure, an affirmation that a person or thing is in a state of purity. When such a static meaning is intended, †uci invariably stands in the position of the predicate rather than a simple attribute12 –thus, †vå mr¢gagrahañe †uciΔ (“in catching a deer a dog is pure” B 1.9.2); ahata∫ våsaså∫ †uci (“new clothes are pure” B 1.13.4); nityam åsya∫ †uci strœñåm (“the mouth of a woman is always pure” M 5.130). The predicative meaning, however, is comparatively infrequent and occurs mostly with reference to objects rather than to persons.

12 It appears that when †uci used adjectivally has a somewhat static meaning, then it refers to moral character (“upright”; meaning # 2) rather than to personal purity.

Caste and Purity

225

Figure 1 A. †uci 1. Person: a. becoming pure: A 2.18.6 (?); G 1.40; 9.2; 26.21; B 1.6.15; 1.8.23; 1.8.24; 1.8.25; 1.11.38; 1.11.41; 2.5.18; 4.2.8; Va 3.41; 10.31; 23.31; 29.21; M 2.51; 2.107; 2.176; 4.35; 5.106; 5.143; 8.87 (twice); Y 1.131; 1.195; 1.225; 2.99 (?); 3.21; 3.26; 3.51; Vi 22.89; 23.49; 23.55; 66.15; N 12.87; 20.17; Vkh 1.3; 2.11.freed from sin: Y 3.245; 3.257; 3.303. b. being pure: B 1.10.5 (twice); M 5.106; Y 1.187; Vi 22.89 (thrice). c. morally upright: A 2.15.11; 2.25.13; 2.26.4; G 11.4; B 2.14.6; Va 2.9; M 2.109; 2.115; 5.106; 7.22; 7.31; 7.38; 7.60; 7.62; 7.63; 7.64; 8.77; 9.188; 9.335; Y 1.28; 1.121; 1.309; 1.312; 1.322; 1.355; 2.191; Vi 3.71; 29.10; N 1.133; 1.209; Vkh 3.5 (?). d. criminally innocent: M 8.115; N 18.49 (twice). e. god: A 1.22.7; B 1.10.4 & 1.13.2. 2. Non-persons: a. becomes pure: Va 12.15. b. is pure: place (†ucau de†e): A 1.11.23; 2.4.23; 2.18.6; G 1.36; B 1.8.11; 2.5.21; M 2.222; 3.206; 5.68; Y 1.18; 1.227; Vi 61.17; Vkh 3.5others: B 1.9.2;1.9.3; 1.9.6; 1.13.1 (re. sacrifice); 1.13.4; 3.5.2; 3.9.2; 3.47; 14.27; 21.14; 28.8; M 5.130; 5.131; 9.70; Y 1.187; 1.191; 1.192; 1.193; Vi 23.49; 23.50; 23.52; N 18.42. B. †auca 1. Purification: a. person: A 2.15.12; G 1.35; 1.42; 9.25; 9.71; 14.44; B 1.6.2; 1.6.15; 1.7.1 (twice); 1.8.1; 1.8.3; 1.8.4; 1.8.11; 1.8.52; 1.8.53; 1.14.19; 2.4.5; 2.11.24; 2.15.11; 3.1.25; 3.1.26 (twice) Va 4.35; 4.37; 6.17; 6.19; 11.28; 12.17; 19.47; 28.6; M 1.113; 2.61; 2.69; 3.126; 3.192; 4.93; 4.148; 4.175; 5.94; 5.97; 5.98; 5.100; 5.106 (twice); 5.110; 5.137; 5.139; 5.140; 5.146; 7.145; 9.11; Y 1.15; 1.17; 1.71; 1.98; 1.209; 1.232; 3.29; Vi 22.26; 22.89 (twice); 22.93; 23.42; 60.24; 60.26; 91.18; Vkh 2.9 (twice); 3.4. b: thing: B 1.14.16; Va 3.48; 3.53; M 5.114; 5.118 (twice); 6.53; Vi 23.34 (divine image). 2. Virtue: G 8.23; 10.51; Va 6.23; 11.35; M 3.235; 6.92; 10.63; 12.31; Y 1.122; 3.66; 3.137; 3.313; Vi 2.16; Vkh 1.4; 2.4. C. a†uci 1. Person: a: impure: G 16.46; B 1.6.14; Va 4.38 (? sinner); 5.6; M 4.71; 4.127; 4.142; 4.143; 5.75; 5.76; 5.79; 5.81; 5.84; 5.86 (?);Y 1.135; 1.149; 3.30; Vi 96.26; Vkh 2.14; 3.3; 3.4. b. morally bad (sinner): A 1.21.12; 1.21.19; 1.29.14 ; 1.29.15; 1.29.17; 1.29.18; 2.12.22; G 9.11; 9.16; 23.22; B 1.10.5; 2.2.15; 2.2.23; 2.2.24; N 14.24. b. guilty: N 18.41; 18.49 (twice). 2. Non-persons: a. impure (adj.): B 1.9.6; Va 4.23; 14.30; M 4.124; Y 1.149; Vi 70.17; 98.70; N 18.42; Vkh 3.5. b. filthy substance (noun): A 1.2.29; M possibly 5.86; Vi 5.106; 60.13; N 5.6. D. å†auca [asauca] Period of Impurity: G 2.3; 14.1; 14.23; 14.29; 16.18; B 1.11.1; 1.11.18; 1.11.19; Va 4.9; 4.16; 4.23; 4.31; 4.34; 4.36; 23.24; M 5.59; 5.61; 5.62; 5.74; 5.80; 5.97 (twice); 11.183; Y 3.6; 3.18; 3.27; Vi 19.13; 19.18; 20.32; 21.1; 22.1; 22.6; 22.7; 22.8(twice); 22.10; 22.11; 22.12; 22.13; 22.14(twice); 22.15; 22.16; 22.17; 22.18; 22.19; 22.21; 22.33; 22.35 (thrice); 22.38; 22.39; 22.40; 22.47; 22.56; 27.5; Vkh 2.11; 2.14; 3.8.

226

LANGUAGE, TEXTS AND SOCIETY

The multiple meanings and nuances of †uci permits our authors sometimes to play with that diversity. A good example is Manu 5.106 (variants in Vi 22.89; Y 3.32): sarveßåm eva †aucånåm artha†aucam para∫ smr¢ta∫ Ù yo ‘rthe †ucir hi sa †ucir na mr¢dvåri†uciΔ †uciΔ ÙÙ “Of all forms of purifications [a], keeping oneself pure [b] in transactions is the best, smr¢tis say, for a man is pure [c] when he makes himself pure [d] with respect to transactions (or procuring wealth); he is not pure [e] by becoming pure [f] using earth and water.” Here [a] has the general meaning of purification (see section 2.1), while [b] already borders on keeping oneself honest; [c] and [e] appear to have the static meaning of being pure (as predicate), but that state is earned through a dynamic process of purification [d] and [f], where [d] again has the meaning of honesty and integrity, while [f] is washing to get rid of stain and smell. As this verse shows, the four meanings I have separated for heuristic purposes are seen by our authors as forming a single spectrum of meanings. 1.1 ‡auca ‡auca is what Sanskrit grammarians call “a vr¢ddhi derivative”, that is, a word derived from another word by strengthening its first syllable. We get †auca from †uci with the strengthening of ‘u’ to the vr¢ddhi grade of the diphthong ‘au’. Words derived in this manner indicate in the most general way a relationship to the meaning expressed by the primary word. When the derived word is a neuter substantive, however, as is the case with †auca, it often expresses an abstract concept. Accordingly, we should expect †auca to mean “purity”. Its primary meaning, however, is not the abstract quality of “purity” but the dynamic process of “purification”. This may well be due to the specialization of meaning that vr¢ddhi derivatives often undergo; but I think it is more likely that this meaning became attached to †auca because, as we saw, the primary meaning of †uci is not simply “being pure” but “becoming pure”. Thus †auca came to mean “what is connected with becoming pure”, that is, the process by which a person becomes pure. Just as †uci acquired the specialized meaning of “becoming pure” with reference to persons, so †auca is used specifically with reference to the purification of persons (see Fig.1: B.1.a). We saw above that the Dharma literature distinguishes the purification of persons from that of things by using †auca for the former and †uddhi for the latter. The rules of personal purification are collectively referred to as †auca: thus, a teacher is instructed to teach †auca to his students (M 2.69; Y 1.15). Although †auca is used as a general term for purification by bathing, washing, sipping water, purificatory rites, and penances, it is used especially with reference to washing the anus, the penis, and the hands with water and earth (used as a cleansing agent) after voiding urine or excrement. Thus, for example, kr¢ta†aucåva†iß™a is earth “left over from a previous purification” after toilet (Va 6.17). Manu (4.93) instructs a person to perform †auca after he has risen in the morning

Caste and Purity

227

and answered the call of nature. And when texts refer to the different degrees of †auca for students, householders, hermits, and ascetics, they are speaking about the purification after toilet (Va 6.19; M 5.137). I have failed to notice a single occurrence of †auca as an abstract noun indicating “purity” in the Dharma literature.13 The closest we come to such a usage is in passages that list †auca among other virtues or habits that a person should cultivate (Fig. 1: B.2). The earliest such occurrence is in Gautama: athåß™åv åtmaguñåΔ Ù dayå sarvabhüteßu kßåntir anasüyå †aucam anåyåso maõgalam akårpañyam aspr¢heti– “Now, the eight virtues/qualities of the self are: compassion toward all creatures, patience, lack of envy, †auca, tranquillity, having an auspicious disposition, generosity, and lack of greed” (G 8.22-23). The commentator Haradatta, citing a series of verses that explain each of these virtues, states that †auca refers to the purification of things, mind, speech, and body. Later, explaining the provision that even ‡üdras should cultivate the virtues of truthfulness, not getting angry, and †auca, Haradatta gives the same explanation. I think the natural tendency of the native commentator to take †auca even within the context of virtues as attention to internal and external purification is correct. Virtues, after all, are not unalterable states, like being white or tall; they are habits to be cultivated. It makes better sense, therefore, to see the virtue of †auca as the habit of engaging in activities of purification.14 Although †auca applies most frequently to persons, it is used occasionally in some texts with reference to the purification of things (see Fig. 1: B.1.b). Thus Baudhåyana uses it for the purification of honey and milk (B 1.14.16), Vasiß™ha for the purification of articles made of ropes, bamboo, and leather (Va 3.53), and Manu for the purification of metal vessels and grain (M 5.118; 6.53). 1.2 A†uci Even though a†uci is the negative of †uci, its earliest usage indicates that it had already acquired a technical meaning connected more with †uci in the sense of “upright, honest, innocent” than in the sense of “pure”. Åpastamba uses a†uci eight times. Certainly in six of these and in all likelihood also in a seventh,15 a†uci is used not as an adjective but 13 Here I take “purity” as an abstract concept connected in some ways to the way it is used by Dumont as a objective state inherent in different castes. ‡auca does refer to “purity” in the sense of concern with and the procedures for “becoming pure”; such concern is evident in the use of the term for a specific virtue or habit. 14 The other option in these contexts is to take †auca as referring to †uci in the sense of upright and honest. If that is the meaning, then the virtue to be cultivated is uprightness or honesty. 15 The doubtful case is the compound a†uci†ukla which can mean “impure semen” or “the semen of an impure man”, meaning the semen of a man who has committed a crime making him an a†uci. The sentence reads: a†uci†ukla∫ yan nirvartate na tena saha sa∫prayogo vidyate –“Likewise, there can be no association with what is produced [i.e., child] by the semen of a sordid man” (A 1.29.14). I think here also a†uci refers to the class of sinners listed earlier, for otherwise it is difficult to explain how the issue of impure semen would

228

LANGUAGE, TEXTS AND SOCIETY

as a substantive with reference to a type of sinner. The sins that create this state are called a†ucikara, “making someone a†uci.” Åpastamba does not know or does not recognize the distinction common in later literature between mahåpåtaka and upapåtaka, grievous and secondary sins causing loss of caste. He begins the section on sins by stating that “social interaction with outcastes (patita) is not permitted, as also with degraded (apapåtra) people (A 1.21.5-6). Then he describes one group of sins which he calls patanœyåni (“causing loss of caste”: A 1.21.7-11), and a second group of sins which he calls a†ucikaråñi (“causing someone to be a†uci: A 1.21.12-19). These two groups must correspond to the two categories of people in the introductory statement; an a†uci, therefore, is an apapåtra, a degraded person with whom social interaction is forbidden. Gautama uses a†uci four times (see Fig. 1: B.1.a, b). Certainly in three of them, and possibly also in the fourth (G 16.46), the term has the same or similar technical meaning. Thus, when someone sees an a†uci, he should look at the sun to regain purity (G 23.22). An a†uci is also listed alongside mleccha and adhårmika (barbarians and the unrighteous) as people with whom one should not speak (G 9.16). Besides Åpastamba’s patanœyåni sins, Baudhåyana is the earliest writer to mention the traditional upapåtakåni, secondary sins causing loss of caste (B 2.2.1-14). Nevertheless, he gives as a third class of sins, the a†ucikaråñi (B 2.2.15). This category of sins disappears from the Dharma vocabulary after Baudhåyana. An element of this meaning, however, is retained in the use of a†uci within moral and criminal contexts (see Fig. 1: B.1.c). The king should punish a†ucœn, “guilty” or “immoral” people (N 18.41); the word of the king can make an innocent person (†uci) guilty (a†uci: N 18.49); and when cattle are stolen, suspicion falls on a village where there are “bad people” (a†ucir janaΔ: N 14.24). In later texts the common meaning of “impure” both as an adjective and as a substantive emerges. Thus, a menstruating woman is a†uci for three days (Va 5.6), menstrual blood is a†uci (Va 4.23), and one is a†uci for ten day upon hearing of a relative’s death (M 5.79). Used substantively with reference to things, a†uci means simply filth, especially bodily excretions. Åpastamba (A 1.2.29) speaks of limbs “smeared by a†uci” (a†uciliptåni), where a†uci, as the commentator Haradatta himself acknowledges, refers to filthy substances such as urine and feces.16 A lavatory is a†ucisthåna (“place of a†uci”: N 5.6). An interesting feature of the negative a†uci, a feature shared by its counterpart a†uddha (see section 2.3), is that in its diverse senses it be subject to such serious social disabilities. The same expression occurs in the very same context also in B 2.2.24. The eighth case is A 1.2.29 discussed later. 16 In a very similar context, Va (3.48) uses the phrase amedhyaliptåni, “smeared with amedhya”, a term more closely identified with bodily excreta (see section 6.1).

Caste and Purity

229

frequently exhibits a static meaning, a feature quite different from the generally dynamic meanings of the positives †uci and †uddhi. 1.3 ņauca Like †auca, the term å†auca is a “vr¢ddhi derivative” with the negative prefix ‘a’.17 Even more than †auca, however, å†auca developed a restricted technical meaning. In the Dharma literature it invariably applies to the period of impurity following a death or a birth (see Fig. 1: D). The period of impurity following a birth is sometimes characterized as sütaka (“caused by childbirth”), while the period following a death is sometimes distinguished from the former with the expression †avå†auca (“å†auca caused by a corpse”).18 The term å†auca (or a†auca) is a strictly technical term in the Dharma literature. It is never used with a more general meaning of “ impurity”.19 2. ‡udhyati The verb †udhyati20 in all its verbal forms has the meaning of “becoming purified” and is used most frequently with reference to a ritual or an act of purification (Fig. 2: 1.A). The verb is used indiscriminately with reference to persons and things. Both references are 17 The grammatical derivation of å†auca is somewhat problematic because of the double vr¢ddhi of both the initial negative ‘a’ and the second vowel. The easiest derivation is the addition of the negative prefix ‘a’ to †auca. This would give a†auca and not å†auca, and Påñini’s rule 7.3.30 is intended to permit the optional strengthening of the initial ‘a’ even in such cases. The alternative is to derive it from a†uci. Påñini 5.1.121 disallows such a derivation if a†uci is taken as a Tatpurußa compound (i.e., meaning “not pure”). Even if we take it as a Bahuvrœhi (i.e., “one who does not possess the quality of pure”), the derivative should be å†uci and not å†auca. Although grammatically the derivation from †auca is easier, semantically the latter appears to make better sense, for then å†auca would mean “a state of being a†uci or impure”. As we have seen, †auca does not simply mean “purity” but “purification”; hence its negative should be “non-purification” rather than “non-purity”. Alternatively, we would have to take å†auca to mean precisely “non-purification”, that is, a period when purificatory rites are disallowed, as is in fact the case when people are in a period of å†auca. The expression sadhyaΔ†auca, “immediate purification” (for example, for a king: M 5.94; G 14.9-12, 44-46), that is contrasted with å†auca does support the latter meaning. I want to thank George Cardona and Madhav Deshpande for their valuable insights on this question. 18 Manu (6.62), for example, clearly distinguishes †åvam å†aucam from sütakam. Space does not permit me to discuss here the unresolved question concerning the different periods of å†auca for different varñas, the length of time increasing for lower varñas. See Dumont (1980: 70); Mines (1990); Orenstein (1970); Tambiah (1973: 208-18). 19 The only instance when such a meaning appears possible is in Gautama: na tadupaspar†anåd å†aucam –“No impurity is contracted through his (i.e., a child before initiation) touch” (G 2.3). Stenzler’s edition reads here å†aucam, while the commentators Haradatta and Maskarin appear to read a†aucam, glossing it with a†ucitvam (“state of being impure”). Here also my suspicion is that the provision refers to a child a relative of whose has died or who has touched a corpse; even in that condition his touch, unlike that of an adult in a similar condition, does not create å†auca. The two forms å†auca and a†auca (both permitted by Påñini 7.3.30) occur in the Dharma texts, although the former is much more frequent; orthographic confusion often makes it difficult to isolate the original reading. But both forms generally have the same meaning. The term å†aucin (“a person in a period of å†auca”) is found only in Vkh 2.14. 20 In the Dharma literature this (the Parasmaipada) is the dominant form. The Åtmanepada form †udhyate is much less common: B 1.13.19; Va 3.67; M 5.108.

230

LANGUAGE, TEXTS AND SOCIETY

found in the same verse of Vasiß™ha: rajaså †udhyate nårœ nadœ vegena †udhyate –“A woman is purified by her menstrual flow, and a river by its current” (Va 3.58; cf. M 5.108). The verb has a wide semantic compass, referring to the purification from bodily stain or ritual impurity, the release from sin, the acquittal from a criminal charge, and the discharge from a crime by paying the penalty. The verb is used very frequently, and this may be due in part to the fact that the term †uci does not have a verbal counterpart with the same meaning.21 Thus, when an author wants to use a verb for purification, his choices are limited to this and to the less frequently used verb püyate (see section 4). When the texts want to indicate an act of purification in a transitive sense, the only term they use is the causative †odhayati (“make someone/thing pure”). It is totally absent in the vocabulary of the older texts, appearing for the first time in Manu (Fig. 2: B). Although used much less frequently, it has about the same semantic range as †udhyati. Some exceptional usages include neutralizing the effects of poison in food (M 7.218), clearing roads of dangers such as robbers (M 7.185), and deducting expenses in legal calculations (Y 2.122; 2.146). ‡odhana, the nominal derivative from †odhayati, appears for the first time in Vasiß™ha (Fig. 2: C) and refers to a means (see pavitra and påvana) or the act of purification with the same semantic range as the verb.

21 ‡uci is, of course, derived from the verbal root †uc, but the verb itself did not undergo the semantic development of †uci and is never used in the Dharma literature for becoming pure.

Caste and Purity

231

Figure 2 A. †udhyati 1. Person a. purification: G 14.6; 14.30; 22.6; B 1.8.2; 1.11.8 (twice); 1.11.41; 3.1.27; 3.8.12; Va 3.58; 3.60; 4.24; 4.30; 18.16; 20.21; 27.12; M 3.132; 5.63; 5.64; 5.65; 5.66; 5.72; 5.76; 5.78; 5.83(twice); 5.85; 5.87; 5.88; 5.99; 5.101; 5.102; 5.103; 5.108 (twice); 5.109 (twice); 11.199; 11.201; 11.202; Y 1.21; 3.20; 3.277; Vi 22.11; 22.39; 22.58; 22.72; 22.74; 22.77; 22.85; 22.87; 22.91; 22.92; 62.9. b. moral/legal: G 20.10; 20.15; 23.1; 23.11; 23.19; B 1.8.2 (twice); 2.1.40; 3.1.27; 3.3.17; 4.1.4; 4.1.21; 4.2.14; 4.5.14; 4.5.27; 4.5.28 (twice); Va 3.60 (twice); 21.12; 25.6; 26.5; 27.14; 27.15; 28.3; 29.16; M 5.107; 5.109; 11.46; 11.83; 11.100; 11.103; 11.123; 11.141; 11.146; 11.149; 11.162; 11.193; 11.249; 11.254; 11.257; Y1.72; 3.246; 3.248; 3.249; 3.258; 3.262; 3.263; 3.280; 3.283; 3.287; 3.289; 3.301; 3.311; Vi 22.90; 22.91; 22.92; 28.50; 35.6; 36.8; 40.2; 50.47; 54.28. 2. Non-person a. purification: B 1.9.7; 1.10.1; 1.13.19; 3.1.27; Va 3.57; 3.58 (twice); 3.59; 3.63; M 5.108 (twice); 5.112; 5.122; 5.123; 5.124; 5.125; 5.126; Y 1.186; 1.194; 1.197; Vi 22.91 (twice); 23.2; 23.38; 23.40; 23.41; Vkh 3.2. B. †odhayati (and verbal derivaties) 1. Person a. purification: Va 27.13; M 11.160; Vkh 2.13; 3.3. b. moral/legal: M 8.202; 11.226 (twice); 11.258; Y 2.269; N Må 2.39; 20.37; 20.40. 2. Non-person a. purification: M 5.108; 7.218; 9.282; 9.283; Y 3.32; Vi 22.91; 23.44; Vkh 3.3 (thrice). b. legal: M 7.185; Y 2.122; 2.146; N Må 3.5; 1.80; 19.26; 19.27. C. †odhana 1. Person a. purification: Va 25.12; M 11.125; 11.143; 11.156; 11.160; 11.165; 11.200; Y 3.24; 3.34; Vi 41.5 (twice); 50.49 . b. legal: M 1.115; 9.253; Y 2.122; 2.146; Vi 11.9; N Må 2.39; 12.76; 15-16.6; 20.6; 20.31. 2. Non-person: N 5.6; 19.17. D. †uddhi 1. Person: a. purification: G 24.4; B 1.8.3; Va 4.20; 27.10; M 5.21; 5.57; 5.61; 5.67(twice); 5.71; 5.100; 5.105; 5.134 (twice); 5.136; 6.30; 6.69; Y 3.14; 3.20; 3.25; 3.31; 3.32; 3.34; 3.62; Vi 22.21; 22.35; 22.88; N 5.33. b. purity: M 9.9; 11.160; Vkh 1.9; 3.3 (four times). c. moral/legal: B 1.8.52; 4.3.7; Va 25.1; 26.15; M 11.53;

232

LANGUAGE, TEXTS AND SOCIETY

11.72; 11.89; 11.117; 11.138; 11.163; 11.164; 11.181; 12.105; Y 1.77; 2.94; 2.95; 2.107; 2.109; 2.111; 3.159; 3.220; 3.243; 3.244; 3.248; 3.250 (twice); 3.253; 3.265; 3.268; 3.274; 3.326; Vi 11.12;13.7; N Må 3.6 (twice); 1.102; 1.222; 6.19;7.2; 7.4;15-16.7; 15-16.25; 20.7; 20.24. 2. Non-person: a. purification: G 1.29; M 1.113; 5.57; 5.110; 5.111; 5.115; 5.116; 5.117; 5.119(twice); 5.121; 5.146; Y 1.183; 1.188; 1.189; 1.190; 1.191; 3.32; 3.60; Vi 22.93; 23.7; 23.39; 23.46; 23.57; 96.8. b. legal: Y 2.92; N Må 1.3; Må 2.44; 1.78 (twice); 18.43 (twice). E. †uddha 1. Person: a. purified: G 20.10; B 1.7.2; 1.11.32; M 2.160; 5.77; 11.190; 11.242; Y 3.159; Vi 22.73; 54.32; 99.18; b. pure: B 1.9.1; Va 27.15; M 5.129;10.76; Vi 23.48; Vkh 3.6; 3.8; 3.11 (four times); 3.12; 4.3. c. moral/legal: B 3.6.10; 3.8.27; 3.9.10; 4.2.14; 4.2.16; 4.5.32; 4.7.2; 4.7.4; 4.8.12; M 7.219; Y 2.102; 2.113; 3.159; Vi 10.12; 11.8; 12.6; 13.5;14.5; N Må 2.39; 1.133; 20.12 (twice); 20.20; 20.26; 20.27; 20.39. 2. Non-person a. purification: Vkh 2.15. b. pure: A 1.19.7; Va 27.15 (white); M 5.128; 10.90; 12.27; Y 3.72; Vkh 2.12; 2.13; 2.14; 3.1; 3.5; 3.8. c. moral/Legal: M 4.11; 8.201; 9.279; N Må 3.6; 1.102; 7.8. F. a†uddha 1. Person: Va 27.15; M 5.58; Vi 11.8; 12.6; 14.5; 57.4; Vkh 3.11 (twice); N 17.5. 2. Non-person: M 4.127; Y 2.266; N 7.8.

2.1 ‡uddhi We have seen above that the texts make a clear distinction between the purification of persons and things by using †auca for the former and †uddhi for the latter. ‡uddhi, the nominal derivative of the verb †udh, however, is used more generally with regard to both persons and things, but with rare exceptions it normally has a dynamic meaning, referring to the act or process of purification/cleansing rather than to a state of purity. The dynamic meaning is even more pronounced in †uddhi than in †uci. Unlike the latter, †uddhi is a noun and refers directly to the act of purification, the getting rid of impurities, whereas †uci indicates a person “becoming pure”. Even in moral contexts †uddhi refers to the elimination of immoral or sinful qualities through an appropriate penance rather than to a state of moral purity. Thus, in introducing the chapter on secret penances, Vasiß™ha (25.1) states: “I will explain fully the purification (†uddhi) of all sinners whose guilt has not been made public”.

Caste and Purity

233

In legal contexts †uddhi is used with reference to the discharge of a debt (Y 2.94; N 1.102), the exoneration of an accused man (Y 2.95; N 1.222; 6.19), the settlement of a dispute (N 1.78), and the establishment of the authenticity of a legal document (Y 2.92)–all appearing in late texts. The fact that †uddhi always indicates purification rather than purity may be the reason why its negative form, a†uddhi, is totally absent in the Dharma vocabulary.22 It is clear that although it is possible to speak of impurity, it makes little sense to speak of non-purification. 2.2 ‡uddha Like †uddhi, the past participle †uddha is used with regard to both persons and things, both in ritual and moral/legal contexts (see Fig. 2: E). Unlike †uddhi, however, it is used in the sense of both “purified” (dynamic) and “pure” (static). The former is the most common. Thus, for example, Baudhåyana (1.11.32) states: “When someone accidentally touches the corpse of an outsider, he becomes pure (†uddha) immediately after taking a bath with his clothes on”; and Gautama (20.10): “An excommunicated man may be purified (†udhyet) by performing a penance, however, and when he has been so purified (†uddha), they should fill a golden pot with water from a very sacred lake or from a river and make the man take a bath with the water from that pot.” The term, however, often simply means pure or white (Fig. 2: E.1.b, 2.b), and in this sense it is used predicatively. Thus, “almsfood is pure” (†uddhå bhikßå) and may be eaten (A 1.19.7); the hand of an artisan is always “pure” (†uddhaΔ kåruhastaΔ, M 5.129). 2.3 A†uddha As in the case of a†uci, the negative a†uddha often has a static meaning, although the dynamic meaning is quite comon. The first use of this negative is found in Vasißtha (25.15) and has a dynamic meaning: “When a man consumes barley grains in accordance with the rules, he becomes visibly pure (†udhyati): if he has become pure (lit. if his being has become pure: vi†uddhabhåva) the grains remain white (†uddha), whereas if he has not become pure (a†uddha) the grains become discolored.” When an initiated child dies, the relatives become impure, a†uddha (M 5.58). Both persons and things are said to be a†uddha, in the sense of both ritual impurity and moral/legal culpability (see Fig. 2: E). 3. Prayata Prayata is the past participle of the verb yam with the prefix pra and in the vedic literature meant simply “outstretched” or “pre22 The negative occurs only once, in the compound †uddhya†uddhi (Vkh 1.9) in the context of holy ascetics (Paramahamsas) who eliminate all binary opposites. This is a unique usage and may be an artificial formulation limited to this context.

234

LANGUAGE, TEXTS AND SOCIETY

sente/offered”. This meaning is still present in the expression prayatå∞jali (“with his cupped hands outstretched”: B 2.1.35), but even here we see that this gesture is done within a ritual context calling for a proper mental and bodily preparation.23 In the Dharma texts, however, the term came to mean “pure” (see Fig 3), although, as Gonda (1961-62) has shown, it has the much wider meaning of proper internal and external preparation for a solemn rite. Figure 3 A. prayata 1. Person: a. purified: A 1.11.23; 1.15.2; 1.15.17; 1.15.23; B 1.11.40; 2.7.2; 2.7.4; 3.6.7; 4.2.13; Va 26.14; M 2.222; 4.145; 4.146; 5.86; Vi 52.2. b. pure: A 1.15.3; 1.15.13; 1.16.9; 2.3.1; 2.17.4; B 2.7.2; M 2.183; 2.185; 3.216; 3.226; 3.228; 3.258; 4.49; 5.145; 8.258; 11.258; Vi 66.15. 2. Thing: a. purified: A 1.17.11. b. pure: B 2.17.37. B. aprayata 1. Person: A1.14.18; 1.14.19; 1.14.20; 1.15.8-9; 1.15.13; 1.15.18; 1.16.14; 1.16.21; 1.16.22; 1.31.4; 2.15.19; B 1.3.29 (twice); 1.3.30; 2.7.2; M 5.142; 11.153; Vi 23.54. 2. Thing: A1.16.21; 1.29.14; B 2.2.22. C. pråyatya 1. Person: A 1.31.19. 2. Thing: B 1.9.11; Va 3.56. D. apråyatya 1. Person: A 1.11.25; Vkh 2.9.

The oldest dharma†åstric writer, Åpastamba, uses prayata, as well as the negative aprayata, much more frequently than any other term with regard both to persons and things. The term is absent in Gautama, used infrequently by Baudhåyana and Vasiß™ha, and falls into relative disuse in later literature, except in standard expressions such as prayatåtman (“with himself ritually prepared or purified”: e.g., Va 26.14). The use of the term with reference to things is limited to the early texts, especially Åpastamba (see Fig. 3: A.2, B.2, C.2). 4. Püyate Derivatives from the verbal root pü are used much less frequently in the Dharma literature than those from the root †udh. The active voice punåti (“purifies”) is never used outside of vedic mantras, except by Vasiß™ha, who uses it no less than five times per23

For an extended discussion of the range of meanings of prayata, see Gonda (1960-61).

Caste and Purity

235

haps with an intent to archaize his language.24 More frequent is the passive püyate (“is purified”), and it is used with the same semantic range as the verb †udh with reference to both persons and things, and with regard to both ritual purification and the expiation of sins and crimes (Fig. 4: A). The past participle püta also has the same broad range of meanings but, unlike its counterpart †uddha, always has the dynamic meaning of “being purified” rather than simply “pure”. This dynamic meaning is evident also in the rarely used negative apüta, which always means something or someone “not purified” or “not yet purified” rather than simply “impure”.25 Figure 4 A. püyate (including derivatives and negatives) 1. Person: Va 23.22; 28.10; B 3.7.1; 3.7.18; 3.9.10; 4.2.12; 4.2.16 (twice); 4.7.10; M 2.62; 8.257; 8.311. punanti: Va 6.3; 6.5; 22.10; 26.4; 28.15. 2. Non-person: B 2.17.18; 3.1.13; Va 3.61; Vkh 3.4. B. pavitra 1. Rite or mantra: A 1.2.2 (thrice); 1.26.7; 1.27.2; G 25.7; 26.10; B 1.2.16; 1.2.17; 2.7.2; 2.7.4; 2.14.5; 4.1.22; 4.2.16 (twice); 4.8.3; 4.8.4; Va 22.10; 23.47; 25.3; 25.4; 28.10; M 3.256; 11.225; Vi 46.25; 56.1; 64.36; 64.40; 72.5 (virtue); 100.2; Vkh 3.8. 2. Object: B 1.9.9; 1.9.10; 2.4.; 2.8.11; 2.11.24; 2.17.11; 2.17.34 (twice); 2.17.37 (twice); 3.2.7; 3.2.17; 3.5.1 (twice); 3.6.5; Va 11.35; 14.24; 25.4; 28.4; M 2.75; 3.210; 3.223; 3.235; 3.256; 5.127; 6.41; 10.102; 11.85; Y 1.240; 3.325; 1.251 ; Y 1.226; 1.230; Vi 23.47; 23.57 (cows); 48.6; 48.17; 79.16; Vkh 2.6; 2.8 (twice); 3.6; 3.6; 3.8. 3. Purity: B 3.7.4. C. påvana 1. Person: A 2.17.22; G 15.27; 15.28; B 2.14.2; Va 3.19; 11.20; M 3.183; 3.184; 3.186; Y 3.306; Vi 1.57; 83.1; 99.4. 2. Non-person: G 19.12; B 1.2.16; 1.3.43; 3.10.10; 4.5.9; 4.5.25; 4.5.29; Va 1.16; 22.9; 25.11; M 2.26; 11.85; 11.177; Y 1.281; 2.83; Vi 8.16.

4.1 Pavitra Of the nominal derivatives of the verb pü, the most common is pavitra (Fig. 4: B). It refers to any agent or instrument of purification. Thus, a water strainer is called pavitra, as also the two blades of Ku†a grass between the fingers during rituals, purificatory verses and rites, and penances. Other means of purification, such as a woman’s menstrual flow, are also called pavitra. 24 25

Va 6.3; 6.5; 22.10; 26.4; 28.15. G 25.6; M 2.40; 8.330 (re. unhusked grain).

236

LANGUAGE, TEXTS AND SOCIETY

Here also we note the focus on regaining lost purity, this time with reference to instruments that impart purification. 4.2 Påvana The term påvana (“purifying” or “imparting purity”) is derived from the causative of pü (påvayati, “to make someone/thing pure”). In the Dharma literature it has a somewhat restricted meaning (Fig. 4: C). Although it is used a few times with reference to actions and rites that impart purity (in the same ways as †odhana: Fig. 4: C.2), påvana is used most frequently as a technical term with reference to specially holy/learned Brahmins who are called paõktipåvana, “purifying those alongside whom they sit” during a meal.26 5. Ucchiß™a Ucchiß™a is the past participle of the verb †iß with the verbal prefix ut and means “left” or “left over”. Its most common meaning is with reference to food that is left over after a person has eaten (Fig. 5: 2). Such food may be either left on the plate or attached to a persons hands and lips. In general ucchiß™a is viewed as impure and causes anyone coming into contact with it to become impure. In this sense, ucchiß™a is most frequently used as a neuter substantive (ucchiß™am).27 In an adjectival sense, however, the term refers to a person who is rendered impure by coming into contact with ucchiß™a food. Thus, after a meal a person remains ucchiß™a until he or she has performed the required purification (†auca). There are, however, other extended meanings of the term. Medhåtithi (on M 4.80), an early commentator of Manu, isolates four possible meanings of ucchiß™a: (a) because of contact with the inside of the mouth while eating, the eater, the eaten food, and the plate from which one eats become ucchiß™a; (b) food left on the plate after someone has eaten off it is ucchiß™a; (c) as also what is left in the dish from which food has been served to someone; (d) and food left in the pot after people have been served; and (e) a person is ucchiß™a after voiding urine or excrement and before purification. According to Medhåtithi, the primary meaning of the term is the first, (a), resulting from food and fingers come into contact with the inside of the mouth.28 26 A 2.17.22; G 15.27-28; B 2.14.2; Va 3.19; 11.20; M 3.183, 184, 186; Vi 83.1. The opposite of paõktipåvana is paõktidüßaña, a man who defiles a company (A 2.17.21). 27 For a detailed discussion of this term, see Malamoud (1972). Some types of ucchiß™a are good and can be eaten. Generally such leftovers belong to a person superior to oneself. Thus, a wife may eat the leftovers of her husband, and a student the leftovers of his teacher (A 1.4.1-11; G 2.31-32). Leftovers of a sacrifice or an offering to a god are especially holy. 28 There is an interesting rule given by Åpastamba (1.31.22): “He should not give his leftovers to someone who is not a Brahmin. If he does so, he should pick his teeth, place what he has picked from his teeth on the leftovers, and then give it.” Here ucchiß™a as (c) or (d) has to be converted to ucchiß™a as (b) if leftover food is to be given to an inferior persons.

Caste and Purity

237

Figure 5 ucchiß™a 1) Person: a. from food: B 1.8.28; M 2.56; Vi 68.36; 70.17. b. from bodily discharge: A 1.16.11; G 1.41; B 1.8.27; 1.8.29 (ucchiß™œ); 1.8.51; 1.10.34; Va 3.27; 3.42; M 4.75; 4.82; 4.109; 4.142; 5.141; 5.143 (both thing and person?); Vi 23.53; 23.55; Y 1.155; Vkh 2.14 (twice); 3.2; 3.3. c. unclear source: G 1.28; B 1.13.26; 1.14.1; 1.14.17. 2) Food: A 1.3.27; 1.3.37; 1.4.1; 1.4.2; 1.4.5; 1.4.11; 1.6.36; 1.7.27; 1.7.30; 1.15.23; 1.17.3; 1.21.17; 1.31.22; 2.9.7; 2.18.11; 2.20.2; G 2.32; 10.59; 17.17; B 1.3.35; 1.3.36; 2.1.26; 2.8.10; 3.6.5; 3.8.10; Va 11.10 (twice); 11.21; 11.22; 11.23; 14.20; 14.21(twice); 18.14; 23.9; 23.11; M 2.56; 2.209; 3.245; 3.249; 4.80; 4.211; 4.212; 10.125; 11.26; 11.152; 11.159; Vi 21.3; 21.15; 28.11; 28.33; 51.10; 51.46; 51.47; 51.50; 51.51; 51.52; 51.53; 51.54; 51.55; 51.56; 54.19; 71.49; 73.17; 73.25; 81.22; 91.18; Y 1.33; 1.154; 1.162; 1.167; 1.168; 1.209; 1.242; 1.257; 3.37; Vkh 3.2; 3.4; N 5.6 3) Objects: B 1.8.32; 1.13.27; 1.14.2; Va 3.43; M 11.148; Vi 48.20; Vkh 3.3; N 1.57.

6. Medhya In its vedic usage the term medhya referred to animals and substances suitable for use in a sacrifice. In the Dharma literature, it is used with the wider meaning of “pure” (Fig. 6: A). The old connection with sacrificial substances, however, is evident because the term is used most frequently with reference to food or sources of food. Other items called medhya includes the bodily parts above the navel, vessels, and soil. Persons are generally not called medhya. The only exception, interestingly, is women, who, like food, are to be enjoyed by men.29 6.1 Amedhya The negative of medhya, just as the negative of †uci, developed a technical meaning. With rare exceptions amedhya is used as a neuter substantive to refer to impure or dirty substances, especially bodily excreta. Baudhåyana (3.8.17-18), for example, says that one should not look at urine and feces, and immediately adds, “if he sees amedhya”, clearly equating the latter with bodily excreta. This specialized meaning is not only documented in the Dharma literature but also singled out for comment by medieval commentators. Vij∞åne†vara, commenting on Yåj∞avalkya 1.191 (see also on Y 2.214), defines amedhya as sarœrajå malåΔ, “dirt originating from the 29 See Va 28.6; Y 1.71. The verb bhuj is used often with regard to both food and women: see, for example, Vij∞åne†vara on Y 1.71.

238

LANGUAGE, TEXTS AND SOCIETY

body”. Anticipating Mary Douglas’s (1966) definition of dirt as substances that violate boundaries of cultural categories (i.e., “matter out of place”), Vij∞åne†vara explains that the impure nature (amedhyatva) of amedhya substances results from the fact that they have come out of the body; only those substances that have come out of the body are amedhya, not those that remain in their respective places (amedhyatva∫ caivam eßå∫ [...] dehacyutånåm eva na svasthånåvasthitånåm). So, a substance is not impure in itself but only in so far as it has been displaced and has crossed a boundary. Figure 6 A. medhya 1. Food sources: A 1.17.31; 2.16.28; B 1.9.1; 1.9.2; Va 14.46; M 5.54; 5.129; 6.5; 6.11; 6.13; 11.153; 6.12; M 5.133. 2. Non-food items: B 1.7.4 (twice); 1.13.19; M 5.133; Vkh 3.4; N 18.41. 3. Parts of human body: B 1.10.19; Va 28.6; 28.8; 28.9; M 1.92; 5.132; Y 1.71 (twice); 1.194; 1.195; Vi 23.40; 23.51. 4. Actions/rites: G 19.13; B 3.10.11; Va 22.11. B. amedhya 1. As substantive: a. Bodily substance: A; 1.16.24; 1.16.25; 1.16.26; G 1.42; 9.12; Va 3.48; M 4.56; 5.126; Y 1.191; Vi 22.39. b. Filth: A 1.16.14; G 9.15; B1.9.4; 1.9.10; 1.10.1; 1.14.18; 1.15.31; 2.2.36; 2.8.6; 3.8.18; M 2.239; 4.53; 5.5; 5.128; 9.282 (twice); 11.96; 12.71; Y 1.148; 2.213; Vi 23.43; 43.41; 51.36; 51.41; 71.32. 2. As adjective a. Re. food/beverage: G 20.4; 23.23; Vi 22.84. b. Re. body below navel: B 1.10.19; M 5.132; Vi 23.51. c. Re. people/animals: A 1.17.5.

Vij∞åne†vara’s astute observation is supported by the usage of the early Dharma texts. Åpastamba (1.16.23-24), for example, says that food “in which there is a hair or some other amedhya” is unfit to be eaten, indicating that amedhya must refer to other bodily substances such as hair. This is confirmed by the fact that commentators instinctively take amedhya to be some bodily substance, as when Såyaña (on Taittirœya Årañyaka 10.1.13) glosses the term with niß™hœvanådi, “substances such as saliva”. 7. Mala The term mala is a noun indicating dirt or impure substance (Fig. 7: A). The early Dharma texts use the term only within the technical expression malavadvåsas, “one with dirty clothes,” with

Caste and Purity

239

specific reference to a menstruating woman.30 The term is used extensively for the first time by Manu. Although it can refer to spiritual “stains” or sins, in its most common meaning mala is a synonym of amedhya and refers to bodily excretions. That is the meaning in the well-known statement listing the twelve malas of human beings: “Oily exudations, semen, blood, fat, urine, feces, snot, ear-wax, phelgm, tears, discharge of eyes, and sweat –these are the twelve malas of men” (M 5.135; Vi 22.81). Vij∞åne†vara (on Y 1.191) cites this list in explaining the meaning of amedhya, indicating that he saw the two terms as synonyms.31 Mala, however, can also refer to the “stain” of moral infractions (Fig. 7: A.1.c). Authors sometimes play on this double meaning. Manu (11.94), for example, shows that both liquor and sin are mala, “dirt” and “sin”, respectively: “Liquor is the dirt (mala) derived from food,32 and sin is also called dirt (mala). Therefore, Brahmins, Kßatriyas, and Vai†yas should not drink liquor.” The negatives amala and nirmala are used as adjectives to indicate the absence of mala especially in a moral sense (Fig. 7: B). Thus nirmala means a persons who is “immaculate”, free from the taint of sin. Figure 7 A. mala 1. As substantive —dirt a. Abodily filth: M 4.220; 5.132; 5.134; 5.135; Vi 22.77; 22.81; 23.1; 23.40; 23.51; Vi 51.2; 64.18; 96.47; Y 1.194. b. other impurities; M 6.71; 11.93; c. moral: M 2.102; 11.70; 11.93; 11.101; 11.107; 11.125; Vi 33.5; 41.4; 41.5; 43.30; 44.9; 99.18; N 15-16.15. 2. As adjective—dirty: A 1.9.13; G 9.3; Va 12.5; M 4.34; Vi 63.35; 69.11; 69.12; 71.9; 71.24; Y 1.70; Vkh 3.2; 3.14; 3.15; N 9.7; 15-16.15. B. amala/nirmala 1. Purified re. person (always moral/legal): Va 19.45; 26.6; M 8.318; 11.250; N 19.55. 2. Pure re. person: Vi 47.10. 3. Bright/sotless re. thing: Vi 99.10 (twice).

30

A 1.9.13; G 9.3; Va 12.5 (malinavåsas); M 4.34. The vedic text Kå™haka Sa∫hitå (34.12) gives a similar list of twelve impurities of a man and calls them amedhya. Såyaña, commenting on Aitareya Bråhmaña 7.13.7, explains mala as †ukra†oñita (“semen and menstrual blood”), seeing in the term a reference to the householder’s å†rama connected with childbirth. 32 Liquor is considered dirt possibly because it results from fermentation. 31

240

LANGUAGE, TEXTS AND SOCIETY

II. Conclusions I started this investigation with four questions: does a term refer: (a) to a person or to a thing; (b) to a condition or to a transition; (c) to purity or to the areas of morality or criminal law; or (d) to an individual or to a group? We are now in a better position to answer these questions. The vocabulary of pure/impure in the Dharma literature makes a clear distinction between persons and things. Although a variety of objects and animals are characterized as impure, the main focus is on bodily discharges, those oozing substances that violate the boundary of the human body. With respect to persons, the vocabulary clearly indicates that the focus is not on any permanent, or even transitory, condition of purity but rather on the transition from impurity to purity, on the recovery of lost purity. The dynamic meaning dominates the use of †uci and †uddhi. There is no discussion about how one may remain in a condition of permanent purity, although the texts do talk about avoiding sources of moral and ritual pollution. To remain permanently pure is not only impossible (one has to eat and go to the bathroom), but also undesirable, for it would entail not having sex or children, both sexual activity and childbirth being sources of pollution. Most of the terms for pure/impure are used both with respect to ritual purity and in the context of moral and criminal law. In the Dharma literature these three areas are not compartmentalized. Finally, we see no instance when a term for pure/impure is used with reference to a group of individuals or to a varña or caste, the only exception being people who have fallen from their caste due to grievous sins; these are often called a†uci.33 Groups outside the established society, especially the Cåñ∂ålas, are subsumed under the category of the “fallen”. In general the Dharma texts place both Cåñ∂ålas and grievous sinners in the same group and treat them alike. These people are impure precisely because they have breached the boundary of varña society; they have been excreted out of the social body and are hence impure, amedhya. Social interaction with these people is forbidden, and any contact makes a person within society impure (morally and ritually). But the overwhelming focus of the vocabulary concerns individuals, irrespective of their caste/varña affiliation, who have become impure and are in need of recovering their lost purity. A person of any varña, including a ‡üdra, can become a†uci, impure, and can recover purity (become †uci) by employing the appropriate purification (†auca). 33 At M 1.92 there is an oblique reference to the Brahmin coming from the mouth (mukha) which is the “purest” (medhyatama) part of the body. But no explicit conclusion is drawn from this fact that the Brahmin is purer than other varñas, only that he is the “lord of the whole creation” and that gods consume offerings through the medium of the Brahmin’s mouth.

Caste and Purity

241

Dumont is correct in his assessment that the ideology of varña is not based on purity. If it were we should expect to find at least some comment on the relative purity/impurity of the different varñas. What is even more important is that the ideology of purity/impurity that emerges from the Dharma literature is concerned with the individual and not with groups, with purification and not with purity, and lends little support to a theory which makes relative purity the foundation of social stratification. The mythical legitimization of the varña division is, of course, contained in the Purußa Hymn of the R¢g Veda (10.90). Some may think that the connection of the different varñas with different parts of the body, especially the association of the ‡üdra with the feet, may support a theory of the relative purity of varñas. But this conclusion is never drawn in the Dharma literature even in contexts when such a conclusion would have buttressed the author’s argument. Baudhåyana, for example, appeals to the Purußa Hymn to demonstrate that ‡üdras should serve the higher varñas, “for they were created from his feet” (B 1.18.5-6). Feet here symbolizes service not impurity. The evidence of these ancient texts gives greater support to the theory of impurity proposed by Mary Douglas (1966) than to Dumont’s theory of caste as based on the gradation of purity.34 Scholars have connected impurity with hygiene, with morality, with the separation of spirit and matter, and so forth. I think Mary Douglas is right to reject such definitions, as well as the temptation to reify the pure and the impure, to see these as somehow descriptive adjectives like heavy or blue, as indicating objective qualities inhering in substances. Pure and impure are relational and evaluative concepts–they are related to an ideology that establishes categories and fixed boundaries and they evaluate actions, objects, persons, and social interactions in relation to those categories. In this sense, with William James (Douglas 1966: 164) and Freud (Kubie 1937: 390), we can define impurity or dirt as “matter out of place”. If we can abstract pathogenicity and hygiene from our notion of dirt, we are left with the old definition of dirt as matter out of place. This is a very suggestive approach. It implies two conditions: a set of ordered relations and a contravention of that order. Dirt then, is never a unique, isolated event. Where there is dirt there is a system. Dirt is the by-product of a systematic ordering and classification of matter. (Douglas 1966: 35)

Concern for impurity translates into concern for maintaining the integrity of boundaries, both physical and classificatory, which in turn relates to the concern for maintaining social boundaries. The human body becomes the locus for expressing all these concerns, especially the concern for maintaining purity. 34 Quigley (1993: 45-46) is right in his assessment: “For Dumont, as we have seen, the opposition of the pure and the impure is the principle of hierarchy in ‘the caste system’. As it stands, this formulation is problematic because the opposition of pure and impure is a universal feature of human societies. The reason for this has been brilliantly explored by Mary

242

LANGUAGE, TEXTS AND SOCIETY

The body is a model which can stand for any bounded system. Its boundaries can represent any boundaries which are threatened and precarious. … We cannot possibly interpret rituals concerning excreta, breast milk, saliva and the rest unless we are prepared to see in the body a symbol of society, and to see the powers and dangers credited to social structure reproduced in small on the human body. (Douglas 1966: 115).

The most common examples of breaching boundaries concern the body. When internal substances break the bodily surface and ooze out, then the person becomes a†uci, impure. Other persons coming into contact with such amedhya or unclean substances also become a†uci. This is the type of impurity that has been subject to most scholarly scrutiny. There are, however, other boundaries, and the impurities associated with their breach provide further insights into the Dharma ideology of impurity. Let me focus on just two: the spatial boundary of the village (although interesting information comes also from the boundary of the house) and the temporal boundaries of day, month, and year. In this context I want to deal with a subject totally ignored by scholars but looming large in the Dharma literature, namely the events causing anadhyåya, or the suspension of vedic recitation. Here are some of Åpastamba’s (1.9.14-19) prescriptions: there is anadhyåya in a village when a corpse is brought or a Cåñ∂åla walks into its boundaries. Likewise when outsiders, even respectable people, come into a village there is anadhyåya for the duration of that day. This prescription mirrors another rule (A 1.10.11) that forbids collective vedic recitation for a day when a fellow student has gone out of the village. When a newcomer arrives or a member of a group leaves, its boundary is breached and the community disrupted; the two events are comparable to a birth and a death, both of which carry a period of impurity (å†auca) and also anadhyåya for those within the affected group. I think death causes a period of impurity to the community of which the dead person was a part precisely because the community has been disrupted and has to be reconstituted over a period of time. This is the reason, I believe, why in the Dharma†åstras death impurity and birth impurity are always treated together in spite of the enormous difference between the two events. This is also the reason why the deaths of some individuals, such as infants, outcastes, and ascetics, do not cause å†auca, because they are not full and integral parts of the community. An analysis of the times when recitation is suspended provides even more interesting data. Whenever normalcy is violated, there is anadhyåya. So, for example, when there is a solar or lunar eclipse, an Douglas in her book Purity and Danger, and it is regrettable that she did not exploit her own insight further when she wrote the introduction to the first English-language translation of Homo Hierarchicus.”

Caste and Purity

243

earthquake, or a whirlwind, recitation is suspended. Interestingly, lightning, thunder, clouds, and rain cause a suspension only when they appear out of season. In the rainy season they are normal and are to be expected; therefore, they do not cause ana- dhyåya. A large number of time-related suspensions happen when one time unit ends and another begins–that is, the time of twilight or sandhyå when time boundaries are breached. Thus, there is ana- dhyåya during the morning and evening sandhyå, on new- and full-moon days, at the beginning of seasons, at spring and summer festivals, and even on the day that opens and concludes the annual course of study. The boundaries between time zones are marginal and thus ambiguous and inappropriate. A final comment on boundary and margin: as the margin–the betwixt and between–is a zone where normal activity is suspended, so people temporarily or permanently in such a marginal zone are not subject to impurity, showing once again that impurity has to do with structure and thus concerns only those within a structure. So, for example, the death of a family member does not make a brahmacårin (student), a dœkßita (a man consecrated for a sacrifice), an officiating priest, or an ascetic impure. Likewise, children before their vedic initiation (upanayana) are exempt from all purity rules (A 2.15.19-25; G 2.1-6). Inversely, the death of a family member who is not fully incorporated into the corporate structure does not cause impurity or cause minimal impurity to its living members–thus there is no impurity when an infant dies. Another interesting point about marginal people is that they are not permitted to engage in purificatory acts while they are in that state: thus menstruating women are not permitted to bathe or comb their hair before the conclusion of that period; people in mourning cannot bathe; and, according to some, a student is not allowed to bathe or brush his teeth (A 1.7.1; G 2.13). These are purificatory acts signaling the end of the period of impurity and, hence, are inappropriate during the period of impurity. What stands out when we examine the Dharma vocabulary on pure/impure is the preoccupation with the body, a preoccupation that borders on scrupulosity and anxiety–†aõkå. Madan (1988, 32, 61) speaks frequently about the †aõkå of his Kashmiri pandits, an anxiety relating primarily to the observance of purity rules. Madhav Deshpande (1993a: 41) refers to a suggestive euphemism prevalent in Maharashtra: going to the bathroom is †aõkå, urinating is called laghu†aõkå (“short anxiety)”) and defecation dœrgha†aõkå (“long anxiety”). Anxiety is connected with intentionality. One recurring problem with scholarly treatments about purity in India is that it is often reified and turned into a self-existing reality. An important component of the Dharma ideology of impurity serves to undermine such a notion –and that is the connection between impurity and intention-

244

LANGUAGE, TEXTS AND SOCIETY

ality. Intention plays a central role in ethics, and some scholars, including Mary Douglas, contrast ethical and purity rules precisely on this point–intention has a role in ethical rules but not in rules of purity (Douglas 1966, 160). This, I believe, is mistaken. In Dharma discourse not only are terms for impurity and immorality often interchangeable, but intentionality is central also to rules of impurity, although its role there is in many ways different from its role in ethics. Let me cite a few examples in support. First, it is a rule repeated frequently in the Dharma†åstras that ignorance makes a thing pure. Let me repeat–lack of awareness (adr¢ß™a) is given as one of the three means that makes a thing pure.35 If I have not seen that my rice has been licked by a dog, then, as far as I am concerned, my rice is pure. That this is not merely a †åstric fantasy is shown by a story that Dumont (1980, 383) himself relates. At a †råddha a boy from the dead man’s family mischievously touches the plate of one of the assembled Brahmins. When the Brahmins are about to abandon the dinner, one of them remarks that there were no witnesses to the touching. So they proceed to eat, and the boy’s mother admonishes him never to tell this to anyone. Surely, if purity were an objective reality my ignorance should have nothing to do with its existence. In a similar manner, the death of a relative away from home makes the relatives impure only when they hear about it, and then for a shorter period of time than if he were to die at home. Second, there is what I would call “stipulative purity”– that is, rules stipulate that certain people or objects are pure by definition. So, the hand of an artisan or cook, items sold in the market, the mouth of a woman, the beak of a bird that makes a fruit fall, a dog when catching a deer, and the excreta of a child, and so forth –these are all reckoned pure by definition (Vi 23.48f). The earth is also pure by definition; so it cannot carry impurity from one person to another. Thus if a pure man and an impure man are sitting on a seat of straw properly arranged then the pure man contracts impurity because that straw constitutes a “seat”, whereas if the straw is strewn haphazardly he does not contract any impurity (A 1.15.13). And finally, there is a wonderful method of purifying a thing suspected of being impure–you get a Brahmin to declare it to be pure. This method is resorted to also when one is anxious about one’s own purity (B 2.12.6). This is rather like going to a confessor or psychologist –you want someone in power to say that everything is OK. The rules and practices relating to impurity in the Dharma texts constitute, I believe, a ritual apparatus rather than a social ideology. These rules do not create social structures but are intended to sustain and reinforce such structures. Within the specifically Indian context, 35 “Gods invented three means of purification for Brahmins: being unaware that something is impure, sprinkling it with water, and getting it verbally declared as suitable.” See Va 14.24; B 1.8.52 1.9.9; M 5.127; Vi 23.47; Y 1.191.

Caste and Purity

245

I do not think that purity was the historical cause of the varña or caste system, nor did it serve as a theodicy to justify that system. As boundaries must precede attempts to sustain and strengthen those boundaries, so the caste system must precede rules concerning the pure and the impure that aim to sustain it. In this sense, Mary Douglas is right in saying that “dirt” is a by-product of a system. We can consider impurity rules as a system of socialization. Individuals within the society must be made to acknowledge and support the social boundaries imposed on them, and this is effected primarily through social rituals. This may be one reason why many of the Dharma†åstric rules on impurity are found in the sections dealing with the vedic student. In any program of socialization you have to start young! And it is within the socializing context that what I said about intentionality makes sense. Socializing involves paying attention, involves anxiety–†aõkå. Rules of purity are meant to cause anxiety, for anxiety creates heightened attention to the boundaries that the rules are meant to uphold. The socializing aspect of these rules also explains another aspect of purity/impurity: the ultimate aim of these rules is not to make people remain constantly pure, which is in principle impossible. Their aim, I believe, is to make people intent on recovering purity. Hence most of the terms used for purity, especially †uddhi and †uci, as we have seen, mean not being pure but the act of becoming pure. It is purification not purity that is at the heart of the system. Attention to the socializing aspect of purity also helps us understand the connection between purity and auspiciousness (†ubha) that has received considerable attention recently (Madan 1988: 48f; Carman and Marglin 1985). The two have been viewed as both contradictory and complementary. I see the two as dealing with very different aspects of human and social life. ‡ubha deals with the flow of time and life and seeks to direct that flow in benign, fertile, and prosperous directions. ‡ubha is not connected with a program of socialization, with the protection of social boundaries and structures. That is the province of purity/impurity. In a totally pure world time would stand still, there will be no change; the world of total purity, ironically, would be a world of death. To be totally and always pure is not only impossible but from a variety of perspectives a highly undesirable condition. But that is pushing the system to its absurd limit. Purity is one among many competing and often contradictory values of human existence and human society. The purpose of rules of impurity is not to ensure permanent purity but to make people anxious about becoming impure and when they become impure, as they must, to make them anxious about recovering their lost purity. This anxiety, finally, is an integral part of the socializing process that sustains and strengthens cultural and social boundaries, including the caste system.

246

LANGUAGE, TEXTS AND SOCIETY

A

247

X Rhetoric and Reality: Women’s Agency in the Dharma†åstras*

The Indian legal treatises called Dharma†åstras, especially that of Manu, have become the focus of attention and controversy, because they are viewed as giving divine sanction and theological justification for the oppression and marginalization of women and the lower classes of society. In 1935 during the first conference of so-called untouchables under the leadership of Dr. B. R. Ambedkar, for example, a resolution was passed to reject Hinduism with the title “To the Untouchable Community: A New Message of a New Manu”. Within a month a group of young Dalit men gathered in Nasik to burn a copy of the Manu’s treatise (the Manu-smr¢ti) ceremonially.1 The eminent women’s rights advocate, Madhu Kishwar, in a recent article has also dealt with Manu and the issues relating to marginalized groups based on gender and caste. Kishwar refers to the burning of copies of the Manu-smr¢ti in the precincts of the Rajasthan High Court on March 25, 2000, and observes: “The protesters believed that the ancient text is the defining document of Brahmanical Hinduism, and also the key source of gender and caste oppression in India.”2 That these ancient texts contain statements and injunctions relegating the majority of individuals in Indian society to a marginal status is undeniable. * Originally published in Encounters with the Word: Essays to Honour Aloysium Pieris, edited by R. Crusz, M. Fernando, and A. Tilakaratne, pp. 489-505 (Colombo: Ecumenical Institute for Study and Dialogue). 1 See Eleanor Zelliot, “The Psychological Dimension of the Buddhist Movement in India,” in Religion in South Asia: Religious Conversion and Revival Movements in South Asia in Medieval and Modern Times, ed. G. A. Oddie, pp. 119-44 (Columbia, Missouri: South Asia Books, 1977). The burning of the MDh was advocated also by other reform activists, such as E. V. Ramasami: see M. R. Barnett, The Politics of Cultural Nationalism in South India (Princeton: Princeton University Press, 1976), p. 37. 2 “Manu and the Brits,” Hinduism Today, January-February 2001, pp. 56-59.

248

LANGUAGE, TEXTS AND SOCIETY

Most of the studies on women either in the Manu-smr¢ti or more generally in the Dharma†åstras, whether they attempt to portray these texts in a positive or a negative light, have looked at the attitudes and pronouncements of their authors; these studies have, for the most part, simply analyzed the gender and caste ideologies at the heart of these texts. That is not my aim. I am here not focusing on what Manu himself, or other authors of legal codes, thought or had to say about women. Indeed, looking at their pronouncements can only produce a rehashing of well-known platitudes and a very boring paper. The aim of this paper, on the contrary, is to see what light these ancient texts can shed on the lived reality of women’s lives in ancient India. Given the constraints of time, I will focus on two issues: the agency of women with respect to economic resources and sexuality. No text, not even one produced by a Brahmin pandit with claims to historical transcendence,3 can escape the historical reality in which it was produced. This is especially true of the Dharma†åstric genre of literature that seeks to provide guidelines for personal and criminal law. At the hands of a historian, therefore, such texts, if used with due diligence, can yield significant historical information. Moreover, recent studies by Richard Lariviere and Albrecht Wezler have argued, convincingly I believe, that the ancient Dharma†åstras are records of laws, customs, and moral norms found in different regions of India at different historical times.4 If behind the theological rhetoric these texts reflect social reality, distorted though it may be, then they can be significant windows to the lived reality of Indian women when they are used with some degree of methodological sophistication. Stephanie Jamison5 in an intriguing study has pointed out that there is a noticeable increase in the vehemence of the anti-women rhetoric between the early Dharmasütra of Åpastamba (circa 4-3 cent. BCE) and Manu (circa 1/2 cent. CE). Jamison ascribes this at least partly to the very structure of the two texts, Åpastamba adopting a more ritual rhetoric and viewing women within a ritual setting, and Manu using a more discursive rhetoric and dealing with real-life situations. She also notes the probability that there was a surge in “independent women”, especially Buddhist and Jain nuns, during the time of Manu. Beyond the literary genre, we can detect a shift in the attitudes towards women from Åpastamba to Manu, although it is unclear whether these differences reflect regional or personal attitudes or whether they indicate a historical shift in attitudes towards women. Åpastamba, for example, not only prohibits polygamy and divorce (ÅpDh 2.11.12-14) but also acknowledges that women and 3

On the deliberately ahistorical nature of †åstric literature, see Pollock 1990. Lariviere 2004; Wezler 2004. 5 In a paper entitled “Women ‘between the Empires’ and ‘between the Lines’” presented at the conference “Between the Empires” held at the University of Texas at Austin, April 10-13, 2003; to be published in Olivelle, Forthcoming-c. 4

Rhetoric and Reality

249

people of lower classes possess a particular knowledge of Dharma: “The knowledge found among women and ‡üdras forms the conclusion [of dharma], and they point out that it is a subsidiary component of the Atharva Veda [...]. According to some, one should learn the remaining Dharmas from women and people of all classes.”6 Such a liberal view of women and the lower classes of society is missing in later texts of the tradition. The Dharma†åstras, nevertheless, present contradictory and often schizophrenic views of female agency. Women as independent agents are acknowledged by Manu within the very rhetoric that attempts to deny them such agency and independence, as in the wellknown proverb repeated in several Dharma†åstras: “As a child, she must remain under her father’s control; as a young woman, under her husband’s; and when her husband is dead, under her sons’. She must never seek to live independently” (MDh 5.148).7 The reality of women owning wealth and property is evident in much of what the Dharma†åstras have to say, even while they deny women the right of ownership, as in another oft-repeated adage: “Wife, son, and slave–all these three, tradition tells us, are without property. Whatever they may earn becomes the property of the man to whom they belong” (MDh 8.416; NSm 5.39).8 It is the lived reality of women’s lives, diverse though they may have been across space and time, that I am seeking to capture by using these same texts as witnesses to get behind them into the reality of women that they were attempting to define, control, and legislate for.

Economic Agency of Women I want to deal with the agency of women with respect to material resources at the outset, because, directly or indirectly, every other kind of agency is ultimately dependent on it. Manu’s pronouncements about women being incapable of owning property, of being economically independent, are belied by the facts. The numerous works of Gregory Schopen have clearly demonstrated that Buddhist women, both laywomen and nuns, were responsible for numerous and expensive donations to Buddhist monasteries in their own names and on 6 så niß™hå yå vidyå strœßu †üdreßu ca [...] strœbhyaΔ sarvavarñebhya† ca dharma†eßån pratœyåd ity eke (ÅpDh 2.29.11, 15). In the context of a funeral, we have a similar provision in the Bhåradvåja Pitr¢medhika Sütra (1.8.8). After the funeral the mourners return to the village “and do what the women may instruct” (yat striyaΔ pråhus tat kurvanti). Likewise, in the context of the ritual parting of the wife’s hair (sœmantonnayana), the ņvalåyana Gr¢hyasütra (1.14.8) says that the young married couple should do whatever old Brahmin women, whose husbands and children are alive, tell them to do. 7 bålye pitur va†e tiß™het påñigråhasya yauvane Ù putråñå∫ bhartari prete na bhajeta svatantratåm ÙÙ 8 bhåryå putra† ca dåsa† ca traya evådhanåΔ smr¢tåΔ Ù yat te samadhigacchanti yasya te tasya tad dhanam ÙÙ

250

LANGUAGE, TEXTS AND SOCIETY

their own initiative. In one set of nineteen donative inscriptions in Mathurå during the Kußåna period studied by Schopen, eight were made by men and eleven by women. And in another set of inscriptions recording donations at Sanchi, 129 were made by monks and 125 by nuns (Schopen 1997, 249). Leslie Orr has examined the gifts women made to temples in Tamilnadu during 9-14th centuries CE and concludes: “Women in the Chola period evidently had more access to property and more autonomy in disposing of their property than normative texts would lead us to expect” (Orr 2000, 71). And examining the corpus of medieval Telugu inscriptions from Andhra Pradesh, Cynthia Talbot comes to a similar conclusion: “Another striking feature of the inscriptional corpus is the relatively large number of women represented within it. Women comprise 11 percent (87 in number) of all individual donors. This is certainly less than a proportional representation, but it is much higher than one might expect given the limits placed on women’s personal property (strœdhana) in the legal literature” (Talbot 2001, 82). She notes that women even alienated immovable property, making land grants to temples. Clearly there is a dissonance between the rhetoric of the Dharma†åstras and the reality of women’s position within the ancient Indian economy. These texts do acknowledge that married women owed property; such property is called either påriñåhya, an ancient term already found in the vedic texts (Jamison 1997), or more commonly strœdhana (“women’s property”).9 Manu (9.194) gives six kinds of such property: “what a woman receives at the nuptial fire, what she receives when she is taken away, what she is given as a token of love, and what she receives from her brothers, mother, and father”.10 Strœdhana is generally viewed as consisting of clothes, jewelry, and the like. The picture gets a bit blurred, however, because in their sections on crime and punishment the same texts impose various kinds of fines on women payable in coins (YDh 2.285; Kåt 487 ). A husband is permitted to fine his own wife: “If, after she is forbidden, a wife drinks liquor, albeit at festivals, or visits shows and fairs, she should be fined 6 Kr¢ßñalas” (MDh 9.84).11 Manu’s (8.369) prescription of a fine even for an unmarried woman engaging in lesbian love indicates that even unmarried women were presumed to have control of money: “If a virgin violates another virgin, she should be 9 Åpastamba, on the other hand, appears to support the position that marital property is owned jointly by the husband and wife; hence there is no partition of property between them: “There is no division of property between a husband and a wife, because from the time of their marriage they are linked together in performing religious rites, as also in receiving the rewards of their meritorious deeds and in acquiring wealth” (ÅpDh 2.14.16-19). 10 adhyagny adhyåvåhanika∫ datta∫ ca prœtikarmañi Ù bhråtr¢ m åtr¢ p itr¢ p råpta∫ ßaÛ∂vidha∫ strœdhana∫ smr¢tam ÙÙ Similar enumerations are given in NSm 12.8-9; YDh 2.14344; Kåt 894-901. 11 pratißiddhå pibed yå tu madyam abhyudayeßv api Ù prekßåsamåjau gacched vå så dañÛ∂yå kr¢ßñalåni ßa™ ÙÙ

Rhetoric and Reality

251

fined 200, pay three times the bride-price, and receive ten lashes.”12 Kåtyånana (571, 574) in dealing with a wife’s obligations after her husband’s death, specifies women who have a large amount of wealth (pradhanå, adhikårthå). Indeed, Kåtyåyana is the one who appears to give us a more realistic picture of women’s property. On strœdhana, for example, he says that “the father, mother, husband, brother, and kinsmen should give strœdhana to a woman up to 2,000 Pañas, except immovable property” (Kåt 902).13 Yet, Kåtyåyana (919) himself, contradicts this provision elsewhere under inheritance when he says that immovable property given by parents to a daughter goes to her brothers if she dies childless. This provision assumes that women did have control over land. The Dharma†åstras viewed a married woman’s property as clearly separate from that of the husband; such property was not part of the common family estate and was given the technial term yautuka (MDh 9.131). Nårada (NSm 13.7) is clear and explicit: “The mother is master (of her property) just as the father is of his.”14 She is thus able to give her property to anyone of her own free will, just as the father. If a husband took a loan from his wife’s strœdhana, he was obliged to pay it back. A husband was permitted to use his wife’s property without repayment only in clearly specified circumstances when other resources were unavailable: a time of famine, for a religious purpose (such as a daughter’s marriage), sickness, and imprisonment (YDh 2.147). There are clear guidelines for inheriting the mother’s separate property: some prescribe that it goes only to her daughters and others to both her sons and daughters, but never to the children of her husband by other wives. A husband inherits his wife’s property only in the absence of her own children. Severe restrictions were placed around the property of women so that it would not be swindled by crafty relatives. Indeed, it is one of only two kinds of property not alienated by possession (enjoyment): “The property of women and kings is never lost, even if it is possessed without title for hundreds of years” (NSm 1.75).15 Although they are called propertyless (nirdhana), women are in line to inherit the property of their male relatives. The texts offer divergent opinions, but the mother first and then the daughters are next in line to inherit the father’s estate in the absence of sons (YDh 2.135-36; Kåt 921, 927), taking precedence over all other heirs, 12 kanyaiva kanyå∫ yå kuryåt tasyåΔ syåd dvi†ato damaΔ Ù †ulka∫ ca triguña∫ dadyåc chiphå† caivåpnuyåd da†a ÙÙ 13 pitr¢måtr¢patibhråtr¢j∞åtibhiΔ strœdhana∫ striyai Ù yathå†akty å dvisåhasråd dåtavya∫ sthåvaråd r¢te ÙÙ 14 måtrå ca svadhana∫ datta∫ yasmai syåt prœtipürvakam Ù tasyåpy eßa vidhir dr¢ß™o måtåpœß™e yathå pitå ÙÙ 15 strœdhana∫ ca narendråñå∫ na kadåcana jœryate Ù anågama∫ bhujyamåna∫ vatsaråñå∫ †atair api ÙÙ

252

LANGUAGE, TEXTS AND SOCIETY

including the father’s parents and brothers. A wife without her own strœdhana, moreover, inherits a share equal to that of her husband’s sons (YDh 2.115). Then there is the well-known institutions of putrikå, that is, the appointment of a daughter to become the son in the absence of a male heir. As Jolly (1885) has clearly pointed out, this daughter has all the rights and privileges of a son; she is not merely holding the place for her own son (her father’s grandson) to inherit the paternal estate. She becomes the sole heir of her father’s estate, and this estate is outside the jurisdiction of her own husband. Reading the Dharma†åstras closely we see that behind the rhetoric lies the reality not only of women who owned property, including land, but also of wealthy women. Their property, moreover, was out of bounds to their male relatives, including husbands, sons, and in-laws. Women undoubtedly had agency with regard to their own wealth as confirmed by the donative inscriptions of women I have referred to earlier. The rhetoric of the Dharma†åstras, once again, clashes with this reality. Women are said to be asvatantra, not independent, especially with regard to law and legal transactions. Nårada is explicit: “Except in distress, transactions undertaken by women are invalid, especially the giving, pledging, or selling of a house or field” (NSm 1.22; so also YDh 2.31).16 In law, as in moral discourse, we know that injunctions are leveled against existing and often prevalent practice; we can, therefore, be certain that these rules presuppose precisely their opposite; women did alienate land as clearly stated in inscriptions. Nårada (1.117) also says that a document executed by a woman is null and void. This theory is belied by the texts themselves. Yåj∞avalkya (YDh 2.46), for example, clearly states that debts may be entered into by the wife and the husband independently of each other: “The wife shall not pay a debt incurred by her husband or son; a son a debt incurred by his father; or the husband a debt incurred by his wife–unless it was incurred for the benefit of the family.”17 He also refers to debts jointly contracted by the husband and wife, debts that the wife is obligated to pay in the absence of the husband (YDh 2.49). Nårada (NSm 1.13; also Kåt 569-70) likewise distinguishes the debts of husband and wife: “A wife does not have to pay a debt made by her husband or by her son unless she agreed to do so or unless she had made the debt jointly with her husband.”18 These explicit rules clearly presuppose the independent agency of women in economic matters, including taking out loans. The Dharma†åstras, however, clearly state that a document with witnesses should be executed for 16 strœkr¢tåny apramåñåni kåryåñy åhur anåpadi Ù vi†eßato gr¢ h akßetradånådhamanavikrayåΔ ÙÙ 17 na yoßit patiputråbhyå∫ na putreña kr¢ta∫ pitå Ù dadyåd r¢te ku™umbårthån na patiΔ strœkr¢ta∫ tathå ÙÙ 18 na strœ patikr¢ta∫ dadyåd r¢ña∫ putrakr¢ta∫ tathå Ù abhyupetåd r¢te yad vå saha patyå kr¢ta∫ bhavet ÙÙ

Rhetoric and Reality

253

each debt; even otherwise, taking such a loan in the presence of witnesses constitutes a legally binding transaction. The ability of women to enter into debt implies their ability to execute legally valid transactions, including documents, as evident in the donative inscriptions and contradicting the explicit pronouncements of the same Dharma†åstras. The longest section in any discussion on lawsuits and legal procedure in the Dharma†åstras pertains to the non-payment of debts (r¢ñådåna). Clearly, this must have been the primary cause of litigation in ancient India. The discussions of this topic, however, presuppose that the litigants are all male; male nouns and adjectives are the ones generally used. If women did take out loans and executed other legal transactions, it seems fair to assume that women must have constituted at least a good segment of litigants appearing before courts of law. This is hinted at by Manu (MDh 8.68) when he says that women may give testimony in cases involving women. The Dharma†åstras also give us some clues as to the disparity among various classes of society with regard to the economic agency of women. According to the principle articulated earlier, neither the husband nor the wife is responsible for debts incurred by the other when the loan was not taken either jointly or for the common good of the entire family. Nårada, however, makes an exception in the case of those groups in which the women are the wage earners and the husbands depend on their wives for an income: “A debt entered into by the wife never obligates her husband [...]. except in the case of the wives of washermen, hunters, cowherds, and liquor merchants; their livelihood and that of their families depend on their wives” (NSm 1.1516).19 Yåj∞avalkya’s (YDh 2.48) list of such men includes: cowherds, liquor merchants, actors, washermen, and hunters. These generalized statements refer to identifiable and probably stereotyped groups in which the women were commonly thought, at least by the Brahmanical authors of these texts, to be the wage earners. Within a totally different context, namely, the allegation of rape, Manu, after giving detailed rules for litigation and punishment of men accused of rape or other sexual offenses against women, gives this exception: “The above rule does not apply to wives of traveling performers or to wives who earn a living on their own, for such men get their women to attach themselves to men and, concealing themselves, get them to have sexual liaisons” (MDh 8.362).20 Another way, as Manu sees it, for men to use their wives to make a living and another stereotype of social groups! The broader implication, however, is that there were segments of ancient Indian society where women were economically more independent and where they supported the male members of their households with their labor and earnings. 19 na ca bhåryåkr¢tam r¢ña∫ katha∫cit patyur åbhavet ÙÙ [...] anyatra rajakavyådhagopa†auñÛikayoßitåm ÙÙ teßå∫ tatpratyayå vr¢ttiΔ ku™umba∫ ca tadå†rayam ÙÙ 20 naißa cårañadåreßu vidhir nåtmopajœvißu Ù sajjayanti hi te nårœr nigüÛ∂hå† cårayanti ca ÙÙ

254

LANGUAGE, TEXTS AND SOCIETY

The best known example of such women, of course, is prostitutes and courtesans. The wealth of courtesans is a recurrent theme in Sanskrit literature. Despite their moral outrage, the authors of Dharma†åstras do provide guidance as to the legal provisions for such women. Under the ground for litigation called “Nonpayment of Wages”, that is, disputes between employers and employees, Nårada brings up the topic of prostitutes. If a prostitute takes the fee but later refuses to provide the required services, she is fined twice that amount; so also a man who refuses to pay after getting the services of a prostitute (NSm 6.20). In spite of the declarations of the Dharma†åstras, then, women from all social groups and all walks of life did possess varying degrees of economic independence, at least some of them earning incomes from their skills and labor. Contrary again to Dharma†åstric pronouncements, women did indeed execute legally binding transactions, including, in all likelihood, legal documents. Not only reading “between the lines” as Jamison proposes, but even by reading closely the lines themselves, we can see that Dharma†åstras contain strong evidence for the agency of women in economic matters.

Sexual Agency of Women The ideology articulated in the Dharma†åstras with regard to female sexuality is paradoxically bifurcated: women are viewed, on the one hand, as passive sexual objects and, on the other, as sexually rapacious. They are either denied sexual agency or ascribed a heightened and predatory sexual agency. The lack of independence of women especially in sexual matters has been articulated forcefully by the authors of Dharma†åstras. Women are under male supervision at every stage of their life: under the father in their youth, under the husband in their adult years, and under the son in their old age (MDh 5.148). Their identity is merged with that of their husbands; marriage defines the female in a more essential and fundamental way than it does the male. Even after death a married woman is expected to wish for the same world as her husband; even in death they cannot be separated, an ideology gruesomely enacted in the ritual of satœ. The passivity of the woman is nicely illustrated in the oft-repeated parable of the seed (bœja) and the field (kßetra). The woman is compared to the field and the man to the seed in the procreation of children. The question is, to whom does a son belong when the mother’s husband is not the son’s biological father? To the owner of the field (the husband of the mother) or to the owner of the seed (the biological father)? The first opinion favors the latter, arguing that “Even when different kinds of seeds are sown by farmers in the very same plot at the right time, they are seen in the world to sprout different-

Rhetoric and Reality

255

ly, each according to its nature” (MDh 9.38). In other words, the field does not contribute anything to the crop; it is the nature of the seed that determines whether the plant is mango or rice. In like manner, the child takes after and belongs to the biological father. The second option asserts that the child belongs to the owner of the field: “When men who have seeds but no fields sow them in someone else’s field, once the crop has grown they do not receive any of the harvest [...]. When men without fields sow their seed in someone else’s field, they create profits for the owner of the field; the owner of the seed reaps no fruit” (MDh 9.49-50).21 In other words, if you plant a mango seed in your neighbor’s yard, the fruit the tree yields as it matures belongs to your neighbor. Note, however, that in this entire debate the woman is a passive bystander, the mother plays no part; the debate is about the male owners of the seed and the field. According to this ideology, the ownership of women passes from the father to the husband. This is clearly articulated by Manu, who says that the essential element of a marriage (kanyådåna) is precisely the “giving” (dåna): “The invocation of blessings and the sacrifice to Prajåpati are performed during marriage to procure her good fortune; the act of giving away is the reason for his lordship over her” (MDh 5.152). According to this dåna ideology, the act of giving transfers the ownership of the gift from the giver to the recipient in the case of both a common gift and the gift of a maiden.22 The same ideology is behind the ancient custom of †ulka, “bride price”, paid by the suitor to the father of the bride. Although Manu, as well as other Dharma†åstric authors vehemently denounce this custom,23 the rules found in their own works show that it was a widespread practice. Thus, in dealing with the negligence of a father to get his daughter married, Manu says; “A man who takes a girl after she has reached puberty shall not pay a bride-price, for the father has lost his ownership of her by frustrating her menses” (MDh 9.93).24 When a father promises his daughter to one man and gives her to another, it falls under the legal provisions for breach of contract, the clear implication being that the man did not get what he paid for. Nårada, in fact, has this interesting provision as to how to strike a better deal: “If the girl has already fetched a bride-price and a better 21 yo ‘kßetriño bœjavantaΔ parakßetrapravåpiñaΔ Ù te vai sasyasya jåtasya na labhante phala∫ kvacit ÙÙ 22 The view that a wife is the property of her husband is also articulated in the Mahåbhårata (2.60.40, 61.32) when the issue of whether Draupadœ had been legitimately won by the Kaurava brothers is taken up. The conclusion is that Draupadœ is part of the property of Yudhiß™hira, and when he waged all that he owned, he also waged Draupadœ. 23 “A learned father must never accept even the slightest bride-price for his daughter; for by greedily accepting a bride-price, a man becomes a trafficker in his offspring.” MDh 3.51. 24 pitre na dadyåc chulka∫ tu kanyåm r¢tumatœ∫ haran Ù sa ca svåmyåd atikråmed r¢tünå∫ pratirodhanåt ÙÙ

256

LANGUAGE, TEXTS AND SOCIETY

bride-groom should happen to come along, then one might stretch the truth about righteousness, wealth, or desire in whatever way is necessary (NSm 12.30).”25 Under breach of contract, Manu has this revealing comment on fraud during a marriage: “If a man shows one girl to the bridegroom and gives another, the groom may marry both for the same price–so has Manu decreed” (MDh 8.204).26 The remaining portion of a bride-price (clearly for a second and younger wife!) is one of the deceased father’s debts that a son is not obliged to pay (MDh 8.159). When someone, man or woman, violates a virgin, he is obliged to pay a bride-price to the father, highlighting the economic loss he has suffered (MDh 8.366, 369).27 On the other hand, we have the portrayal of women as sexually obsessed; only constant male vigilance is able to control them: “Day and night men should keep their women from acting independently; for, attached as they are to sensual pleasures, men should keep them under their control” (MDh 9.2).28 The reason for this vigilance is obvious: “They pay no attention to beauty, they pay no heed to age; whether he is handsome or ugly, they make love to him with the single thought, ‘He’s a man!’29 Lechery, fickleness of mind, and hardheartedness are innate in them; even when they are carefully guarded in this world, therefore, they become hostile towards their husbands” (MDh 9.14-15).30 There is a deep fear of female sexuality exhibited in the provisions of the Dharma†åstras, and I think it is this fear that has given rise to misogyny and the regulations about guarding women. If we look at the reality of women’s lives from a perspective other than that of the Dharma†åstric authors, if we get to the other side of this looking glass, however, the story may be told differently, the story of women acting independently not just in the economic sphere but also in the realm of sexuality. Sexual independence does not necessarily mean sexual license, as assumed in the Dharma†åstras; it is primarily related to independence from the male control of female sexuality exercised within the context of marriage and family. The refusal to get married 25 kanyåyå∫ pråpta†ulkåyå∫ jyåyå∫† ced vara åvrajet Ù dharmårthakåmasa∫yukta∫ våcya∫tatrånr¢ta∫ bhavet ÙÙ 26 anyå∫ ced dar†ayitvånyå vo∂huΔ kanyå pradœyate Ù ubhe te eka†ulkena vahed ity abravœn manuΔ ÙÙ 27 The sale of the girl is explicitly recognized even in Vedic texts; see, for example, MS 1.10.11, where adultery on the part of the wife is condemned because she is cheating her husband who has bought her. See Thieme 1963, 208. 28 asvatantråΔ striyaΔ kåryåΔ purußaiΔ svair divåni†am Ù vißaye sajjamånå† ca sa∫sthåpyå hy åtmano va†e ÙÙ 29 That this image of female infidelity and a woman’s proclivity to grab the nearest man is illustrated in the proverb cited in the Indian book of animal fables, the Pa∞catantra (I. verse 20): “Kings, women, and vines, do for the most part, Cling to whatever they find close at hand.” 30 naitå rüpa∫ parœkßante nåså∫ vayasi sa∫sthitiΔ Ù surüpa∫ vå virüpa∫ vå pumån ity eva bu∞jate ÙÙ pau∫†calyåc cålacittyåc ca naiΔsnehyåc ca svabhåvataΔ Ù rakßitå yatnato ‘pœha bhartr¢ßv etå vikurvate ÙÙ

Rhetoric and Reality

257

by entering female voluntary institutions such as Buddhist and Jain monasticism was one way in which ancient Indian women exercised such independence. Courtesans represent another form of independent woman. What influence the different waves of foreign migrants and invaders, from the Persians and Greeks in the 4-3rd centuries BCE to the Kushans of the 1-3rd centuries CE, had on Indian society with respect female roles in family and society is impossible to gauge. As I have shown elsewhere, however, Manu and the other moral philosophers and jurists who wrote on Dharma were at least in part responding to the influence of foreigners.31 One effect of the new situation is reflected in the injunction by Manu and subsequent authors that women should be guarded.32 They divide women into two groups, the guarded (gupta) and the unguarded (agupta). Manu (MDh 9.5-12) advises his male readers: “Women in particular should be guarded against even the slightest evil inclination, for when they are left unguarded, they bring grief to both families.33 Seeing that this is clearly the highest Law of all social classes, even weak husbands strive to guard their wives; for by carefully guarding his wife, a man guards his offspring, his character, his family, himself, and the Law specific to him [...]. No man is able to thoroughly guard women by force; but by using the following strategies, he will be able to guard them thoroughly. He should employ her in the collection and the disbursement of his wealth, in cleaning, in meritorious activity, in cooking food, and in looking after household goods. When they are kept confined within the house by trusted men, they are not truly guarded; only when they guard themselves by themselves are they truly well guarded.”34 The dual strategy of guarding them both externally and socializing them to be on guard internally appears to be Manu’s strategy for dealing with the unbridled sexuality he imputes to women. The dual category of women, the guarded and the unguarded, appears again in the context of criminal justice. In the case of both rape and consensual sex, the penalties imposed on the man is dependent on whether the woman was guarded or unguarded, the punishment being more severe in the case of guarded women (MDh 8.374-85). There are women of certain groups, however, who are stereotyped as being sexually promiscuous. After giving rules for deal31

See the introduction to my critical edition and translation of Manu: Olivelle 2004a. In an interesting and possibly significant remark, the Mahåbhårata (1.113.4-21) states that in the olden days women went about freely and in the open, were their own mistresses, and acted as they pleased. It was the sage ‡vetaketu who, after his wife was raped, decreed that women should be cloistered. 33 That is, the natal family and the family of the husband. 34 sükßmebhyo ‘pi prasaõgebhyaΔ striyo rakßyå vi†eßataΔ Ù dvayor hi kulayoΔ †okam åvaheyur arakßitåΔ ÙÙ ima∫ hi sarvavarñånå∫ pa†yanto dharmam uttamam Ù yatante rakßitu∫ bhåryå∫ bhartåro durbalå api ÙÙ [...] na ka†cid yoßitaΔ †aktaΔ prasahya parirakßitum Ù etair upåyayogais tu †akyås tåΔ parirakßitum ÙÙ arthasya sa∫grahe cainå∫ vyaye caiva niyojayet Ù †auce dharme 32

258

LANGUAGE, TEXTS AND SOCIETY

ing with rape and adultery, Manu makes an exception: “The above rule does not apply to wives of traveling performers or to wives who earn a living on their own, for such men get their women to attach themselves to men and, concealing themselves, get them to have sexual liaisons. When someone engages in secret conversations with such women, as also with female slaves serving a single master and with female wandering ascetics, he shall be compelled to pay a small fine” (MDh 8.362-63).35 Here again we encounter the category of women who earn a living on their own; economic independence is here equated with sexual laxity. The reality behind these prescriptions probably is that within certain classes of society there were strong and independent women, both economically and sexually. They were not under the guardianship of the male members of their households. Women acting independently with regard to sex is noted also in other contexts. We have two classical examples. The first is the institution of the self-choice of a husband (svaya∫vara) celebrated in the well-know story of Nala and Damayantœ (Schmidt 1987, 76-109). Interestingly, this institution is completely absent in the Dharma literature and appears to have been limited to royalty. The father announces the ritual of self-choice, and the young woman chooses a man among those assembled. A similar institution is that of the Gandharva form of marriage, celebrated in the story of Dußyanta and ‡akuntalå. Here consensual sex between a woman and a man is viewed as the consummation of their marriage. The Dharma literature, again, restricts this form of marriage to Kßatriyas, hinting that sexual mores may have been different in different segments of ancient Indian society. That women did exercise sexual independence is also indicated in the list of the twelve kinds of sons. A son born in secret, a son born to an unmarried woman, and a son received with marriage are all children born through sex outside of marriage. Manu defines these three: “When in someone’s house is born a son whose father is unknown, he is a son born in secret within the house; and he belongs to the man whose wife gave birth to him [...]. When an unmarried girl gives birth to a son secretly in her father’s house, one should call him by the name son born to an unmarried woman; the offspring of an unmarried girl belongs to the man who marries her. When a pregnant woman is married off, whether her condition is disclosed or not, the child in the womb belongs to the man who marries her and is called son received with marriage” (MDh 9.170-73).36 The jurists are ‘nnapaktyå∫ ca påriñåhyasya cekßañe ÙÙ arakßitå gr¢he ruddhåΔ purußair åptakåribhiΔ Ù åtmånam åtmanå yås tu rakßeyus tåΔ surakßitåΔ ÙÙ 35 naißa cårañadåreßu vidhir nåtmopajœvißu Ù sajjayanti hi te nårœr nigü∂hå† cårayanti ca ÙÙ ki∫cid eva tu dåpyaΔ syåt sa∫bhåßå∫ tåbhir åcaran Ù preßyåsu caikabhaktåsu rahaΔ pravrajitåsu ca ÙÙ 36 utpadyate gr¢he yas tu na ca j∞åyeta kasya saΔ Ù sa gr¢he gü∂ha utpannas tasya syåd yasya talpajaΔ ÙÙ [...] pitr¢ve†mani kanyå tu ya∫ putra∫ janayed rahaΔ Ù ta∫ kånœna∫ vaden

Rhetoric and Reality

259

here attempting to provide rules for dealing with children born out of wedlock. Clearly, some women refused to be guarded! Note, however, that to the question “Whose is the child?”, the mother is never mentioned in the answer except in relationship to a male to whom she belongs. Another aspect of the sexual agency of women is revealed in the rules with respect to lesbian love found in the section on criminal law dealing with sexual assault on women. Here Manu presents two possibilities: a virgin (kanyå) violating another virgin, and an adult woman violating a virgin. The translation of the term kanyå is problematic; it may indicate a virgin or simply a young girl. The term strœ, on the other hand, clearly refers to a mature woman, possibly a married woman. If a virgin violates another virgin, she should be fined 200, pay three times the bride-price, and receive ten lashes. When a woman violates a virgin, however, her head ought to be shaved immediately–alternatively, two of her fingers should be cut off–and she should be paraded on a donkey. (MDh 8.369-70)37

A significant and, perhaps a primary, way in which women become agents of their own sexuality is through the control of their reproductive cycle. The moral conflicts in contemporary society with regard to contraception and abortion are indicators of how emotionally charged this issue is even in otherwise sexually liberated societies. Male fear of female sexuality was especially evident in the area of menstruation and abortion. To some degree, males viewed menstruation as a failed pregnancy, as a sort of an abortion. Thus, a father who fails to give his daughter in marriage before the onset of menstruation is said to be guilty of an abortion every time his daughter has her period in his house.38 That women induced abortions, possibly through medication, is apparent in the severe punishments prescribed in the Dharma†åstras for such women. Funerary offerings are not made for women who have caused abortions or have killed their husbands. Significantly, these two crimes are juxtaposed, indicating the close connection between the two, given that the child is often said to be the husband reborn in his wife (MDh 5.90; NSm 12.92). That there may have been professional abortionists is hinted at by the frequent condemnation of such people.39 There are also indications that women did practice nåmnå vo∂huΔ kanyåsamudbhavam ÙÙ yå garbhiñœ sa∫skriyate j∞åtåj∞åtåpi va satœ Ù vo∂huΔ sa garbho bhavati saho∂ha ity cocyate ÙÙ 37 kanyaiva kanyå∫ yå kuryåt tasyåΔ syåd dvi†ato damaΔ Ù †ulka∫ ca triguña∫ dadyåc chiphå† caivåpnuyåd da†a ÙÙ yå tu kanyå∫ prakuryåt strœ så sadyo mauñÛ∂yam arhati Ù aõgulyor eva vå cheda∫ khareñodvahana∫ tathå ÙÙ 38 See MDh 9.4; BDh 4.1.11-14; NSm 12.26-27; VaDh 17.69-72. This is the one circumstance under which our authors permit the girl to find a husband and get married on her own. 39 See ÅpDh 1.19.15; 1.21.8; 1.24.8; 1.28.21; GDh 17.11; 20.1; 21.9; 24.6-9; BDh 1.10.23; 4.1.12-13, 17, 20, 29; VaDh 2.42; 20.23-24; 28.7

260

LANGUAGE, TEXTS AND SOCIETY

some sort of contraception that would suppress their menstrual period. One law book provides this severe punishment for such a woman: “When a wife suppresses her menstrual periods out of antipathy towards her husband, he should bring her to the middle of the village, declare her to be an abortionist, and drive her away from his house” (BDh 4.1.20).40 Reality of women’s life in ancient and medieval India was, as expected, quite different from what the misogynist rhetoric of normative texts would have us believe. What I have attempted to show here, however, is that the normative texts themselves provide evidence that the reality was not what the texts themselves attempt to portray, bringing up a question of method. It is easy to dismiss textual evidence as of little value for historical purposes. This, I think, is mistaken and misguided. Texts are as much a part of the available historical data of India’s past as are archeological remains, inscriptions, coins, and art historical material. Unlike other forms of historical data, however, texts have a more intense form of authorial intentionality. The mediation of the author’s mind and hand often obscures the view as much as it reveals. Different genre of data and different genre of texts require different methodologies to uncover historical reality. Normative texts, whether they are theological, legal, didactic, or moral, need to be handled with special care if they are to yield historically useful data. As this small study has demonstrated, normative texts and legal literature, at the same time as they place limits on women’s rights to self and property, also reveal that women did enjoy agency and independence that would be unimaginable if we were to take their normative rhetoric at face value.

40 bhartuΔ pratinive†ena yå bhåryå skandayed r¢tum Ù tå∫ gråmamadhye vikhyåpya bhrüñaghnœ∫ nirdhamed gr¢håt ÙÙ

A

261

XI Manu and Gautama: A Study in ‡åstric Intertextuality*

Early explorations of the issue of intertextuality in the Månava Dharma†åstra centered on the hypothesis first articulated by Max Mller in a letter to one Mr. Morley on July 29, 1849 that the extant †åstra was a versified version of a lost Månava Dharmasütra.1 This hypothesis was given strong support by George Bhler (1886, xviiif) in the introduction to his renowned translation of Manu, although it has been abandoned by and large in recent scholarship and vigorously refuted by Kane (1962-75, I: 143-49).2 The focus on this issue, however, has obscured the very real textual connection between Manu and one of the extant Dharmasütras, the Gautama Dharmasütra. It is this †åstric intertextuality that is the subject of this paper. It studies the textual and thematic dependence of the Månava Dharma†åstra on the Gautama Dharmasütra amounting in several instances to the versification of the sütras of Gautama. This analysis also throws some light on the process of composition undertaken by the author of Manu. The textual parallels between Gautama and Manu are so close and so numerous that it is safe to conclude that the author of Manu used Gautama as one of his primary sources; the frequency of these parallels makes it unlikely that the authors of both texts were drawing from a common source. What is given below is not an exhaustive * To be published in Wilhelm Halbfass Commemoration Volume edited by K. Preisendanz and E. Franco. Vienna: Verlag der Österreichische Akademie der Wissenschaften. 1 This letter is cited in full by Bhler (1879, ix-xi) and referred to in his translation of Manu (Bhler 1886, xviii). 2 The abandonment of the Månava Dharmasütra hypothesis does not mean that the extant Manu is independent of the vedic †åkhå tradition. Stephanie Jamison (2000) has recently shown some interesting textual connections between Manu and the Maitråyañœya †åkhå of the Yajurveda.

262

LANGUAGE, TEXTS AND SOCIETY

list of all the parallels between the two texts.3 These examples, however, permit us to draw some significant conclusions concerning both the sources of and the process entailed in the composition of Manu. Manu 2.6 vedo ‘khilo dharmamüla∫ smr¢ti†œle ca tadvidåm Ù The root of dharma is the entire Veda, and the tradition and practice of those who know the Veda.4

Gautama 1.1-2 vedo dharmamüla∫ tadvidå∫ ca smr¢ti†œle Ù The root of dharma is the Veda, and the tradition and practice of those who know the Veda.

The dependence of Manu on Gautama here is evident; the additon of (a)khilo in påda-a and the change in word order in påda-b convert the prose into a †loka. What is less clear, however, is whether the sütra of Gautama itself is a prose rendering of a verse original. No other dharma text has a formulation quite like this. The ceremonies involving the teacher and the vedic student at the beginning and end of vedic instruction are recorded only in Gautama and Manu, and the dependence of Manu on Gautama is clear. Manu 2.73 adhyeßyamåñas tu guru∫ nityakålam atandritaΔ Ù adhœßva bho iti brüyåd viråmo ‘stv iti cåramet ÙÙ5 When he is ready for vedic recitation, he should say to the teacher, “Teach, Sir (bho)!”, without being lazy at any time; and when commanded “Stop!”, he should terminate.

Gautama 1.46 påñinå savyam upasa∫gr¢hyånaõguß™ham adhœhi bho ity åmantrayeta gurum Ù Clasping (the teacher’s) left (hand) — excluding the thumb — with his right, he should address the teacher “Teach, Sir!”

Gautama is the only sütra that specifies the length of time for the evening sa∫dhyå and that notes the time when stars become visible. 3 For further examples, see M 2.14 and G 1.4; M 2.15 and G 1.3; M 2.73 and G 1.46; M 2.101 and G 2.8; M 4.112 and G 16.17; M 5.66 and G 14.15; M 8.112 and G 23.29; M 9.112 and G 28.5; M 9.123-24 and G 28 14-15. Note the term pakßiñœ that occurs only in M 5.81, 4.97 and in G 14.19. 4 Although I have translated smr¢ti as “tradition”, it has become abundantly clear to me as I have worked through the Dharma†åstric material that the term is much more ambivalent and complex. For its early semantic history, see Klaus, Konrad, “On the Meaning of the Root smr in Vedic Literature”, Wiener Zeitschrift fr die Kunde Sdasiens und Archiv fr indische Philosophie: 36 (Sup 1992): 77-86. 5 The original reading of this verse, as also its exact meaning, are unclear. Gautama helps us to determine the original reading. Most mss. read adhyeßyamåña∫ tu guruΔ, making the teacher the subject of brüyåt. In my edition (Olivelle 2004a), however, I have

Manu and Gautama

263

Yåj∞avalkya (1.24-25) has a similar statement, but it is derivative of Manu. Manu 2.101 pürvå∫ sa∫dhyå∫ japa∫s tiß™het såvitrœm årkadar†anåt Ù pa†cimå∫ tu samåsœnaΔ samyag r¢kßavibhåvanåt ÙÙ At the morning twilight, he should stand reciting softly the Såvitrœ verse until the sun comes into view; but at the evening twilight, he should remain seated until the Big Dipper becomes clearly visible.

Gautama 2.11 tiß™het pürvå∫ åsœtottarå∫ sajyotißyå jyotißo dar†anåd vågyataΔ Ù He should stand at the morning (twilight) from the time the stars are still visible until the sun comes into view, and he should sit at the evening (twilight) from the time the sun is still visible until the stars come into view.

In their respective lists of persons unfit to attend a †råddha, Manu (3.150-166) and Gautama (615.16-19) show close similarities. Even though these similarities may result from the wide circulation of such lists, the near identity of some expressions makes it likely that Manu, although its list is much longer than that of Gautama, is dependent on the latter, especially because such an extensive list is lacking in any other Dharmasütra. The beginings of the two lists are nearly identical: Manu 3.150 ye stenåΔ patitåΔ klœbå ye ca nåstikavr¢ttayaΔ Ù Thoses who are thieves, outcastes, and impotent men, and those who follow the livelihood of infidels ...

Manu 3.158 agåradåhœ garadaΔ kuñ∂农 somavikrayœ Ù an arsonist, a poisoner, someone who eats from the son of an adulteress, a seller of Soma ...

Gautama 15.16 na bhojayet stenaklœbapatitanåstikatadvr¢ttiHe should not feed a thief, an impotent man, an outcaste, an infidel, a man who follows the livelihood of infidels ...

Gautama 15.18 kuñ∂å†isomvikrayyagåradåhigaradasomeone who eats from the son of an adulteress, a seller of Soma, an arsonist, a poisoner ...

adopted the reading adhyeßyamåñas tu gurum, supported by the commentator Nåråyaña. Nandana also comments: brüyåt brahmcårœ; thus the subject is not the teacher but the pupil, thereby supporting the adopted reading. Several reasons prompt me to adopt this reading. The parallel in the GDh (1.46), furthermore, reads adhœhi bho ity åmantrayed gurum. Here the subject is clearly the pupil and the words are addressed to the teacher. Gautama’s reading is supported by ‡åõGr¢ (4.8.12), which contains the identical words adhœhi bho. In the TU (3.1-6) also these words are put in the mouth of the pupil. The request adhœhi

264

LANGUAGE, TEXTS AND SOCIETY

In the above list, only the order of the items is changed by Manu, possibly due to exigencies of meter. The expression of Gautama: upapatir yasya ca saΔ (“a wife’s paramour and her husband”, 15.17), moreover, has its parallel in Manu: yasya copapatir gr¢he (“a man in whose house lives his wife’s paramour”, 3.155). This expression is unique to these two †åstras. Likewise, the somewhat ambiguous compound gañåbhyantara (“someone linked to a guild”) of Manu 3.154 parallels gañapreßya (“someone who is in the service of a guild”) of Gautama 15.18, expressions that are again unique to these two †åstras and probably refer to a Brahmin who does contractual work (most likely of a ritual nature) for a guild. One of the most significant parellels occurs in the section on impurity caused by the birth of a child. The vulgate version of Manu 5.61-62 contains two verses, and they are supported by most manuscripts of Manu: yatheda∫ †åvam å†auca∫ sapiñ∂eßu vidhœyate Ù janane ‘py evam eva syån nipuñå∫ †uddhim icchatåm ÙÙ 61 ÙÙ sarveßå∫ †åvam å†auca∫ måtåpitros tu sütakam Ù sütaka∫ måtur eva syåd upaspr¢†ya pitå †uciΔ ÙÙ 62 ÙÙ As this period of death-impurity is prescribed for those who belong to the same ancestry, so the same holds true at a birth for those who desire perfect purity. Death-impurity affects all, but birth-impurity affects only the mother and the father. The mother alone is subject to the period of birth-impurity; the father becomes pure by bathing.

In my critical edition of Manu, the pådas given in bold are retained and the rest is omitted, resulting in a single verse: janane ‘py evam eva syån måtåpitros tu sütakam Ù sütaka∫ måtur eva syåd upaspr¢†ya pitå †uciΔ ÙÙ The same holds true at a birth, but the birth-impurity affects only the mother and the father. The mother alone is subject to the period of birth-impurity; the father becomes pure by bathing.

I believe that the first half-verse of 61 was introduced by a later editor into the text of Manu. The reason for its introduction was probably the intervention of verse 60, which defines sapiñ∂a, between verses 59 and 61, thus breaking the natural continuity between the latter two. The significant pådas a-b of verse 59 define the period of impurity at the death of a relative: da†åha∫ †åvam å†auca∫ sapiñ∂eßu vidhœyate –“a ten-day period of death-impurity is prescribed for those who belong to the same ancestry”. If this verse came immediately before, then evam in the phrase janane ‘py evam of verse 61 becomes clear; the period of (changed to the middle voice adhœßva in Manu) is made by the pupil to the teacher in all these sources.

Manu and Gautama

265

impurity after a birth is the same as that after a death. The intervention of verse 60 prompted a later editor to insert pådas a-b of verse 61, making the connection between 61 and 59 explicit. A similar commentarial intrusion occurs in the parallel passage of the Vasiß™ha Dharmasütra (4.16-22) between sütra 16 (which is identical to Manu 5.59 pådas a-b) and sütra 20 (which is identical to Manu 61 pådas c-d). But in Vasiß™ha, because the commentarial portion is in prose, the connection between the verses 16 and 20 remains transparent. However, both Manu and Vasiß™ha are dependent, I think, on Gautama 14.14-16. This section of Gautama on impurity begins with the determination of the time of impurity following a death (4.1-12). The section begins: †åvam å†auca∫ da†aråtram anr¢tvigdœkßitabrahmacåriñå∫ sapiñ∂ånåm –“there is a ten-night period of death-impurity for those belonging to the same ancestry, except an officiating priest, one consecrated for a sacrifice, and a vedic student”– a provision that parallels Manu 5.59 and Vasiß™ha 4.16. This section in Gautama concludes (4.13) with a brief definition of sapiñ∂a. Then Gautama (14.14-16) deals with impurity resulting from the birth of a child, two causes of impurity generally treated together in the †åstras: janane ‘py evam Ù måtåpitros tat Ù måtur vå Ù

The same holds true at a birth. It affects the father and the mother. Or just the mother.

It is evident that Gautama is giving here three opinions regarding impurity resulting from childbirth. The first treats it exactly the same as impurity resulting from a death. The second restricts it to the parents of the newborn child. The third restricts it even further to only the mother. That these were opposing views is made clear in the parallel version of Vasiß™ha (4.20-22), which adds reasons for the three opinions (given below in italics): janane ‘py evam eva syån nipuñå∫ †uddhim icchatåm Ù måtåpitror vå bœjanimittatvåt Ù måtur ity eke — [after which a verse is cited in support of this view] The same holds true at a birth for those who desire perfect purity. Or it affects the father and the mother, because [the birth] is caused by the seed. Or just the mother, according to some.

The same three opinions are found in Manu 5.61-62, but they are difficult to discern because of the expansion of an original single verse into two. When we look at the single verse of the critical edition, we see the same three opinions clearly stated: janane ‘py evam eva syån måtåpitros tu sütakam Ù

266

LANGUAGE, TEXTS AND SOCIETY

sütaka∫ måtur eva syåd upaspr¢†ya pitå †uciΔ ÙÙ

The textual dependence of both Manu and Vasiß™ha on Gautama becomes clear when we strip the former of their commentarial accretions. The expansion of the first opinion into a full half-†loka with the additon of the reason nipunå∫ †uddhim icchatåm occurs both in Manu and in VaDh. Given the textual problems inherent in Vasiß™ha (see Olivelle 2000, 631-32), it is possible that the extant version may have been influenced by the revised version of Manu, which contains this expansion. It is also possible that the expansion took place independently and was absorbed into both Manu and Vasiß™ha. There is a similar proximity between Manu and Gautama in the rule of purification after a person has touched an impure substance while holding something in his hand. Both Baudhåyana 1.8.27-29 and Vasiß™ha 3.43 have similar provisions;6 but their formulations are not similar to that of Manu and they instruct the person to place the article he is carrying on the ground, purify himself, and pick the article up again. The provision of Manu is quite the opposite; the person should purify himself without placing the article on the ground. Manu 5.143 ucchiß™ena tu sa∫spr¢ß™o dravyahastaΔ katha∫cana Ù anidhåyaiva tad dravyam åcåntaΔ †ucitåm iyåt ÙÙ But a man who, while carrying something in his hand, is touched by a sullied person/thing, becomes pure by sipping some water without laying that thing down.

Gautama 1.28 dravyahasta ucchiß™onidhåyåcåmet Ù Someone who becomes sullied while holding something in his hand should sip water without [after] laying it down.

I have deliberately left no space between ucchiß™o and nidhåyåcåmet in the text of G. The complication created by Sanskrit sandhi makes the reading ambiguous: we can read the text as either ucchiß™o ‘nidhåya (where the negative ‘a’ is elided and in manuscripts often left unmarked without an avagraha) or ucchiß™o nidhåya (without the negative particle). Thus the latter term may be read as nidhåya or anidhåya. The commentator Maskarin reads it as a positive statement and interprets it to mean that one should place what is in the hand on the ground and then sip water, in accordance with the provisions of 6 The provision of the Baudhåyana Dharmasütra is long and complex and could not have been the source of Manu; whereas the wording of the Vasiß™ha Dharmasütra is very differnt from that of Manu: taijasa∫ ced ådåyocchiß™œ syåt tad udasyåcamyådåsyann adbhiΔ prokßet Ù atha ced annenocchiß™œ syåt tad udasyåcamyådåsyann adbhiΔ prokßet Ù atha ced adbhir ucchiß™œ syåt tad udasyåcamyådåsyann adbhiΔ prokßet Ù Baudhåyana 1.8.27-29; pracarann abhyavahåreßücchiß™a∫ yadi sa∫spr¢†et Ù bhümau nidhåya tad dravyam åcamya pracaret punaΔ ÙÙ Vasiß™ha 3.43.

Manu and Gautama

267

Baudhåyana and Vasiß™ha. Likewise, Medhåtithi, commenting on Manu 5.143, reads it without the negative, and takes this sütra to be in conflict with the provision of Manu. Haradatta, on the other hand, reads it with the negative particle, but says that one should place any food on the ground but not other articles, such as clothes. Whatever the interpretation, it is clear that the †loka of Manu is an expansion of the brief sütra of G. It is also quite likely that the author of Manu read the text of Gautama with an avagraha and understood it to recommend that the article carried in the hand be not placed on the ground before purification. The provisions of Manu and Gautama regarding the fate of lost property also show remarkable similarities that cannot be accidental: Manu 8.30 prañaß™asvåmika∫ riktha∫ råjå tryabda∫ nidhåpayet Ù arvåk tryabdåd dharet svåmœ pareña nr¢patir haret ÙÙ A property whose owner is lost should be kept in deposit by the king for three years. Before the lapse of three years, the owner can claim it; after that the king may take it.

Gautama 10.36-37 prañaß™am asvåmikam adhigamya råj∞e prabrüyuΔ Ù vikhyåpya sa∫vatsara∫ raj∞å rakßyam Ù ürdhvam adhigantu† caturtha∫ råj∞aΔ †eßaΔ Ù If someone finds lost property without an owner, he should disclose it to the king. The king should have it publicized and keep it safely for a year, after which time a quarter goes to the finder and the rest to the king.

There is some textual confusion here between prañaß™asvåmikam of Manu and prañaß™am asvåmikam of Gautama. At least one manuscript of Manu reads prañaß™åsvåmikam (“lost and without an owner”), which would agree with the provision of Gautama. Lakßmœdhara’s [Kr¢tyakalpataru 12, p. 554] citation of Gautama, on the other hand, has the reading prañaß™asvåmikam, agreeing with the reading of Manu. The dependence of Manu on Gautama, however, is unmistakable. Another provision relates to the establishment of legal ownership as a result of the continuous use of a property. Both texts have very similar rules regarding the right of a person to claim ownership by his unobstructed use of a property for a given period of time: Manu 8.147-48 yat ki∫cid da†avarßåñi sa∫nidhau prekßate dhanœ Ù bhujyamåna∫ parais tüßñœ∫ na sa tal labdhum arhati ÙÙ aja∂a† ced apogañ∂o vißaye cåsya bhujyate Ù bhagna∫ tad vyavahåreña bhoktå taddravyam arhati ÙÙ When an owner looks on as something is being enjoyed by others in his presence for ten years, he is not entitled to recover it. If something is

Gautama 12.37 aja∂åpogañ∂adhana∫ da†avarßabhukta∫ paraiΔ sa∫nidhau bhoktuΔ Ù When others make use of the property of a person who is neither mentally incapacitated nor a minor in his presence for ten years, it belongs to the user.

268

LANGUAGE, TEXTS AND SOCIETY

enjoyed within his own locality and he is neither mentally incapacitated nor a minor, he loses any legal right to it; the user is entitled to that property.

Here we have the expansion of a single brief sütra of Gautama into two †lokas by Manu; but the identity of the provison and the terminology makes the dependence of Manu on Gautama unmistakable. Several rules of Manu relating to interest charged on loans parallel those of Gautama. The basic rules regarding the rate of interest, the length of time during which interest accrues, and the classiciation of loans are similar in both and show clear dependence of Manu on Gautama:7 Manu 8.153 nåtiså∫vatsarœ∫ vr¢ddhi∫ na cådr¢ß™å∫ punar haret Ù cakravr¢ddhiΔ kålavr¢ddhiΔ kåritå kåyikå ca yå ÙÙ He must not charge interest beyond one year or what is unauthorized. Cyclical interest, periodic interest, contractual interest, and manual labor [are the kinds of interest].

Gautama 12.30, 34-35 nåtiså∫vatsarœm ity eke Ù cakrakålavr¢ddhiΔ Ù kåritåkåyikå†ikhådhibhogå† ca Ù According to some, [interest does not accrue] beyond one year. Cyclical interest, periodic interest, contractual interest, manual labor, daily interest, and use of the collateral [are the kinds of interest].

The laconic nature of Manu’s †loka has mislead the commentators, who take the verse as a syntactic unit with the negative na governing also the second half of the verse. According to this interpretation, the four types of interest given in the second half are also prohibited. All the translators follow this interpretation. Bhler: “Let him not take interest beyond the year, nor such as is unapproved, nor compound interest, periodical interest, stipulated interest, and corporal interest.” When we look at the source of this †loka in Gautama, we can see clearly that the second half verse merely enumerates the kinds of permitted loans carrying different rates of interest. Another rule relates to a loan taken after a pledge has been given to the creditor. Both Manu and Gautama forbid any interest on such a loan if the creditor makes use of the pledge: Manu 8.143

Gautama 12.32

na tv evådhau sopakåre kausœdœ∫ bhuktådhir na vardhate Ù vr¢ddhim åpnuyåt Ù

7 See also the parallel rules regarding interest not exceeding five time the loan on certain items: M 8.151 and G 12.36.

Manu and Gautama

269

If a pledge together with its use has No interest accrues [on a loan] whose been furnished, however, he shall not pledge has been used. receive any interest

I give below further parallel passages that show clear dependence of Manu on Gautama. These require little comment. Manu 4.50 mütroccårasamutsarga∫ divå kuryåd udaõmukhaΔ Ù dakßiñåbhimukho råtrau sa∫dhyayo† ca yathå divå ÙÙ During the day, he should void urine and excrement facing the north, at night facing the south, and at the two twilights in the same way as during the day.

Manu 4.34 na jœrñamalavadvåså bhavec ca vibhave sati Ù He should not wear old or dirty clothes, if he has the means.

Manu 4.57

Gautama 9.41-43 ubhe mütrapurœße divå kuryåd udaõmukhaΔ Ù sa∫dhyayo† ca Ù råtrau tu dakßiñåmukhaΔ Ù8 He should void both urine and excrement facing the north during the day and at the two twilights, but facing the south at night.

Gautama 9.3 sati vibhave na jœrñamalavadvåsåΔ syåt Ù If he has the means, he should not wear old or dirty clothes.

Gautama 9.54

yaj∞a∫ gacchen na cåvr¢taΔ Ù na yaj∞am avr¢to gacchet Ù He should not go to a sacrifice unin- He should not go uninvited to a sacrivited. fice.

Manu 4.63

Gautama 9.56

notsaõge bhakßayed bhakßån Ù na bhakßån utsaõge bhakßayet Ù He should not eat food placed on his Food placed on his lap, he should not lap. eat.

Manu 5.81 †rotriye tüpsa∫panne Ù But at [the death of] a vedic scholar living near by [the impurity lasts for three days].

Gautama 14.22 †rotriye copsa∫panne Ù And at [the death of] a vedic scholar living near by [the impurity lasts for one day].

8 The text of G here is probably based on a verse original. Vasiߙha 6.10 is probably also based on the text of G.

270

LANGUAGE, TEXTS AND SOCIETY

Here Manu follows Gautama verbatim, but the rule is different; Manu has a three-day period of impurity, whereas Gautama requires only a single day. The author of Manu does not follow Gautama slavishly; he shows independent thinking here as in other areas, such as meat eating and the rules on niyoga. Manu 8.337-38 aß™åpadya∫ tu †üdrasya steye bhavati kilbißam Ù ßo∂a†aiva vai†yasya dvåtri∫†at kßatriyasya tu ÙÙ bråhmañasya catuΔßaß™iΔ pürña∫ våpi †ata∫ bhavet Ù dviguñå vå catuΔßaß™is taddoßaguñavid dhi saΔ ÙÙ With respect to theft, the liability for a ‡üdra is eight times; for a Vai†ya, 16 times; for a Kßatriya, 32 times; and for a Brahmin 64 times, or fully 100 times, or twice 64 times, for he knows what is good and bad.

Gautama 12.15-17 aß™åpådya∫ steyakilbißa∫ †üdrasya Ù dviguñottaråñœtareßå∫ prativarñam Ù vidußo ‘tikrame dañ∂abhüyastvam Ù With respect to theft, the liability for a ‡üdra is eight times. It is progressively doubled for those belonging to each of the prior classes. The punishment is more severe when a learned man commits the crime.

Here the author of Manu has expanded on the brief sütras of Gautama, not only specifying the fines for each varña but also explicitly indicating that it is the Brahmin who deserves to be punished more severely because he is expected to know what is good and bad and not just any learned man. A similar expansion is found in the following parallel dealing with the time and manner of partitioning the paternal estate. Manu also adds the provision that both the father and the mother must be deceased for partition to take place, stating explicitly that the sons are incompetent when either the father or the mother is alive. Manu 9.104-05, 111 ürdhva∫ pitu† ca måtu† ca sametya bhråtaraΔ samam Ù bhajeran paitr¢ka∫ riktham anœ†ås te hi jœvatoΔ ÙÙ jyeß™ha eva tu gr¢hñœyåt pitrya∫ dhanam a†eßataΔ Ù †eßås tam upajœveyur yathaiva pitara∫ tathå ÙÙ pr¢thag vardhate dharmas dharmyå pr¢thakkriyå Ù

Gautama 28.1-3 ürdhva∫ pituΔ putrå riktha∫ bhajeran Ù nivr¢tte rajasi måtur jœvati vecchati Ù sarva∫ vå pürvajasyetarån bibhr¢yåt pitr¢vat Ù

tasmåd vibhåge tu dharmavr¢ddhiΔ Ù

After the father and mother have passed on, the brothers should gather together and partition the paternal estate evenly; for they are incompetent while those two are alive.The eldest alone, on the contrary, ought to take the entire paternal estate, and

After their father has passed on, the sons should partition the estate; or, if the father so wishes, even during his lifetime but after their mother has reached menopause. Alternatively, the entire estate goes to the eldest, and he should maintain the others just as the father.

Manu and Gautama

271

the others should live as his depend- When the estate is partitioned, howevents just as they did under their er, dharma increases. father. Living separately increases dharma; therefore, the act of separation is dharmic.

Manu 11.122

Gautama 25.1-2

marutaΔ puruhüta∫ ca guru∫ påvakam tad åhuΔ katidhåvakœrñœ pravi†atœti Ù eva ca Ù caturo vratino ‘bhyeti bråhma∫ marutaΔ pråñenendra∫ balena br¢hastejo ‘vakœrñinaΔ ÙÙ pati∫ brahmavarcasenågnim evetareña sarveñeti Ù When a votary breaks his vow of So, they ask: “Into how many does chastity, the vedic energy within him someone who has broken his vow of enters these four: Maruts, Indra, chastity enter?” – “Into the Maruts Teacher, and Fire. with his breaths; into Indra with his strength; into Br¢ h aspati with the splendor of his vedic learning; and into just the Fire with everything else.”

Manu 11.134

Gautama 22.23

palålabhåraka∫ ßañ∂he saisaka∫ caiva ßañ∂he palålabhåraΔ sœsamåßa† ca Ù måßakam Ù For (killing) a eunuch, a load of straw For (killing) a eunuch, a load of straw and a Måßa of lead. and a Måßa of lead.

There is, furthermore, a structural parallel between Manu and Gautama at the beginning of their sections on penance. In both, the authors state that they have completed their discourse on the dharma of varñas (Manu 10.131; Gautama 19.1), before they embark on the discourse on penance. The clear distinction between the sections on the varñå†ramadharma and on penance is a feature common to both and a structural innovation continued in later Dharma†åstras. In the section on penance, moreover, both use the technical term anirde†ya to describe a sin for which there is no penance (Manu 11.147; Gautama 21.7), a term that is unique to these two †åstras and does not occur anywhere else. Indeed, Gautama ascribes the rule trœñi prathamåny anirde†yåni to Manu, hinting at a tantalizing possibility of a connection between Gautama and the dharma tradition of the Månavas. What insights can we draw from these parallels with reference to the textual history and composition of the Månava Dharma†åstra?

272

LANGUAGE, TEXTS AND SOCIETY

First, for reasons that have been fully spelled out in the introduction to my critical edition, I believe that Manu was composed by a single author, although some accretions appear to have occured after its initial composition.9 Once we put an author and authorial agency behind the text, as opposed to conceiving the creation of the text as a gradual and almost unconscious accumulation parallel to the formation of an iceberg, then we can ask significant and pertinent questions about authorial intent and his use of sources.10 It is evident that the author of Manu conceived of his †åstra as a charter applicable to all and transcending the narrow boundaries of vedic †åkhås. That Manu is not limited to any †åkhå is clearly articulated by Kumårila (on PMS 1.3.15). The author of Manu also introduced a significant and drastic innovation: unlike the Dharmasütras–which were located within the give and take of an expert tradition, offer glimpses into the divergent views within that tradition, and do not pretend to be anything other than humanly authored works–the Månava Dharma†åstra is presented as a treatise composed and handed down by none other than the creator god Svaya∫bhü. He taught it to his son Manu, who transmitted it to his disciples, including Bhr¢gu, who is made the spokesman and promulgator within the treatise. All this raises interesting questions about the social and political cirmumstances and motivations behind the composition of Manu, questions that are beyond the compass of this paper. I have argued elsewhere (Olivelle 2000, 8) that, contrary to the opinion of Kane and others, Gautama is not the oldest Dharmasütra: “The fact that Gautama is composed entirely in prose sütras that are frequently very brief, thus conforming to the aphoristic ideal, has been considered by some as arguing for its antiquity. I would argue that, on the contrary, the omission of all cited verses, a practice common in all other Dharma texts, argues for the author’s deliberate attempt to produce an ideal sütra work” along the lines of Påñini’s grammar. That the prose of Gautama is probably dependent on verse originals is also indicated by many sütras that scan as pådas from †lokas, especially when some inserted words are removed (e.g., 1.38, 40; 4.2, 8; 8.1; 9.41; 14.22; 22.27). Unattached as it was to a larger Kalpasütra, Gautama may have been conceived as a true †åstra in the manner of Påñini’s grammar, a †åstra that was not confined to a particular †åkhå. If this is true, then we can see how Manu, another such †åstra with a universal application, may have drawn on the text of his predecessor. There is also evidence that Gautama had risen to prominence as the first and perhaps the paradigmatic Dharma†åstra. Kumårila, writ9

I have pointed out above one such accretion at Manu 5.61-62. We see an attempt to put authorial intent back into even the large epic, Mahåbhårata, in several recent studies by Alf Hiltebeitel, Rethinking the Mahåbhårata (Chicago, 2001) and James Fitzgerald in several forthcoming articles (cited by Hiltebeitel) and the introduction of the translation of the RåjadharmakåñÛ∂a (Fitzgerald 2004). 10

Manu and Gautama

273

ing in the 7th century, takes Gautama as the first and foremost of the Dharmasütras, using the expression gautamasütrådi and again listing Gautama first in the list gautama-vasiß™ha-†aõkhalikhita-hårœta-åpastamba-baudhåyanådi. As Kane (1962-75, I: 25) observes: “The Gautama Dharmasütra appears to have been held in high esteem by Kumårila, as in his Tantravårtika he quotes or clearly refers to Gautama Dh. S. at least a dozen times, but quotes Åp. Dh. S. and Baudhåyana Dh. S. only a few times.” The fame of Gautama in the 7th century is also confirmed by a Buddhist source. Dharmakœrti in his Nyåyabindu also places Gautama as the first among the writers of Dharma†åstras: gautamådayo dharma†åstråñå∫ prañetåraΔ (“The composers of Dharma†åstras, beginning with Gautama”).11 It is difficult to assess when Gautama rose to prominence as the premier Dharmasütra, and even more difficult to know whether it had reached that prominence during the time when the Månava Dharma†åstra was composed, at least four or five centuries before Dharmakœrti and Kumårila. Given the numerous parallels between Manu and Gautama that I have discussed above, however, and the evidence several centuries later about the prominence of Gautama, I want to present the hypothesis that Gautama had assumed prominence as the chief Dharmasütra by the time Manu was composed. If this is true, then we can understand how Gautama exerted the kind of influence on the author of Manu that we see reflected in the text. Even though Manu purports to be a divine revelation, we must assume that the author operated within the expert tradition of dharma in a way similar to his predecessors. This becomes evident in many instances when the author somehow forgets the divine angle and resorts to common pandit discourse of citing opinions and arguing against opponents.12 The author of Manu was influenced by two expert traditions. We have looked at the Dharma tradition represented principally by Gautama. He was also influenced by the Artha tradition especially in the long chapters 7-9 on Råjadharma. The dependence of Manu on the extant Artha†åstra is also evident when we analyse the two texts, and this will be the focus of one of my forthcoming studies. The Månava Dharma†åstra, however, is not simply a patchwork of material borrowed from different sources. The author integrated what he borrowed into an overall scheme that is very much his own. 11 Dharmakœrti, Nyåyabindu, ed. Chandra†ekhara ‡åstrœ; Kashi Sanskrit Series, 22 (Banaras: Chowkhamba Sanskrit Series, 1954), p. 99. 12 There are, of course, the numerous verses that ascribe a rule to Manu, even though the entire text is ascribed to Manu (e.g., 8.139, 279; 9.239; 10.63, 78). Now, it is certainly true that such ascriptions to the reputed author are found in other †åstras as well, including the Artha†åstra. But these texts, unlike Manu, do not pretend to be anything other than humanly authored compositions. The author of Manu also falls into the habit of using the pandit idiom of iti cet occasionally: 9.122; 10.82. Reference is also made to the views of others: 9.31, 158.

274

LANGUAGE, TEXTS AND SOCIETY

His intent was clearly to produce a †åstra that would be superior in structure, style, and content to all that preceded him. In this he was clearly successful, a success evidenced by the extraordinary reception the text has received in the Dharma tradition spanning nearly two millennia.

A

275

XII Manu and the Artha†åstra: A Study in ‡åstric Intertextuality*

An individual belonging to and writing within a tradition of expert knowledge (†åstra) is likely to compare and contrast his or her views to other exponents of that tradition. Modern scholars do this by means of bibliographical notes. Ancient Indian scholars resorted to several strategies, including citation of authoritative works, as well as presenting and combating opposing (pürvapakßa) views. Within the expert tradition of dharma, the earliest extant texts, the Dharmasütras, frequently cite opinions of other experts with which the authors often disagree. This strategy is also evident in the Artha†åstra (A‡) and must have been a common practice among the early †åstric writers. The overlap between topics dealt with in the †åstric texts devoted to dharma and artha, especially with regard to the king and government, has been noted in previous scholarship.1 This paper, however, explores a different kind of †åstric intertextuality: an author’s use of and dependence on a pre-existing textual tradition in the creation of a new text. The author of the Månava Dharma†åstra evidently used pre-existing sources in composing his treatise.2 Here I explore the connections between Manu and the Artha†åstra of Kau™alya.3 Originally pulished in the Journal of Indian Philosophy. 32 (2004): 281-91. See Jolly 1913; Winternitz 1926-28. The relationship of the Kåtyåyanasmr¢ti to the A‡ was dealt with briefly by Kane (p. xiii) in the introduction to his edition of that smr¢ti (Poona: Oriental Book Agency, n.d.). 2 There has been an ongoing debate about whether the metrical text we possess was a versification of an earlier Månava Dharmasütra, a theory first proposed by Max Mller and given wide publicity by Bhler. The theory was rejected by Kane and has now few supporters. See Olivelle forthcoming-b for Manu’s dependence on the Gautama Dharmasütra, and also my introduction to Olivelle 2004a. 3 Kau™alya refers to a school of Månavas to which he ascribes several opinions. It is doubtful that this Artha†åstric school of Månavas has any connection to the Månava Dharma†åstra: see Kangle 1964, 49-50; 1960-65: III: 80. * 1

276

LANGUAGE, TEXTS AND SOCIETY

Given the problems inherent in the dating of these two texts, it is not possible to assert with a high degree of confidence who is borrowing from whom. I do agree with Kangle, however, that it is most likely that at least sections of the Artha†åstra are older than Manu and are the source for some of the passages and vocabulary I will discuss below.4 The vocabulary analysis does not support Jolly’s (1923) and Fezas’s (1996) contention that the A‡ is contemporaneous with the late smr¢ti texts of Yåj∞avalkya and Nårada. My intent, however, is to show that Manu borrowed from the Artha†åstric tradition in his treatment of the king and civil administration, and of criminal and civil law (råjadharma and vyavahåra) rather than to demonstrate his direct dependence on the extant Artha†åstra, although this appears likely in at least some passages. This study will also show that it is not possible to understand the vocabulary of Manu without reference to the technical terms developed within the Artha†åstric tradition. Perhaps the clearest interconnection between Manu and the Artha†åstra at the level of both text and vocabulary is found in the following passages dealing with judicial procedure and reasons for losing a case.5 Manu 8.52-57 apahnave ‘dhamarñasya dehœty uktasya sa∫sadi Ù abhiyoktå di†ed de†a∫ karaña∫ vånyad uddi†et ÙÙ 52 ade†a∫ ya† ca di†ati nirdi†yåpahnute ca yaΔ Ù ya† cådharottarån arthån vigœtån nåvabudhyate ÙÙ 53 apadi†yåpade†ya∫ ca punar yas tv apadhåvati Ù samyak prañihita∫ cårtha∫ pr¢ß™aΔ san nåbhinandati ÙÙ 54 asa∫bhåßye såkßibhi† ca de†e sa∫bhåßate mithaΔ Ù nirucyamåna∫ pra†na∫ ca necched ya† cåpi nißpatet ÙÙ 55 brühœty ukta† ca na brüyåd ukta∫ ca na vibhåvayet Ù na ca pürvåpara∫ vidyåt tasmåd arthåt sa hœyate ÙÙ 56 såkßiñaΔ santi mety uktvå di†ety ukto di†en na yaΔ Ù dharmasthaΔ kårañair etair hœna∫ tam api nirdi†et ÙÙ 57 52When the debtor, told in court to pay up, denies the charge, the

plaintiff should produce a document (de†a) or offer some other evidence. 53When the plaintiff produces something that is not documentary evidence (ade†a); produces and then disavows it; does not realize that his earlier points contradict the ones he makes subsequently; 54states his case and then backs away from it; does not acknowledge under questioning a point that has been clearly established; 55secretly discusses with witnesses a document (de†a) which is prohibited from being discussed; objects to a question clearly articulated; retreats; 4 See Trautmann 1971, 1850-86; Kangle 1964, 1960-65, III: 80-83. I do not accept all of Kangle’s arguments, some of which are rather weak. Nevertheless, a good case can be made, I believe, for the Arth†åstra being prior to Manu. At least in the sections on the king and jurisprudence, Manu appears to have leaned heavily on the Artha†åstric tradition. 5 Kangle 1964 already drew attention to this parallel, but he did so only briefly and without analysing the text or the vocabulary.

Manu and the Artha†åstra

277

56does not speak when he is ordered “Speak!”; does not prove what he

asserts; and does not understand what goes before and what after –such a plaintiff loses his suit. 57When a plaintiff says “I have people who know”, but when told “Produce them” does not produce them, the judge should declare him also the loser for these very reasons. A‡ 3.1.19 (a) nibaddha∫ vådam utsr¢jyånya∫ våda∫ sa∫kråmati, (b) pürvokta∫ pa†cimenårthena nåbhisa∫dhatte, (c) paravåkyam anabhigråhyam abhigråhyåvatiß™hate, (d) pratij∞åya de†a∫ nirdi†ety ukte na nirdi†ati, (e) hœnade†am ade†a∫ vå nirdi†ati, (f) nirdiß™åd de†åd anya∫ de†am upasthåpayati, (g) upasthite de†e ‘rthavacana∫ naivam ity apavyayate, (h) såkßibhir avadhr¢ta∫ necchati, (i) asa∫bhåßye de†e såkßibhir mithaΔ sa∫bhåßate, (j) iti paroktahetavaΔ ÙÙ (a) When someone casts aside the plaint as recorded and moves on to another plaint; (b) does not make a point made later accord with what was stated before; (c) after challenging an unchallengeable statement of the opponent, remains (stubborn); (d) promises to produce a document (de†a), but when told “Produce it”, does not produce it; (e) produces a defective document (hœnade†a) or something that is not documentary evidence [or a false document: ade†a]; (f) puts forward a document (de†a) different from the document (de†a) specified; (g) denies a significant statement in the document (de†a) he has put forward, saying “It is not so”; (h) does not accept what has been ascertained through witnesses; (i) secretly discusses with witnesses a document (de†a) which is prohibited from being discussed. (j) These are the reasons for loss of suit.

Manu’s discussion is clearly longer and includes items not noted by the Artha†åstra; but the connection between the two passages is unmistakable. The closest parallels are: Manu 52-53 and A‡ (d)-(f), apahnave ‘dhamarñasya dehœty uktasya sa∫sadi Ù abhiyoktå di†ed de†a∫ karaña∫ vånyad uddi†et ÙÙ ade†a∫ ya† ca di†ati nirdi†yåpahnute ca yaΔ Ù ya† cådharottarån arthån vigœtån nåvabudhyate ÙÙ

pratij∞åya de†a∫ nirdi†etyukte na nirdi†ati, hœnade†am ade†a∫ vå nirdi†ati, nirdiß™åd de†åd anya∫ de†am upasthåpayati

Manu 54 (pådas b-c) and A‡ (i): asa∫bhåßye såkßibhi† sa∫bhåßate mithaΔ Ù

ca

de†e

asa∫bhåßye de†e såkßibhir mithaΔ sa∫bhåßate

What binds the two passages together, however, is the obscure and obsolete word de†a, whose meaning appears to have been long forgotten by the time of Manu’s commentators. Indeed, many commentators of Manu, including Kullüka, and most printed editions substitute the

278

LANGUAGE, TEXTS AND SOCIETY

reading de†ya for de†a.6 Following this reading, Bhler translates pÂdas cd of verse 52: “the complainant must call (a witness) who was present (when the loan was made), or adduce other evidence.” The many parenthetical words indicate the problem Bhler had with the term. He takes it as a roundabout way of referring to an eye witness, without explaining why Manu did not use the perfectly clear word såkßin. In a footnote, Bhler notes the variant reading de†a of Medhåtithi and others and states that the meaning would then be that the plaintiff “(must point out) the place”, giving de†a its common meaning. The latter meaning is ascribed to the term also by Derrett (1975), and here he follows Bhåruci’s explanation of the term.7 The commentators generally follow the one or the other interpretation: “place” in Bhåruci, Sarvaj∞a-Nåråyaña, and Nandana; “witness” in Medhåtithi, Govindaråja, Kullüka, Råghavånanda, Råmacandra, and Mañiråma. In the A‡, the reading is clearly de†a and the context provides better clues as to its meaning. That the term refers to some form of evidence is clear. Kangle (1960-65) takes it to mean evidence in general8 and translates the sections (d)-(g) as follows: “after making an affirmation does not indicate the evidence when asked to do so, indicates weak evidence or false evidence, produces evidence other than that indicated, when evidence is produced denies a statement in the matter saying ‘it is not so’.”9 I think Kangle is right in taking de†a to be a term referring to evidence; but he is wrong in taking it to mean evidence in general. In sütra (d), for example, Kangle takes pratij∞åya as standing alone; I think it should govern the following accusative de†am: “having promised a de†a.” It is superfluous to note that a plaintiff would promise to produce evidence in support of his claim; this would be expected by the normal rules of procedure. Further, taking de†a as evidence in general makes it difficult to understand the meaning of ade†a and hœnade†a occurring in the very next sütra (e). Kangle’s translation “weak evidence” and “false evidence” does not seem plausible. Under his interpretation of de†a, the expressions should mean “non-evidence” and “deficient evidence”, both somewhat problematic. A comparison with Manu 8.52 helps us narrow down the meaning. abhiyoktå di†ed de†a∫ karaña∫ vånyad uddi†et ÙÙ 6 The reading de†a is adopted by Jolly (1887) and by me in my critical edition (Olivelle 2004a). 7 The same meaning is found also in the translations by Burnell (1884) and Doniger (1991), although she takes it to be a reference to a witness who was at the place where the transaction took place. 8 See the note to his translation of sütra 3.1.15: “In s. 19 below, de†a seems distinguished from såkßin. It may be understood as evidence in general.” The term de†a occurs at A‡ 3.16.29; 4.6.9 as title of ownership, and here also a written proof of ownership may have been meant. See Kangle 1960-65, III: 218. 9 Shamasastry (1915) clearly did not understand the meaning of the term and translates it as “question” at issue.

Manu and the Artha†åstra

279

The plaintiff should produce a de†a or offer some other evidence.

The plaintiff is required either to produce a de†a or to offer another karaña. It is clear that de†a is viewed here as a sub-category of karaña.10 The latter term, of course, has many meanings; but within the judicial context it refers to an evidentiary instrument, including a written document, as seen in many A‡ passages.11 I believe that in this context of evidentiary instruments, de†a means a piece of documentary evidence. The term lekhya for a document produced as evidence, a term that becomes standard in later dharma texts,12 is not found in either the A‡ or Manu within the context of judicial proceedings, even though the A‡ uses the term for other kinds of writing.13 The conclusion then is that de†a was an old term for a document produced in a court of law, possibly related to di† in the sense of pointing out; see the expression di†et desam in Manu 8.52. Within this context it is easier to understand the meanings of ade†a and hœnade†a. The first is a document that does not count as a proper legal document, and the second is a document that is in some way deficient, for example, without the signatures of the witnesses or damaged at some significant point. We cannot say more than this without a clearer idea about the physical makeup of early Indian court documents. This passage is also the only place where Manu uses the Artha†åstric term dharmastha for a judge; elsewhere he uses the normal Dharma†åstric term prå∂viveka.14 In this passage, we see a clear dependence of Manu on the A‡ both textually and in terms of vocabulary. Several other textual parallels can be cited with regard to the dependence of Manu on the A‡. Manu 8.279 yena kenacid aõgena hi∫syåc cec chreß™ham antyajaΔ Ù chettavya∫ tat tad evåsya tan manor anu†åsanam ÙÙ When a lowest-born man uses a particular limb to injure a superior person, that very limb of his should be cut off –that is Manu’s decree. A‡ 3.19.8 †üdro yenåõgena bråhmañam abhihanyåt tad asya chedayet Ù When a ‡üdra uses a particular limb to injure a Brahmin, he [the judge] should have that limb of his cut off. 10 That karaña could also refer to documentary evidence is indicated by its use at A‡ 3.1.16, where a later karaña annuls an earlier one. See Kangle 1960-65, III: 218. 11 In the A‡ 3.1.17 the term appears to mean title of ownership; in 3.1.15-16 documents drawn up during a transaction; and in 3.12.37-38, 4.8.13 evidence of a general sort. 12 See, for example, Nårada Smr¢ti 1.114-26; Yåj∞avalkya Smr¢ti 2.84-94. 13 See A‡ 1.16.25; 2.7.28; 5.2.9; 11.1.52; 12.2.21. The term is missing in the MDh, although we have lekhitam in MDh 8.168. 14 See MDh 8.79, 181; 9.234.

280

LANGUAGE, TEXTS AND SOCIETY

In the above passage, one can see that Manu has given greater rhetorical elaboration to the terse sütra of the A‡. Manu 8.242 anirda†åhå∫ gå∫ sütå∫ vr¢ßån devapa†ü∫s tathå Ù sapålån vå vipålån vå na dañ∂yån manur abravœt ÙÙ A cow within ten days after giving birth, bulls, and animals dedicated to gods are not subject to punishment, whether they are attended by a herdsman or not –so has Manu declared. A‡ 3.10.24 gråmadevavr¢ßå vå anirda†åhå vå dhenur ukßåño govr¢ßå† cådañ∂yåΔ Ù Bulls belonging to a village or temple, a cow within ten days after giving birth, and stud bulls are not subject to punishment.

The verbal correspondences between the texts in these two examples are clear, but especially noteworthy here is that in both cases the author of Manu thought it necessary to explicitly ascribe this to Manu, with the two standard phrases manur abravœt and manor anu†åsanam. Manu 8.332

syåt såhasa∫ tv anvayavat prasabha∫ karma yat kr¢tam Ù niranvaya∫ bhavet steya∫ kr¢tvåpavyayate ca yat ÙÙ When an act is committed with force and in the presence (of the victim), it is “violence”; when it is committed outside his presence, it is “theft”, and so is an act that someone commits and then denies.

A‡ 3.17.1-2

såhasam anvayavat prasabhakarma Ù niranvaye steyam apavyayane ca Ù When an act is committed with force and in the presence (of the victim), it is “violence”; when it is committed outside his presence, it is “theft”, and also when it is denied.

In this passage also the A‡ is terse and precise and Manu, as usual, is more verbose. We have here a clear statement defining two forms of stealing: robbery that involves violence in the presence of the owner (what we would call today “mugging”), and theft that is done clandestinely when the owner is not present. The two significant and obscure terms, anvaya and niranvaya, are used only here in the A‡, whereas in Manu we have two other occurrences: 8.198, 331. These terms have been subject to much misunderstanding; they clearly mean “in the presence of” and “outside of presence of”, respectively. The occurrence of these two terms in Manu and the A‡ is even more significant, because they fell out of use in later texts.15 15 Only the Kåtyåyanasmr¢ti (796) has this term, in a verse that is probably a paraphrase of Manu: sånvayas tv apahåro yaΔ prasahya haraña∫ ca yat Ù såhasa∫ ca bhaved eva∫ steyam ukta∫ vinihnavaΔ ÙÙ

Manu and the Artha†åstra

281

Textual convergence is also found in the discussion of the two forces that impact on the success of enterprises, fate and human effort. Manu 7.205 sarva∫ karmedam åyatta∫ vidhåne daivamånuße Ù tayor daivam acintya∫ tu månuße vidyate kriyå ÙÙ All activities here depend on divine and human dispensations. Of these, however, the divine is inscrutable; action is possible only with respect to the human. A‡ 6.2.6-12 månußa∫ nayåpanayau daivam ayånayau Ù daivamånußa∫ hi karma loka∫ yåpayati Ù adr¢ß™akårita∫ daivam Ù [...] dr¢ß™akårita∫ månußam Ù [...] tac cintyam Ù acintya∫ daivam ÙÙ Good policy and bad policy are acts of human agency; good fortune and bad fortune are acts of divine agency, for acts of divine and human agency are what keeps the world going. The divine is that which is caused by an invisible agency... The human is that which is caused by a visible agency. The latter can be scrutinized, while the divine is inscrutable.

Manu here appears to be a metrical synopsis of the longer A‡ statement, with the significant terms cintya/acintya occurring in both. The connection between Manu and the A‡, however, goes deeper than the mere presence of parallel texts.16 They share a common and unusual vocabulary. I have already noted the term de†a for documentary evidence presented in a court of law, dharmastha for judge, and the terms anvaya and niranvaya in the context of robbery and theft. Another significant term in the vocabulary unique to the two is the verb pravåsayet and the related nouns pravåsa/pravåsana. These terms are employed frequently in the A‡ both with the ordinary meaning of sending someone into exile and with its more technical, perhaps euphemistic, meaning of putting someone to death. The use of these terms in their technical meaning has been overlooked by the translators and commentators of Manu alike, creating much confusion and misunderstanding. Take, for instance, Manu 8.123: kau™asåkßya∫ tu kurvåñå∫s trœn varñån dhårmiko nr¢paΔ Ù pravåsayed dañ∂ayitvå bråhmaña∫ tu vivåsayet ÙÙ When individuals of the three classes give false testimony, a righteous king should first fine them and then execute them; a Brahmin, on the other hand, he should send into exile. 16 See also the parallels at Manu 7.99, 101 and A‡ 1.4.3 (see Scharfe 1993, 46-7); Manu 8.87-101 and A‡ 3.11.34-37; Manu 8.299-300 and A‡ 3.3.8-9; Manu 9.153 and A‡ 3.6.17; Manu 8.224 and A‡ 3.15.14; Manu 8.159 and A‡ 3.16.9; Manu 8.367 and A‡ 4.12.3; Manu 9.217 and A‡ 4.11.9; Manu 9.277 and A‡ 4.10.1; Manu 9.279 and A‡ 4.11.17; Manu 9.282 and A‡ 2.36.26-

282

LANGUAGE, TEXTS AND SOCIETY

The major difficulty in interpreting this verse lies in the distinction between pravåsa (prescribed for the three lower classes) and vivåsa (prescribed for Brahmins). Bhler, Doniger, and Burnell take both to mean “banish”; the difference then is that Brahmins are only banished, whereas the others are both fined and banished. This is the interpretation offered by the commentators Nåråyaña, Kullüka, and Råghavånanda. Medhåtithi and Govindaråja, on the other hand, take vivåsa to mean depriving of clothes, i.e., making him naked. Medhåtithi thinks that the term may also mean depriving a man of his house, takign våsa in the sense of a residence. I think all these are mistaken. The correct interpretation is offered by the commentators Bhåruci and Nandana, both significantly representing the southern tradition where, I believe, the Artha†åstric traditions survived longer than in the north. The term pravåsayet in this verse of Manu has the same meaning as parallel statements in the Artha†åstra,17 where it refers to execution. Indeed, Medhåtithi (on 8.284) calls this meaning an Artha†å†tra usage (artha†åstrayå). Like the modern militaryinspired term “liquidate” or the more common “get rid of”, pravåsayet may have been an euphemism for imposing the death penalty. This meaning of the term is supported by the very next section (8.124-30) that deals with corporal punishment. The only other time that vivåsa is used by Manu (9.241) it means exile; there also it deals with Brahmins who should be sent into exile without confiscating their property. The use of the term pravåsayet with the technical meaning of capital punishment becomes obsolete in later texts.18 Another technical term common and unique to Manu and the A‡ is prakr¢ta. The term is used in the A‡ at 2.7.10 and 2.8.24 with the clear meaning of a government official or appointee of the king. pracåracaritrasa∫sthånåny anupalabhamåno hi prakr¢taΔ Ù (A‡ 2.7.10) For an officer not conversant with the activities, customs, and rules [...] pracåre cåvaghoßayet “amunå prakr¢tenopahatåΔ praj∞åpayantu” iti Ù (A‡ 2.8.24) And he (the king) should make a proclamation within the (region of his) activity: “Those wronged by that officer should make it known (to me).”

This is quite an unusual and possibly technical meaning of this term. It occurs nowhere else in the legal literature except in Manu. At Manu 8.11 we have the expression raj∞a† ca prakr¢taΔ with the meaning “an officer/appointee of the king”, here the judge. The inability 7; Manu 9.294-95 and A‡ 8.1.19; Manu 11.14 and A‡ 3.14.37; Manu 7.105 and A‡ 1.15.60; Manu 11.14 and A‡ 3.14.37. 17 See A‡ 1.18.16; 4.13.8, 9, 20, 21; 7.7.13; 11.1.33, 47; 12.3.4; 12.4.4; 12.5.23; 13.4.29. Kangle correctly interprets this term to mean execution. 18 Indeed, even the modern Sanskrit dictionaries do not give this as one of the meanings of the verb. They give the meaning of killing for pravåsana, however, Böhtlink-Roth and Monier-Williams ascribing this meaning simply to lexicons.

Manu and the Artha†åstra

283

to understand this meaning of the term has led scribes and commentators to change it to adhikr¢ta, which Bhler translates as an adjective “appointed” with an implied noun: “the learned (judge) appointed by the king”.19 Manu also uses the verb prakurvœta with the technical meaning of “appointing officials” at 7.54, 60, 61, 63, an expression that is absent in the A‡: maulå∞ chåstravidaΔ †ürå∫l labdhalakßån kulodgatån Ù sacivån sapta cåß™au vå prakurvœta parœkßitån ÙÙ 7.54 The king should appoint seven or eight counselors. They must be individuals who are natives of the land, well-versed in the Treatises, brave, well-accomplished, and coming from illustrious families, individuals who have been thoroughly investigated. anyån api prakurvœta †ucœn pråj∞ån avasthitån Ù samyagarthasamåhart≤n amåtyån suparœkßitån ÙÙ 7.60 He should also appoint other officials. They must be individuals who are honest, intelligent, steadfast, and able to collect revenues properly, individuals who have been thoroughly investigated.

Once again it appears that this technical meaning is derived from the Artha†åstric vocabulary. Another Artha†åstric term used by Manu (7.207) and by no other Dharma†åstric writer is pårßñigraha, “the heel-catcher”, the technical term for a king’s ally at the rear of an enemy who is attacking that king. This ally can be called upon to launch an attack at the rear of the enemy, thus relieving some of the pressure on his own forces. The expression †uddhavadha (“clean killing”) as a technical term for beheading is, in all probability, also from the Artha†åstric vocabulary. Among the authors of Dharma†åstras, it is used only by Manu at 9.279. Its meaning becomes clear only in its use within the A‡,20 where this expression is contrasted to citravadha (A‡ 4.11.1), used once by Manu (9.248) also, which refers to the execution of a criminal using torture, dismemberment, and impalement, described in A‡ 4.11.7, 11, 13, 19. The A‡ does not deal with ordeals at all in connection with deciding cases brought before a court. Neither does Manu, except that he permits the use of †apatha or oath. The A‡ uses the term †apatha seven times21, mostly with the meaning of simply oath. But at A‡ 3.1.46, where five means of deciding a lawsuit are given, †apatha appears to imply also ordeals. Manu also uses the term to cover some forms of ordeal. In the four verse (8.109-12) he clearly employs the term sim19 In my critical edition of Manu (Olivelle 2004a), the original reading of prakr¢ta has been restored. 20 A‡ 4.9.2; 4.10.16; 4.11.2, 15, 26. 21 A‡ 1.10.3; 3.1.46; 3.20.17; 7.17.3, 5, 7, 8.

284

LANGUAGE, TEXTS AND SOCIETY

ply for an oath. At 8.115, however, he appears to refer to the fire and water ordeals with the term †apatha. yam iddho na dahaty agnir åpo nonmajjayanti ca Ù na cårtim r¢cchanti kßipra∫ sa j∞eyaΔ †apathe †uciΔ ÙÙ When the blazing fire does not burn a man, the water does not push him up to the surface, and no misfortune quickly strikes him, he should be judged innocent by reason of his oath.22

Ordeals may also be implied in the use of the term at 8.190. It is remarkable, however, that neither Manu nor the A‡ uses the word divya, which becomes the standard term for ordeals in later legal literature.23 On a broader structural level, furthermore, it is clear that the eighteen grounds for litigation (vyavahårapadas) that appear in Manu (Ch. 8-9) for the first time within the Dharma†åstric tradition are derived from the Artha†åstric tradition. The direct connection between the A‡ and Manu, however, is less evident, because both the number and the order in which the vyavahårapadas are enumerated in the two texts differ greatly. The A‡ (1.19.6-25), furthermore, recommends that the king develop a routine for the day and the night. He is asked to divide the day and the night into eight parts each and to perform specific tasks during each period. For example, during the first part of the day he reviews matters of defense and revenue; during the second, looks into the affairs of the citizens; and during the third; takes his bath and the midday meal. Likewise, during the first part of the night he interviews secret agents. This admonition of the A‡, perhaps, was the inspiration for the structure that Manu gives to his central section (Manu 7.145226) on the duties of a king (råjadharma). In this section, Manu devices a narrative scheme to span a single day, from the morning when the king wakes up until nightfall when he goes to bed. Manu squeezes into a single day the description of all the duties of a king spread over 182 verses: meeting with counselors, political strategies, conduct of war and peace, strategies in the aftermath of a victory and conquest, as well as the more commonplace matters of exercise, eating, recreation, and sleep. The morning routine extends from 7.145 to 7.215; the afternoon routine from 7.216 to 7.222; and the evening routine from 7.223 to 7.226. This part concludes with the king going “to bed at the proper time and rise up refreshed.” Although it is clear that Manu depended heavily on the Artha†åstric tradition for his material on the king, government, law, and the judiciary, Manu did not slavishly borrow from his sources. As 22 23

For a description of these ordeals, see Nårada Smr¢ti 20.1-24. TheA‡ uses the term divya at 13.1.8 but with a very different meaning.

Manu and the Artha†åstra

285

I have shown elsewhere,24 the Månava Dharma†åstra was written by a single individual of great talent. He gave a deliberate and unique structure to his treatise. The same attention to organization and systematic presentation is evident also in his treatment of the material that overlap with the concerns of the Artha†åstric tradition. Manu borrowed no doubt, but he integrated the material he borrowed into his own organizational scheme, presenting thus quite a unique text within the literary tradition of Dharma†åstra.

24 See Olivelle 2002a and the introduction to my critical edition of Manu (Olivelle 2004a).

286

LANGUAGE, TEXTS AND SOCIETY

287

XIII Unfaithful Transmitters: Philological Criticism and Critical Editions of the Upanißads*

Since the nineteenth century there has been among western scholars a pervasive mistrust of ancient Indian interpreters and commentators, especially the much-maligned ‡a∫kara, as reliable guides to understanding ancient Indian texts. Early scholars were confident –to modern eyes, overconfident– of their ability to uncover “original” meanings through philological acumen unmediated by native gloss or comment. The observation of Dwight Whitney (1890, 407) in his review of Otto Böhtlingk’s (1889a, 1889b) editions and translations of the two major Upanißads, the Br¢hadårañyaka and the Chåndogya, typifies this attitude: “And the translation is of that character which I pointed out in a paper in this Journal1 some years ago as most to be desired–namely, simply a Sanskrit scholar’s version, made from the text itself, and not from the native comment, and aiming to represent just what the treatises themselves say, as interpreted by the known usages of the language.” Although due to theoretical advances in many fields, we are today, on the one hand, less confident of our ability to recover “original” meanings of ancient documents and, on the other, more aware of the importance of the history of the reception, understanding, and interpretation of texts within the native traditions, the bias against commentators persists. To some degree this bias is justified, not because of some perversity or ignorance on the part of the commentators, but because their goals were different from those of modern scholars; they were primarily theologians and apologists strugOriginally published in the Journal of Indian Philosophy. 26 (1998): 173-87. American Journal of Philology 7(1886): 1-26. This statement is all the more striking because, as I will indicate below, Whitney was one philologist who did have a healthy mistrust of philological reconstructions divorced from or contrary to manuscript evidence. * 1

288

LANGUAGE, TEXTS AND SOCIETY

gling to discover theological truths in their authoritative scriptures. To criticize them for what they did not set out to do is misplaced,2 but they may not always be suitable guides for the work of the textual scholar. The distrust of the commentators’ interpretations, however, spilled over into doubts about the reliability of the textual transmission mediated by these commentators and more broadly into a mistrust of the scribal tradition as such. This mistrust is most evident in the case of the Upanißads, a group of texts that came under close scholarly scrutiny both because of their centrality in later Indian theological discourse and because of the perceived philosophical importance of their message. To restore these texts to their presumed pristine state prior to the corrupting intervention of scribes and commentators, European Sanskritists of the nineteenth and early twentieth centuries undertook to create “critical editions” of several Upanißads, the most famous of which were Böhtlingk’s editions of the Br¢hadårañyaka and the Chåndogya. Strange though it may seem, none of these so-called “critical editions”3 used manuscript material; even where manuscript differences are noted, the editions were not based on a careful sifting and collation of all the available manuscript evidence and the application of recognized editorial principles. There is no evidence that even a thorough search for manuscripts was ever undertaken. Indeed, even modern “critical editions” of the Upanißads, such as those of Limaye and Vadekar 1958, Frenz 1968-69, Maue 1976, PérezCoffie 1994, and Oberlies 1995, are not based on a methodical search for and collation of manuscripts. Most are based on comparisons of previously printed editions, often perpetuating printing errors and conjectural readings of these earlier versions. Frenz’s edition of the Kaußœtaki Upanißad, for example, uses only a single manuscript conveniently located in Tbingen; all other “variants” are derived from previously printed editions. Maue’s (1976) edition of the first chapter of the Br¢hadårañyaka Upanißad (Kåñva recension) has greater claim to be a critical edition. He uses six manuscripts (three from India and three from Europe) but seven printed editions; the limited use of manuscript material may be explained in part as due to his use of only manuscripts containing accentuation. Pérez-Coffie (1994) follows Maue closely in his “critical edition” of the second chapter of the Br¢hadårañyaka (Kåñva recension) and adds only a single manuscript 2 The stark truth, however, is that most, if not all, western translators and interpreters look to ancient commentators and modern pandits for guidance, especially in difficult passages and in technical matters. Their assistance, like that of informants in the case of ethnographies, is often left unacknowledged. The commentators were clearly closer to the traditions within which the texts were composed and are often indispensable guides especially in technical matters, such as grammar and ritual techniques. 3 The reader may consult the appended bibliography to see the frequency with which the epithet “Kritische Ausgabe” is included in the title of these editions.

Unfaithful Transmitters

289

to Maue’s six. In both these editions, however, there is no evidence of a thorough search for manuscripts especially in India. The “critical” in many of these so-called “critical editions”, especially the older ones of Böhtlingk (1889a, 1889b), Hertel (1924), and Hauschild (1927), consists principally in the application of philological acumen to editorial reconstruction. Philological conjectures are an important part of scholarly investigation of texts. The question I want to address in this paper, however, is whether it is legitimate to incorporate such conjectures into the very body of the edited text. I will show that such emendations, especially when they are not clearly demarcated as conjectures, create a new and unfaithful textual transmission that often misleads later scholars and stifles scholarly debate about the received text. As Salomon (1991, 48) says, “Whether a true critical edition would clarify the textual and linguistic questions about this [i.e., Pra†na] Upanißad (and, again, other Upanißads as well), or whether such an edition is even feasible, remains to be seen.” Yet, the feasibility of a critical edition cannot be determined unless a manuscript search and collation is undertaken.4 If after such a search one finds only a few variants of significance, that itself will open important and interesting questions regarding the textual history and the transmission of the Upanißads.

I Böhtlingk’s 1889 editions and translations of the Br¢hadårañyaka and the Chåndogya Upanißads were a landmark in Upanißadic scholarship and exerted considerable influence on later scholars. He followed an unconventional – the less charitable but more accurate term may be outrageous – editorial practice, however, especially in the Chåndogya. He gave within the body of the edited text all his conjectural readings and philological improvements to the received text. The corresponding readings in the received text, on the other hand, he placed either in footnotes at the bottom of the page or, even more 4 Some scholars have opined that establishing the root text on which ‡a∫kara’s commentaries were based will provide a better text than one that could be arrived at by sifting manuscript evidence, principally because ‡a∫kara pre-dates any manuscript we are likely to discover. Max Mller (1879, lxxi) proposed this over a century ago: “I therefore hold that when we succeed in establishing throughout that text which served as the basis of ‡aõkara’s commentaries, we have done enough for the present, and have fulfilled at all events the first and indispensable task in a critical treatment of the text of the Upanishads.” Rau (1960, 299) has also expressed a similar opinion with reference to the Kåñva recension of the Br¢hadårañyaka Upanißad. Clearly the text that served as the basis of ‡a∫kara’s commentary is of vital importance for establishing a critical edition; but I think that it is not sufficient. There may well have been other lines of transmission preserved in the manuscript tradition; lateness of a manuscript does not always mean that it is not a witness to a reading that may well be very ancient. Further, ‡a∫kara’s commentaries themselves have not been subject to critical editing, and, to complicate matters further, we are not even sure that all the commentaries ascribed to ‡a∫kara were written by him.

290

LANGUAGE, TEXTS AND SOCIETY

problematically, in endnotes wedged between the edition and the translation.5 The text itself contains no mechanism to warn the reader that the editor has emended the received text. This editorial practice is especially problematic in the case of wholesale changes that Böhtlingk made to certain words. Thus, for example, in the Chåndogya he changes aitadåtmyam to etadåtmakam; saumya to somya; and adhidaivatam to adhidevatam throughout the document without any marker or note to indicate that these are his own conjectures. Even the philological reasons offered in support of these wholesale changes are dubious, as Whitney (1890, 412) observes: “to say that such forms ‘make their first appearance in the epics’ sounds curious; it is equivalent to saying that they are not to be retained in the Upanishads because they do not occur in the Upanishads”. Böhtlingk’s treatment of iti (marking quotations and direct speech) in the Chåndogya Upanißad is also idiosyncratic. Sometimes he drops it even though it is found in the received text, and at other times he inserts it where it is absent in the received text. The reason for this practice appears to be the editor’s own judgment as to whether a particular passage calls for a concluding iti. These and similar editorial practices caught the eye of Whitney (1890, 409): “Least of all to be approved, perhaps, is the tampering with the traditional text [...] without any note to inform the reader of the change.” Although such tampering reflects an arrogantly imperious attitude toward the text, at least these changes for the most part do not seriously affect its meaning. In other instances, however, Böhtlingk’s tampering not only mutilated the text, but frequently also misled later scholars who used his edition. It is not possible, nor is it necessary, to give an exhaustive list of such tampering. I will cite only a few egregious examples. Thus, for instance, Chåndogya 1.4.1 reads: om ity etad akßaram udgœtham upåsœta (“O· – one should venerate the High Chant as this syllable”). Böhtlingk’s edition drops udgœtham and explains the omission not in a footnote, where it would be readily noticed by the reader, but in an endnote, saying that manuscripts and previous editions insert udgœtham possibly under the influence of Chåndogya 1.1.1. Senart (1930), who followed Böhtlingk’s edition closely, reproduces Böhtlingk’s emended version. He probably did not read the endnote and accepted it as the received reading, apparently unaware that Böhtlingk had emended the text here. So in his translation, Senart is forced to add a parenthetical remark: “Il faut connaitre que om est la syllabe (par excellence ou l’«impérissable»)”, reflecting his discomfort that there was no qualification of akßara. Senart, who had already 5 More recently, Alsdorf (1950, 627, n. 1) has endorsed such a method with regard to the Ka™ha Upanißad: “I can see no cogent reason why K[a™ha] should not be printed in its true, reconstituted metrical form with the traditional readings given in the footnotes.” The confusions that Böhtlink’s edition created should serve as a warning against such hubris and such questionable editorial practices.

Unfaithful Transmitters

291

written an article on the topic (Senart 1909), knew that the verb upåsin the vedic literature does not simply mean “to venerate” but indicates equivalence, i.e., that something is equivalent to something else; indeed, in the Upanißads this is the term of choice to indicate various correspondences existing between the bodily, ritual, and cosmic spheres. Accordingly, at Chåndogya 1.1.1 Senart translates the identical phrase as “Il faut savoir que la syllable om est l’udgœtha” and adds the footnote: “C’est-à-dire: est équivalent à l’udgœtha lui-même.” The absence of the term at Chåndogya 1.4.1 in Böhtlingk’s edition, which Senart accepted as the received text, thus misled him and caused a somewhat contorted translation. The insertion of conjectural readings into the body of the edited text becomes even more troubling when the editor later happens to change his mind. Böhtlingk recorded these changes of heart in a series of articles written in the somewhat obscure journal Berichte ber die Verhandlungen der königlich sächsischen Gesellschaft der Wissenschaften zu Leipzig, Philologisch-historische Classe, articles that were unavailable to or unnoticed by later scholars, who continued to rely on his editions. A couple of examples will suffice. The traditional reading of Chåndogya 8.15.1 is: åcåryakulåd vedam adhœtya yathåvidhåna∫ guroΔ karmåti†eßeña abhisamåvr¢tya –“Having returned from the teacher’s house, where he learned the Veda in the prescribed manner during his free time after his daily tasks for the teacher”. Böhtlingk (1889b) found the expression karmåti†eßeña unacceptable and changed it to karma kr¢tvåvi†eßeña.6 Senart (1930) is troubled by the reading of the received text but uncertain of Böhtlingk’s emendation. He ends up keeping kr¢två and putting ati†eßéña within brackets, noting: “Le texte est ici troublé. Je ne puis considérer comme admissible l’interprétation de yathåvidhåna∫ guroΔ karmåti†eßeña que suggère Caõkara [et d’après laquelle le disciple n’étudierait qu’à ses instants de loisir]. Il faudrait au moins que toute la locution fût ramassée en seul composé, et cela même serait difficilement admissible. J’ai donc traduit avec l’addition de kr¢två, introduit par Böhtlingk; mais je ne saurais dire que l’altération supposée d’une lecture si facile me semble plausible. En tout cas cette hypothèse implique la correction de atiçeßeña qui ne donne aucun sens en aviçeßeña; mais cet aviçeßeña est lui-même bien faible et bien superflu, et devrait en tout cas beaucoup plutôt porter sur la suite: 6 Böhtlingk based his emendation partly on ‡a∫kara, whose text, he believed, contained the term kr¢två. I do not think that is the case. ‡a∫kara’s commentary of this expressions reads: guroΔ karma yat kartavya∫ tat kr¢två karma†ünyo yo ‘tißiß™aΔ kålaΔ tena kålena vedam adhœtyety arthaΔ. It is unlikely that kr¢två in this gloss is taken from the root text, for karma kr¢två is so simple as not to require such a lengthy explanation. The point of the gloss is to show that karmåti†eßa is not a simple tatpurußa compound but has a more complex meaning referring to the time left over after performing the chores for the teacher. One may make a stronger argument that ‡a∫kara read ati†iß™a in place of ati†eßa in his root text, although even this is uncertain, because ati†iß™a may be just a gloss on ati†eßa; commentators frequently give a gloss without repeating the word of the root text.

292

LANGUAGE, TEXTS AND SOCIETY

bref, je ne puis rien faire ni de atiçeßeña ni de aviçeßeña abhisamåvr¢tya” (Senart 1930, 121). Senart’s observation about how such a simple reading as karma kr¢två could be changed into such a difficult one points to a principle problem with many of Böhtlingk’s emendations: he rejects difficult readings in favor of easy ones, violating the cardinal principle of lectio difficilior, the bedrock of textual criticism that more difficult readings (in terms of grammar, meaning, orthography, etc.) are to be preferred over easier ones. What Senart did not know, however, is that, in an article published eight years (Böhtlingk 1897a, 92) after his edition of the Chåndogya, Böhtlingk had changed his mind and returned to the traditional reading. The reason for this turn around is probably because Böhtlingk found the same reading karmåti†eßeña also in the Gautama Dharmasütra (3.6).7 Although Stenzler’s edition of Gautama appeared in 1876, thirteen years before Böhtlingk’s edition of the Chåndogya Upanißad, he probably saw it only after its preparation, indicating once again the perils of philological hubris. Likewise, in the Chåndogya passage (3.11.6) etad eva tato bhüya iti, Böhtlingk drops iti in his edition (1889b) but wants to retain it in a later article (1897a, 82); iti is clearly required here because the phrase bracketed by it gives the reasons for the previous statement: a man should not impart the teaching to anyone even if he is given the whole world, “because that (formulation) is far greater than (all) that”. Senart (1930), once again, was unaware of Böhtlingk’s change of mind and follows his edition in omitting iti. The verbal form pradhmåyœta occurring in Chåndogya 6.14.1, which I have translated “he would drift about” (Olivelle 1996a,155), did not please Böhtlingk, who gives the conjectural reading pradhåveta in his edition (Böhtlingk 1889b; rejected by Whitney 1890, 413, who prefers the traditional reading) but proposes something totally different, prahvayœta, in a later study (Böhtlingk 1897b, 128).8 A conjecture first suggested by Deussen (1897) took a somewhat different route to enter the modern textual transmission as a variant reading. The traditional reading of Chåndogya 4.9.2 is bhagavå∫s tv eva me kåme brüyåt –“But, if it pleases you, sir, you should teach it to me yourself”. Böhtlingk (1889b) found kåme (locative singular) to be unsuitable and ‡a∫kara’s explanation unacceptable9 and presented 7 There is no indication that Senart was aware that the identical expression with which he had such problems also occurred in the Gautama Dharmasütra making it unlikely that it was a scribal error in the Chåndogya. 8 There are numerous other examples of Böhtlingk’s changes of mind after the publication of his editions. I have recorded all these in my edition of the Upanißads (Olivelle, 1998a), which lists all the variants recorded in editions, as well as scholarly conjectures. At Chåndogya 1.6.7, for example, Böhtlingk changes kapyåsam to kapilåsam in his edition (rejected by Whitney 1890, 413), only to change it later to kalmåßam (Böhtlingk 1897b, 127). At Chåndogya 3.11.6 he omits in his edition the two iti but later prefers to retain them (Böhtlingk 1897a, 82). 9 ‡a∫kara glosses me kåme mamecchåyåm. Böhtlingk (1889b, 102) observes: “me kåme

Unfaithful Transmitters

293

a clever but suspect emendation by dividing the words differently:10 bhagavå∫s tv evam eko me brüyåt (“Jedoch könnte der Erhabene, aber auch nur er allein, mir es auf diese Weise verknden”). Deussen (1897, 124, n. 1), likewise, was dissatisfied with ‡a∫kara’s explanation but could not accept Böhtlingk’s emendation. So, quite appropriately in a footnote, Deussen suggested emending kåme to kåma∫,11 translating the term as “bitte” (please). Senart (1930, 53) also rejected Böhtlingk’s emendation, but could not decide between Deussen’s suggestion and the traditional reading. In 1958 Limaye and Vadekar published an edition of eighteen Upanißads that has become the standard edition used by scholars today.12 Within the body of the text, they give the traditional reading kåme, but they add a footnote giving kåmam not as Deussen’s conjecture but as a variant reading. Now, Limaye and Vadekar’s edition is not based on manuscript evidence;13 their variants are gleaned from published material and emendations suggested by scholars. Indeed, immediately after they note kåmam as a variant, they give Böhtlingk’s conjecture. It appears to me very probable that they got kåmam from Deussen’s suggestion. Now, in 1980 Deussen’s German translation of the Upanißads was retranslated into English by Bedekar and Palsule. They reproduce (1980, 126) Deussen’s footnote about kåmam, to which they add a further note of their own: “In the later edition [sic], e.g. in the Eighteen Upanißads edited by R.D. Vadekar and V.P. Limaye the reading is ‘kåmam’.” This completes an interesting and instructive circle of modern textual transmission, where a scholarly suggestion in a footnote becomes finally the “accepted” reading. The verses contained in the older prose Upanißads and in a special way the later metrical Upanißads pose a different type of problem for the editor and the scholar when the traditional text violates the meter. It is a prima facie rule of editing that the original text could not have viosoll nach ‡[aõkara] = mamecchåyåm sein; aber mir zu Liebe heisst im Sanskrit me (mama) kåmåya. Aus der Uebersetzung wird man ersehen, dass ich mit einer geringen Aenderung wohl das Richtige getroffen habe.” 10 Sanskrit manuscripts do not leave a space between words (cf. note 18). So eva me kåme will appear as evamekåme, which Böhtlingk divides as evam ek(o) me. The confusion between ‘å’ and ‘o’ can be explained orthographically at least in the Devanågarœ script. 11 In the Devanågarœ script, the pure nasal anusvåra (∫), written as a dot above the preceding letter, can be mistaken for a final ‘e’, written as a stroke above the preceding letter. But this confusion should not occur in South Indian scripts where the two are clearly distinguished. 12 Salomon’s (1981, 1991) important studies of Upanißadic language, for example, are based on this edition, even though Salomon himself expresses some unease about using an uncritical edition (1991, 48). 13 They claim that they consulted some “old MSS of the Br¢hadårañyaka and the ‡vetå†vatara but our collation was not found to be useful; in many cases we found the MSS more faulty than the printed editions” (p. vi). It would have been a great service to scholarship, however, if they had recorded the “faulty readings” of these old manuscripts in their edition.

294

LANGUAGE, TEXTS AND SOCIETY

lated metrical rules, and therefore editors regularly adopt one manuscript reading and reject another on the basis of meter (metri causa). This principle is also frequently invoked in conjectural emendations to the traditional reading of a verse, emendations not based on manuscript evidence. We see this principle applied repeatedly in the “critical editions” of the Muñ∂aka and ‡vetå†vatara Upanißads by Hertel (1924) and Hauschild (1927), respectively. Yet, it is well to remember the salutary warning of Max Mller (1879, lxxii): “The metrical emendations that suggest themselves are generally so easy and so obvious that, for that very reason, we should hesitate before correcting what native scholars would have corrected long ago, if they had thought there was any real necessity for correction.” The problems inherent in restoring the meter through conjecture are exemplified by the variety of scholarly opinion on how to restore the meter of the third påda of Ka™ha Upanißad 1.19: etam agni∫ tavaiva pravakßyanti janåsaΔ –“People will proclaim this your very own fire”, a påda that contains fourteen syllables instead of the required eleven. Mller (1884), Böhtlingk (1890, 134), and Charpentier (1928-29) take tavaiva to be an interpolation distorting the meter; but this emendation is rejected by Garbe and Alsdorf (1950, 627). Alsdorf, followed by Rau (1971, 173), drops agni∫ and eva. Likewise, the second half-verse of Ka™ha Upanißad 2.11 is metrically incorrect: stomamahad urugåya∫ pratiß™hå∫ dr¢ß™vå dhr¢tyå dhœro naciketo ‘tyasråkßœΔ –“Great and widespread praise is the foundation. These you have seen, wise Naciketas, and having seen, firmly rejected.” Böhtlingk (1890, 142) proposes deleting dr¢ß™vå to restore the meter, whereas Alsdorf (1950, 628) thinks that Böhtlingk has made the “wrong choice” and suggests deleting the next word, dhr¢tyå. The problem is not such scholarly disagreement, which is fruitful; but when a particular conjecture is reproduced in the edited text, especially if it is an influential edition, then the conjecture becomes accepted passively by later scholarship and preempts that very scholarly give and take. An issue peculiar to verses has to be weighed in any philological reconstruction of broken meters, and that is pronunciation in terms of both sandhi and pråkr¢tic ways of pronouncing words.14 Alsdorf (1950, 623) has shown that numerous irregular meters can be regularized by simply dropping external sandhi. The first two pådas of Ka™ha Upanißad 1.7, for example, read vai†vånaraΔ pravi†atyatithir bråhmaño gr¢hån (“A Brahmin guest enters a house as the fire in all men”), with seven syllables in the first påda, instead of the regular eight. This can be remedied by dissolving the sandhi between the two pådas and reading pravi†ati atithir. This operation is not a textual 14 The influence of modern Indian languages, such as Marathi and Hindi, on Sanskrit meter in verse compositions of modern authors has been noted by Madhav Deshpade in “On Vernacular Sanskrit: The Gœrvåñavåõma∞jarœ of DhuñÛiråja Kavi” in Deshpande 1993a, 36.

Unfaithful Transmitters

295

emendation because strict sandhi rules are applied consistently mostly in printed texts; most manuscripts are quite free with their sandhi, and when speaking or chanting sandhi is broken whenever there is a pause. Then there are the pråkr¢tic pronunciations of certain words that would restore the meter. Some common words listed by Alsdorf (1950) include iva, iti, bhavati, which were pronounced as va, ti, and bhoti. Thus, Ka™ha 1.9 sasyam iva martyaΔ pacyate sasyam ivåjåyate punaΔ (“A mortal man ripens like grain, and like grain is born again”) has an extra syllable in both pådas. This can be eliminated by using va for iva (sasya∫ va). Likewise, Ka™ha 3.5 yas tv avij∞ånavån bhavati (“When a man lacks understanding”) has an extra syllable, which can be eliminated by reading bhavati as bhoti.

II I have already made reference to the editorial principle of lectio difficilior. Simply stated, this principle holds that more difficult readings (in terms of grammar, meaning, orthography, etc.) are to be preferred over easier readings. This is based on the reasoning that scribes and readers down the centuries were more likely to replace readings they found difficult or did not understand with easier readings; the inverse, namely the replacement of easy readings with more difficult ones, is much less likely. This principle holds especially in the case of “ungrammatical” forms. After all, Indian pandits probably knew their Pañinian grammar a bit better than modern western scholars. It would have been a simple matter for them to replace the suspect readings with the correct forms. But we find repeatedly that they did not, and this is a clear indication that ancient Indian commentators did not take liberties with their received Upanißadic texts; they were, by and large, faithful transmitters. The importance of the principle of lectio difficilior and the reliability of the texts transmitted by ‡a∫kara have been highlighted by the recent work of Richard Salomon (1981, 1991). He has studied the linguistic peculiarities of two late Upanißads, the Muñ∂aka and the Pra†na, and seen in them the vestiges of a Kßatriya dialect of Sanskrit. Arguing for the superiority of ‡a∫kara’s15 version of the Pra†na Upanißad, Salomon (1991, 49) remarks: “The example of pra/pråtiß™hante,16 just cited, for example, clearly points toward a lectio 15 In this paper I follow the traditional ascription of the commentaries on this and other Upanißads to ‡a∫kara; it should not be viewed as support for that ascription. For my purposes “‡a∫kara” may stand for any ancient Indian commentator/theologian. 16 The verbal form one should expect here is pratitiß™hante, and the traditional reading pråtiß™hante, Salomon (1991, 66) argues, is not a scribal mistake but a haplology introduced by the composer(s) of the text themselves. Salomon rightly rejects Böhtlingk’s (1890) “arbitrary emendations” of this and other non-standard terms into “proper Sanskrit”. “The very fact that these ‘wrong’ forms are preserved in the textual and commentatorial tradition argues for, not against, their originality. In other words, I would propose that we have in these verbs a matter, not of haplography, i.e. of scribal error, as

296

LANGUAGE, TEXTS AND SOCIETY

facilior in the Raõgaråmånuja text,17 suggesting the superiority of the ‡aõkara version. And indeed, it is precisely the large number of nonstandard forms in the latter text which is the strongest argument for its integrity. For as a matter of general principle, such nonstandard forms–especially when they can be corroborated by parallels in other texts, as is most often the case–should be considered a priori as stronger readings than the ‘grammatically correct’ variants … In other words, the fact that such unexpected forms as pråtiß™hante or avedißam, which could have been so easily emended to the normal forms, are preserved in the vulgate says much both for its reliability and for their originality.” When European scholars change an unusual form to its “correct” grammatical form, we lose much of the dialectical variations evident in old Sanskrit. A case in point is Muñ∂aka 3.2.2, which contains the non-standard instrumental plural kåmabhiΔ (“with desires”). In his edition of this Upanißad, Hertel (1924) emends the term to the regular karmabhiΔ (“with actions”) thereby not only changing the meaning radically but also obliterating this significant dialectical variant (see Salomon 1981, 94). The replacement of difficult readings with more standard conjectures also preempts further scholarly meditation on those difficult readings. At Muñ∂aka 3.1.4, for example, we have the expression vijånan vidvån bhavate nåtivådœ, translated by Hume (1931) as “Understanding this, one becomes a knower. There is no superior speaker.” Böhtlingk (1901, 8) emends the second half to bhavati tenåtivådœ, following the reading of Chåndogya 7.15.4. Following the traditional text, however, and without a need for emendation, Rau (1965) has pointed out a far superior reading: bhava tenåtivådœ,18 according to which the phrase can be translated: “Be a man who perceives, who knows this, and thereby a man who out-talks” (Olivelle 1996, 274). ‡vetå†vatara Upanißad 6.11 contains the word cetå, which ‡a∫kara19 and, following him, nearly all modern scholars have taken to be an agent noun derived from the root cit, “to perceive, to observe”. If that were the case, the standard nominal derivative should have been cettå, and Hauschild (1927),20 in his edition of this Upanißad, emends cetå to the “grammatically correct” cettå. It is important to note that even though ‡a∫kara explains cetå to be an agent noun of cit, nevertheless he preserved the non-standard form, Böhtlingk would have it, but rather of haplology, i.e. an original dialectal usage on the part of the composer(s) of the text itself” (Salomon 1991, 66). 17 Raõgaråmånuja is a late (possibly 15-16th century) commentator on the Upanißads. 18 It must be remembered that Sanskrit manuscripts do not divide words; there are no “white spaces” between words. Letters follow each other in unbroken sequence. Thus, taking te from the end of bhavate to the beginning of nåtivådœ requires no emendation. 19 ‡a∫kara glosses cetå with cetayitå. 20 In a note to his translation, Hume (1931, 409) also suggests the same emendation. Hauschild, however, introduces his conjecture into the body of the text.

Unfaithful Transmitters

297

whereas Hauschild changes the form in his edition. Rau (1964), however, relying on the traditional text, sees cetå as an agent noun derived not from cit but from the root ci “to avenge” and offers the very plausible translation “avenger”, an interpretation that fits the context better and that would have been foreclosed if ‡a∫kara, like Hauschild, had emended his text. The Upanißadic texts as traditionally handed down also preserve expressions that make no linguistic sense; they are “nonsense” phrases. One example is tajjalån, which occurs in Chåndogya 3.14.1: sarva∫ khav ida∫ brahma tajjalån iti †ånta upåsœta.21 ‡a∫kara found the expression jalån problematic. Instead of emending it, however, he gives an explanation that takes it to be an acronym. According to his interpretation, the term means that everything proceeds from (ja = jan, “to be born”), dissolves into (la = lœ, “to dissolve”), and lives by (an, “to breathe”) Brahman. A modern scholar may not agree with ‡a∫kara’s overly theological explanation, but at least the theologian did not deliberately change the text. In contrast, Böhtlingk (1889b) emends the expression in his critically edited text to read taj jånånœti †ånta upåsœta, “Der zur Ruhe Gelangte verehre es als das, was er kennen möchte.” He does not explain, however, how such a common verb form as jånåni (first person singular present subjunctive) could have become garbled into jalån; again Böhtlingk has rejected the lectio difficilior in favor of the easier reading.22 Even when commentators encounter an ungrammatical form, they usually preserve it. Generally they explain the irregularity as a vedic form (chåndasa); it was believed that certain Sanskrit forms found in the vedic texts were unusual or irregular and did not conform to Pañinian grammar.23 This belief, I think, and the innate conservatism (in the best sense of the word) of Brahmin scholars especially with regard to vedic texts prevented them from changing the received text deliberately. They did not, however, naively take every aberrant form to be simply vedic; they also suspected that some forms may be the result of textual corruption. To give but one example, Chåndogya 1.12.3 reads: tån hovåcehaiva må pråtar upasamœyåteti, “He told them: ‘Come and meet me at this very spot in the morning’.” The long “œ” of 21 “Brahman, you see, is this whole world. With inner tranquillity, one should venerate it as jalån.” Here I take the initial tat as a separate word; but even that is uncertain. 22 Tradition also has preserved the most celebrated of Upanißadic sayings: tat tvam asi. Vedic and Upanißadic usage shows that if tat (taken as a nominative singular neuter pronoun) were connected to tvam (“thou”) by the copula asi (“art”), then it should have been in the same gender, masculine, as the latter, i.e., saΔ. Even though native commentators translated the phrase as “Thou art that”, they preserved the anomalous form. This permitted Brereton’s (1986) new and brilliant interpretation of the phrase (taking tat as adverbial akin to tena or tasmåt) as meaning “That is how you are”, i.e., that is how you came into being and that is how you continue to live. 23 For a fine explanation of how ancient Indian grammarians viewed the landscape of actual Sanskrit, see Madhav Deshpande, “Historical Change and the Theology of Eternal (Nitya) Sanskrit,” in Deshpande 1993, pp. 53-74.

298

LANGUAGE, TEXTS AND SOCIETY

upasamœyåta is irregular; the standard form is upasamiyåta. ‡a∫kara notes this irregularity and explains it: dairghya∫ chåndasam … pramådapå™ho vå, “The long (œ) is either a vedic form or an erroneous reading [i.e., a reading caused by carelessness].” In contrast to ‡a∫kara’s careful attention to manuscript evidence,24 Böhtlingk (1989b) summarily changes this word in his edition to the standard upasamiyåta.

III Western, especially European, philologists, we have seen, were often less faithful transmitters of Upanißadic texts than the Indian scribes and commentators they so often criticized. Native commentators and theologians did not, as often assumed, carelessly or deliberately change the received texts to suit their doctrinal or grammatical tastes. Indeed, it is the modern philologists who are often guilty of changing the texts to suit preconceived notions of correctness, whether grammatical or otherwise. Lest I be misunderstood, my criticism is not against philology as such but against the substitution of arm-chair philology for the tedious but important examination of manuscript evidence and the resultant improper application of philology to mutilate texts. Philology is the indispensable bedrock of any serious study of texts for any purpose, especially for the purpose of historical reconstruction. But philology is not a substitute for critical edition of texts based on manuscript evidence; it is not possible to reconstruct texts by brain power or philological training alone. We have seen confusion reign especially when scholarly conjectures are introduced into the body of edited texts, especially those calling themselves “critical editions”, often with little warning to readers that they are dealing with conjectures and not received readings. In ancient and medieval India texts were transmitted and preserved by copying onto manuscripts and by memorization, which is a lost art today. The copying of texts introduced errors through negligence or misreading the exemplars and sometimes through deliberate emendations. The existence of numerous manuscripts of a single work, however, permits the careful editor to detect such errors and emendations. Error or emendation was limited to a single manuscript and to others for which it served as exemplar. In the modern age the printed book (and today its electronic counterpart) is the unique medium of textual transmission and preservation. The printed book comes in thousands of identical specimens. An error or emendation introduced into a printed edition, unlike its manuscript counterpart, is reproduced in every single specimen. Given the expense of publishing, moreover, once an ancient text has been published, it is unlikely 24 It is, of course, uncertain whether ‡a∫kara was working from manuscripts or memorized texts, but the reference to negligence strongly hints at his use of manuscripts.

Unfaithful Transmitters

299

that a new edition would be forthcoming soon or ever. The responsibility, therefore, of a modern editor to ensure the faithfulness of transmission is thousand times greater than that of a scribe. Philological hubris, I have tried to demonstrate in this paper, has made many modern editions of the Upanißads unreliable. Modern editors may take note of this set of verses often appended by Indian scribes at the end of their manuscripts. bhagnapr¢ß™ika™igrœvaΔ stabdhadr¢ß™ir adhomukhaΔ Ù kaß™ena likhita∫ grantha∫ yatnena pratipålayet ÙÙ With great trouble I have written this book, My head bent low, with unwavering eyes, I have broken my back, my hips and neck; So be diligent and take care of it. yådr¢†a∫ pustaka∫ dr¢ß™vå tådr¢†a∫ likhita∫ mayå Ù yadi †uddham a†uddha∫ vå mama doßo no vidyate ÙÙ I copied exactly What I saw in the book; Whether it’s right or wrong, I am not to be blamed.

XIV Sanskrit Commentators and the Transmission of Texts: Haradatta on Åpastamba Dharmasütra*

In my recent article “Unfaithful Transmitters” (Olivelle 1998; above pp. 287-289) I drew attention to the pervasive mistrust of ancient Indian commentators as reliable guides to understanding ancient Indian texts prevalent among western scholars, a mistrust that spilled over into doubts about the reliability of the textual transmission mediated by these commentators and more broadly into a mistrust of the scribal tradition as such. Drawing on examples of “critical editions” of Upanißadic texts, especially Böhtlingk’s (1889b) edition of the Chåndogya Upanißad, and the readings preserved by the commentator ‡a∫kara, I tried to show there that western, primarily European, philologists were often less faithful transmitters of Upanißadic texts than the Indian scribes and commentators they so often criticized. Native commentators and theologians did not, as often assumed, carelessly or deliberately change the received texts to suit their doctrinal or grammatical tastes. In this paper I return to that theme and this time examine closely the manner in which Haradatta, the commentator of the Åpastamba Dharmasütra,1 explained and transmitted that ancient text. Just as it is unfair to indict all western scholars because of the excesses of some, so it is not my intention to present Haradatta as typical of all Indian commentators. If the “Orientalist” debate has taught us anything, it is to treat traditional Indian authors as individuals, to restore “agency” to them. They are not all alike; there are good and not so Originally published in the Journal of Indian Philosophy 27 (1999): 551-74. Haradatta is also credited with commentaries on Åpastamba Gr¢hyasütra, Åpastamba Mantrapå™ha, ņvalåyana Gr¢hyasütra, and Gautama Dharmasütra. He is also the author of the Padama∞jarœ, a commenatary on the Kå†ikå commentary on Påñini. The editor of this text (Hyderabad: Sanskrit Academy, Osmania University, 1981, p. ix), P. Sri * 1

302

LANGUAGE, TEXTS AND SOCIETY

good commentators. Haradatta is one of the best. Yet, I do not think that he is unique or atypical; he is good, but he also represents well the tradition from which he comes. Haradatta is what we would call today a “close reader” of the text. He does not let even the slightest irregularity, peculiarity, or quirk go unnoticed. He points out the presence or the absence of a visarga or an anusvåra (something even those of us who dabble in collating manuscripts are prone to overlook), the shortening or lengthening of a vowel, whether an “n” is dental or retroflex, whether a letter is a “v” or “b” or “p”, and so on. In short, he takes the text he received from the tradition of scribes and reciters seriously and reverentially, even when he happens to disagree with it. This stands in sharp contrast, once again, to Böhtlingk (1885b), who wrote an article reviewing Bhler’s edition (1st ed.,1868) and translation (1882) of Åpastamba, an article that also contains another “close reading” of the text. I do not want to demonize Böhtlingk; in many instances he suggests prudent and sometimes ingenious emendations. I have given in the notes his observations and emendations on the passages commented on by Haradatta. One can readily see from them that on the whole the cavalier attitude of Böhtlingk towards the received text stands in sharp contrast to that of Haradatta. This is evident also from his Euro-centric remarks in the introduction to his article: Before starting to discuss the individual sütras, I am obliged to state my position about the commentator Haradatta. In spite of all of his scholarship, he cannot make any claims of authority in matters of language, because he lacks the European critical attitude. Bhler, a scholar and grammarian of the first rate, reveres Haradatta to the extent that he follows him unquestioningly. And that even when he finds himself in a conflict with his own knowledge of the language. I am glad to acknowledge archaisms, provided that it is possible to find an analogy from other older or contemporary writings. In the case of an alleged archaism that is completely isolated and for which a single manuscript offers the correct grammatical form, whether it be of a different manuscript family or whether it be resolved by itself only with a slight change, I would not hesitate to endorse it. My conscience does not allow me to attribute a gross grammatical error to an old author arbitrarily. It has been established that many mistakes adhere even to the oldest Indian texts, in spite of the agreement with manuscripts and commentators. Why then should the younger texts, which were not kept so scrupulously, be so inviolable that a European, who has devoted himself for over fifty years to Sanskrit studies, should not counter Haradatta in matters of language?2 Ramachandrudu assigns Haradatta to 1100 CE, whereas Kane (1962-75, I: 746) dates him 1100-1300 CE. In the Padama∞jarœ Haradatta gives some details of his life. His father was Padmakumåra, his mother ‡rœ, and his older brother Agnikumåra. His vidyåguru was Aparåjita. He belonged to the Tamil region. 2 Böhtlingk 1885b, 517. Translated with the assistance of Edeltraud Harzer.

Sanskrit Commentators and the Transmission of Texts

303

To give some coherence and structure to my analysis, I will present the material I have gleaned from Haradatta’s commentary under three rubrics: A) variant readings preserved by Haradatta; B) Haradatta’s explanations of difficult and possibly erroneous readings that he encountered in the transmitted text; and C) Haradatta’s broader observations and explanations.

A. Variants Preserved by Haradatta I I have collected 31 variant readings of the text preserved by Haradatta. First, I present a set of texts where Haradatta presents a reading which he presumably prefers, but acknowledges that it is a minority reading by giving the more common reading. 1) Åp 1.5.23: sarvåhña∫ suyukto ‘dhyayanåd anantaro ‘dhyåye (“He shall remain fully attentive all day long and at the time of vedic study never let anything distract him from his lesson”). Haradatta notes that the common reading for adhyåye is adhyåyet, calling the final ‘t’ either erroneous or a vedic peculiarity: adhyåyed iti pråyeña pa™hanti Ù tatra takåro ‘papå™ha† chåndaso vå. Bhler (1932, 11) notes, correctly I believe, that the ‘t’ was probably introduced by the doubling of the initial ‘t’ of tathå, which is the first word of the next sütra. In many manuscripts sütras are neither numbered nor separated by dañ∂as but run continuously. Haradatta’s explanation may be faulty, but he is careful to note the variant. 2) Åp 1.8.2: måly åliptamukha upaliptake†a†ma†rur akto ‘bhyakto veß™ity upaveß™itœ kå∞cuky upånahœ pådukœ (“He may wear a necklace; apply lotions on his face, oil on his hair and beard, collyrium [on his eyes], and oil [on his body]; and wear a turban, a lungi, a jacket, sandals, and shoes”). Haradatta notes the variant compound form ka∞cukopånahœ, providing a Påñinian (5.4.106) explanation of that form,3 but he calls thenon-compound form the common reading and provides an explanation of the long initial vowel of kå∞cukœ: prasiddhe på™he ka∞cukam eva kå∞cuka∫ tadvån kå∞cukœ. 3) Åp 1.31.19: divådityaΔ sattvåni gopåyati nakta∫ candramås tasmåd amåvåsyåyå∫ ni†åyå∫ svådhœya4 åtmano guptim icchet pråyatyabrahmacaryakåle caryayå ca (“During the day the sun protects creatures, and during the night, the moon. Therefore, on the night of the new moon he should try his very best to guard himself by keeping himself pure and chaste and by performing rites appropriate for the occasion”). Haradatta sees clearly the grammatical problems posed by the last phrase containing an impossible compound. Indeed, all the manuscripts of Bhler’s edition have this reading; 3 Haradatta comments: ka∞cuka∫ copånac ca ka∞cukopånaham Ù dvandvåc cudaßahåntåd (Påñini 5.4.106) ity ac såmåsåntaΔ Ù tad asyåstœti ka∞cukopånahœ Ù 4 For an explanation of this word, see below B: 40.

304

LANGUAGE, TEXTS AND SOCIETY

only one gives Haradatta’s preferred reading written above the line. Haradatta prefers the grammatical pråyatyabrahmacaryåbhyå∫ kåle caryayå ca,5 calling it the reading that accords with the meaning: aya∫ tåvad arthånurüpaΔ på™haΔ. But he confesses that the more difficult reading is the the way the sütra is actually recited, calling it either a careless mistake or a vedic form: adhœyamånas tu pramåda† chåndaso vå. 4) Åp 2.5.2: adhœtya cåviprakramaña∫ sadyaΔ (“After completing the vedic study he should not go away immediately”). Haradatta observes that the more common form has ‘i’ after ‘m’ (i.e., aviprakramiñam), which, he says, is either a vedic form or an error, but in any case has the same meaning: pråyeña makåråt param ikåram adhœyate Ù tatråpy eßa evårthaΔ Ù ikåras tu chåndaso ‘papå™ho vå. Note the care with which Haradatta records this and similar minute variants that we will encounter later. He does not simply write it but spells out the difference in words. 5) Åp 2. 6.13: adhyayanaså∫vr¢tti† cåtrådhikå (“In his case [i.e., when a guest is a vedic student], however, there is the additional requirement to perform the vedic recitation along with him”). Haradatta prefers the reading sa∫vr¢tti6 with a short ‘a’ but acknowledges that in the common reading the prefix sam has a long ‘å’, which he takes to be a vedic peculiarity: prasiddhe tu på™he pürvapadåntasya samo ‘kårasya chåndaso dœrghaΔ. 6) Åp 2.7.7: yad anutiß™haty udavasyaty eva tat (“When he rises [as his guest gets up to leave], it constitutes the final rite of the Soma sacrifice”). Haradatta thinks that the reading should have the prefix ut after anu (i.e., anüttiß™hati), but acknowledges that experts (†iß™a) read it differently: pråyeñocchabda∫ †iß™å na pa™hanti kevalam anu†abdam eva pa™hanti Ù tatråpy arthaΔ sa eva. 7) Åp 2.26.8: tatra yan mußyate tais tat pratidåpyam (“[Security officers] must be forced to make good anything that is stolen within those [limits]”). Haradatta notes that generally the last word was pronounced with a ‘v’ rather than a ‘p’ (i.e., pratidåvyam): pråyeña dantyoß™ya∫ vakåra∫ pa™hanti. 8) Åp 2.27.7: niyamårambhaño hi varßœyån abhyudaya evamårambhañåd apatyåt (“for the happiness resulting from following this restriction is far greater than that resulting from children obtained by following that custom [i.e., levirate]”). Haradatta notes that åpatyåt is the common reading, although apatyåt is his preferred reading: apatyåd iti på™haΔ Ù åpatyåd iti pråyeña pa™hanti. II Then there are the occasions when Haradatta records variants without indicating whether they were common or not. Here he 5 6

This reading is adopted by Böhtlingk (1885b), but rightly rejected by Bhler (1886b). Böhtlingk (1885b) takes this to be “correct” reading.

Sanskrit Commentators and the Transmission of Texts

305

appears to accept the possibility that the variants are also correct and frequently explains their formations. 9-11) Åp 1.7.21: we have here the form dhårmyam. Haradatta appears to prefer the reading dharmyam,7 but explains the other form: dhårmyapå™he svårthe ßya∞ (i.e., the ‘ya’ suffix with vr¢ddhi of the initial vowel and having the same meaning as the original word: Påñini 5.1.123-4). A similar preference for dharmya is seen at Åp 1.15.22 and 2.26.9, where Haradatta glosses dhårmya with dharmya. 12) Åp 1.10.2: vairamaño gurußv aß™åkya aupakaraña iti tryahåΔ (“[Vedic recitation is suspended] for three days after the conclusion of the annual course of vedic study, the death of an elder, an ancestral offering made on the eighth day after the full moon, and the commencement of the annual course of vedic study”). Haradatta prefers the locative vairamañe,8 but notes that in the variant the nominative stands for the locative: prathamåntapå™he saptamyarthe prathamå. 13) Åp 1.31.4: devatåbhidhåna∫ cåprayataΔ (“[Let him refrain from] pronouncing the name of a god while he is impure”). Haradatta gives the variant apidhånam, which he considers to have an identical meaning: apidhånam ity api på™he eßa evårthaΔ. 14) Åp 1.32.27: nåvå∫ ca så∫†ayikœm (“[He should avoid] using an unsafe boat”). Haradatta prefers the readings with the normal accusative nåvam (from nau), yet gives the alternate form nåvåm, which he takes to be the genitive plural, with the explanation that “among boats” one should avoid one that is unsafe: nåvåm iti ßaß™yantapå™he nåvå∫ madhye så∫†ayikœ∫ nåva∫ varjayet. The form nåvåm, however, as Bhler (1892, 54) has pointed out, is probably the accusative of the vedic/prakritic nåvå. 15) Åp 2.11.17: bråhme vivåhe bandhu†œla†rutårogyåñi buddhvå prajåsahatvakarmabhyaΔ pratipådayec chaktivißayeñåla∫kr¢tya (“At a ‘Brahma’ marriage, he should inquire about the groom’s family, virtue, learning, and health; adorn the girl with jewelry to the best of his ability; and give her for bearing children, for companionship, and for carrying out rituals”). Haradatta’s reading, also recorded in several manuscripts of Bhler’s edition, is: prajå∫ sahatvakarmabhyaΔ. But he acknowledges the reading adopted by Bhler, saying: prajåsahatvakarmabhya iti på™he prajårtha∫ sahatvakarmårtha∫ ceti. 16) Åp 2.16.2: prajåniΔ†reyaså ca (“[Manu created the †råddha ritual] also for the prosperity of the people”). Haradatta reads prajåniΔ†reyasåya ca,9 in keeping with classical Sanskrit; -†reyaså, howev7

Böhtlingk (1885b) accepts this reading. Böhtlingk (1885b) also emends to vairamañe, objecting to a nominative in the middle of locatives. Bhler (1886b), however, shows that such mismatches of cases are common in Åpastamba, as at 1.11.31 and 2.26.20, where we have the substantive and its adjective in different cases. See also the different numbers in subjects and verbs at 1.1.27-29 and 1.2.5. 9 Böhtlingk (1885b) accepts this reading without hesitation and sees no reason to accept the difficult reading. 8

306

LANGUAGE, TEXTS AND SOCIETY

er, as Bhler has pointed out, is a vedic instrumental possibly giving the reason. Haradatta gives this lectio difficilior also, giving his own explanation of this strange formation, as well as that of another commentator: prajåniΔ†reyasåya tådarthye caturthœ Ù prajånå∫ niΔ†reyasårtham Ù niΔ†reyaså cetipå™he chåndaso yakarasya cakåraΔ Ù apara åha chåndaso liõgavyatyayaΔ Ù prajåniΔ†reyasa∫ cåsya karmañaΔ phalam iti. We have here two explanations of the difficult form. Haradatta takes the final ca to be part of the word (niΔ†reyasåca), and explains that the ya of niΔ†reyasåya is represented here as ca, a vedic peculiarity! The other opinion he cites takes the word to be a feminine nominative; the change of gender being attributed again to vedic speech. 17) Åp 2.17.1: kha∂gopastarañe kha∂gamå∫senånantya∫ kålam (“With the meat of a rhinoceros offered on a rhinoceros skin, the gratification [of ancestors] lasts an unlimited time”). Haradatta prefered a different reading for ånantyam; different manuscripts of the commentary give his reading as either atyanta∫ kålam or ananta∫ kålam (given his explanation of the variant, I think his reading was the latter). He records, however, also the other reading: ånantyam iti på™he svårthe ßya∞ (see above A: 9-11). 18-19) Åp 2.17.22: here we have a reference to jyeß™hasåmagaΔ (“a man who sings the Jyeß™ha Såmans”) in a list of “people who purify a row of eaters” (paõktipåvana). Haradatta records the variant jyeß™hasåmikaΔ with the comment: jyeß™hasåmika iti på™he vrœhyåditvåt ™han [-ika suffix of possession for words beginning with vrœhi; Påñini 5.2.116]. Some versions of this sütra also contains the compound †rotriyaputraΔ. Haradatta notes this as an addition: †rotriyaputra ity api pa™hanti Ù tad ådarårtha∫ draß™avyam. 20) Åp 2.19.1: gaurasarßapåñå∫ cürñåni kårayitvå taiΔ påñipåda∫ prakßålya mukha∫ karñau prå†ya ca ... (“[A man who wants to be prosperous] should get some white mustard seeds made into powder, rub it on his hands, feet, face, and ears, eat it ...”). Haradatta records a variant of prå†ya with a dental ‘s’: pråsyeti på™he pråsyed vikiret (“scatter”). 21) Åp 2.19.4: na cånyenåpi bhoktavyaΔ (“and no one else should eat from [one’s eating bowl]”). Haradatta prefers the neuter reading bhoktavyam, but gives also the masculine reading adopted by Bhler: bhoktavya iti pu∫liõgapå™he ‘py eßa evårthaΔ. 22) Åp 2.20.15-16: nakhai† ca nakhavådanam Ù spho™anåni cåkårañåt (“[He should refrain from] making noises by striking the nails against each other and cracking the finger joints without a good reason”). Haradatta records a variant (also found in some manuscripts) that joins the two sütras by making a compound of the final and initial words of the two: vådanaspho™anånœti samåsapå™he ‘py eßa evårthaΔ. III Finally, there are numerous other occasions when Haradatta gives variant readings but clearly states which of them is correct and which is wrong.

Sanskrit Commentators and the Transmission of Texts

307

23) Åp 1.5.16: dakßiña∫ båhu∫ †rotrasama∫ prasårya bråhmaño ‘bhivådayœta [...] prå∞jalim10 (“With joined hands, let a Brahmin greet by stretching his right hand level with his ears [...] “). Haradatta prefers the reading prå∞jaliΔ, calling it the correct reading: yuktaΔ på™haΔ. 24) Åp 1.6.28: ceß™ati ca cikœrßa∫s tacchaktivißaye (“And when the teacher is doing something, he should, if he is capable, offer to do it himself”). Haradatta records the participial reading cikœrßan, but also gives the optative reading, which he takes to be the correct one: cikœrßed iti tu yuktaΔ på™haΔ.11 25) Åp 1.17.37: pa∞canakhånå∫ godhåkacchapa†våvi™chalyaka [...] (“Animals with five nails are forbidden, with the exception of the monitor lizard, tortoise, porcupine, hedgehog, [...]”). Haradatta sees †våvi™†alyaka (or -†aryaka) as the correct form but records what he takes to be the erroneous reading: eke tu chakåra∫ pa™hanti chakåråt pürvam ikåram. 26) Åp 2.10.11: nyaståyudhaprakœrñake†aprå∞jaliparåõåvr¢ttånåm åryå vadha∫ paricakßate (“Åryas condemn the killing of those who have thrown down their weapons, who have disheveled hair, who fold their hands in supplication, or who are fleeing”). Haradatta accepts this as the correct reading but records what he regards as an erroneous variant that dissolves the long compound with visargas at the end of each word: nyaståyudhaΔ prakœrñakeßa iti visarjanœya∫ kecid vadanti [or pa™hanti] Ù so ‘papå™haΔ. 27) Åp 2.17.22: In this list of individuals who are paõktipåvana (cf. A: 18-19) we read: vedådhyåyy anücånaputraΔ (“a man who recites the Veda, a son of a vedic savant”). Haradatta finds an erroneous variant which adds iti between the two words: vedådhyåyœty asyåpy anta [variant asyånantaram] iti†abda∫ pa™hanti Ù so ‘papå™haΔ. 28) Åp 2.20.18: yoktå ca dharmayukteßu dravyaparigraheßu ca (“He should be a man who applies himself to acquiring wealth in righteous ways”). Haradatta observes that one ca is meaningless and that it is omitted by some: eka† ca†abdo ‘narthakaΔ Ù kecin naiva pa™hanti. All the manuscripts used by Bhler (1932) have both ca, and Haradatta does not make it clear which of the two should be deleted. Since all five sütras preceding this (2.26.13-17) and the four following it (2.26.1922) have ca as the second word, I would conclude that the offending ca is the last. 29) Åp 2.26.6: sarvato yojana∫ nagara∫ taskarebhyo rakßyam (“They must protect a town from thieves up to a yojana on all sides”). Haradatta notes the incorrect variant: rakßyann ity apapå™haΔ. 30) Åp 2.27.5: avi†iß™a∫ hi paratva∫ påñeΔ (“[Levirate is forbidden] for with respect to the husband all are equally outsiders”). Haradatta notes the incorrect variant: ava†iß™am ity apapå™haΔ. 10 It is unclear from the manuscript tradition whether Haradatta’s reading of the variant was prå∞jalim or prå∞jali. 11 Böhtlingk (1885b) proposes to emend the text here to reflect this “correct” reading.

308

LANGUAGE, TEXTS AND SOCIETY

Perhaps the greatest insight into the research conducted by Haradatta in writing his commentary is provided by the following passage, and for that reason I want to discuss it in greater detail. 31) Åp 2.17.25: anyatra råhudar†anåt (“except when there is a [lunar] eclipse”). Haradatta has a long comment on the authenticity and the meaning of this sütra given at the end of his comments on sütra 24. I quote it in full: anantaram “anyatra råhudar†anåd” iti pa™hanti Ù “na ca naktam” [sütra 23] ity asyåpavådaΔ råhudar†ane naktam api kurvœteti Ù udœcyås tv etat pråyeña na pa™hanti Ù tathå ca pürvair na vyåkhyåtam Ù pratyuta “na ca naktam” ity etat somagrahañavißayam iti vyåkhyåtam Ù pa™hyamåna∫ tu “na ca naktam” ity asyånantara∫ pa™hitu∫ yuktam. “Immediately after this [i.e., after sütra 23], (some) read “except when there is a (lunar) eclipse”. This is an exception to “An ancestral offering should not be performed after nightfall” [sütra 23]; that is, when there is a (lunar) eclipse, one may perform it even at night. The northerners for the most part, however, do not read this. Accordingly, it has not been commented on by earlier (commentators). On the contrary, they have explained the (sütra) “An ancestral offering should not be performed after nightfall” as referring to a lunar eclipse. If it is to be read, however, it should properly be read immediately after the (sütra) “An ancestral offering should not be performed after nightfall” [that is, if we accept sütra 25, then it should come after sütra 23 and not after 24, which reads “and once it is started the performer should not eat until it is completed”]. This comment shows that Haradatta had taken the trouble to check the readings of Åpastamba preserved in the north. It is unclear what part of India “north” refers to, and whether Haradatta is referring to northern manuscripts or to actual recitations in northern schools, a question I will return to in the conclusion. He does, however, appear to privilege the northern tradition at least in this case. The comment also shows that Haradatta had access to and utilized earlier commentaries.12 The expression tathå ca (“accordingly”) in the statement that the earlier commentators have not commented on this sütra appears to indicate that the reason for their ignoring it is that fact that it was absent in the northern recension. If this is true, then we can draw a couple of conclusions. First, Haradatta locates these “early commentators” in the north. Second, he thus privileges the northern tradition of transmission over the southern, with which he was probably more familiar. 12 Because of the use of the plural here, Bhler (1892, vi) concludes that Haradatta had access to at least three previous commentaries; if there were only two he would have used the dual. This may be somewhat of a stretch, since commentators often use the honorific plural. But Bhler is on firmer ground when he shows that Haradatta refers to at least three opinions of his predecessors, using the words aparaΔ, anye, and anye at Åp 1.5.2, and anye, eke, and anye at Åp 2.17.22.

Sanskrit Commentators and the Transmission of Texts

309

B. Explanations of Difficult or Ungrammatical Readings Haradatta was a close reader of the text, whether it was in manuscript form or orally recited by experts, a question I will return to later. Whenever he found an unusual or ungrammatical reading, he did not, as Böhtlingk often does in his critique of the critical editions of Åpastamba and Vasiß™ha prepared by Bhler and Fhrer, change the received readings to suit a preconceived notion of grammatical correctness. Instead, Haradatta attempted to explain the difficult readings. Sometimes he calls them vedic peculiarities (chåndasa) or erroneous readings (apapå™ha), but he never changes them. I have noted 58 such instances. 1) Åp 1.1.27: å ßo∂a†åd bråhmañasyånåtyayaΔ (“In the case of a Brahmin there is no lapse [in postponing vedic initiation] until the sixteenth year”). Haradatta comments on the irregular long nå of anåtyaya, calling it either an accidental or a secondary formation.13 We have two somewhat different versions of Haradatta’s text in the two editions: Bhler (1892) anatyayo ‘natikramaΔ sa evånåtyayaΔ Ù yådr¢cchiko dœrghaΔ Ù åõgo vå prayogaΔ; Kashi Sanskrit Series edition: atyayo ‘tikramaΔ Ù sa evåtyayaΔ tadabhåvo ‘nåtyayaΔ Ù yådr¢cchiko dœrghaΔ åõgo vå pra†leßaΔ. 2) Åp 1.2.13: pådünam ([An initiated person should reside as a student in his teacher’s house for forty-eight years or] “for one-quarter of that time.”). The normal form with regular sandhi would be pådonam (påda + ünam), and Haradatta explains the irregularity: pådenona∫ padünam Ù pararüpa∫ †akandhvåditvåt (cf. Kåtyåyana’s Vårttika and Pata∞jali on Påñini 6.1.94; see also B: 6 for pararüpam).14 3) Åp 1.2.21: adhåsana†åyœ (“[A student should] occupy a lower seat and bed [than his teacher].”). The grammatically correct form with regular sandhi should be adhaåsana†åyœ, and there appears to be a second sandhi of a + å in the received form [cf. B: 23]. Haradatta thinks this is a mistake or a vedic peculiarity: adhaΔ†abdasya savarñadœrghatva∫ [cf. Påñini 6.1.101] chåndasam apapå™ho vå. 4) Åp 1.3.22: aglå∫snuΔ (“[He shall] not be lethargic”). Haradatta notes the irregular insertion of an anusvåra before the sibilant, calling it again either an error or a vedic peculiarity:15 atrånusvåraΔ chåndaso ‘papå™ho vå. 5) Åp 1.4.1: yad ucchiß™a∫ prå†∞åti (“When he eats [his teacher’s] leftovers”). The correct form should be prå†nåti, and Haradatta calls the irregular palatal “∞” a vedic form, since the dental “n” and “t” are 13 Böhtlingk (1885b) dismisses this reading, taking it to be merely a scribal error for anatyaya, and cites BaudhÂyana Dharmasütra 1.3.12 in support. Bhler (1886b) defends the reading anåtyaya, explaining that there is a negative particle ana besides a, an, and na. 14 Böhtlingk (1885b) dismisses this, saying: “pådonam wird wohl die richtige Lesart sein.” 15 Böhtlingk (1885b) dismisses the reading with an anusvåra, saying: “aglå∫snuΔ kann auch nach Haradatta eine falsche Lesart sein, und darin hat er gegen Bhler Recht.” Böhtlingk does not note that Haradatta thinks that it could also be a vedic form.

310

LANGUAGE, TEXTS AND SOCIETY

not changed into palatals after “†”: ∞akårapå™ha† chåndasaΔ †åditi [Påñini 8.4.44] †cutvapratißedhåt.16 6) Åp 1.5.7: yat ki∫ ca samåhito ‘brahma py åcåryåd upayuõkte ... (“Whatever other science besides the Veda a steadfast man learns from his teacher [...]”). Haradatta notes the elision of the initial “a” in pi (= api) and seeks to explain it in two ways: a) the final vowel of the preceding word and the initial vowel of api coalesce to take just the form of the latter vowel (pararüpa: thus abrahma + api = abrahmapi rather than abrahmåpi), as in the example karka + andhu = karkandhu (“well”, rather than karkåndhu); b) the loss of the initial “a” of api as in the examples pihita etc.: pararüpa∫ karkandhuvat Ù aper akåralopo vå pihitapinaddhådivat (see B: 2).17 7) Åp 1.6.28 [cf. A: 24]: Haradatta notes the parasmaipada form of the present participle locative ceß™ati, which he considers irregular: vyatyayena parasmaipadam. 8) Åp 1. 6.36: bhuktvå cåsya sakå†e nånütthåyocchiß™a∫ prayacchet (“After he has eaten in his [teacher’s] presence, moreover, he should not give away his leftovers without first getting up.”). Haradatta notes the irregular lengthening of “ü” in anütthåya calling it a vedic peculiarity: anütthåya chåndaso dœrghaΔ. 9) Åp 1.7.11: rajasvalo raktadan satyavådœ syåd iti hi bråhmañam (“A Bråhmaña states: ‘He shall keep his body dirty, his teeth stained, and his speech true’ “). Haradatta notes the rare form raktadan, calling it a vedic form for the common raktadanta (see also B: 53): chåndaso dannåde†aΔ paõkiladanta ity arthaΔ. 10) Åp 1.7.27: Haradatta notes the singular form dåre in this sütra and considers it an irregular vedic form: dåra ity ekavacana∫ chåndasam (see also B: 45). 11) Åp 1.8.15: The text contains the term vyupajåva (“whispering”), with several manuscripts recording the more regular form vyupajåpa.18 Haradatta accepts the former reading and considers the “v” to be a vedic peculiarity or an error: vyupajåvaΔ karñayor muhur muhur jalpanam Ù vakåra† chåndaso ‘papå™ho vå. 12) Åp 1.8.19: This sütra contains the expression pråcåryåyopasa∫gr¢hya (“having clasped the feet of his teacher’s teacher”). Haradatta notes the dative case when we should have expected the accusative, remarking: pråcåryåya dvitœyårthe caturthœ.19 16 Böhtlingk (1885b) dismisses this as a false archaism and wants here to follow the majority opinion of the manuscripts. But Haradatta is older than our oldest manuscript, and Bhler (1892, 8) is right in following his testimony about this difficult reading, saying: “The form is Prakr¢tic. Similarly the Påli pa∞ho presupposes a Sanskrit form pra†∞a.” 17 Here too Böhtlingk (1885b) dismisses Haradatta’s explanation and reverts to the regular reading: “(a)brahma pi is trotz Haradatta fehlerhaft fr (a)brahmåpi.” 18 Böhtlingk (1885b) accepts this as the correct reading, noting “vypajåva is ja auch nach Haradatta möglicher Weise eine falsche Lesart.” Once again Böhtlingk ignores the fact that Haradatta considers the possibility that it is a vedic form. 19 Böhtlingk (1885b) cannot accept this anomaly: “Dass upasa∫graha in demselben

Sanskrit Commentators and the Transmission of Texts

311

13) Åp 1.8.22: muhü∫† cåcåryakula∫ dar†anårtho gacchet (“He shall go frequently to visit his teacher’s family”). The term muhü∫s has two irregularities: long “ü” and the anusvåra before the sibilant (see above B: 4).20 Haradatta thinks that this form is vedic and further that the usual repetition of this term has been omitted here: muhü∫† cety anusvåradœrghau chåndasau Ù vœpsålopa† cåtra draß™avyaΔ Ù muhur muhur iti vivakßitam. 14) Åp 1.9.14-15: antaΔ†avam Ù anta†cåñ∂ålam (“[Vedic recitation is suspended in a village] in which there is a corpse or a Cåñ∂åla”). Haradatta notes the anomalous nominative endings of the two compounds, where we should have expected locative endings,21 and explains that the nominatives are used here with a locative meaning or the compounds are adverbial: ubhayatra prathamå saptamyarthe avyayœbhåvo vå vibhaktyarthe draß™avyaΔ. 15) Åp 1.9.21: svapnaparyånta∫ vidyuti (“When there is lightning, [vedic recitation is suspended] until he has slept”). Haradatta thinks that the correct form should be the feminine svapnaparyantåm, agreeing with råtrim of the previous sütra, and considers the transposition of the short and long vowels between the final and the penultimate syllables an error or a vedic irregularity:22 antyo dœrghaΔ Ù upåntyo hrasvaΔ Ù viparyåsa† chåndaso ‘papå™ho vå. 16) Åp 1.10.18: adhœyåneßu vå yatrånyo vyaveyåd etam eva †abdam utsr¢jyådhœyœta (“Likewise, when someone comes when they are engaged in vedic recitation, he [they] may continue the recitation only after that person utters the same words [i.e., ‘adhœhi bho’]”). Haradatta notes that the plural in adhœyåneßu is superfluous or nonessential (the rule applies even when a single person is reciting) and that the verb, which should have been in the plural, is in the singular because the rule is directed at each reciter individually: bahuvacanam atantram Ù pratyekam upade†åd ekavacanam Ù adhœyœran. 17) Åp 1.10.19: This sütra contains the word salåvr¢kœ. Haradatta notes that the normal spelling is sålåvr¢kœ, but accepts the short initial vowel here as an exception: indro yatœn sålåvr¢kebhya ityådau [TS 6.2.7.5] dar†anåt sarvatrådimaΔ svaro dœrghaΔ Ù sa evåya∫ hrasvaΔ [variant vikr¢taΔ] prayuktaΔ. It is remarkable that Haradatta does Sütra das eine Mal mit dem Acc. und das andere Mal mit den Dativ construirt wrde, ist doch beinahe unglaublich. Man lese pråcåryam upasa∫gr¢hya.” 20 Böhtlingk (1885b) dismisses this form, saying “muhü† ca sicherlich fehlerhaft fr muhu† ca.” 21 Böhtlingk (1885b) emends the nominatives to locatives antaΔ†ave and anta†cåñÛ∂åle, saying that the anusvåra and “e” sign are easily confused! Bhler (1886b) rightly observes that such a mistake is only possible in Devanågarœ and not in the southern scripts. 22 Böhtlingk (1885b) takes -paryantam to be the correct form and, as Bhler (1886b) has pointed out, erroneously thinks that Haradatta’s reading is -paryantåm rather than -paryåntam. When Haradatta says that “the last is long and the penultimate is short” he is talking about what the reading in his eyes should be; the reverse is the actual reading which Haradatta found in his sources and which he took to be a vedic peculiarity or a mistake.

312

LANGUAGE, TEXTS AND SOCIETY

not simply make this easy change but takes the trouble to note this insignificant variant. 18) Åp 1.11.13: yathå pådaprakßålanotsådanånulepañånœti (“[During private activities] such as washing the feet, massaging, and applying oil [one should not recite the Veda]”). Here we have the irregular retroflex “ñ” in anulepañåni.23 What a careful reader Haradatta must have been to note this minor irregularity! (see also B: 20, 42) He considers it either “fortuitous”24 or an error: ñatvam åkasmikam apapå™ho vå. 19) Åp 1.11.31: Haradatta notes the irregular masculine gender of the samåhåradvandva compound pratisüryamatsyaΔ (“a parhelion and a comet”), which should have been in the neuter:25 samåhåradvandve chåndaso liõgavyatyayaΔ. 20) Åp 1.12.8: åryasamayo hy agr¢hyamånakårañaΔ (“for that accepted practice of the Åryas has no tangible motive”). Here again Haradatta carefully notes the irregular dental “n” in agr¢hyamåna and sees it as a vedic peculiarity:26 sütre ‘gr¢hyamånkåraña iti ñatvåbhåva† chåndasaΔ. 21) Åp 1.16.2: åsœnas trir åcåmed dhr¢dayaõgåbhir adbhiΔ (“Seated on his haunches, let him sip three times with water sufficient to reach his heart”). The grammatical object here is put in the instrumental, and Haradatta sees it as irregular: adbhiΔ tr¢tœyå dvitœyårthe. 22) Åp 1.16.3: trir oß™hau parimr¢jet (“He should wipe his lips three times”). Haradatta is uncomfortable with the form parimr¢jet [6th class verb] and glosses it with parimr¢jyåt [2nd class]. 23) Åp 1.19.8: sarvatopeta∫ vårßyåyañœyam (“According to Vårßyåyañi, [food] that one receives [unasked] from anybody [is pure]”). Haradatta notes the double sandhi [cf. B: 3] in sarvatopetam and attributes it to a vedic peculiarity: sarvata upeta∫ sarvatopeta∫ chåndoso guñaΔ. 24) Åp 1.19.12: nånaniyogapürvam iti hårœtaΔ (“[One may eat food received unasked] but not if it is received subsequent to an invitation, according to Hårœta”). Haradatta notes the irregular double negative 23 Bhler (1892) in his edition reads anulepanåñi. Haradatta’s comment on Åp 1.32.5 (see B: 42), where there is no ambiguity since the reading is the singular anulepañaΔ, makes it clear that the retroflex should be the penultimate “n” and not the final. 24 The term åkasmika appears to have a technical meaning with reference to the peculiar pronunciation of certain letters, especially the substitution of retroflex “ñ” for the dental, by Brahmins of the Taittirœya branch. See M. Deshpande 1993a, p. 210, n. 204; Bhartr¢hari, Mahåbhåßyadœpikå, Fasc. IV, Ahnika I, ed. J. Bronkhorst (Poona: Bhandarkar Oriental Research Institute, 1987), p. 1. K. V. Abhyankar in his Veda-Padapå™ha-Carcå (Pune: Bhandarkar Oriental Research Institute, 1974), p. 35 comments: taittirœyå nimitta∫ vinå nakårasya sthåne ñakåram uccårayanti. I want to thank Madhav Deshpande for this information. 25 Böhtlingk (1885b) dismisses Haradatta’s explanation and emends to the regular neuter form, because -matsya∞ca can easily be read as -matsya†ca (in the sütra ca follows). Here again Böhtlingk is thinking only of Devanågarœ; in Grantha the two forms are quite distinct. 26 Böhtlingk (1885b) wants to adopt the regular reading with the retroflex “ñ”, refer-

Sanskrit Commentators and the Transmission of Texts

313

ananiyoga, which is used, he thinks, to emphasize the original positive meaning:27 nivedana∫ niyogas tadabhåvo ‘niyogaΔ Ù punar na∞samåsaΔ Ù dvau na∞au prakr¢tårtham ati†ayena gamayataΔ. 25) Åp 1.19.15: this sütra contains the expression yåcann anr¢tasa∫kare (“A supplicant [rubs his sin off] on the man who makes false promises”). Haradatta notes the anomalous sa∫kara for sa∫gara, taking it to be a vedic form:28 kakåras tu chåndasaΔ. 26) Åp 1.20.6: na dharmådharmau carata åva∫ sva iti (“Dharma and Adharma do not go around saying, ‘Here we are!’”). Haradatta notes the vedic form åvam for the regular åvåm and comments: åvam iti chåndasa∫ rüpam Ù bhåßåyå∫ tu prathamåyå† ca dvivacane bhåßåyam iti (Påñini 7.2.88) åtva∫ pråpnoti. 27) Åp 1.21.8 contains the compound brahmojjham. Haradatta thinks that this should be in the masculine and comments: brahmojjham ujjha utsarge Ù bhåve gha∞ Ù chåndaso liõgavyatyayaΔ. 28) Åp 1.21.9: A list of women with whom sex is forbidden contains the compound gurvœsakhi (“female friend of a female guru”) with the irregular short final “i”, which Haradatta takes to be vedic:29 sakhœ†abdasya chåndaso hrasvaΔ. 29) Åp 1.22.1: contains the expression adhyåtmikån yogån (“disciplines pertaining to the inner self”). Haradatta takes the short initial vowel to be vedic:30 adhyåtmikån chåndaso vr¢ddhyabhåvaΔ. 30) Åp 1.23.6 contains the expression sårvagåmœ. Haradatta explains this unusual form:31 sårvagåmœ sarvasmai hitaΔ sårvaΔ åtmå ta∫ gacchati pråpnoti. 31) Åp 1.24.11 contains the expression †ava†iradhvajaΔ (“a man who carries a skull as a banner”) with the irregular elision of the final “s” of †iras. Haradatta takes this also to be a vedic peculiarity: †ava†iro dhvajo yasya sa †ava†iradhvajaΔ Ù salopa† chåndasaΔ. 32) Åp 1.28.20: striyås tu bhatr¢vyatikrame kr¢cchradvåda†aråtråbhyåsas tåvanta∫ kålam (“Women who abandon their husbands, on the other hand, should perform the twelve-day arduous penance for the same length of time”). Most extant manuscripts have regularized the reading to either bhartr¢vyatikrame32 or bhartur vyatikrame. ring the reader to Pata∞jali on Påñini 6.3.109. Grammatical correctness was always uppermost for him. 27 Böhtlingk (1885b) objects to the double negative and suggests nånanuniyogapürvam. Bhler (1886b) correctly points out that this would give a meaning opposite to the one that is intended. 28 Böhtlingk (1885b) rejects this claim: “Dass sa∫kara die ältere Form fr sa∫gara sein sollte, will mir nicht einleuchten.” Could this confusion between “g” and “k” be influenced by Tamil, which does not have the soft “g”? 29 This is rejected by Böhtlingk (1885b), who wants to accept the regular sakhœ as at Baudhåyana Dharmasütra 2.2.13. 30 Böhtlingk (1885b) calls this form “die schlechtere Lesart”. 31 Böhtlingk (1885b) calls it “wohl nicht richtig”. 32 Böhtlingk (1885b) rejects the irregular reading of Bhler’s edition and accepts the regular bhartr¢vyatikrame.

314

LANGUAGE, TEXTS AND SOCIETY

Haradatta’s comment on this is ambiguous, Bhler’s edition reading: bhatr¢vyatikrama iti chåndaso rephalopaΔ, whereas the Kashi Sanskrit Series ed. reads: bhartuvyatikrama iti chåndaso rephalopaΔ. Haradatta simply states that there is an elision of “r” without specifying whether it is that before “t” in the stem bhartr¢ or after “u” of the genitive bhartur. Given that all the Grantha manuscripts (which, as Bhler acknowledges, are more trustworthy than the Devanågarœ ones) used by Bhler, as well as the parallel passage in Hirañyake†i Dharmasütra, contains the repha above “t”, I feel that the form intended by Haradatta is probably bhartuvyatikrame, and I have adopted this in my new edition of Åpastamba (Olivelle 2000). Further, in Grantha, even more than in Devanågarœ, the signs for “u” and “r¢” can be confused with each other. In any case, Haradatta carefully notes the loss of the repha, calling it a vedic peculiarity. 33) Åp 1.29.9 contains the unusual verbal form sa∫pratyapatsyata. Haradatta takes this as an aorist expressing the hope for the future, noting the irregular insertion of “y” after the sibilant, which he takes to be a vedic peculiarity or an error: sa∫pratyapatsyata Ù å†a∫såyå∫ bhütavac ceti [Påñini 3.3.132] bhavißyati luõ Ù sakåråt paro yakåra† chåndaso ‘papå™ho vå. 34) Åp 1.29.15 contains the word abhœcåra (“sorcery”) with the irregular lengthening of “i”. Haradatta explains this formation by citing Påñini, according to whom the final vowel of a prefix is lengthened diversely before a word formed with the kr¢t suffix “a”: abhicåra eva abhœcåraΔ Ù upasargasya gha∞y amanußye bahulam iti [Påñini 6.3.122] dœrghaΔ. 35) Åp 1.30.2 likewise contains parœmåña with the irregularly lengthened “i”. Haradatta explains this as a vedic peculiarity: parimåñam eva parœmañam Ù chåndaso dœrghaΔ. 36) Åp 1.30.3: vidyå vratena cety eke (“Some say: ‘[He should bathe] after learning the Veda and completing the vow.’“). Haradatta recognizes vidyå as a vedic instrumental:33 vidyeti tr¢tœyaikavacanasyåkårasya supå∫ suluk ityådinå [Påñini 7.1.39] luk. 37) Åp. 1.30.12: anüdbhåsi våso vasœta (“He should wear clothes that are not shiny”). Haradatta notes the irregular lengthening of “u” in anüdbhåsi and calls it a vedic form:34 anüdbhåsi chåndaso dœrghaΔ. 38) Åp 1.30.17: svå∫ tu chåyåm avamehet (“He may, however, discharge urine in his own shadow”). Haradatta is quite observant here in noting the omission of “c” before “ch” (i.e., tucchåyåm) and sees it as a vedic peculiarity: chåndasas tugabhåvaΔ.35 39) Åp 1.31.12 contains the word pra†åstam with the irregular long 33 Böhtlingk (1885b) rejects this interpretation and wants to make a compound of the two words: vidyåvratena. 34 Böhtlingk (1885b), on the other hand, calls it merely an error for the regular anudbhåsi. 35 Böhtlingk (1885b) dismisses Haradatta’s comment: “Die Bemerkung Haradatta’s, das tu chåyåm ein Archaismus fr tu cchåyåm sei, ist ganz werthlos.”

Sanskrit Commentators and the Transmission of Texts

315

“å”, which Haradatta takes to be vedic: pra†åsta∫ pra†astam Ù chåndaso dœrghaΔ. 40) Åp 1.31.19: This sütra, which we have already examined (A: 3), contains the irregular form svådhœyas in place of sådhœyas (“exceedingly”). Haradatta notes the insertion of “v” and calls it vedic:36 svådhœyaΔ vakåra† chåndasaΔ. 41) Åp 1. 31.22 contains the irregular gerund skuptvå (probably meaning to pick the teeth). If the verb is √skubh, the gerund should be skubdhvå; and if the verb is √sku, it should be skutvå.37 Haradatta notes the anomaly and offers two possible solutions: a vedic peculiarity of substituting the surd “p” for “bh” (if the root is √skubh), or the addition of the letter “p” (if the root is √sku): skuptveti skubhnoteΔ ktvåpratyaye chåndasa∫ bhakårasya cartvam Ù skunoter vå pakåra upajanaΔ. 42) Åp 1.32.5: anåviΔsraganulepañaΔ syåt (“Let him not appear in public wearing a garland or anointed with oil”). Here again, as in Åp 1.11.13 (see B: 18), we have the irregular retroflex “ñ” in anulepaña.38 Haradatta refers the reader back to his earlier explanation: ñatva∫ pürvavat. 43) Åp 1.32.16: kåmam apa††ayœta (“He may, if he so wishes, rest leaning against something”). Haradatta takes the regular form of the verb to be †rayœta and notices two irregularities: the doubling of “†” and the omission of “r”: tatra rephalopa† chåndasaΔ tathå †akårasya dvirvacanam. 44) Åp 1.32.26 spells the name dharmaprahråda, instead of the more common dharmaprahlåda. Haradatta notes the irregularity as vedic: prahråda†abde hakåråt paro repha† chåndasaΔ. 45) Åp 2.1.17: Here, as earlier at Åp 1.7.27 (see B: 10) Haradatta takes the singular dåreña as a vedic peculiarity, since the word dåra is normally declined in the plural: chåndasam ekavacanam Ù nitya∫ bahuvacanånto hi dåra†abdaΔ. 46) Åp 2.5.9 contains the compound bhrukßepaña with the irregular short “u”, which Haradatta explains as a vedic peculiarity: bhrukßepaña∫ bhrükßepaΔ chåndaso hrasvaΔ. 47) Åp 2.5.17: r¢tve vå jåyåm (“[He may optionally have sex] with his wife during her season”). Haradatta notes the irregular r¢tve in place of the normal r¢tvye, taking the omission of “y” to be vedic: r¢tvya iti rüpasiddhiΔ Ù atra yalopa† cåndasaΔ. 48) Åp 2.6.17: uddhr¢tåny annåny avekßeteda∫ bhüyå3 idå3m iti (“When the food has been dished out, he should look at it, thinking: ‘Is this portion larger or this?’”). Haradatta notes that when there is an act of reflection between two alternatives, in ordinary Sanskrit 36

Böhtlingk (1885b) rejects this explanation as “laughable” and emends it to the regular sådhœyaΔ. Bhler (1886b), however, points out that Winternitz also found a parallel form in the Åpastamba Gr¢hyasütra. 37 This is adopted by Böhtlingk (1885b); see also Böhtlingk 1885d. 38 Once again, Böhtlingk (1885b) rejects the retroflex.

316

LANGUAGE, TEXTS AND SOCIETY

(bhåßå) only the first alternative is subject to pluta (prolation indicated by the number 3). He explains the two plutas in the sütra as a vedic peculiarity: vicåre plutaΔ [cf. Påñini 8.2.97] Ù pürva∫ tu bhåßåyam iti [Påñini 8.2.98] etad avekßitam Ù chåndasa evåya∫ prayogaΔ. 49) Åp 2.7.13: vråtya tarpaya∫stv iti (“Vråtya, let this refresh you”). Haradatta sees the insertion of the anusvåra and the sibilant into the regular tarpayatu as a vedic peculiarity:39 anusvårasakårau chåndasau. 50) Åp 2.8.11 contains the word †œkßå (“phonetics”) with a long “œ”.40 Haradatta explains this, citing Påñini 6.3.109 that permits the elision and mutation of letters when they are found in vedic usage: pr¢ßodaråditvåd dœrghaΔ. 51) Åp 2.9.4 contains the word apratœbhåyåm (“forgetting”) with a long “œ”, which Haradatta explains as vedic: prater dœrgha† chåndasaΔ. 52) Åp 2.10.11 contains the word paråõåvr¢tta (“fleeing”) with the unusual insertion of “õ”, which Haradatta again explains as vedic: paråõåvr¢tta iti õakåra† chåndasaΔ. 53) Åp 2.12.22 contains the compound †yåvada (“man with black teeth”) without the usual final “n” (see B: 9), an elision that Haradatta takes to be vedic: vibhåßå †yåvårokåbhyåm iti [Påñini 5.4.144] datråde†aΔ tasya lopa† chåndasaΔ. 54) Åp 2.13.6 contains the irregular nominative plural of the present participle bibhyantaΔ for the regular bibhyataΔ. Haradatta takes the insertion of “n” to be vedic:41 bibhyantaΔ chåndaso num bibhyataΔ. 55) Åp 2.21.4 contains the word anütsargaΔ with the irregular long “ü”, which Haradatta takes to be vedic: anütsargaΔ chåndaso dœrghaΔ. 56) Åp 2.23.12: tataΔ param anantya∫ phala∫ svargya†abda∫ †rüyate (“Thereafter, the Vedas declare, they obtain an eternal reward designated by the term ‘heaven’”) Haradatta sees the insertion of “y” into anantya and svargya as either a vedic peculiarity or an error:42 anantya∫ svargyam iti yakåra† chåndasa upajano ‘papå™ho vå. 57) Åp 2.27.17 contains the compound cakßunirodha for the regular cakßurnirodha (“blindfolding”) with the irregular omission of “s” of cakßus, which Haradatta takes to be vedic: cakßunirodha iti rephalopa† chåndasaΔ. 58) Åp 2.29.13: lakßañakarmañåt tu samåpyate (“But by acting according to the markers one can master [dharma]”). Haradatta gives two explanations for the irregular karmañåt, which, as Bhler points out, was probably caused by a Prakritic doubling of the “t” of the following tu. The doubling of “t”, according to Haradatta, is 39 Böhtlingk (1885b) rejects this reading and emends the text against the testimony of all the manuscripts: “Wolte ich auch dem klugen Haradatta den Unsinn nachsehen, dass tarpaya∫s ein Archaismus fr tarpaya sei, so msste ich doch fragen wie tarpaya zu der Bedeutung tarpayasva oder tarpayåtmånam käme, und was das tu am Schlusse hier bedeuten solle. Die richtige Lesart tarpayes tveti oder tarpayasveti liegt doch wahrlich nahe genug.” 40 Böhtlingk (1885b) takes this to be simply an error for †ikßå. 41 Böhtlingk (1885b) restores the regular bibhyataΔ. 42 Böhtlingk (1885b) prefers the regular ananta and svarga.

Sanskrit Commentators and the Transmission of Texts

317

either a vedic peculiarity or there is the insertion between karmañå and tu of the particle åt (“immediately”): karmañåt tv iti dvitakårapå™ho ‘yam årßaΔ Ù åd iti vå nipåtasya pra†leßaΔ Ù sa ca sadya ity asyårthe draß™avyaΔ.

C. Explanations without Emendations The care with which Haradatta transmits and explains the received text is demonstrated also in numerous explanations where the received text is not to his liking. But even here Haradatta does not simply emend the text but tries his best to draw out from the text what he perceives to be its proper meaning.43 Some of these explanations we have already seen in the preceding discussion. I cite here 15 additional cases. 1) Åp 1.8.26: antevåsy anantevåsœ bhavati vinihitåtmå guråv anaipuñam åpadyamånaΔ (“A pupil ceases to be a pupil when he is inattentive to his teacher and so becomes a dolt”).44 Haradatta understands this text differently, and his interpretation requires the final participle to be in the causative at least implicitly: åpadyamåna ity antarbhåvitañyarthaΔ [...] guråv anaipuñam åpådayati. 2) Åp 1.13.16: vr¢ddhånå∫ tu (“[Obedience is not required] also towards older [fellow students]”). Haradatta, as the context clearly requires, wants ca (“and”, “also”) in place of tu (“but”, “however”) and comments: tu† cårthe Ù vr¢ddhånå∫ cåntevåsinå∫ na gatir vidyate. 3) Åp 1.14.4: na cåsmin doßa∫ pa†yet (“[When someone asks him for instruction, he should not spurn him] provided he does not see any fault in him”). The context clearly calls for “if” and Haradatta interprets ca (“and”) to mean ced (“if”), citing grammatical precedence: cañ iti nipåto ‘sti Ù nipåtair yadyadihantakuvinnecceccañkaccidyatrayuktam iti [Påñini 8.1.30] Ù sa cedarthe vartate. 4) Åp 1.32.24: müla∫ tüla∫ vr¢hati (“It tears up his root and sprout [...]”). Haradatta’s comment on this passage shows how careful he was in preserving the text, here the danger being that “v” be changed to “b” or a similar cognate letter. He carefully notes that the reading should be “v”: vr¢hati utpå™ayati dantoß™hyo vakåraΔ. 5) Åp 2.2.4: yathaußadhivanaspatœnå∫ bœjasya kßetrakarmavi†eße phalaparivr¢ddhir evam (“This is similar to the way the seeds of plants and trees, when they are sown on a well-plowed field, increase their fruit”). The context is how the transmigratory process works. Haradatta wants a ca after yathå so as to connect this to the preceding and thinks that it has been dropped: calopo ‘tra draß™avyaΔ Ù yathå coßadhadhœnå∫ [...]. 43 As we have seen, Haradatta knows his Påñini well and cites his rules in support of his explanations. Besides those cited here, Haradatta cites Påñini or grammatical rules at Åp. 1.1.28; 1.2.21; 1.3.26; 1.4.15; 1.4.19; 1.5.1; 1.5.23. 44 I follow Böhtlingk’s (1885b, 519) convincing argument. Haradatta has misunderstood this passage, and Bhler follows him in this labored translation: “That pupil who, attending to two (teachers), accuses his (principal and first) teacher of ignorance, remains no (longer) a pupil.”

318

LANGUAGE, TEXTS AND SOCIETY

6) Åp 2.3.8: api våß™amœßv eva parvasu vå vaperan (“Alternatively, they may shave only on the eighth day of each fortnight or on newand full-moon days”). Now, normally one does not shave oneself but gets a barber to do it. So, generally the verb is in the causative. Haradatta explains that the causative is implied here, especially because the causative was used in the preceding sütra: vaperann iti antarbhåvitañyarthaΔ Ù våpayerann ity arthaΔ Ù tathå ca lomanakhavåpanam iti [Åp 2.3.7] pürvatra ñicprayuktaΔ. 7) Åp 2.4.24: bråhmaña åcåryaΔ smaryate tu (“Tradition says that only a Brahmin can be a teacher”). Haradatta points out that tu has here a restrictive meaning and further that it is in the wrong place, since it should come immediately after the first word: tu†abdo ‘vadhårañartho bhinnakrama†ca Ù bråhmaña eva [...]. 8) Åp 2.5.7: anya∫ vå samudetam (“[He should do the same] also for any other distinguished guest”). Haradatta notes that vå here should have the meaning of combination (i.e., api, ca): vå†abdaΔ samuccaye Ù anyam apy evam evåcamayet. 9) Åp 2.6.5: svadharmayukta∫ ku™umbinam abhyågacchati dharmapuraskåro nånyaprayojanaΔ so ‘tithir bhavati (“When such a man comes to the home of a householder devoted to the dharma proper to him –and he comes for no other purpose than to discharge the dharma– then he is a ‘guest’”). Haradatta sees the need for the correlative yaΔ at the beginning to grammatically complete this sentence: ådito yacchabdo draß™avyaΔ ante sa iti dar†anåt. 10) Åp 2.8.6: åcårya r¢tvik snåtako råjå vå dharmayuktaΔ ([Among people deserving a cow and honey mixture:] “a teacher, an officiating priest, a bath-graduate, and a king who follows the Law”). Here also Haradatta sees vå as having the meaning of ca: vå†abdaΔ samuccaye. 11) Åp 2.13.2: dåyenåvyatikrama† cobhayoΔ (“And neither parent may deprive such a son of his share in the estate”). Haradatta understands this sütra differently: “[Sons have the right] to the estate, if they do not sin against either parent.”45 Hence he has to interpret ca to mean “if” (see above C: 3): ca iti cedarthe Ù avyatikrama† cet Ù yadi te måtara∫ pitara∫ ca na vyatikrameyuΔ. 12) Åp 2.13.6: apramattå rakßatha tantum etam (“Diligently guard this progeny of yours”). Haradatta carefully notes the indicative when we should expect the verb to be in the imperative, even though the difference in the two forms is minimal (the aspirate “tha” for the nonaspirate “ta”), and states that the indicative here carries an imperative meaning:46 rakßatha Ù lo∂arthe la™ Ù rakßatetyarthaΔ. 13) Åp 2.15.12 contains the words de†ataΔ kålataΔ †aucataΔ etc. Haradatta explains that these ablative/adverbial endings carry a locative meaning: saptamyarthe tasil. 45 This interpretation of Haradatta is followed by Bhler. My interpretation of the sütra is based on the view of another commentator cited by Haradatta. 46 Böhtlingk (1885b) simply changes it to rakßata.

Sanskrit Commentators and the Transmission of Texts

319

14) Åp 2.20.2: samutedå∫† ca bhojayen na cåtadguñåyocchiß™a∫ dadyuΔ (“He should feed only individuals who possess the required qualities, and not give any leftover food to anyone who does not possess the same qualities”). Here we have a verb in the singular followed by one in the plural. Haradatta explains: dadyur iti bahuvacana∫ tathåvidhakartr¢bahutvåpekßam. 15) Åp 2.27.1: carite yathåpura∫ dharmåd dhi sa∫bandhaΔ (“Once [the expiation] has been performed, [the guardians should treat them] as before, for their relationship is based on dharma”). Haradatta notes the ablative in place of the normal instrumental: tr¢tœyårthe pa∞camœ Ù dharmeña sa∫bandho bhavati.

D. Concluding Observations Haradatta was a careful and observant reader of the text–that much is clear. It is also clear that the tradition of textual transmission, whether it was oral, written, or a combination of both, from which Haradatta received his text was an extremely faithful one, preserving not only the substance of Åpastamba’s text but also the unusual and extraordinary readings. It is apparent that Haradatta had conducted meticulous research prior to writing his commentary; he not only notes variant readings from different sources from his southern homeland but also consulted the text as handed down in the north. It is, therefore, remarkable that except for a single sütra (Åp 2.17.25; A: 31), Haradatta did not find substantial differences in the various recensions to which he had access. Haradatta was also a faithful transmitter of the text that he received from a tradition of reciters and scribes. Unlike Böhtlingk, Haradatta never changes a reading, even when he suspects that it may simply be an error. He notes the erroneous reading even when he presents the correct one, the yuktaΔ på™haΔ. Only in the eight cases listed under section A-III, however, does Haradatta state outright that a particular reading is wrong (apapå™ha). When he encountered other variant or difficult readings, Haradatta employs a variety of strategies. When he records the common readings which were at variance with what he himself preferred (see A-I), he gives no explanation four times, calls a reading either “erroneous” or “vedic” three times and simply “vedic” once. In other variants listed under section A-II, Haradatta gives no explanation twelve times and only once calls a reading “vedic”. It is in his treatment of difficult or ungrammatical readings where he found no variants that Haradatta resorts to the “vedic” explanation most frequently. Of the 58 cases noted in section B, he calls 37 “vedic”, 6 “vedic” or “erroneous”, provides explanations for 11, calls 3 “accidental” or “erroneous”, and leaves one unexplained. Without the benefit of historical linguistics, this is the best Haradatta could do. Some of the difficult forms can be better explained as dialectical variants, an area of Sanskrit studies that has

320

LANGUAGE, TEXTS AND SOCIETY

seen spectacular growth in recent years.47 Haradatta’s faithfulness to the received text, however, has preserved these unusual forms for such scholarly scrutiny. In researching the transmission of Åpastamba’s text, did Haradatta use manuscripts or oral transmission among experts or within vedic †åkhås? This is a question that is significant not only for this text but also for the broader issue of textual transmission in medieval India. The term used for “variant reading” is på™ha. Although this term directly indicates an oral reading, like its English counterpart, the term acquired a broader meaning including a “reading” in written form in a manuscript. So, when Haradatta says adhyåyed iti pråyeña pa™hanti (see A: 1), it is not clear whether he is referring to oral recitation or to manuscripts. Haradatta’s use of the verb adhi√i to refer to variant readings, however, appear to suggest that at least sometimes he is referring to oral transmission. Thus, at Åp 1.31.19 he says about the usual reading adhœyamånas tu pramåda† chåndaso vå (see A: 3). Likewise, at Åp 2.5.2 we have pråyeña makåråt param ikåram adhœyate (see A: 4). The use of eke pa™hanti (A: 25), †iß™åΔ na pa™hanti (A: 6), and udœcyås pråyaña na pa™hanti (A: 31) with reference to variant readings also suggest orality.48 Taking all this into consideration, it appears that Haradatta gathered his readings at least partly from the oral traditions found either among †åÂstric experts or in vedic †åkhås. This is the conclusion drawn also by Bhler (1892, vii-viii): “Haradatta certainly consulted for particularly difficult passages men acquainted with oral tradition of the Åpastambœyas and able to recite their Sütras [...] No ingenuity of interpretation can convert ‡iß™as into palm-leaves.” Critical edition of texts is a modern concept, and certainly Haradatta was not attempting such a project. Yet he appears to share the major goal of critical editions–to present the best possible reconstruction of the text, including the preservation of difficult and variant readings. Haradatta’s aim was probably to help his readers sort out the different readings that may have been present in both manuscripts and in oral traditions and to establish an authoritative text by explaining the difficulties he found in the received text. But in doing so he has also preserved for future scholars invaluable data regarding the state of Åpastamba’s text around 1100 C.E.

47

See, for example, Salomon 1981, 1991, 1997; Witzel 1987, 1989; Deshpande 1993; Cardona 1997; Norman 1997. 48 The last example raises the question as to how Haradatta knew that the sütra in question was not found among the northerners. Aside from the memory of experts from the north whom Haradatta could have consulted, the only other way he could have come to this conclusion was by knowing that certain manuscripts were from the north, either because they were located there or were written in the northern script, such as Någarœ or ‡åradå.

321

XV Hair and Society: Social Significance of Hair in South Asian Traditions*

The human body has become in recent years the subject of renewed interest across a spectrum of disciplines, from sociology to literary theory. Approaches to its study vary, of course, with each discipline. Since the groundbreaking study “Techniques of the Body” by Marcel Mauss (1935), however, an underlying assumption in the human sciences has been that the human body is not merely a physical and biological reality confronting human consciousness as an external and independent entity, but primarily a cultural construct carrying social and cultural meanings and messages. Attention has also been drawn by many sociologists and social-anthropologists to a central dimension of the cultural construction of the body: the human body stands as the primary symbol of the social body, or the body politic (Turner 1984). Mary Douglas posits the interrelationship between the two types of bodies in clear terms: The social body constrains the way the physical body is perceived. The physical experience of the body, always modified by the social categories through which it is known, sustains a particular view of society. There is a continual exchange of meanings between the two kinds of bodily experience so that each reinforces the categories of the other. (Douglas 1982: 65)

Berger and Luckman (1967), furthermore, have drawn our attention to a central dimension of culture: all cultural creations, including the human body, have a dialectic nature. On the one hand, it is a human product, is nothing but a human product, and is continuously changed and recreated by human activity. On the other hand, cul* Originally published in Hair: Its Power and Meaning in Asian Cultures edited by Alf Hiltebeitel and Barbara D. Miller, pp. 11-49. New York: State University of New York Press.

322

LANGUAGE, TEXTS AND SOCIETY

ture stands against the individual as a reality that imposes its own logic on individual consciousness, even though cultural grammars, just as those of languages, are very elastic, and individuals continuously change them in the very process of using them. This study deals with just one aspect of the cultural creation of the body—the symbolic use of hair, especially the hair of the head and face— within the cultural history of just one region—South Asia.1 Yet, attention to the dialectic nature of this symbol is essential to my approach. Just like language, hair symbolism imposes its own grammar on the individuals in a given period of a given society; an individual is unable to produce an entirely new symbolic value of hair from his or her own subjective consciousness and still be able to communicate with the rest of that society. Hence, we can justifiably seek to understand the grammar of this symbol. On the other hand, being a cultural product, the grammar of this symbol is not rigid; it is elastic and subject to diverse individual appropriations and uses. Such individual uses will, over time, change the very grammar of the symbol. We should, therefore, also seek to understand how that grammar may have changed over time.2 The comparison with language is instructive. Although English imposes itself on my will, and I am not free to use English in any way I want and still expect to be understood, yet my own usage will change the very language that imposes its rules on me. The point I want to make is that searching for the underlying grammar of hair symbolism, as I will do in this study, does not imply some form of social determinism. I will examine some of the ways hair is used as a public symbol to communicate a variety of socially significant meanings –in a special way, to demarcate the interstices within the complex South Asian societies, to mark their internal boundaries. But my interest in hair symbolism goes beyond the merely descriptive. I want to find out some of the reasons why humans, especially South Asians, have placed and continue to place so much value and significance on hair? To twist Levi-Strauss’s expression, why have humans found hair something so good to think and to communicate with? What patterns emerge from this “thinking with hair” and how do they relate to broader issues of individual and social existence?3 1 My principal focus will be on traditional India, although I will comment on some modern practices in several South Asian societies. 2 This diachronic and historical aspect of hair symbolism is often ignored by anthropologists and psychoanalysts. Obeyesekere (1990: 39) notes the need for uncovering the historical/genetic etymology of a symbol. Truly diachronic study of a symbol within a given culture can advance the study of its etymology, just as much as psychoanalytic investigations. Such a historical study (with a keen sense of humor) of the beard in Europe has been done by Reynolds (1949). 3 I am not searching for a single and universally applicable reason, because the reasons are multiple even within the same culture. Further, I am not searching for a single meaning of all hair practices or even of a single practice. I am convinced that this is a complex symbol with a spectrum of meanings and significances both public and private, both conscious and, often, unconscious (Obeyesekere 1981).

Hair and Society

323

Symbols, like words, do not operate in isolation but within a web of relations and oppositions to other symbols. This context within which alone a symbol can be adequately understood, is what I would call the “grammar of a symbol.” To understand why, for instance, a Sikh male is required to wear his hair long and to cover it with a turban, we need to examine the other customs that demarcate a Sikh male from Sikh females and from non-Sikhs, as well as the historical context in which the custom was created (Uberoi 1967). It is this symbolic grammar –in both its synchronic and diachronic dimensions– of Sikh maleness that will provide us with the context for teasing out the social meanings of Sikh hair. Our starting point, however, must include (1) as full and accurate a description as possible of the customs, practices, and rituals concerning hair found within South Asian culture and history, (2) an examination of the broader symbolic grammar within which these practices are located, and (3) a study of the explanations and exegeses of these practices offered by native sources, whether they be informants or texts. Given the limits of space and ability, it will be presumptuous of me even to contemplate the completion of such an enormous task. I will therefore limit myself to describing in greater detail some ritual uses of hair and to drawing attention in passing to others, in the hope that we will have sufficient evidence to tease out some broader social meanings of hair symbolism in South Asia.

HAIR PRACTICES With the help of Table 1, I will explore first the broad spectrum of South Asian rites, customs, and institutions involving some form of hair manipulation. This chart clearly does not contain an exhaustive listing; hair manipulation, especially shaving, pops up in the most unexpected of places. Alter (1992: 322), for example, has drawn our attention to a rather unusual group, the wrestlers: “Like some sannyasis, wrestlers shave their heads completely or at least have their hair cut very short [...]. Sannyasis and wrestlers alike are distinguished from other men by their radical attitude toward hair as a symbol of identity.” In India symbolic manipulations of hair appear as variations of three central themes: (1) the groomed control of hair, (2) shaving the hair of the head (in the case of adult males this involves also the shaving of the beard), and (3) the neglect of hair resulting in either loose unkempt hair or dirty matted hair, often accompanied by the neglect of nails, and, in the case of males, of the beard. Without denying the possibility of personal meanings –which, after all, are only to be expected, given the dialectic nature of cultural products– all these types of hair manipulation, I hope to show, communicate deeply social meanings, placing the individual whose hair is so manipulated in different relationships both to the broader society and to the segment of that society to which that individual belongs.

324

LANGUAGE, TEXTS AND SOCIETY

PHYSICAL SEPARATION

SOCIETY

RITUAL SEPARATION

Matted Hair and Beard

Controlled Hair

Uncontrolled Hair

Forest hermit Exile of the aged Political exiles

ADULT MALE

Pollution Separation

Hair, beard, nails left to grow without any grooming.

Hair groomed: cut or long. —arranged close to head and/or covered by turban

Women in mourning Menstruating women Vows of Vengence

ADULT FEMALE e.g. Draupadœ and Cåñakya Hair groomed: always long. —arranged either close to head or braided and left hanging; —never covered by turban; but may be covered in other ways

Shaving TEMPORARY SEPARATION Initiatory Separation

____________________________ ____________________________ LONG-HAIR ASCETICS Avdhüta ‡aivite ascetics Sri Lankan female/male ascetics

____________________________ ____________________________

First Hair Cut (caula) ____________________________ First Beard/Hair Cut (gondåna) ____________________________ Vedic Initiation (upanayana) Sacrificial Consecration SIKH MALE Royal Consecration Pilgrimage Unshaven head and beard hair enclosed in turban. Reintegration into Society

____________________________ ____________________________

Oucastes Lepers End of studentship King after Consecration Other impure people Pollution Separation Mourning son Penetential Separation Prior to vows and penances PERMANENT SEPARATION Pollution Separation Widow Corpse Penal Separation People guilty of major crimes Ascetic Separation Hindu/Buddhist/Jain Ascetic —both male and female [Wrestlers]

Hair and Society

325

Groomed Hair Groomed and controlled hair is the hallmark of people with publicly recognized roles within society, in a special way of adult males and females (Hallpike 1969). I believe that the “controlled social hair” of such individuals, especially of married males (given the patriarchal nature of traditional South Asian societies), is the point of reference of most–although not necessarily of all–other hair manipulations from which they derive their meaning and significance. This is simply to say that society is the ultimate point of reference even in its critique and rejection. The hair of an adult male in modern India is usually short and combed and his beard and mustache shaved or trimmed. The medical treatise of ‡u†ruta advises a man to keep his hair and nails trimmed, and to oil and comb his hair regularly.4 The latter advice, as well as trimming the nails,5 clearly applies to women as well. Customs regarding male hair may have varied according to caste.6 Brahmin men, for example, were expected to shave their heads but leave a tuft of hair, the topknot, unshaved. This topknot was generally kept tied in a knot when a Brahmin appeared in public.7 The hair of an adult female, especially a married woman, is long but restrained by a knot, by one or several braids, or by some other means; some women may even cover their hair, especially when they appear in public. The distinctive ways in which hair is worn by adult males and females clearly symbolize their different gender roles. Even though short hair appears to be distinctive of the male in modern South Asia, the picture is less clear in ancient and medieval periods.8 Both males and females are depicted in Indian art and sculpture, for example, with long hair but with distinctive coiffures (Padma 1991). It will be an interesting study to detect gender differences in these modes of coiffure, but one distinctive element is the long braided hair 4 Su†ruta Sa∫hitå, Cikitsåsthåna, Ch. 24. 29, 73-75, 89. Manu (Manusmr¢ti 4.35), likewise, advises a young adult who has completed his vedic studies and is about to get married (snåtaka) to keep his hair, beard, and nails trimmed (k¬ptake†anakha†ma†ruΔ). 5 Most sources regard hair and nails as a pair. When hair is cut, as we shall see, so are the nails, and when hair is left unattended, nails are also left to grow. 6 The Vasiß™ha Dharmasütra (2.21), for example, states that people should wear their hair according to the custom of their castes. 7 An ancient text states: “He shall not untie his topknot on a road” —na pathi †ikhå∫ visr¢jet. Hårœta quoted in Aparåditya’s commentary on the Yåj∞avalkyasmr¢ti (Ånandå†rama Edition), p. 225. On the practice of keeping a certain number of topknots to indicate one’s lineage, see Kane 1963-75, II, 263-64. 8 Even in modern times long haired men were a common sight in Sri Lankan villages until quite recently; the practice has not completely disappeared even today. Sikh men, of course, do not cut their hair or beard. The length of male and female hair is subject to cultural determination. In Tikopia, for example, women’s hair is short, while men’s hair is long (Firth 1973: 272). It is thus not possible to make a universal symbol out of the customs of modern South Asia and Europe, as Hallpike (1969) appears to do. Nevertheless, within each society gender difference is expressed in the prescription of gender specific coiffure.

326

LANGUAGE, TEXTS AND SOCIETY

of females.9 What is common to both genders, however, is that their hair is groomed and controlled.

Loose Hair The groomed control of hair is especially demanded when people present themselves in public. Thus, when a person appears in public with loose and uncontrolled hair, it carries a variety of meanings and messages. Untying the hair before a king, for example, is regarded in the legal literature as an insult subject to punishment (Lingat 1973: 239). On the other hand, legal authorities uniformly affirm that a thief should run to the king, with his hair loose (muktake†a) and carrying a club on his shoulder, to confess his crime.10 Here loose hair in the presence of the king appears to indicate the thief’s recognition of his status as a sinner and an outlaw removed from the bounds of society. Loose hair, especially of women, is a sign of domestic informality and even of sexual intimacy. In sculpture, for example, erotic couples are depicted with loose and falling hair (Padma 1991, 266-67). In iconography disheveled and flying hair may indicate the demonic and the female outside of male control, as in representations of Kålœ.11 Marglin (1985: 54) observes: “The single goddesses are often represented iconographically with loose flowing hair, which signals their celibate state.” Indeed, it may signal even more their liminal and dangerous status. An early medieval ascetic text, for example, warns mendicants not to beg from a muktake†inœ (“a woman with loose hair”), a term which could indicate either that there is sexual intimacy or, as we shall presently see, that she is having her monthly period.12 What is clear, however, is that males and in a special way females are not expected to present themselves in public with loose and ungroomed hair. If they do, their actions carry publicly recognized meanings—they are making a public statement about their social status. Loose and especially disheveled hair is associated in a special way with temporary ritual separations from society. The most common instance of such a separation is that of women during their menstrual period, when the hair is left unbraided and unwashed. As Hershman (1974: 278), in his detailed study of hair among the 9 On the distinctive patterns of braiding called ekaveñi (single braid or clasping all the hair once in the back and letting the rest fall loose), triveñi (triple braid), and the like, see Hiltebeitel 1981, 184-86. On the way hair is worn by contemporary women in Punjab, see Hershman 1974. 10 See Manusm®ti, 8.314; Åpastamba Dharmasütra, 1.25.4; Gautama Dharmasütra, 12.43; Baudhåyana Dharmasütra, 2.1.16; Vasiß™ha Dharmasütra, 20.41. The king beats the thief over the head with the club. Whether he dies or survives the blow, the thief is freed from his crime and sin. 11 See Kinsley 1975: 81-159; Hiltebeitel 1981: 206. 12 Yådavaprakå†a, Yatidharmasamuccaya, ed. P. Olivelle (Albany: State University of New York Press, 1994), 6.145.

Hair and Society

327

Punjabis, has shown, the expression “I have to wash my head,” is used euphemistically by even contemporary women to indicate the onset of their menstrual period. Their ritual separation makes them untouchables; no social intercourse with them is permitted, including touching. Menstruating women do not cook or even sit with the rest of the household to eat. Loose and disheveled hair of women, but sometimes also of men,13 is also a sign of mourning, another ritually impure state when normal ritual activities and social relations are suspended.14 Loose hair on all these occasions of ritual separation tells the world “I cannot be approached.” There are also two prominent cases in Indian literature where the hair is left loose until a vow of vengeance has been fulfilled. Literary sources depict Draupadœ, the wife of the Påñ∂ava brothers in the Mahåbhårata, as leaving her hair loose after she was insulted in public by the Kauravas until their final defeat and death (Hiltebeitel 1981). Her hair was, in fact, already loose when the outrage occurred, because she was then having her period. Cåñakya, the prime minister of Candragupta Maurya, provides the other example. Vi†åkhadatta, in his Sanskrit play Mudråråkßasa (Act 1, verse 9), depicts him as keeping his Brahmanical top knot untied until he had fulfilled his vow of placing Candragupta securely on the throne and vanquishing completely the dynasty of the Nandas. In both these cases the vow of vengeance suspends the normal social roles of Draupadœ and Cåñakya until the completion of their vows, a feature such vows share with penitential vows I examine below that require the shaving of the head. People who display loose and uncontrolled hair in public, therefore, appear to have temporarily suspended –for a variety of reasons and with a variety of consequences– their normally assigned roles in society.

Shaving Clearly the most common and possibly the most significant manipulation of hair in South Asian societies is the shaving of the head.15 It occurs so frequently in ritual settings that space does not permit me to fully explore individual instances. Instead, I derive the symbolic grammar of shaving from a brief survey of the broad spectrum and a closer examination of a few of the rituals of shaving. The dominant social 13 The loose topknot of a Brahmin, for example, may be a sign of mourning: ņvalåyana G®hyasütra, 4.2.9; ‡rautako†a, English Section, ed. R. N. Dandekar (Poona: Vaidika Sa∫†odhana MañÛala, 1958-73), I.2, p. 1079. The disheveled hair of a woman in mourning is recorded already in the Atharvaveda, 9.9.7. Some sources instruct the mourners to ruffle their hair and to put dust on it: Åpastamba Dharmasütra, 2.15.7; ‡rautako†a, I.2, p. 1052. 14 The relatives of a person excommunicated from the caste are also expected to let their hair hang loose during the rite: Vasiß™ha Dharmasütra, 15.13. 15 For a study of cross-cultural symbolism of shaving, see Firth 1973: 287-91.

328

LANGUAGE, TEXTS AND SOCIETY

meaning of shaving in South Asian traditions that emerges from this examination is the separation of the shaven individual from society, a separation certainly more profound and often more permanent than that signaled by loose hair. For heuristic purposes I distinguish different types of social separation signaled by shaving, in terms of both duration and the type of separation involved.

Temporary Separation Initiatory Separation: The most common temporary separation occurs in initiation ceremonies. Since van Gennep’s (1960) groundbreaking work Rites de Passage, initiation rites are commonly recognized as having three moments: separation, liminality, and integration. The initiate is first ritually separated from society and from his or her social role and rank and left in an ill-defined marginal state. The initiatory rite concludes with the reintegration of the initiate into his or her new status within society. In South Asian traditions almost every initiatory separation is accompanied and signaled by the ritual shaving of the initiate. When a young boy undergoes vedic initiation (upanayana), when a sacrificer is consecrated (dœkßå) prior to his performing a vedic sacrifice,16 when a king is anointed (abhißeka) –at all these initiatory rites the subject is first shaved. Indeed, these ceremonies are presented expressly as new births of the individuals (Gonda 1965b, 331). Many explicit statements and symbolic enactments of the initiates’ return to the womb are found in these ceremonies:17 “The sacrificial priests make into an embryo again the man whom they prepare for the sacrificial consecration (Aitareya Bråhmaña 1.3.1). Shaving of the initiate clearly belongs to the same symbolic grammar. Shaving reduces the individual to the state of an embryo or an infant –the asexual and hairless condition. The first cutting of a child’s hair (caula or cü∂åkaraña) is also a rite (sa∫skåra) that marks a transition. The ceremony is performed generally when the child is about three years old. The mantra accompanying the shaving states that the shaving is intended to assure a long life. The connection with fertility is implicit in the places where the cut hair is buried, for example, in a cow pen or at the foot of an Udumbara tree.18 Another life-cycle rite is connected with the first 16 The Taittirœya Bråhmaña (1.5.6.1-2) provides the interesting detail that at a sacrifice the shaving should be done the “divine” way: one first shaves the hair of the armpits, then the beard, and then the head. The demons (Asuras), on the other hand, did it the opposite way, which was the reason for their defeat at the hands of the gods. The ‡atapatha Bråhmaña (3.1.2.3) alludes to a custom of shaving the entire body of a man at his sacrificial consecration. 17 For a detailed examination of the symbols of rebirth, See Gonda 1965b: 284-462, especially 337. 18 For a study of the connection between hair and fertility, see Lincoln 1977. At least

Hair and Society

329

shaving of the beard (ke†ånta, also called godåna, at which the head is also shaved), performed at age sixteen and associated with sexual maturity (Pandey 1969, 143-45). The final rite of passage in the Hindu liturgy is the funeral. It is regarded by the Brahmanical tradition as the last sacrifice of the deceased at which his own body becomes the victim offered in the cremation fire. Here too the individual is reborn through the sacrifice. Now, according to most ritual texts the head or even the entire body of the corpse is shaved prior to cremation.19 I believe that initiatory shaving, especially the shaving of the boy at his vedic initiation and of the sacrificer at his consecration, was paradigmatic and influenced the ritual articulation of most ritual separations in South Asian societies. Initiatory Reintegration: The other side of the coin of initiatory separation is the reintegration into society of people who have been separated for a considerable period of time due to a variety of factors. Outcasts, the polluted, students at the completion of their period of study, the king after his year-long seclusion following his consecration –all are reintegrated into their respective social ranks through ceremonies that feature the ritual shaving of the head. During the yearlong separation the king leaves his hair uncut and ungroomed; he does not bathe and sleeps in the shed where the sacred fire is kept. The student, on the other hand, either lets his hair grow into a matted condition (ja™ila) or shaves the head but keeps his topknot unshaved.20 In these ceremonies of reintegrating people after protracted periods of social separation, shaving appears to mark the conclusion of that period –a kind of separation from their liminal state– and their assimilation into their new social roles. Noteworthy is the absence of shaving during the marriage rite, which is the most central life cycle ritual within the Brahmanical system. The reason for this absence is unclear, but it appears that for the adult male the marriage ceremony is the final act of a process that starts at the conclusion of his vedic studies. The final bath and the other ceremonies associated with his return from his teacher’s house remove him from the ascetic, celibate, and mendicant life of a student. Such an individual, technically called snåtaka (“bathed”), is then decorated with garlands, ornaments, and finery. His status is said to be higher than that of a king. It is significant that in the legal literature the provisions for a snåtaka often overlap those for a marin modern times, the hair is often offered to the goddess or put in a river: see Hershman 1974; Freed & Freed 1980: 396-97. 19 For primary sources from the G®hyasütras, see the ‡rautako†a, I.2, pp. 1033, 1039, 1070, 1071, 1074, 1080, 1083. See the ņvalåyana G®hyasütra, 4.2.9, for shaving the whole body. 20 See Manusm®ti 2.219; Åpastamba Dharmasütra, 1.2.31-32; Gautama Dharmasütra, 1.27; Vasiß™ha Dharmasütra, 7.11.

330

LANGUAGE, TEXTS AND SOCIETY

ried householder, who is also often referred to as snåtaka. The interval between the completion of studies and marriage is supposed to be relatively short. We may thus view the shaving prior to the ritual bath at the conclusion of his studentship as the first step and the marriage itself as the final chapter of the reintegration of the student into his new social role and status. Pollution Separation: Social intercourse is forbidden with people who are tainted with ritual pollution. Such people are ritually separated during the period of impurity. Some of these temporary periods of separation, such as those created by the death of a close relative, can also be marked by shaving. A son, for example, is expected to shave his head at the death of his father or mother.21 Penitential Separation: A person undergoing a penance or vow (vrata) also is separated from society, and many of the major penitential practices of Hinduism are preceded by the shaving of the penitent. Some sources give a reason for this practice: sins become lodged in the hair. Thus a person who wishes to expiate sins should shave the hair.22 People also shave when they go to a place of pilgrimage (tœrtha), an act which may be regarded as either an initiatory or a penitential separation from society.23

Permanent Separation Pollution Separation: A permanent ritual separation from society occurs in the case of a widow. The social position of a widow has undergone repeated changes in Indian history. There is at least one period when the ritually impure, inauspicious, and unmarriageable state of a widow was signaled by the shaving of her head. A frequently cited verse states: “The long hair of a widow’s head grow in order to bind her husband. A widow should, therefore, always keep her head shaved.”24 The permanence of this condition, moreover, required 21 For the general behavior expected of people in mourning, see Kane 1962-75, IV, 238ff. See also above, n. 13. 22 See Kane 1962-75, IV, 122 for sources of this belief and for other customs involving shaving when a person performs a vow or a penance. 23 Some sources, such as the Padma Puråña and the Skandha Puråña, make shaving obligatory prior to a pilgrimage. See Kane 1962-75, IV, 573-76 for further details and sources of this practice. A frequently quoted verse states: “One should shave one’s hair at Prayåga, before a pilgrimage, and at the death of the father or mother; one should not shave without cause” (Kane 1962-75, IV, 574, n. 1300). Kane (IV 1962-75, 575) notes that some texts make a distinction between the technical terms kßaura and muñ∂Ûana, the former referring to the shaving of the head and the beard, and the latter indicating the shaving of only the head. Val Daniel (1984: 245-87), in describing the pilgrimage to the shrine of Ayyappan, notes the initiation rite prior to departure, a rite that amounted to the ascetic renunciation of the pilgrims. He does not mention shaving explicitly, but the context strongly suggests it. 24 The K农khañÛ∂a (4.75) of the Skandha Puråña. See Våsudevå†rama, Yatidharmaprakå†a (ed. P. Olivelle; Vienna: 1976-77), 71.96-97; Kane 1962-75, IV, 585.

Hair and Society

331

that she keep her head permanently shaven, and in this and other customs a widow often resembled an ascetic.25 Penal Separation: Major crimes, such as murder, were punished by death, but when the capital punishment was not meted out, as when the criminal happened to be a Brahmin, the criminal was shaved and lived the life of a beggar outside of society.26 Ascetic Separation: The best known ritual shaving associated with permanent separation from society is that of the Hindu sa∫nyåsin or renouncer, the Buddhist and Jain monk, and their female counterparts. A central feature of the rites of initiation into the ascetic life in all these traditions is the removal of head and facial hair. Throughout their life these ascetics keep their head and face clear of hair by periodic shaving.27 Even though, as I will argue, the central social meaning of ascetic shaving, just as the shaving of students, sacrificers, and widows, is that of separation from society, sexual symbolism is not lacking. Not just ascetics, but all people ritually shaven are forbidden to engage in sex. For most this is a temporary condition required by a rite of passage or necessitated by ritual pollution, but for the ascetic (and often also for the widow) it is permanent, and therein lies the difference between ascetic and other forms of ritual shaving. Social control, after all, is primarily sexual control, and the controlled hair of social individuals symbolizes their participation in the socially sanctioned structures for sexual expression, especially marriage (Hallpike 1969). Removal of hair separates the individual from that structure and from the legitimate exercise of sexual activity. Shaving for the ascetic, I believe, indicates his or her removal from socially sanctioned sexual structures, and, a fortiori, also from other types of social structures and roles.28 In the Indian context, this implies loss of caste, inability to own property, and lack of legal standing in a court of law for most purposes (Olivelle 1984: 140-51). 25

See Kane 1962-75, II, 583-98 for the customs and duties relating to a widow. For further details see Kane 1962-75, III, 396-97. Other punishments subject to shaving include adultery and incest with the wife of one’s teacher. Shaving associated with punishment and reduction to slavery is a widespread practice cross-culturally: Firth 1973: 289-90; Hallpike 1987, 155. 27 Hindu renouncers shave at the beginning of each of the five Indian seasons, thus shaving approximately every two months. Periodic shaving (or uprooting) of head hair is a feature also of Buddhist and Jain monks. One rule permits Buddhist monks to keep only one and a half inches of hair on their head: see Cullavagga, 5.2.2-3. 28 This is the opposite of what Eilberg-Schwartz (1990: 145) points out in the case of circumcision: “Since circumcision exposes a boy’s sexual organ, it is also a natural symbol of his readiness for social intercourse. Sexual intercourse, after all, is one of the most powerful symbols of social intercourse.” As Eiberg-Schwartz (1990: 145) himself notes, “when a person is outside or in transition between recognized social positions, sexual intercourse is prohibited.” 26

332

LANGUAGE, TEXTS AND SOCIETY

Elements of the ascetic initiatory ritual also indicate that shaving symbolizes the return to the sexually and socially undifferentiated status of an infant. During the Hindu ritual, for example, the shaven ascetic takes off all his clothes. The naked renouncer is significantly called jåtarüpadhara, which literally means “one who bears the form he had at birth.” The ascetic is not just naked; he is reduced to the condition in which he was born, to the state of a new-born infant. I believe that shaving is part of the symbolic complex that signifies his return to “the form he had at birth.” The absence of hair, just as much as nakedness, takes the initiate back to the prepubertal state of infancy. The sexual symbolism of hair also helps explain some interesting features of ascetic behavior toward hair. It is well known that Jain monks at their initiation and periodically throughout their life remove their head hair by tearing them by the roots, a painful procedure I believe. That this custom was not limited to the Jains is demonstrated by its presence in a somewhat abbreviated form in the Hindu ritual of ascetic initiation. Here the ascetic’s hair is first shaved, but five or seven hairs at the crown are left uncut.29 At the conclusion of the rite, the ascetic plucks these few hairs from the roots. Although one may attribute these practices to the common ascetic propensity to bodily torture and pain, this literal eradication of hair, especially viewed in the light of the broader grammar of ascetic bodily symbols, can be seen as a symbolic and ritual uprooting of sexual drives and attachments. That shaving is the opposite of sexual engagement is also brought out in the head-shaving rites of Hindu ascetics during the annual liturgical cycle. They are not allowed to shave any time they may want. Rather the prescribed time for shaving is at the junctures between the five Indian seasons: spring, summer, rains, autumn, and winter. Now the Sanskrit term for season is r¢tu, the same term that is used to indicate the monthly menstrual cycle of a woman. Brahmanical law and ethics require a husband to engage in sexual intercourse with his wife in her r¢tu, that is, soon after the end of her period when a new “season,” a new fertile period, begins for his wife (Manusmr¢ti 3.45). I think it is not farfetched to see a correspondence between the husband approaching his wife at the beginning of her fertile season (r¢tugamana), and the ascetic shaving his head at the beginning of each calendrical season (r¢tuvapana). The r¢tugamana is thus transformed into r¢tuvapana. This shaving appears to symbolize an ascetic’s renunciation of sex precisely at the time —at least in a terminological sense— when the ethics of society requires a married man to engage in it. Significantly, it is this very need for periodic shaving that is denied in the case of the 29

These hairs represent the Brahmanical topknot. Its uprooting may thus also indicate the ascetic’s abandonment of the ritual religion represented by the topknot and the sacrificial string, both of which are abandoned by some types of Brahmanical ascetics. See the Yatidharmaprakå†a, ed. cit., 21.39, 104.

Hair and Society

333

Buddha. According to a Jåtaka account, the Buddha cut his hair with his sword to the length of two fingers-breath. His hair remained the same length for the rest of his life, signifying, it would seem, the total extinguishing (nirvåña) of his sexual fires.30 A closer examination of the three institutions involving either permanent or extended periods of separation from society marked by head-hair shaving–the vedic student, the widow, and the world renouncer–indicates their structural similarity. Indeed, the Brahmanical legal literature frequently brackets these three institutions together because many legal provisions are common to all three.31 They share similar characteristics: all are shaven-headed, all are forbidden to have sexual relations, all receive their food from others, all are expected not to adorn themselves or to participate in amusements, and all have a marginal legal status–they do not own property, for example, and are not permitted to enter into contracts or to take part in legal proceedings, such as being a witness or a surety in a court of law. They lead a penitential life, sleeping on the floor, not chewing betel, not anointing their bodies, and eating little. Students are reduced to the level of servants of their teachers. Both students and ascetics move out of their homes and are reduced to the status of beggars; neither is affected by pollution at the death of a relative.32 Indeed, when we look at the other prolonged states of separation signaled by shaving, we detect many of these same features. I want to focus here especially on food. People who are either in a permanent or a prolonged state of ritual separation, including ascetics, vedic students, widows, and criminals, do not own food; they have to obtain their food from people in society.33 During shorter periods of separation, people either fast or eat food cooked and given to them by people within society. There is a parallel between the restrictions with regard to food and sex, both being derived from their removal from social structures and roles. I want, however, to highlight one aspect of the food habits of shaven individuals. Apart from caste endogamy, the most distinctive feature of Hindu society consists of dietary restrictions that limit the exchange 30 See Zimmer 1962: 160. See H. C. Warren, Buddhism in Translations (New York: Atheneum, 1962), p. 66. The belief that the hair and nails of a liberated individual (kevalin) do not grow is found also in the Jain tradition (Dundas 1985: 179 and n. 142). 31 One source, for example, states: “Ascetics, vedic students, and widows should abstain from chewing betel leaves, from anointing their bodies with oil, and from eating out of brass plates.” Yådava Prakå†a, Yatidharmasamuccaya, ed. cit., 7.140. For the legal status of a renouncer, see Olivelle 1984, and for a discussion of the similarities and differences between a widow and an ascetic, see Leslie 1991. 32 See Kane 1962-75, IV, 298, where other instances are given where ascetics and students are treated alike with reference to periods of impurity. With regard to food, the ascetic resembles a man consecrated for a vedic sacrifice (dœkßita): the food of both causes impurity and are not to be eaten by others. On the ascribed purity of begged food, see Yådavaprakå†a, Yatidharmasamuccaya, ed. cit., 6.95, 109-10. 33 On the symbolism of food in the ascetic traditions, see Olivelle 1991.

334

LANGUAGE, TEXTS AND SOCIETY

of food across caste boundaries. The purest food is your own food. Those who are ritually separated from society, however, live in a liminal state defined by the absence of boundaries. This absence is symbolized, I believe, by their acceptance of food from others –in the case of ritual beggars, from people whose level of purity cannot be easily determined; hence the legal fiction that begged food (bhikßå) is always pure. Begging is the paradigmatic opposite of the restrictive laws of food exchange. The one establishes and reinforces social boundaries; the other symbolizes the lack, or, as in the case of ascetic ideology, the transcendence, of such boundaries. Significantly, Hindu law forbids householders from accepting cooked food from other people.34 The food of shaven people, likewise, become unfit for others: people are instructed, as we have noted, not to touch the food of ascetics and people consecrated for a sacrifice (dœkßita). Shaven individuals of widely different sorts, from ascetics to criminals, are excluded from the two central institutional spheres of society: the sexual and the economic. Celibacy and mendicancy are the results of the separation of shaven individuals from social structures, whatever the cause and motive of that separation.

Neglected Hair Finally, we have a unique manipulation of hair by refusing to manipulate it at all –that is, the utter neglect of hair. The most common instance of neglected hair is the so-called matted hair (ja™å) associated with forest hermits.35 At least in its early history, neglected and matted hair symbolized ideally and typically an individual’s physical separation from society and civilized living, even though there are instances when the ja™å is recommended for other individuals separated from social living but not necessarily from social geography, such as vedic students (see above n. 20). To understand the symbolism of matted hair it is necessary to locate it within the larger grammar of the symbols associated with physical withdrawal from society in ancient India. Besides long and matted hair, bodily symbols of forest living included a long and uncut beard in the case of males, long and uncut nails, eating only uncultivated forest produce, clothes of tree bark or animal skin, and frequently also bodily uncleanliness.36 People with matted hair are 34 “By giving cooked food to a householder one goes to the Raurava hell.” Våsudevå†rama, Yatidharmaprakå†a, ed. P. Olivelle (Vienna, 1976-77), 68.62-63. 35 Matted hair is generally caused by the neglect of hair. I have found one instance, however, where the matted coiffure is artificially created using the juice of a banyan tree: Råmåyaña, 2.46.55-56. For a study of modern examples from Sri Lanka, see Obeyesekere 1981; the subjects of his study uniformly viewed the appearance of matted locks as sudden and interpreted it as a divine gift. 36 See Gautama Dharmasütra 3.26-35; Åpastamba Dharmasütra, 2.22.1-11; Baudhåyana Dharmasütra, 2.11.5; 3.3.1-22; Vasiß™ha Dharmasütra, 9.1-12. A significant incident in the epic

Hair and Society

335

required to live in the forest or wilderness; they are repeatedly admonished “not to step on plowed land,” the prime symbol of civilized geography.37 They are often said to imitate the habits of wild animals. One can decipher from this symbolic grammar the following statement: a matted-hair individual withdraws from all culturally mediated products and institutions and from all culturally demarcated geographical areas and returns to the state of nature, the condition of the wild, to the way of life of animals. Not grooming the hair, not controlling it in any way, letting it grow naturally into a wild and matted condition–all this appear to symbolize a person’s total and absolute withdrawal from social structures and controls and from human culture as such.38 In Indian history we can identify at least three distinct types of matted-hair people who have withdrawn or have been forced to withdraw from society. First, there are the forest hermits called Vaikhånasa or Vånaprastha, and second, the aged. Old people, especially old kings, both within the Hindu institution of the four orders of life (å†ramas) and outside that structure, were expected to leave their family and society and assume a forest mode of life.39 These two classes –the hermits and the retirees– are often collapsed into a single category in Indian legal literature. The third class consists of political exiles. The epic heroes of the Mahåbhårata and the Råmåyaña, the five Påñ∂ava brothers and Råma, for example, are all sent into political exile. Significantly, political exiles assume the bodily symbols and the mode of life of forest hermits, just as criminals assume the life style of shaven ascetics. People and groups viewed by the mainstream of society as standing outside social boundaries, such as the tribal and forest peoples of India and the Sri Lankan Veddas, are also depicted in art and popular imagination with long and unkempt hair. story of the Mahåbhårata relates to the ugliness and smell of an ascetic. While accepting the invitation to father a son for his deceased brother, Vyåsa tells his brother’s widow that she should “bear with my ugliness. If she bears with my smell, my looks, my garb, and my body, Kau†alyå shall straightway conceive a superior child” (MBh 1.99.42-43; trans. of van Buitenen). His odor and sight are so overwhelming, however, that the woman is forced to close her eyes when he comes to her bed, resulting in the blindness of her son, Dh®taråß™ra. 37 See Gautama Dharmasütra, 3.26-35; Baudhåyana Dharmasütra, 2.11.15. 38 A man with matted hair (ja™ila) is among those forbidden at a funerary offering (†råddha): Manusm®ti 3.151. There is nothing in this symbolic structure to suggest the unrestrained sexuality proposed by Leach 1958: 154. The operational meaning of neglected and long hair in this context is far removed from what I call its “root” meaning (see below). It is, however, clear that even though the operational meaning is about withdrawal from society, the rules of life of hermits are very ambivalent regarding their sexuality, some prescribing celibacy and others suggesting a non-celibate life style. Celibacy is not a hallmark of matted-hair ascetics in quite the same way as it is of shaven-headed ascetics. 39 See Olivelle 1993; Sprockhoff 1979, 1981, 1984, 1991. Both the Råmåyaña (2.20.21) and the Mahåbhårata (3.186.2-3) ascribe the institution of this practice to ancient seers or to the very first king. Buddhist sources ascribe its foundation to the universal emperor Da¬hanemi (Dœgha Nikåya, III, 60-64) or to King Makhådeva of Mithilå, who was the Buddha in one of his former births (Majjhima Nikåya, II, 75-82).

336

LANGUAGE, TEXTS AND SOCIETY

Exceptions to Ritual and Physical Separation In concluding this overview of hair manipulation, I want to examine briefly some instances of long hair that do not fit into the scheme I have outlined. Although I call these ‘exceptions’ for the sake of convenience, they illustrate rather the extremely loose nature of any cultural grammar. They demonstrate the ease with which individuals and groups can cross from one symbolic domain to another. The first example is the Sikh male.40 He is not permitted to cut any of his hair –head hair or beard– from birth until death. An adult Sikh male is distinguished by his long hair and beard. He is, however, married and part of the social fabric. To understand the hair symbolism of Sikh males we have to locate it within the historical context –North India between the fifteenth and the eighteenth century– that gave birth to the Sikh tradition. Coming from the background of devotional religion (bhakti) and saints (sant) of northern India, the early Sikh gurus deliberately drew a contrast between the Sikh bodily symbols and those of both the traditional Hindu renouncer with his shaven head and the Moslem with his circumcised penis. The Sikh holy man, by contrast, has long hair, is married, and is uncircumcised. There is a structural inversion between Hindu renouncers and Sikhs. The Sikh householder, the representative of a new form of holiness, stands, in structural opposition to two Hindu institutions: the ordinary householder because the Sikh is a holy man, and the renouncer because the Sikh affirms holiness within marriage and society. This dual opposition is symbolized on the one hand by the Sikh long hair and on the other by the turban that encloses and controls the hair and by the wellgroomed and waxed beard. An interesting historical point is that the Sikh long hair is itself a symbol borrowed from another and by then obsolete form of separation from society, the uncut hair and beard of the forest hermit. Yet, as part of the social fabric, the uncut hair of the Sikh is not neglected; it is washed, oiled, combed, and enclosed within a turban –a traditional way of hair control in India. As Uberoi (1967: 96) has pointed out, the symbolism of the five k’s expresses power and its control. As the comb controls the power of the hair, the steel bangle controls the power of the sword, and the underwear (kach) the power of the uncircumcised penis. As the Sikh male crosses the boundary between society and ritual separation from it, so the long matted hair of certain types of Hindu ascetics living within society stands at the boundary between ritual and physical separation from society. There were and are a variety of such ascetics, including those of some ‡aivaite sects and those known in the ascetic literature as Avadhütas. We must include within this class the Sri Lankan male and female matted hair ascetics described 40

For extensive studies of hair among the Sikhs, see Uberoi 1967 and Hershman 1974.

Hair and Society

337

by Gananath Obeyesekere (1981) in his Medusa’s Hair. The literature makes it clear that the Avadhütas, and possibly also other ‡aivaite ascetics, let their hair grow and become matted to symbolize their liberated status and freedom from normal ascetic rules. Thus they let their hair grow to show their transcendence of the shaven-headed state of the ordinary ascetic. They, like the Buddha, have no need to periodically shave their head. The Avadhüta symbol of matted hair, therefore, should be understood not in relation to the adult male in society but in relation to the mainstream asceticism of shaven monks, an institution these ascetics claim to have transcended.

A SEARCH FOR MEANINGS The Native Exegesis It would be equally naive to limit scholarly investigations to the meanings assigned to rites by the actors themselves or by the native tradition and to ignore the meanings and interpretations offered by that tradition. Both the rites and the indigenous interpretations of the rites constitute the data that the scholar must take into account. As with most condensed and central symbols of a society, indigenous exegesis of hair is neither extensive nor frequent. The ascetic literature, for example, never tells us why an ascetic must shave his head. This block in native exegesis, which generally waxes eloquent in most other areas of ritual practice, suggests some level of discomfort in dealing explicitly with this symbol.41 One aspect of hair that stands out clearly in native exegesis, however, is its impurity.42 Most ancient Indian sources require that people throw away any food contaminated by hair.43 Hair in this sense is equal to excrement. Seneviratne (1992: 181) refers to a Sinhala belief that hair and nails are made from the impure waste produced in the process of digesting food, and that cutting hair and nails is similar to voiding excrement. Ancient vedic texts share this belief; the reason for shaving before a ritual is to remove the impurity of hair and nails from the body. The ‡atapatha Bråhmaña (3.1.2.2), for example, explains why a sacrificer must shave before his consecration: 41 On the possible reasons for such blocks in native reflections on rites and symbols, see Obeyesekere 1990: 43. 42 For an examination of this attitude toward cut or fallen hair crossculturally, see Leach 1958: 156-57, and Firth 1973: 287. Hershman (1974: 292-94) objects, I think wrongly, to Leach’s and Douglas’s (1966) cross-cultural study of “dirt” as applied to hair. 43 Åpastamba Dharmasütra, 1.16.23. Yådavaprakå†a (Yatidharmasamuccaya, ed. cit., 10.74) tells an ascetic to discard both his bowl and the food in it when water from the hair or beard falls into it. In a significant juxtaposition, he immediately goes on to say what an ascetic should do if he happens to purge, to vomit, or to void urine or excrement while he is begging. The Mahåbhårata (1.3.126) records the story of one Utaõka, who curses King Paußya because the king gave him food with a hair in it.

338

LANGUAGE, TEXTS AND SOCIETY

He [the sacrificer] then shaves his hair and beard, and cuts his nails. For impure, indeed, is that part of man where water does not reach him. Now at the hair and beard, and at the nails the water does not reach him: hence when he shaves his hair and beard, and cuts his nails, he does so in order that he may becoming pure before he is consecrated.44

According to this interpretation, hair and nails are impure because they do not absorb water, the ultimate means of purification, whereas according to the Taittirœya Sa∫hitå (6.1.1.1-2) they are impure because they are dead skin. A significant and informative contradiction within the native tradition occurs, however, when what is said to be equivalent to feces is offered ceremonially to gods and goddesses (Hershman 1974). This happens, as we have seen, especially when young children are shaved for the first time. How can the same substance be regarded as excrement in one ritual setting and as a substance fit for the gods in another? The sexual symbolism of hair that I discuss below may provide one clue. If at some level of its symbolic complex hair represents the fertile sexuality of it owner, then we can see how it can be at the same time both a sacred offering and excrement. Indeed, sexual fluids, especially male semen, are at one time said to be the most refined part of the body and of food, even the carrier of personality from one birth to the next,45 and at other times bracketed with urine and feces as impure substances. A common way to indicate the depravity of a particular act, for example, is to say that if a man does it “he, in fact, offers to his ancestors semen, urine and excrement.”46 Another element of the native exegesis of hair is its frequent connection in both myth and ritual to grass and plants, emphasizing thereby its relationship to fertility. A couple of R¢gvedic verses (8.91.5-6) connects three areas of hair/grass growth: head, pubic region, and fields. In this hymn Apålå, a young girl, prays to Indra to make the hair grow on her father’s head, on her own pubic region, and in her father’s fields.47 The connection between hair 44 The block in native exegesis that I referred to earlier is borne out by the contradictory reasons given in the ‡atapatha Bråhmaña itself. In giving the reason why a king after his consecration remains unshaved for a year, it says: “The reason why he does not shave his hair (is this): that collected essence of waters wherewith he is then sprinkled (anointed) is vigor, and it is the hair (of his head) that it reaches first when he is sprinkled; hence were he to shave his hair, he would cause that glory to fall off from him and would sweep it away: therefore he does not shave his hair” (5.5.3.1; Eggling’s trans.). So one shaves because the water does not reach the hair, and one refrains from shaving because the water reaching it first! 45 The B®hadårañyaka Upanißad, 6.4.1, states: “Earth is the essence of these beings; water is the essence of earth; plants are the essence of water; flowers are the essence of plants; fruits are the essence of flowers; man is the essence of fruits; and semen is the essence of man.” See also Pra†na Upanißad, 1.14. For the semen as the carrier of personality, see B®hadårañyaka Upanißad, 6.2.9-16; Chåndogya Upanißad, 5.4-10. 46 See, for example, Varada’s Yatiliõgasamarthana (ed. in P. Olivelle, Renunciation in Hinduism: A Medieval Debate, Vol. 2; Vienna, 1987) III.43. 47 For an examination of this text, see Schmidt 1987; Vajracharya 1988.

Hair and Society

339

and grass/plants is well established in the vedic literature (Gonda 1985). The Taittirœya Sa∫hitå (7.5.25.1) states quite simply that “vegetation is hair.” Another text records the myth that the hair that fell from the creator god Prajåpati’s body turned into vegetation (‡atapatha Bråhmaña, 7.4.2.11). Other myths connect the creation of grasses and plants to the hair of Vißñu incarnate as a boar (Gonda 1985, 63-64). This correlation is brought out nicely in the Upanißadic verse: As a spider spins out a thread, then draws it in; as plants sprout out from the earth; as on body and head hair grows from a living man; So from the imperishable all things here spring.48

Gonda (1981) has objected to Lincoln’s (1977) claim that haircutting rituals of Brahmins establish a clear association between hair and vegetation. Even though Gonda is right that there was no fixed rule regarding the disposal of ritually cut hair as suggested by Lincoln,49 it is nevertheless clear that shaving the head is related to prolongation of an individual’s life-span and that many elements of the rite, including grass, sesame seed, ghee, and the burial of the hair in a place covered with herbs, posit a clear relationship of hair cutting with fertility. Indeed, Gonda (1956) himself has suggested a similar correlation in another hair ritual, the parting of a pregnant wife’s hair (sœmantonnayana). This correlation is brought out clearly in two riddles and their solutions recorded in the ‡atapatha Bråhmaña (11.4.1.6-7, 14-15). The riddles ask how one knows from an analysis of the full-moon and new-moon sacrifices (1) why people are born with hair, why hair grows a second time on the face, armpits, and other parts of the body, and why in old age the hair of the head first turns gray and finally one becomes gray all over; and (2) why the semen of a boy is not fertile, that of a man in midlife is fertile, and that of an old man is again infertile. Here are the solutions: Inasmuch as he spreads a cover of sacrificial grass (on the altar), therefore, creatures here are born with hair; and inasmuch as he for the second time, as it were, spreads the Prastara-bunch of grass, therefore, for the second time, as it were, the hair of the beard and the armpits, and other parts of the body grow; and inasmuch as at first he only throws the Prastara-bunch after (the oblations into the fire), therefore it is on the head that one first becomes gray; and inasmuch as he then throws after it all the sacrificial grass of the altar, therefore, in the last stage of life, one again becomes gray all over. 48 Muñ∂aka, 1.1.7. The Aitareya Upanißad (1.1.4) makes following sequence in the way the original being (man) gave rise to the creation: first the skin; from skin, the body hairs; and from the hairs, the plants and trees. 49 The list of places cited by Kane (1962-75, II, 263)—near water, under an Udumbara tree, in a bunch of darbha grass, in a wooded area—indicate a clear association with fertility.

340

LANGUAGE, TEXTS AND SOCIETY

And inasmuch as the fore-offerings have ghee for their offering-material, a boy’s seed is not productive, but is like water, for ghee is like water; and inasmuch as, in the middle of the sacrifice, they sacrifice with sour curds and with cake, therefore it is productive in the middle stage of life, for thick-flowing, as it were, is that offering, and thick-flowing, as it were, is seed; and inasmuch as the after-offerings have ghee for their offeringmaterial, it again is not productive in his last stage of life, and is like water, for ghee, indeed, is like water. (Eggeling’s trans. slightly modified)

Here is an interesting and informative juxtaposition of hair growth, ritual use of grass, and fertility of semen. Hair on the head of children produce weak and infertile sexuality. The second birth of hair on the face and body produces fertile semen. The graying of hair produces a second childhood when semen becomes weak and infertile.50

The Social Meanings of Hair Manipulations There is no single and unique meaning to be discovered within this vast range of hair rituals. Further, as we have seen in the case of Sikhs and Sri Lankan ascetics, historical contexts and individual decisions can give new meanings to traditional symbols.51 It is in the very nature of the dialectic character of a cultural product, moreover, that the same fact or act may carry different meanings to different individuals or groups of individuals, creating what Obeyesekere (1981) calls “personal symbols.” Nevertheless, a set of related symbols of a society—in this case the ritual manipulations of hair—cannot exist in total isolation from each other. Just as a word in a language, so a symbol operates within a broader grammar within which alone it becomes meaningful. And just as it is heuristically profitable to search for the root meaning or the etymology of a linguistic symbol, not because it will exhaust the meanings available in actual usage but because such a meaning permits us to discover the relationships among those operational meanings and thereby further our understanding of those very meanings, so also is it useful to search for a root meaning, or a cluster of such meanings, of a symbol such as hair. Such a root meaning will not exhaust the multiplicity and 50 In this context it may also be worthwhile to note that bald-headed men were considered ritually handicapped and bracketed with others with similar disabilities, such as eunuchs, lepers, cripples, and the blind. The Gautama Dharmasütra (15.18) forbids the feeding of a bald man at a †råddha. Bald men also, along with eunuchs, are among those not permitted to become Brahmanical ascetics: Nåradaparivråjaka Upanißad, 136; B®hat-Sa∫nyåsa Upanißad, 251-52 (both in Sa∫nyåsa Upanißads, tr. P. Olivelle; New York: Oxford University Press, 1992). A person who is ke†ava, probably meaning a man with long hair, is said in the ‡atapatha Bråhmaña (5.1.2.14) to be neither a woman, because he is a man, nor a man, because he has long hair. The reference may be to an eunuch. On the relation between hair and sex within the Indian context, see also Hara 1986. 51 We see a similar profusion of significations in modern fads and practices. There have been and are actors and athletes who have shaved their heads. And there is the motley group of young men called “skin heads”. It is difficult to see a uniform social meaning or message underlying all these practices.

Hair and Society

341

the variety of operational meanings, but it may allow a deeper understanding of those meanings and their interrelationships and of the enormous power of this ubiquitous symbol. The validity and usefulness of such a search is also indicated by the relative uniformity of the modes of life signaled by shaving and the neglect of hair in South Asian traditions. Why, for example, are all shaven people, whether they are ascetics or criminals, forbidden to engage in sex or to eat their own food? Much of the theoretical work on the symbolism of hair has been carried out thus far by scholars in the fields of anthropology and psychoanalysis. It may be useful here to review briefly some of the major contributions. James Frazer in The Golden Bough (1913: II, 25287) was one of the earliest to pay systematic attention to the customs relating to hair. Although his focus is on the reasons for the methods various peoples use to dispose of cut hair and nails, Frazer also deals with the significance of the head, head hair, and hair cutting. The reason why people of widely different cultures consider the head and head hair sacred and taboo, Frazer (1913: II, 252-53) argues, is because they believe that various spirits or divinities reside in the head and hair. “When the head was considered so sacred,” Frazer concludes, “it is obvious that the cutting of the hair must have been a delicate and difficult operation” (258). A fascinating, though no doubt extreme, example of the precautions taken at cutting one’s hair is that of the chief of Namosi in Fiji, who, according to a letter Frazer received from a Christian missionary, “always ate a man by way of precaution when he had his hair cut. There was a certain clan that had to provide the victim, and they used to sit in solemn council among themselves to choose him” (264). There are two major problems with Frazer’s analysis. The first, common among early anthropologists, is that he cites examples of hair customs from around the world without regard either to the accuracy of description or to the social and cultural context within which those customs are located. He cites indiscriminately from travellers’ diaries, from the writings and letters of Christian missionaries, and from ethnographies of varying degrees of reliability. He makes little attempt, moreover, to understand the customs he mentions within their contexts. The meanings of symbols, like those of words, can be studied adequately only if they are located within the broader grammar of the society. In the case of the Fijian chief, for example, we do not know whether he ate the man because he wanted to shave, or whether he shaved in order to eat the man, or, what is more likely, whether both acts were part of a larger ritual undertaking and liturgical calendar within which alone the two acts may reveal their significance. This brings us to the second problem: what Frazer offers as a theory is in fact merely a generalized account of the native exegesis. Hair is sacred because a god dwells in it. A theory, if it is to have cross-cultural validity and usefulness, has to go beyond the native exegesis, which is part of our data and not a substitute for a theory.

342

LANGUAGE, TEXTS AND SOCIETY

Moving beyond Frazer, theories of hair symbolism fall broadly into two categories: psychological and sociological. The psychological, or, more accurately, the psychoanalytic theory sees hair symbolism among humans as derived from the workings of the unconscious. Clinical observations reveal that through the mechanism of displacement the head often stands for the penis. The phallic symbolism of the head, it is argued, is transferred to the hair, where hair itself may stand for semen or the phallus. Cutting of hair or shaving is thus viewed by proponents of the psychoanalytic theory as equivalent to castration.52 Edmund Leach (1958) in his influential essay “Magical Hair” examines the psychoanalytic theory and finds that the ethnographic evidence by and large corroborates the sexual symbolism of hair. He concludes that “an astonishingly high proportion of the ethnographic evidence fits the following pattern in a quite obvious way. In ritual situations: long hair = unrestrained sexuality; short hair or partially shaved hair or tightly bound hair = restricted sexuality; close shaven hair = celibacy” (1958: 154). Leach, however, wants to keep the psychoanalytic and the anthropological fields separate, the former dealing with individual motives, and the latter, with social meaning. He does so by neatly dividing symbols into private and public. A private symbol “reflects the actual psychological state of the actor” (1958: 153, original italics), whereas a public symbol is merely a means of communication with publicly recognized meanings and does not necessarily correspond to the “psychological state of the actor.” Hallpike rejects even the somewhat circumspect acceptance of the unconscious association of hair and sexuality of Leach in his response to Leach appropriately entitled “Social Hair” (1969). He rejects the association of shaving with castration for the simple and obvious reason that it is inapplicable in the case of women, whose head may also be shaved in ritual contexts such as mourning. He also asks why, if head hair equals male genitals, so little regard is paid to beards, the symbol par excellence of masculinity, in ritual contexts. The hair of the beard, after all, is physically more similar to pubic hair than the hair of the head. Finally, he finds it very strange that, if long hair equals unrestrained sexuality, celibate ascetics wear long hair. Not surprisingly, Hallpike rejects the psychoanalytic theory and offers instead a thoroughly sociological one. His own theory boils down to this: “Long hair is associated with being outside society and [...] the cutting of hair symbolizes re-entering society, or living under a particular disciplinary regime within society” (1969: 260). He accepts dressing the hair as a ritual equivalent of cutting. Thus we have the equation: cutting or dressing the hair places a person within society and social control, while long and loose hair places one outside such control.53 Hallpike appears to 52

See Berg 1951; Leach 1958; Obeyesekere 1981 & 1990; Hershman 1974. Mary Douglas (1982: 72, 85, 89) also supports such a distinction, where the smooth stands for control and the shaggy, for the informal, the anti-social, and the prophetic. 53

Hair and Society

343

include within one category both the cutting of hair by ordinary people in society and the close shaving of the head associated with monks and ascetics. “Thus the cropped head or tonsure in all three cases of monk, soldier and convict,” Hallpike (1969: 261) argues, “signifies that they are under discipline.” The manipulative potential of hair makes it suitable for use in ritual,54 but Hallpike never shows why hair and only hair has become almost universally a powerful symbol of the relationship between individuals and society. The psychoanalytic theory provides a reason for this association, and if we are to reject that theory then we must be prepared to offer an alternative. Further, Hallpike’s identification of the shaven head and the ordinary cutting of hair and grouping them together under the category of social control and discipline are unconvincing; the evidence points in a different direction. Two anthropologists, Hershman (1974) in his excellent essay “Hair, Sex and Dirt” on hair symbolism among the Hindu and Sikh Punjabis and Gananath Obeyesekere (1981, 1990) in two of his recent works, Medusa’s Hair on modern Sri Lankan matted-hair ascetics and the more theoretical The Work of Culture, have attempted to bridge the divide between the psychoanalytic and the sociological viewpoints. Hershman (1974: 274) does so by establishing “a connexion between the symbolism of the individual subconscious and that of the collective consciousness.” “It is my contention,” he argues, that this connexion lies in the fact that bodily symbols gain their emotive power through being subconsciously associated with the anal-genital organs and processes, but that they are then used to spell out cultural messages, where the message is something quite separate and apart from the symbols which are transmitting it. It follows that a message becomes empowered by the subconscious associations of the symbols in whose terms it is expressed, but that its communication content remains something entirely different.

I agree that there is a distinction between the emotive power of a symbol and its social message; but the two, surely, cannot be “entirely different.” A theory of hair must address the problem of their connection: how is the socially accepted message related to the original unconscious symbolism of hair? A way toward a solution is pointed out by some important concepts put forward by Obeyesekere, who objects both to Leach’s watertight division between private and public symbols and to the psychoanalytical assumption that all symbols must have deep motivational significance. He suggests a distinction between personal symbols, where deep motivational significance is involved, and “psychogenetic symbols.” 54 Hallpike’s (1969: 257) analysis of hair symbolism is based on nine special characteristics of hair that he isolates: “1. Like the nails it grows constantly. 2. It can be cut painlessly, again like the nails. 3. It grows in great quantity, such that individual hairs

344

LANGUAGE, TEXTS AND SOCIETY

Psychogenetic symbols originate in the unconscious or are derived from the dream repertoire; but the origin of the symbols must be analytically separated from its ongoing operational significance. This is often the case in myths and rituals: symbols originating from unconscious sources are used to give expression to meanings that have nothing to do with their origin. (1981: 13-14; original italics)

Obeyesekere’s distinction between the genesis and the operational significance of symbols is similar to, but expresses more clearly, Hershman’s distinction between emotional power and cultural message. But is it possible that the operational significance of a symbol could “have nothing to do with their origin?” In his later book, The Work of Culture, Obeyesekere presents a more systematic and theoretical discussion of the phenomenon that he earlier referred to simply as the distinction between a symbol’s unconscious genesis and its operational significance. He calls this distinction “symbolic remove.” A symbol may operate at different levels of symbolic remove from its genesis in deep motivation “producing different levels of symbolization, some closer to, some more distant from the motivations that initially (psychogenetically) triggered the symbolic formation” (1990: 57). The theory of symbolic remove, I believe, is an important contribution to our understanding of the formation and function of symbols, and in what follows I will examine the symbolic remove that takes place in the case of hair symbolism of South Asia. Let me make some preliminary observations. Although Obeyesekere’s theory of “symbolic remove” is a rich heuristic device, at least implicitly he appears to acknowledge only a single source –namely, the unconscious– for the origin of a symbol. Such a position could be called the “monogenesis” of symbols. But I think we should posit that at least some symbols, including hair, are polygenetic –they originate from a multiplicity of sources. This suggests at least two consequences. First, polygenesis may create the polysemy of a symbol; that is, the same symbol either simultaneously or in different contexts may contain more than one meaning.55 Second, not all the meanings of a symbol may be reducible to a single root meaning or to the same source. An adequate theory, moreover, should offer a root meaning (or a cluster of such meanings) of a symbol—that is the meaning least removed from its genesis—and should delineate the process of symbolare almost numberless. 4. Head hair is apparent on infants of both sexes at birth. 5. Genital/anal hair appears at puberty in both sexes. 6. In some races, males develop facial hair after puberty, and also body hair. 7. Hair on different parts of the body is of different texture, e.g. eyelashes, pubic hair, head hair. 8. In old age hair often turns white and/or falls out. 9. Hair is a prominent feature of animals, especially monkeys, man’s analogue in the animal kingdom.” At least some of these features, as I shall point out, clearly played a role in the genesis of hair symbolism. 55 Indeed, polysemy of symbols is at the very heart of Obeyesekere’s (1990: 56) theory. Although polysemy is possible in the context of symbolic remove even within a monogenetic scheme, I think the parallel notion of polygenesis will strengthen the case for polysemy.

Hair and Society

345

ic remove that gives rise to its operational meanings. Now a root meaning of a symbol, like the root form of a verb, is ultimately a fiction; it is abstracted or extrapolated from observed meanings and forms. Yet, as in the case of words, it helps us establish the various symbolic removes that occur in actual practice and understand the relationship between various operational meanings of the same symbol that may appear on the surface to have little in common. A root meaning, in other words, helps us formulate the grammar of a symbol. As a fiction, the root meaning is neither true nor false, but more or less adequate or useful. The adequacy and usefulness of the root meaning we ascribe to a symbol can be validated only by comparing it with the actual and operational meanings available through ethnographic and historical study. This may appear to be a vicious circle, the root meaning being abstracted from operational meanings and the root meaning in turn validating the operational meanings. Just as in philology, however, this circle can be converted from a vicious to a hermeneutical circle (Obeyesekere 1990: 93). The concrete uses of a symbol yield its root meaning; the root meaning will reveal further levels of meaning of the symbol; as more operational meanings from the same culture (and from other cultures, if one is engaged in a cross-cultural study) are analyzed, they will help us further refine the root meaning. In explicating my theory, I first examine the sources from which the symbolism of hair is derived and then proceed to analyze the root meaning and the levels of symbolic remove. At least three sources are significant for the development of the symbolic meanings of hair. First, in humans there is a clear and visible association between the growth of body hair—especially axillary, pubic, and, among males, facial hair—and the onset of puberty and sexual maturity. As far as I know, this is a developmental feature unique to human beings; at least it is absent in animals likely to influence the human creation of symbols. This curious biological fact is also the likely foundation of the second source,56 namely the unconscious association between head hair and sexuality; this association has been sufficiently demonstrated both clinically and ethnographically. One may observe, however, that this association can stand on its own and apart from the related unconscious association of the head with the penis suggested by psychoanalysts. This dissociation of head and hair symbolism is also necessitated by the fact noted by Hallpike that female hair is treated in ways very similar to male hair. The third source is the biological fact that hair and nails are unique among body parts in that they grow continuously and they grow back when they are cut. On the one hand, they can be trimmed, shaved, and otherwise manipulated in ways that are impossible with other bodily parts. On the other hand, they may be viewed as imbued 56 I say the foundation, because the unconscious after all must receive the grist for its mill from the conscious and the sensory spheres.

346

LANGUAGE, TEXTS AND SOCIETY

with extraordinary vitality. Hair on the human body, as we have seen, bears a striking parallel to grass on earth; both grow again when they are cut, and both testify to the vitality and the fertility of their respective hosts. Hair and grass even show a certain physical resemblance. All three sources contribute to the complex of hair’s symbolic meanings. There may be others. Some symbolic uses of hair that consider it as excrement may derive that meaning from the fact that hair resembles dead matter, without blood or sensation, and often falls from the body on its own. The South Asian materials we have examined, however, point to the association of hair with puberty and sexual maturity as the primary source of the root meaning of hair symbolism at least within the South Asian context. Now to the root meaning. I posit that the root meaning from which most, though not necessarily all, operational meanings of hair is derived is a multifaceted complex consisting of sexual maturity, drive, potency, and fertility. For the sake of brevity, I shall henceforth refer to the root meaning simply as sexual maturity. The adequacy of this root meaning can only be gauged by examining how the operational meanings can be derived from or related to it, and how it enhances our understanding of those operational meanings. The root meaning I have assigned to hair is significantly different from the one proposed by the psychoanalytic school. In some contexts, as in neurotics, in the dream repertoire, and when hair is used as what Obeyesekere calls “a personal symbol” by an individual, hair may stand for the sex organ in general or the phallus in particular. That psychoanalytic meaning, however, cannot explain the variety of socially established and ritually enacted forms of hair symbolism we have encountered in South Asia; indeed, even apart from female shaving, in many cases the equation of shaving with castration is simply inadmissible. But at least one portion of the psychoanalytic theory is essential for explaining the process of symbol formation in the case of hair, and that is the theory of displacement. Simply stated, displacement occurs when the unconscious substitutes the entity X for the entity Y, thus permitting individuals at the conscious level to speak about and to manipulate X which at a deeper level are statement about and the manipulation of Y. The need for displacement arises because of a block at the conscious level that prevents individuals from dealing with Y directly. In the case of hair, it is not the pubic, axillary, or even facial hair—that is, hair associated with sexual maturation—that is the focus of attention in most rituals but the hair on the head. This hair at the source level is unrelated to puberty, appearing as it does at birth or soon thereafter. In ritual contexts the root meaning of hair as sexual maturity appears to be displaced from hair with clearer sexual connotations to the hair on the head without such connotations.57 57 This displacement is clear in the case of Brahmanical ascetics who are not only enjoined to shave their heads, but also positively forbidden from shaving their armpits and

Hair and Society

347

The root meaning of hair as such, however, does not occur in actual ritual or social settings. A basic symbolic remove intervenes between the root and operational meanings. This symbolic remove consists in the transformation of the root meaning of sexual maturity into its operational meanings relating to the status and role of an adult within the structures of society. This transformation of meaning is a fairly simple operation, because the central social structure within which adults are co-opted into and operate within society is marriage, the structure that controls the adult sex drive. This assumption is validated by, among others, the fact that, as we have observed, the ritual separation from society associated with shaving the head invariably entails the suspension of sexual activity. Having isolated the root meaning and its main operational derivative, we must nevertheless recognize that the meaning of a symbol, by its very nature, is rich, nuanced, and multifaceted, that several meanings normally inhere simultaneously, and that individuals participating in a ritual performance or social custom may be aware of or place emphasis on different aspects of that complex of meanings. Using this basic operational meaning, let me now delineate what I consider to be the main features of hair symbolism manifested within the broad spectrum of hair rituals in Indian religious history that we have examined. Hair in ritual has no inherent or absolute meaning; its meaning or meanings are derived always from its relationship or opposition to other ritual functions of hair existing within the same society. Thus, to understand the meaning of shaving the head or letting the hair fall loose and uncontrolled we must locate them in relationship to others in society who do not shave their head or who keep their hair braided or under control. Now, it is not necessary that all hair rituals obtain their meanings in relationship to a single point of reference. Thus, as we have seen, the long hair and beard of Sikh males derive their primary meaning in relation to the shaven head of a Hindu ascetic, whereas the clean and trimmed hair of an adult in society may be the point of reference for the long matted hair of hermits. Nevertheless, I believe that the most significant and central person with reference to whom most other hair rituals within the South Asian social context, and probably in most other societies, derive their meaning is the adult male, and to a lesser degree the adult female living within society. Their status and role within the social structure, their submission to and participation in the structures of social control, including structures for sexual control, are symbolized by the public control of their hair, a control that can be exercised in a variety of ways and not just by cutting, as supposed by Hallpike.58 There are cultural and historical variations in the pubic regions; appropriate penances are prescribed for those who do (see Yådava Prakå†a, Yatidharmasamuccaya, ed. cit., 8.9-11). 58 The Sikh males are a good example from India. Rivière (1969) has shown how the

348

LANGUAGE, TEXTS AND SOCIETY

the methods of hair control, but the most common method, at least in contemporary South Asia,59 is for an adult male to trim and dress the hair and to shave or at least to trim the beard and for an adult female to braid or to tie the hair in a knot. Thus we arrive at our first principle: control of hair by cutting, grooming, braiding, enclosing in a turban, or other means indicates an individual’s participation in social structures within a publicly defined role and that individual’s submission to social control. Such a submission assigns the subject clear social roles and grants him or her rights and privileges. We have seen that the most significant and widespread ritual use of hair in India is shaving, most frequently the shaving of the head and face, but sometimes also of the entire body. If we return to our root meaning of hair as sexual maturity, removal of hair would mean the denial or suspension of sexual maturity. The shaven individual is ritually reduced to the level of an infant, that is, to a sexually, and therefore socially, undifferentiated status. At a level symbolically removed from this, the primary meaning of shaving, I posit, is that the shaven individual is placed outside the social structures and denied a social status and role; hence, the almost universal association of shaving with rites of passage. Since permitted sexual activity is restricted to precisely such a social structure, namely marriage, shaven separation from society invariably involves celibacy. Thus we arrive at our second principle: shaving the head amounts to the ritual separation of an individual from society either for a temporary period or permanently. Shaving relating to reintegrating into society of persons who have been separated for a relatively long period of time shows an inversion of that process. Here the separation is not from society but from a liminal condition; the individual is released from that condition to assume his or her new social role. Variant forms of ritual separation from society are expressed not through shaving but through lack of control of hair. We have seen that periods of separation resulting from pollution such as menstruation and death are often marked by women leaving their hair unbraided, uncombed, and unwashed. Shaving and leaving hair uncontrolled, therefore, in their own way oppose the controlled public hair of people in society. In both cases there is some form of interdiction of social intercourse between people in society and those ritually separated from it. Related to such lack of control over hair is the third type of ritual use of hair in India: the neglect of hair associated with the physical separation of an individual from society. When an individual is ritually separated from society, he or she continues to live within the geographical boundaries of society and often in close relationship with people in hair tube is used by the tribes in the northern Rupununi district of Guyana to enclose and control the long hair of adult males. Cf. n. 8 on Tikopia. 59 As I have noted above, the picture is much less clear in ancient times.

Hair and Society

349

society, whereas in physical separation the rite of separation culminates in the individual’s departure from the social geography into the uncivilized realm of the forest or the wilderness. In India this type of separation is symbolized by long and uncontrolled hair that is left unkempt and unattended. The meaning of matted hair must be seen in its relation to the controlled hair of the adult within society. The total lack of hair control and the resultant long, ugly, matted hair (accompanied by long nails) indicates an individual’s utter separation from civilized structures and controls and his or her integration to the uncivilized realm of the wild and the beast. The third principle, therefore, is: matted hair indicates an individual’s physical separation from society and civilized structures. In medieval and modern South Asia, however, such physical separation has not been a lived reality. Thus, we have seen the diverse operational meanings given to matted hair in South Asian traditions. The case of the Sri Lankan male and female ascetics studied by Obeyesekere falls into the latter category. Once the institution of forest hermits disappeared, it was easier to give new operational meanings to this practice.60

CONCLUSION Once a particular social meaning has been assigned to a form of hair manipulation within a specific institutional or ritual setting –shaving as ascetic separation, for example– that same symbol may acquire new meanings for the participants, meanings that may go beyond, and thereby transform the earlier meaning. Thus ascetic shaving has acquired the meaning of “belonging” to a particular community as opposed to separation from society, in a way similar to that of the Sikh hair. In this way hair becomes a symbol that demarcates new boundaries –the monastic community or the Sikh community. Its new conventional meanings may thus hide to a large degree some of the basic meanings that I have attempted to uncover. In the new conventional settings a particular type of hair manipulation may become a “condensed symbol,” that is a symbol so powerful that it encapsulates all the diverse aspects of the symbolized, which under normal circumstances would require separate symbolic expressions. A flag or a national anthem may become such a condensed symbol; the rule against pork for the Jews and Friday abstinence from meat for traditional Catholics are similar condensed symbols that signify the essence of being a Jew or a Catholic. I think that the shaven head for the Buddhist monk and the uncut hair enclosed in a turban for the Sikh are such condensed symbols. This is nicely illustrated by 60 I wonder, however, whether within the mainland of India, where the legends of matted-hair ascetics and seers are still alive, an ascetic would be able to publicly claim that his or her matted hair was the gift of a god, as the Sri Lankan ascetics regularly claim. In Sri Lanka where the matted hair tradition and legends are less frequent, it is evidently easier to ascribe deeply personal meanings to this symbol because it lacks an articulated public meaning.

350

LANGUAGE, TEXTS AND SOCIETY

an ethnographic observation of H. L. Seneviratne,61 who found that one way contemporary Buddhist monks in Sri Lanka display their independence from monastic authority and their modernist outlook is to let their hair grow, now only to the length of a crew-cut, but who knows what is in store for the future? Hair thus remains both a means of strong institutional control and an instrument of liberation from and critique of social and institutional controls.

61

Personal communication.

XVI Abhakßya and Abhojya: An Exploration in Dietary Language*

If the way to a persons’s heart is through the stomach, then the way to the soul of a civilization may be through its dietary practices. Examining the food habits of a people has been a staple among anthropologists, some of whom, like Mary Douglas (1966), in her piece on Jewish dietary laws “The Abominations of Leviticus,” have ventured into interpreting the food taboos and dietary restrictions encoded in ancient texts. The ancient Indian literature on dharma devotes considerable attention to matters of food: what kinds of animals and vegetables may or may not be eaten, from what sorts of people one may or may not receive food, what types of conditions make food unfit for consumption, and so on. Such practices have drawn considerable attention among scholars; what has been ignored, however, is the vocabulary used to indicate food prohibitions, a vocabulary that may give us new insights into the ancient Indian world.1 And that is the focus of this paper. The dharma vocabulary of food proscriptions contains four words: abhakßya, abhojya, anådya, and apeya. In this paper I will focus on the first two, abhakßya and abhojya, which alone underwent significant semantic developments and assumed technical meanings. Apeya is restricted to liquids, principally milk. Anådya is, relatively speaking, the most frequent term in the vedic literature occurring a total of nine times, often in the metaphorical sense that the Bråhmaña should not be eaten by the king: bråhmaño ‘nådyaΔ.2 This term occurs Originally published in the Journal of the American Oriental Society 122 (2002): 345-354. I did a similar study a couple of years ago on the vocabulary of the pure/impure: Olivelle 1998b. 2 AV 5.18.1, 3; 8.2.19; TS 5.7.9.2; ‡B 5.3.3.12 (=5.4.2.3); 9.1.2.5; 9.4.3.16; 12.9.1.1; TB 1.4.3.6(2). The meanings in these usages include both items that ought not to be eaten, such as Bråhmañas and cows, as well as things that are simply inedible, such as clay and * 1

352

LANGUAGE, TEXTS AND SOCIETY

with some frequency in the dharma literature, but it did not develop the kind of technical meaning that the other two did.3 Abhakßya and abhojya are, of course, the negative forms of bhakßya and bhojya. The positive forms of the words have been studied in detail by Toru Yagi (1994). I will only mention that these two terms, even though they are gerundives, for the most part lack any prescriptive or permissive meaning: they do not mean “what should be eaten” or “what may be eaten” but are simply types of food. It is within the context of the negative forms, abhakßya and abhojya, that the terms assumed a strong prescriptive, or more precisely proscriptive, meaning. Of the two, abhakßya occurs only once in the vedic literature in a somewhat obscure passage in the Kå™haka Sa∫hitå (35.16), and abhojya occurs twice, once in the Såmavidhåna Bråhmaña (1.5.13) and once in the Gopatha (1.3.19).4 The emergence of these forms and their semantic development occur principally within the context of lists containing items of foods that are either totally forbidden or for some reason have become unfit for consumption. These lists are absent in the vedic literature and in the ‡rauta and Gr¢hya Sütras. They make their first appearance in the Dharmasütras.5 These lists must have become sufficiently standard by Pata∞jali’s time (2nd cent. BCE) for him to use a stock example repeatedly: abhakßyo gråmyakukku™o ‘bhakßyo gråmyasükaraΔ –“it is forbidden to eat a village cock; it is forbidden to eat a village pig” (1.1.1 [5:16]; 1.1.1 (8:10); 7.3.14 [320:22]). A close reading of these lists in the dharma literature and the use of the two terms within them permit us to draw the following conclusions: A) Abhakßya refers to items of food, both animals and vegetables, that are completely forbidden; they cannot be eaten except under the most dire circumstances.6 Generally, these lists refer to food sources rather than cooked food served at a meal. Thus, carnivorous animals, web-footed birds, garlic, red resins of trees are all abhakßya. I translate abhakßya as “forbidden food”. stone. On the metaphorical use of “food” and “eating” to express power relationships, see Smith 1990. 3 Anådya occurs in ÅpDh 1.17.17; 1.19.14; GDh 25.7; BDh 4.1.6; 4.2.13, 14; VaDh 1.45 [=‡B 5.4.2.3]; MDh 11.57, 146, 162. The term does not occur in ViDh or YDh. 4 The Kå™haka instructs the priest to put abhakßañœya into the holes over which the Soma is crushed: prajåpataye svåhety abhakßañœyån uparaveßv avanayet. The Såmavidhåna gives a purification for eating abhojya; and Gopatha says that before the Soma has been purchased, the man consecrated for the sacrifice (dœkßita) is abhojanœya, which has the same meaning given to abhojya in the dharma texts. 5 The likelihood is high that even though we encounter them for the first time in the Dharmasütras, these lists, at least in informal ways, must have existed before them. Stephanie Jamison (1991: 212-221; 1996b: 197-200) has argued for the existence of law codes in fixed verbal form in vedic times. 6 Manu (MDh 10.104-108) permits a man facing starvation to eat anything at all, giving the examples of Ajœgarta, Våmadeva, Bharadvåja, and Vi†våmitra. This falls within Manu’s description of åpaddharma, rules for times of adversity.

Abhakßya and Abhojya

353

B-i) Abhojya, on the other hand, refers to food that is normally permitted but due to some supervening circumstances has become unfit to be eaten. These lists contain not food sources but food that is actually served at a meal. Thus, food contaminated by hair or insects, food touched by an impure man or woman, food given by a person from whom food cannot be accepted, food that has turned sour or stale are all abhojya. I translate abhojya as “unfit food”. B-ii) Abhojya takes on a secondary meaning referring not directly to food but to “a person whose food one is not permitted to eat.” This meaning is sometimes very clear and explicit; sometimes, as in the common compound abhojyånna, the meaning is ambivalent, especially in cases other than the nominative where the masculine marks the compound as a Bahuvrœhi. Thus, the statement abhojyånna∫ nå†nœyåt, may, as a Karmadhåraya, mean: “He should not eat unfit food,” or, as a Tatpurußa, “He should not eat the food of a man whose food one is not permitted to eat.” The latter, I think, is the meaning in most instances. An examination of the dharma†åstric lists of forbidden and unfit foods, I believe, supports these conclusions.

Gautama Gautama is the most clear and systematic. He begins his discussion of food (GDh 17.1-8) with a list of people from whom food and other articles may be accepted: 17.1 pra†astånå∫ svakarmasu dvijåtœnå∫ bråhmaño bhu∞jœta. A Brahmin may eat food given by twice-born men renowned for their devotion to their respective duties.

Note the use of bhuj. He goes on to say: 17.6-7 pa†upålakßetrakarßakakulasa∫gatakårayitr¢paricårakå bhojyånnåΔ vañik cå†ilpœ. A man who looks after his animals or plows his fields, a friend of the family, his barber, and his personal servant –these are people whose food he may eat, as also a merchant who is not an artisan.

Bhojya for Gautama is a person whose food one may eat; and conversely abhojya means the opposite. Gautama concludes the discussion with sütra 8, where abhojya has the primary meaning of food that is not to be eaten. 17.8 nityam abhojyam. Their food is not fit to be eaten everyday.

354

LANGUAGE, TEXTS AND SOCIETY

In the next twelve sütras (GDh 17.9-21) Gautama lists items of food that should not be eaten due to some circumstance. Food into which hair or an insect has fallen; what has been touched by a menstruating woman, a black bird, or someone’s foot; what has been looked at by an abortionist or smelt by a cow; food that looks revolting; food that has turned sour, except curd; re-cooked food; food that has become stale, except vegetables, chewy or greasy foods, meat, and honey; food given by someone who has been disowned by his parents, a harlot, a heinous sinner, a hermaphrodite, a law enforcement agent, a carpenter, a miser, a jailer, a physician, a man who hunts without using the bow or eats the leftovers of others, a group of people, or an enemy, as also by those listed before a bald man as people who defile those alongside whom they eat;7 food prepared for no avail; a meal during which people sip water or get up against the rules, or at which different sorts of homage is paid to people of equal stature and the same homage is paid to people of different stature; and food that is given disrespectfully.

There is no verb or verbal equivalent in this list, and it is clear that the term abhojyam of sütra 8 is carried over into these sütras (anuvr¢tti), especially since most of these items have neuter singular endings. This is confirmed by both Haradatta and Maskarin, who comment: abhojyam iti sarvatrånuvartate (“the term abhojyam is supplied everywhere from the previous sütra”). This list contains, on the one hand, food that has been spoilt for a variety of reasons and, on the other, food that is unfit because it was given or touched by the wrong individual. The next five sütras (GDh 17.22-26) deal with milk: The milk of a cow, a goat, or a buffalo, during the first ten days after it gives birth; the milk of sheep, camels, and one-hoofed animals under any circumstances; the milk of an animal from whose udders milk flows spontaneously or of an animal that has borne twins, gives milk while pregnant, or has lost her calf.

These sütras, likewise, lack a verb or verbal equivalent. In other texts, the term apeyam (“not to be drunk”) would be found within such a list. We must assume that here the anuvr¢tti of abhojyam in sütra 8 continues. Note that Gautama here deals both with animals whose milk can normally be drunk and those, such as sheep, camels, and one-hoofed animals, whose milk is completely forbidden. Lists of drinks, especially milk, tend to include both unfit and forbidden milk; apeya appears to cover both categories. Most lists treat milk in a somewhat different way than food. Then Gautama (GDh 17.27-34) turns to forbidden foods. Here the term abhakßyåΔ comes at the very end of the list; but it is clear that it governs the entire list: Animals with five nails, with the exception of the hedgehog, hare, por7

See GDh 15.16-18.

Abhakßya and Abhojya

355

cupine, Godhå monitor lizard, rhinoceros, and tortoise;8 animals with teeth in both jaws, with a lot of hair, or without any hair; one-hoofed animals; Kalaviõka sparrows; Plava herons; Cakravåka geese; Ha∫sa geese; crows; Kaõka herons; vultures; falcons; water birds; red-footed and red-beaked birds; village cocks and pigs; milch-cows and oxen; meat of animals whose milk-teeth have not fallen and of animals that are sick or wantonly killed; young shoots; mushrooms; garlic; resins; red juices flowing from incisions on trees; woodpeckers; Baka egrets; Balåka ibis; parrots; Madgu cormorants; ~i™™ibha sandpipers; Måndhåla flying foxes; and night birds –these are forbidden foods (abhakßyåΔ).

In this list we can see very clearly the distinction between abhojya food items given previously and abhakßya items. On the one hand, abhakßya foods are forbidden because of their very nature; on the other, as one can see from the list, the reference is to food sources –animals and plants– and not to prepared food items presented at a meal. Abhojya, on the contrary, refers to prepared food and frequently the term annam is used, a term that generally refers to cooked food, especially rice. Finally, Gautama (GDh 17.35-37) turns to food sources that are permitted, using the term bhakßyåΔ to refer to them. The last word in the previous list (GDh 17.34) is abhakßyåΔ; and Gautama presents a deliberate contrast by beginning the next list with bhakßyåΔ: bhakßyåΔ pratudavißkiråjålapådåΔ/ matyå† cåvikr¢tåΔ/ vadhyå† ca dharmårthe Ù These may be eaten: birds that feed by thrusting their beaks or scratching with their feet and that do not have webbed feet, fish that are not grotesque, and animals that have to be killed for the sake of the Law.

Here we have an interesting usage of bhakßya not in the normal positive sense of “food” as studied by Toru Yagi (1994) but as the opposite of abhakßya; that is, foods that are permitted to be eaten, paralleling the use of bhojya to refer to a person whose food may be eaten (GDh 17.6). This meaning of bhakßya, moreover, falls within the Mœmå∫så definition of parisa∫khyåvidhi, that is, positive injunctions whose principle aim is to prohibit what is not enumerated. Thus, one is not obliged to eat the listed animals, but one is forbidden to eat animals that are not listed. This meaning is also evident in Pata∞jali (1.1.1 [5:14-17]), who says that when some things are called abhakßya it implies that the opposite is bhakßya, and vice versa. Thus, the prohibition of village cocks and pigs implicitly permits the eating of wild cocks and pigs; and the permitting of the five fivenailed animals implicitly forbids the eating of five-nailed animals other than those listed: 8 For a detailed discussion of these exception to the prohibition against eating animals with five nails or claws, see Jamison 1998.

356

LANGUAGE, TEXTS AND SOCIETY

bhakßyaniyamenåbhakßyapratißedho gamyate Ù pa∞ca pa∞canakhå bhakßyå ity ukte gamyata etad ato ‘nye ‘bhakßyå iti Ù abhakßyapratißedhena vå bhakßyaniyamaΔ Ù tad yathå Ù abhakßyo gråmyakukku™o ‘bhakßyo gråmya†ükara ity ukte gamyata etad årañyo bhakßya iti Ù Pata∞jali (1.1.1 [5:14-17]) By specifying the bhakßya the prohibition of the abhakßya is understood. When it is said, “The five five-nailed animals are bhakßya, “it is understood that animals other than these are abhakßya. Alternatively, by prohibiting the abhakßya, the bhakßya is specified. So, when it is said, “the village cock is abhakßya; the village pig is abhakßya”, it is understood that a wild one is bhakßya.

Åpastamba The usage of abhakßya and abhojya in other dharma texts supports the very clear distinction established by Gautama. Åpastamba, in my estimation the author of the oldest Dharmasütra (Olivelle 2000, 410), is much less consistent in his terms for prohibited food items, using anådyam (ÅpDh 1.17.17; 1.19.14) as a synonym of abhojya, or simply saying nå†nœyåt (ÅpDh 1.17.14). Nevertheless, it is clear that he too sees abhojya and abhakßya as technical terms referring to two distinct classes of proscribed foods. He begins his list of unfit food (ÅpDh 1.16.16-32) with the term abhojyam. At one point (ÅpDh 1.16.21), however, he makes the interesting distinction between food that is aprayata (“unclean”) and food that is abhojya (“unfit to be eaten”); the former can be purified and then eaten, whereas abhojya food cannot be eaten at all. 1.16.16 hi∫sårthenåsinå må∫sa∫ chinnam abhojyam. Meat cut with a slaughering knife is abhojya. 1.16.21 aprayatopahatam annam aprayata∫ na tv abhojyam. Food that has been touched by an impure person becomes impure but is not rendered abhojya.

Åpastamba next turns to rules about eating and utensils for eating (ÅpDh 1.16.33; 17.1-13), and only then turns to forbidden foods (ÅpDh 1.17.14-39); but this is a sundry list containing a variety of items, from food bought in the market or turned sour to forbidden animals and plants. He begins with the phrase na a†nœyåt (“he must not eat”), and uses the terms akhådya, apeya, anådya, and abhakßya. Åpastamba uses abhojya once with respect to a forbidden food: kyåkv abhojyam iti hi bråhmañam (“ ‘Mushrooms should not be eaten,’ so states a Bråhmaña” ÅpDh 1.17.28), but here it may be simply a citation of an older text. It appears, nevertheless, that the vocabulary is still somewhat non-technical in Åpastamba. A little later (ÅpDh 1.17.30), however, he used bhakßyam as permitted food:

Abhakßya and Abhojya

357

dhenvana∂uhor bhakßyam, “It is permitted to eat the meat of milch cows and oxen.”9 He devotes two chapters (ÅpDh 1.18-19) to the issue of people from whom food may or may not be accepted. Here we find the term bhojya, bhoktavya, and abhojya used with consistency. At 1.26.7, furthermore, Åpastamba distinguishes three categories of food in the expression abhakßyåbhojyåpeyaprå†ane. Bhler translates this as “(He who has been guilty of) eating or drinking things forbidden” thus reducing the three categories to two and taking abhakßa and abhojya as synonyms. Åpastamba is here signaling the three kind of food that should not be eaten: forbidden (abhakßya), unfit (abhojya), and, significantly, what should not be drunk (apeya); this category generally includes milk of certain forbidden animals (broadly falling within the abhakßya category), as well as milk that has become contaminated or in some other way (e.g., first ten days after giving birth) made unfit.

Baudhåyana Baudhåyana begins his list of forbidden animals and plants (BDh 1.12.1-15) with the word abhakßyåΔ, which governs the rest of the paragraph (anuvr¢tti). 1.12.1 abhakßyåΔ pa†avo gråmyåΔ. It is forbidden to eat village animals.

Immediately after this list of forbidden foods, Baudhåyana, just like Gautama, gives a list of permitted foods, again beginning this list with the word bhakßyåΔ: 1.12.5 bhakßyåΔ †våvi∂godhå†a†a†alyakakacchapakha∂gåΔ kha∂gavarjåΔ pa∞ca pa∞canakhåΔ. It is permissible to eat the following: porcupine, Godhå monitor lizard, hare, hedgehog, tortoise, and the rhinoceros –these, excluding the rhinoceros, are the five five-nailed animals.

Baudhåyana has no list of abhojya foods.10 Nevertheless, at 1.9.8 he uses abhojya in the sense of a “person whose food one may not eat”: 9 There appears to have been a deep split among authorities regarding the meat of cows and oxen. Åpastamba and Vasiß™ha (VaDh 14.46) permit it, but their unease is evident in that both authors have to cite the opinion of Våjasaneyins in support of their position. We have seen that Gautama (GDh 17.27-34) includes these two animals in his long list of prohibited food. 10 Baudhåyana’s text was subjected to more tampering and interpolations than perhaps any other early dharma text. The lack of a list of abhojya foods, therefore, does not permit us to make any general conclusions regarding the original sütra. See Kane 1962-75, I: 42-44; Olivelle 2000, 191. Baudhåyana (4.1.6) specifies a penance for abhakßyåbhojyåpeyånådyaprå†aneßu. It is unclear what the distinction here is between anådya and the

358

LANGUAGE, TEXTS AND SOCIETY

khalakßetreßu yad dhånya∫ küpavåpœßu yaj jalam Ù abhojyåd api tad bhojya∫ yac ca goß™hagata∫ payaΔ ÙÙ Grain from a threshing floor and water from a well or reservoir, as also milk from a dairy farm –these may be consumed even if they are given by someone whose food one is not allowed to eat.

Vasiß™ha Vasiß™ha, the author of the last Dharmasütra, begins his discussion of unfit food (VaDh 14.1-32) with the statement: 14.1 athåto bhojyåbhojya∫ ca varñayißyåmaΔ. Next we will describe food that is fit and food that is unfit to be eaten.

He then lists people from whom food may not be accepted (annam abhojyam), people from whom food may be accepted, and finally foods that have become spoilt by contact or gone stale. All this is just what we would have expected under abhojya food. Immediately after his discussion of bhojyåbhojya food, he turns to forbidden foods (VaDh 14.33-48). The very first sütra lists forbidden foods indirectly by indicating the penance for eating them: 14.33 la†unapalåñ∂ukyåkugr¢∞jana†leßmåntakavr¢kßaniryåsalohitavra†cana†vakåkåvalœ∂ha†üdroccheßañabhojaneßv atikr¢cchraΔ. For eating garlic, onions, mushrooms, Gr¢∞jana onions, ‡leßmåntaka fruits, tree resins, red juices flowing from incisions on tree barks, anything licked by a dog or a crow, or the leftovers of a ‡üdra he should perform the very arduous penance.

We should have expected bhakßaña here rather than bhojana. This usage may indicate that, as in any specialized use of words, technical terms continue to be employed with a broader meaning. However, this passage begins with abhakßya types of food (garlic etc.) but concludes with abhojya types (licked by a dog etc.). I think the use of bhojana here may have been influenced by the abhojya items at the end of the list. When Vasiß™ha turns to listing forbidden foods in 14.34-38, the governing verb he uses is varjayet (“he should avoid”). In sütra 39, however, he does use bhakßya regarding permitted animals in the common list of the five five-nailed animals. Next, Vasiß™ha gives a list of forbidden animals (sütras 40-45) without any verb or verbal equivalent to indicate whether they are forbidden or not. They are forbidden only by implication. Sütra 46 begins with bhakßyau with reference to the milch cow other terms; perhaps in this interpolated chapter the author is covering all his bases and listing every term for proscribed food and drink.

Abhakßya and Abhojya

359

and draft ox (dhenvana∂uhau), thus bracketing the abhakßya animals between two lists of bhakßya animals (see above note 9). The long list of 36 forbidden animals contained in the last sütra (48) of this section, likewise, has no verb or verbal equivalent; these are also forbidden only by implication. It is clear, however, that Vasiß™ha recognized the two categories of food represented by abhojya and abhakßya, as evident in the following verse (VaDh 27.10) given within the context of remedies when one suspects that one has eaten such food: 27.10 †aõkåsthåne samutpanne abhojyåbhakßyasa∫j∞ake Ù When a doubt has arisen bearing the name “unfit to be eaten” or “forbidden to be eaten”

Here also Bhler translates “whether it may be called fit to be eaten or not” thus eliminating the distinction between the two terms.

Manu The use of the two terms by Manu, the author of the first dharma text composed completely in verse, also demonstrates their continued employment as technical terms for the two broad categories of proscribed food. At the end of the first chapter of Manu there is a table of content.11 It lists bhakßåbhakßyam as one topic, and Manu deals with this topic at 5.4-26. At the very beginning of this list, Manu uses abhakßyåñi to refer to the forbidden foods. la†una∫ gr¢∞jana∫ caiva palåñ∂u∫ kavakåni ca Ù abhakßyåñi dvijåtœnåm amedhyaprabhavåñi ca ÙÙ Garlic, leeks, onions, and mushrooms are foods forbidden to twiceborn persons; and so is anything growing in an impure medium.

This long list of forbidden food does not contain any item that would fall under abhojya, such as contaminated food. All items are sources of food rather than food in the sense of cooked food served at a meal. Three times in this list (MDh 5.17, 18, 23) bhakßya is used with reference to permitted food. Manu concludes this section with a “transition verse” (MDh 5.26) introducing the next topic; significantly, this verse refers to the two categories of food dealt with in the previous section as bhakßyåbhakßya: etad ukta∫ dvijåtœnå∫ bhakßyåbhakßyam a†eßataΔ Ù må∫sasyåtaΔ pravakßyåmi vidhi∫ bhakßañavarjane ÙÙ 11 As I have worked on the critical edition of Manu for the past several years, it has become more and more clear that this table of contents, as well as a large chunk of the first

360

LANGUAGE, TEXTS AND SOCIETY

I have described above everything that the twice-born may and may not eat. I will now explain the rule on eating and on avoiding meat.

Manu does not deal with the abhojya type of foods at all in chapter five; he does so in chapter four (MDh 4.205-224) under rules for a snåtaka. Manu’s rules on abhojya deal mostly with people whose food should not be eaten. He begins the lists (MDh 4.205) with the expression na bhu∞jœta and repeats it at the beginning of his long list (MDh 4.207-217) containing a motley collection of people who are all abhojya: He must also never eat (na bhu∞jœta) the following: food given by drunkards, the quick-tempered, or the sickly; food contaminated with hair or insects, touched deliberately with the foot, looked at by a murderer of a Brahmin, touched by a menstruating woman, pecked by a bird, or touched by a dog; food smelled by a cow; especially, food given after a public announcement; food given by a group or by a prostitute; food that is despised by learned men; food given by a thief, a musician, a carpenter, an usurer, a man consecrated for a sacrifice, a miser, a prisoner, a shackled man, a heinous sinner, a eunuch, a promiscuous woman, or a hypocrite; food that has turned sour or is stale; leftovers of a ‡üdra; food given by a physician, a hunter, a cruel man, someone who eats leftovers, or an Ugra; food of a woman impure by reason of childbirth; food served at a meal where someone sips water during the meal; food given during the ten days of impurity resulting from a birth; food given without respect; meat prepared for a futile reason; food given by a woman without a husband, an enemy, a city dweller, or an outcaste; food someone has sneezed upon; food given by a slanderer, a liar, a ritual merchant, an actor, a tailor, an ingrate, a blacksmith, a Nißåda, a theatrical performer, a goldsmith, a basket-weaver, an arms merchant, those who raise dogs, liquor merchants, a washerman, a dyer, a heartless man, someone who lets his wife’s paramour live in his house or who condones a paramour, or someone who is bossed by his wife; food of persons during the first ten days after a death in their family; and unappetizing food.

Towards the end of his discussion on unfit food (MDh 4.221), Manu calls these people abhojyånnåΔ, people whose food should not be eaten; and at MDh 4.253 lists some ‡üdras who are bhojyånnaΔ, that is, persons whose food may be eaten. årdhikaΔ kulamitra∫ ca gopålo dåsanåpitau Ù ete †üdreßu bhojyånnå yå† cåtmåna∫ nivedayet ÙÙ A sharecropper, a friend of the family, and one’s cowherd, slave, and barber are the ‡üdras whose food is fit to be eaten, as also a person who has presented himself.

Nowhere in this entire discussion on unfit food are the words bhakßya or abhakßya used. chapter, is a later interpolation. This interpolation must have been made quite early in the life of the text, because it is found in every manuscript and is noted by all the commentators.

Abhakßya and Abhojya

361

Vißñu The Vißñu Dharmasütra, although it lacks the sustained treatment of food found in other dharma texts, contains interesting and significant uses of abhakßya and abhojya that support my thesis. At ViDh 37.7 and 44.11 he uses an identical expression clearly establishing the distinction between the two terms: 37.7 abhojyånnåbhakßyabhakßañam. Eating forbidden food or the food of some-one whose food one is forbidden to eat. 44.11 abhojyånnåbhakßy农 kr¢miΔ. A man who eats forbidden food or the food of someone whose food one is forbidden to eat becomes a worm.

The first (ViDh 37.7) is part of a long list of sins for which the Cåndråyaña (lunar) penance is prescribed. It is noteworthy that the first Bahuvrœhi compound abhojyånna contains the word anna, which refers generally to cooked food. In a section (ViDh 57.1-5) dealing with people who should be abandoned or avoided (tyåjyåΔ), we have again the use of abhojya in its technical meaning: 57.5 sarva evåbhojyå† cåpratigråhyåΔ. All (these) are people whose food one is forbidden to eat and from whom one is forbidden to accept gifts.

At ViDh 57.16, on the other hand, in a verse identical to MDh 4.253, Vißñu lists certain types of ‡üdras, such as a friend of the family and the barber, whose food may be eaten, calling them bhojyånna. Clearly, abhojya and abhojyånna are synonymous terms. Vißñu also reserves abhakßya to refer to foodstuffs that are forbidden. In his extensive discussion on crimes, he (ViDh 5.98) imposes a fine of sixteen Suvarñas (gold coins) on a man who taints a Bråhmaña by feeding him abhakßya food (abhakßyeña bråhmañasya düßayitå). Jolly glosses abhakßya as “uneatable food such as excrements and the like, or forbidden food, such as garlic.” I think this is unwarranted; in Vißñu, as in other Dharma†åstras, the term refers to forbidden foods. This meaning is quite evident at ViDh 5.175 (in Jolly 5.174), where the highest fine is imposed on a group of offenders, including abhakßyavikrayœ, “a man who sells abhakßya food”; and here Jolly correctly translates it as “he who sells forbidden food.” Vißñu makes an interesting distinction within the abhakßya class of foods. After sütra 5.98 cited above, which deals with tainting a Bråhmaña with abhakßya food, he goes on to impose a larger fine of one hundred Suvarñas if

362

LANGUAGE, TEXTS AND SOCIETY

that food was jåtyapahårin, “causing a man to lose his caste.” A similar class of abhakßya food is referred to later in the same section of crimes (5.174; in Jolly 5.173): 5.174 jåtibhra∫†akarasyåbhakßyasya bhakßayitå vivåsyaΔ. A man who makes someone eat food that causes loss of caste should be sent into exile.

Given the extreme nature of the punishment, something not imposed even on a man who destroys a boundary marker or breaks an idol, it appears that this class of abhakßya must be foods that are at the top of the list of forbidden foods. Jolly translates bhakßayitå in this sütra as “he who eats”; I think we are justified in seeing here the causative of bhakß given the serious nature of the crime and its parallel with ViDh 5.98, which also deals with feeding rather than eating abhakßya food. In two sütras (ViDh 66.12-13), Vißñu juxtaposes abhakßya and bhakßya as opposing categories.12 66.12 nåbhakßya∫ naivedyårthe. He should not give forbidden food for use at a divine offering. 66.13 na bhakßye ‘py ajåmahißœkßœre. Nor the milk of a goat or buffalo, even though it is a permitted food.

The milk of goats and buffalos can ordinarily be consumed; but they are excluded from a naivedya. This is the only occasion to my knowledge that bhakßya is used in this literature with reference to milk of forbidden animals.

Yaj∞avalkya The last Dharma†åstra that I will investigate is Yåj∞avalkya.13 He devotes six verses (YDh 1.161-166) to unfit food,14 eight and a half verses (YDh 1.170-178ab) to forbidden foods, and another three and a half (YDh 1.178cd-181) to meat. Significantly this section is prefaced by the commentator Aparåditya with the words adhunå bhojyåbhojyabhakßyåbhakßyaprakarañam årabhyate Ù tatråbhojyam åha, “Now, he begins the topic of fit and unfit foods, and permitted and forbidden foods. Of these, he lists (first) unfit food.” 12 The term bhakßya in its technical meaning is also used by Vißñu (ViDh 22.70) with reference to the well-known category of the five five-nailed animals (pa∞capa∞canakhåΔ). 13 Other texts (mülasmr¢ti) continued to be composed, but we have only fragmentary evidence of them. The texts of Nårada, Br¢haspati, and Kåtyåyana, on the other hand, deal exclusively with law and legal procedure. 14 The section actually begins, according the Aparåditya, at YDh 1.160cd.

Abhakßya and Abhojya

363

After enumerating the usual types of abhojya persons, Yåj∞avalkya concludes (YDh 1.165): eßåm anna∫ na bhoktavyam, “The food of these should not be eaten”. The next verse (YDh 1.166) –nearly identical to MDh 4.253 and ViDh 57.16 cited above– specifies certain types of ‡üdras who are called bhojyånnåΔ, “persons whose food may be eaten”. Stale food, which is normally abhojya, is said to be bhojya (YDh 1.169) when it is mixed with ghee. In his section on forbidden food, Yåj∞avalkya does not use the term abhakßya probably because the first part of the list (YDh 1.175) ends with the appropriate penance for eating such food (jagdhvå sopavåsas tryaha∫ vaset), and the second part (YDh 1.176) with another penance (jagdhvå cåndråyaña∫ caret).15 The section ends (YDh 1.177) with the by now familiar refrain about the five five-nailed animals that are permitted to be eaten (bhakßyåΔ). Finally, in his chapter on law (vyavahårakåñ∂a, YDh 2.296), Yåj∞avalkya has a provision identical with that of ViDh 5.98: anyone who taints a Brahmin by feeding him abhakßya food is subject to the highest fine (abhakßyeña dvija∫ düßyo dañ∂ya uttamasåhasam).

Mahåbhårata In the Mahåbhårata, a list of proscribed foods occurs, as far as I know, only once; it is in the Råjadharmakåñ∂a 12.37.1-26. The discussion begins with the question (2): ki∫ bhakßya∫ kim abhakßya∫ ca (“What is permitted to be eaten and what is forbidden?”). The list proper begins at verse 16; in this and in the next verse forbidden items and animals are called abhakßya. At verse 21 begins a list of unfit foods which are called abhojya. Likewise, the food of people whose food cannot be eaten is called abhojyam annam (22). The list ends in a somewhat unclear verse (26): påyasa∫ kr¢sara∫ må∫sam apüpå†ca vr¢thå kr¢tåΔ Ù abhojyå† cåpy abhakßyå† ca bråhmañair gr¢hamedhibhiΔ ÙÙ I am unsure whether the reading here is accurate, but one thing is clear: the list concludes with the two types of foods listed above –the unfit or abhojya and the forbidden or abhakßya. An interesting light is thrown on the distinctive meanigs of abhakßya and abhojya by a story recorded in the Mahåbhårata (1.166.132). King Kalmåßapåda tells his cook to prepare a meat dish for a Bråhmaña guest. The cook is unable to find meat and the furious king orders him to get human flesh, which the cook does. The Bråhmaña’s powerful vision permits him to recognize the dish in front of him as human flesh and exclaims: abhojyam idam (“this is abhojya”). Here, interestingly, human flesh which should be abhakßya is called abhojya; I believe the usage here is proper because the term refers not to human flesh in general but to the food set before him, 15 The very first verse of this section (YDh 1.170) dealing with milk contains the term varjayet (“he should avoid”).

364

LANGUAGE, TEXTS AND SOCIETY

which significantly is called annam, a term that, as we have seen, is used for food that is cooked and ready to be eaten. As the types of food that one should avoid became the focus of an expert tradition, they were bound to be classified and re-classified. Bhakßya and bhojya represent one of the earliest and clearly the broadest of such classifications. Vij∞åne†vara, in his Mitåkßarå commentary on Yåj∞avalkya (YDh 3.289), elaborates this twofold classification into an elevenfold classificatory system. 1) jåtiduß™a: examples onions and milk of cow in heat. 2) svabhåvaduß™a: forbidden animals. 3) a†ucisa∫spr¢ß™a: food touched by a crow or dog, or leftovers of a ‡üdra. 4) a†ucidravyasa∫spr¢ß™a: food contaminated with hair or insects. 5) bhåvaduß™a: revolting by reason of color, shape, etc. 6) kåladuß™a: stale food, milk during first ten days after birth of the calf. 7) guñaduß™a: †ukta food gone sour or acidic; food of bad people. 8) hastadånådikriyåduß™åbhojya: food given by hand, ghee dripping from nails. 9) parigrahåbhojya: given by wrong kind of individual: eunuchs, thieves, etc. 10) å†aucaparigr¢hœtånna: given by people during a time of impurity caused by death etc. 11) aputrådyanna: from people without children, Brahmacarins, and ascetics. Within its elaborate structure, however, we still detect our two basic categories. The first two fall into the category of abhakßya (with milk, as usual, being an exception), and the remaining nine belong to the category of abhojya, both in the sense of food that is unfit (e.g., with hair or insect in it) and in the sense of persons whose food is not to be eaten. A significant point to note finally is how the specialization of meaning of the negatives abhakßya and abhojya influenced the meaning of their positive counterparts. The terms bhakßya and bhojya, often in the compound form bhakßyabojya, refer to food in general, often demarcating a distinction between hard and soft, or basic food and delicacies. But within the semantics of abhakßya and abhojya, the positives begin to connote the types of food sources (vegetables and meats) that are permitted and the types of “served food” that may be eaten and types of persons from whom it may be accepted.

Conclusion It appears to me incontrovertible that a specialized linguistic usage indicates an underlying intellectual structure or ideology.

Abhakßya and Abhojya

365

The larger question, which must await a longer and more detailed study of ancient Indian food prohibitions, is what bearing this distinction at the linguistic level has on the way ancient Indians in general, and more specifically Bråhmañas, viewed and constructed their world16 and society. I agree with Mary Douglas that a bearing it must have. As a preliminary guide, let me propose the following thumbnail sketch. The category of abhakßya, within whose purview I have not seen the pure/impure play any significant role, refers to the physical and biological world –the cosmology– constructed by ancient Bråhmañas.17 In the classification of abhakßya animals, we see the distinctions between atmosphere, earth, and water; between village and wilderness (gråma/årañya); between cloven-hoofed farm animals based on the paradigmatic cow and the uncloven work animals (horse); between paradigmatic birds that live up in the air and those that live and feed in water (web-footed); between carnivores and herbivores. The category of abhojya, on the other hand, is closely connected with social boundaries and with the pure/impure distinctions governing social relationships.18 Much of the anthropological work on food prohibitions in modern India has focused on food transactions –the bhojya-abhojya– from whom can food be accepted, and what types of food can be accepted from which types of people. This brings us to the very heart of the construction of Indian society, a society based on an intricate web of internal boundaries, the most prominent of which is caste; boundaries, however, that are much more complex than simply caste. Consider, on the one hand, the provision that allows a Bråhmaña to eat the food of his barber or a ‡üdra who is a family friend, and on the other, the one that forbids him from eating the food given by a Bråhmaña without children, a relative whose son has died, or his own wife during her menstrual period. Watching the CBS Morning News one lazy Sunday morning, I was struck by the sheer beauty of the creations made with flowers, leaves, and twigs by the noted Belgian florist Daniel Ost. But something this man, who probably had never even heard of Levi-Strauss, said struck me even more. “Once a flower or a leaf is taken out of its natural habitat, it becomes an object of thought; it becomes a medium of expression.” Human food, likewise, is a cultural construct and 16 I use “world” here as a shorthand for the way humans understand themselves and their experiences within the physical and the social world. I use it with the meaning given to it by Berger and Luckmann 1967, and Berger 1969. 17 This is the most we can say, because all our sources come exclusively from Bråhmañas. Given their leading role in the intellectual life of ancient India, however, we must assume that at least some features of the world they constructed were shared by others. 18 On the meanings of the terms for “pure/impure” and their connection to social boundaries, see Olivelle 1998b; see above, pp. 217-245. On the issue of caste and purity, see Dumont 1966; Quigley 1993.

366

LANGUAGE, TEXTS AND SOCIETY

a medium of thought and expression for individuals and cultures. Food, like animals, is good to think with –and the categories of bhakßya/abhakßya and bhojya/abhojya represent one example of the ways ancient Indians thought with food.

A

367

XVII Food for Thought: Dietary Rules and Social Organization in Ancient India*

The year was 1988. The South Korean government banned the slaughtering of dogs and the serving of dog meat in restaurants. The Olympics were coming to town and the South Korean government did not want to project to the world the image of an uncivilized dogeating nation. A few years ago in England demonstrators prostrated themselves in front of trucks carrying British horses to France for slaughter. Then there are the cows in India: even Macdonald’s had to change its menu when they went to India –lamb patties instead of the traditional beef. In insults too food looms large: the British call the French “froggies”, the Indians call despised groups “dog-cookers” (Leach 1964; Wijewardene 1968). Humans do take their food seriously, not just in a nutritional or medical sense but also in a cultural sense. Poor Koreans, a people with a long and distinguished culture, are forced to prove that they are “civilized” to the rest of the world by changing their food habits. Anthropologists have long noted that among the wide variety of edible vegetables and especially animals available to a given social group, only a small fraction is actually consumed. Prohibited foods, whether by injunction or by custom, far outnumber permitted ones in most societies. Even in modern times, entire ethnic and religious groups are often identified by what and how they eat, whether it is Kosher or Halal; and even in the ultra-modern and culturally liberated American society, you can go to any grocery store and find usually only two kinds of bird meat (chicken and turkey) and three kinds of animal meat (beef, pork, and, perhaps, lamb). The question the cultural historian, just as the social-anthropologist, asks is: Why? * Originally published as the 2001 Gonda Lecture. Amsterdam: Royal Netherlands Academy of Arts and Sciences, 2002.

368

LANGUAGE, TEXTS AND SOCIETY

The French anthropologist Lévi-Strauss (1963:89) put it in a nutshell when he said that animals are “good to think”. They are food good not only for eating but also for thinking. The pioneering studies of the British social-anthropologists Edmund Leach (1964) on the English and the Kachin and Mary Douglas (1966, 2000) on Jewish food prescriptions of the Leviticus have taken Lévi-Strauss’s work forward. The Sri Lankan anthropologist Stanley Tambiah (1969: 423-4) summarizes their conclusions: Lévi-Strauss (1966: 104) has formulated a view ... that ‘there is an empirical connection between marriage rules and eating prohibitions’. ... The theory [of Mary Douglas] argues that dietary prohibitions make sense in relation to a systematic ordering of ideas (a classification system) as exemplified for example by the abominations of Leviticus. Leach, partly building on Douglas, ... demonstrates for the English and the Kachin the correspondence between three scales of social distance from Ego, pertaining respectively to marriage and sex relationships, spatial categories, and edibility of animals.

I think there is broad agreement among scholars that A) animal categories and food prescriptions have deeper meanings in relation to a society’s construction of its cosmic and social world, and B) piecemeal attempts to justify the prohibition of individual animals and especially the use of what Mary Douglas (1966, 29) has called “medical materialism” –trichinosis, for example, to explain the prohibition of the pig– are inadequate as explanatory tools. Tambiah himself has provided a rich ethnography of food prohibitions in a village in northeastern Thailand, an ethnography that underscores the parallel between animals that can and cannot be eaten, persons one can and cannot marry, and the spatial organization of the family residence. Animals are such good food for thought because they lend themselves to classification. And classification, the placing of our experiences into distinct categories, is basic to human thought and rationality. The late E. H. Gombrich, the eminent art historian, postulates that “all thinking is sorting, classifying”; and the philosopher George Lakoff concludes that “An understanding of how we categorize is central to any understanding of how we think and how we function, and therefore central to an understanding of what makes us human.”1 For the anthropologist living in a village, it is somewhat easy to discover the classificatory systems underlying the dietary practices and social interactions in that village. He or she can observe the villagers, can talk to them, and if some points are unclear, can question them. Those of us studying ancient cultures do not have that luxury; our knowledge is limited by what people of those cultures chose to tell us in the extant historical documents they have left behind. Our knowledge is further complicated by the nature of those documents: 1

Cited by Smith 1994: 3-4.

Food for Thought

369

most are normative, most are written by an élite segment of society (in India the Brahmins), and almost all give the perspective of the male half of that society. To complicate matters further, while the anthropologist takes a snapshot of a particular time and place, our data come from different historical periods and often from unidentifiable geographical locations; we have to work diachronically. Heeding all these caveats, I want to do three things in this paper. First, I will examine the animal classifications in the early vedic texts, in the dietary rules of the later legal texts (the Dharma†åstras), and in the medical treatises.2 Second, I will present as systematically as possible the dietary regulations and animal prohibitions that are largely based on these classifications. Third, I will propose a hypothesis that relates these regulations to broader social categories and relationships, especially those relating to social hierarchy and marriage.

1. ANIMAL CLASSIFICATIONS Paralleling the cosmological classification of earth, water, and atmosphere, the early vedic sources already contain a threefold classification into land animals, birds, and fish, a classification that is maintained in the later dietary regulations (see Chart 1).3 The wellknown creation hymn, the Purußasükta (R¢gveda 10.90.8), on the other hand, presents a different classification: from the clotted butter of the primordial sacrifice were made the three kinds of beasts (pa†u): those of the air, those of the wilderness, and those of the village.4 The distinction between the village and the wilderness will become significant in the dietary regulations.

2 The early texts on Dharma go back probably to about the 3rd century BCE, the later ones to the middle of the first millennium CE. The two medical treatises, Caraka and Su†ruta, probably come from the first half of the first millennium c.e. For the sake of simplicity, I will call the Dharma†åtras “legal texts” even though their scope is much broader than the merely legal. 3 Later texts provide a comprehensive taxonomy of all living entities based on their modes of propagation: those born from a womb or placenta (jaråyuja), those born from eggs (añÛ∂aja), those born from warm moisture (svedaja), and those born from sprouts (udbhijja). See Månava Dharma†åtra 1.43-46; Chåndogya Upanißad 6.3.1. 4 “From that sacrifice, fully offered, was collected the clotted butter; it made the beasts of the air and of the wild, as well as those of the village.” tasmåd yaj∞åt sarvahutaΔ sa∫bhr¢ta∫ pr¢ßadåjyam Ù pa†ün tå∫† cakre våyavyån årañyån gråmya† ca ye ÙÙ. In this classification, birds are separated from the animals of the wild and the village. In the later dietary regulations, however, birds also are divided into the categories of wild and village. In the

370

LANGUAGE, TEXTS AND SOCIETY

Chart 1. ANIMAL CLASSIFICATIONS HABITAT

I.

Land Animals

II.

Village

Birds

Fish Wild

TEETH

Incisor Teeth in Both Jaws [ubhayatodat]

Incisor Teeth in One [lower] Jaw [anyatodat]

FEET

I.

Two-footed [dvipad]

II.

Hooves

Single-hoofed [eka†apha/ekakhura]

Four-footed [catußpad] Nails/Toes Double-hoofed [dvi†apha/dvikhura]

Five-nailed [pa∞canakha]

The most significant classifications, however, pertain to land animals, a category that also comprehends human beings. These classifications, moreover, have a ritual significance, because they specify which animals may be offered in sacrifice (medhya). In this they parallel the Jewish classifications in Deuteronomy and Leviticus. We have noted the distinction between village animals (gråmya) with the generic designation pa†u, and wild animals (årañyaka) with the generic designation mr¢ga. Only the former were medhya, fit to be sacrificed. Although wild animals and even birds are included in the list of animals at a horse sacrifice, they are not killed but released and their inclusion may have been purely for the sake of completeness so that the rite includes the sacrifice of all beings.5 Classifications pertaining to dental and foot structures (see Chart 1), however, are the most significant, and once again these parallel Purußa hymn, the term pa†u is given its widest application; it includes both birds and land animals. The term is generally restricted to land animals, however; and even among them, pa†u in the restricted sense applies to domestic animals, especially farm animals such as cows, goats, and sheep. The term mr¢ga, specifically meaning the deer or antelope, is used to cover the spectrum of wild animals, and thus stands in contrast to pa†u. The hymn is silent with regard to fish, which are a marginal category in most classificatory systems. For a detailed account of the vedic animal classifications and their connection to social structures, see Smith 1994: 241-86. 5 See, for example, Taittirœya Bråhmaña 3.9.1-4, where wild animals are said to be nonpa†u; they are not fit to be sacrificed.

Food for Thought

371

the Deuteronomic classifications based on the hoof and the chewing of the cud. The broadest of the two relates to dental structure: animals are divided into anyatodat (or ekatodat), literally those that have teeth only in one jaw but referring actually to animals that have incisor teeth only in the lower jaw, and ubhayatodat, those that have incisor teeth in both jaws. In the Purußa hymn it is said: “From it [the sacrifice] were born horses, as also whatsoever have incisor teeth in both jaws. Cows were born from it. From it were born goats and sheep.”6 Even though the text does not mention the class with incisors in the lower jaw, the anyatodat, it is clearly implicit; we have here a division of animals into two classes –those with two rows of incisor teeth headed by the horse and those with a single row of incisors, whose major representatives are cows, goats, and sheep, the most common farm animals. Again, although left unstated, the vast number of carnivorous animals with paws rather than hoofs also belong to the category of animals with two rows of incisors. With reference to the feet, we have the ancient distinction between the two-footed, that is, human beings, and the four-footed;7 but this did not play a significant role in ritual or dietary classifications. Far more significant is the classification based on the structure of the feet, according to which animals fall into two broad categories: those that have five toes with nails (the pa∞canakha) and those that have hoofs. Hoofed animals are further classified into those that are double-hoofed (or the cloven-footed; dvi†apha, dvikhura) such as cows, goats, sheep, and deer, and those that have single or unparted hoofs (eka†apha, ekakhura), namely horses, donkeys, and mules.8 A point to be noted is that the members of the single-hoofed category also fall into the category of animals having two rows of incisor teeth and are thus linked by their dental structure to the five-nailed carnivorous animals. The overlapping nature of the different classifications (see Chart 3) is deliberate, I think, and plays a significant role in the dietary regulations. With regard to birds, the early vedic literature, to my knowledge, offers no classificatory system; it appears only in the later legal and medical literature. Two classifications that birds share in common with land animals are the division into the village and the wild, and into the carnivorous and the non-carnivorous. “Village” in this context, howev6 tasmåd a†vå ajåyanta ye ke cobhayatodataΔ Ù gåvo ha jaj∞ire tasmåt tasmåj jåtå ajåvayaΔ ÙÙ. R¢gveda 10.90.10. See Månava Dharma†åtra 1.39-42 for a similar but more elaborate classification of fauna and flora within the context of a creation account. The ruminant animals, such as bovines and antelopes, have only a thick dental layer called “the dental pad” in the upper jaw. 7 See ¥gveda 3.62.14; Atharvaveda 2.34.1; Taittirœya Sa∫hitå 4.3.4.3; 5.2.9.4-5. 8 It is interesting to note that modern scientific classifications of mammals are also based on foot structures. The one-hoofed animals fall into the class Perissodactyla, animals with an uneven number of toes; but most members of this class, except for the Tapir and the Rhinoceros, have only one toe. The two-hoofed animals fall into the class Artiodactyla, animals with an even number of toes.

372

LANGUAGE, TEXTS AND SOCIETY

er, does not refer to domesticated or pet birds but birds that normally inhabit villages, such as crows.9 A classification based on their feeding patterns divides birds into those that feed by scratching with their feet (vißkira) and those that feed by pecking (pratuda). The former are ground feeders such as partridge, quail, and fowl; the latter probably refer to tree feeders, and medical texts associate them with fruit eaters. Another broad classification is based on foot structure; it divides birds into the web-footed (jålapåda) and those with unwebbed feet, the former clearly encompassing all water birds. Less common classifications include night birds, and red-footed and red-beaked birds.10 The two early medical texts, the Caraka Sa∫hitå and the Su†ruta Sa∫hitå, also contain animal classifications, again within the context of dietary regulations; but for them these regulations are related to the health benefits of various meats rather than to socio-religious prescriptions. These texts include all animals –land animals, birds, and fish– within their comprehensive classificatory systems. These have been examined in great detail by Zimmermann (1987), and I will present here only a brief summary. Caraka (1.27.35-87), the older of the two texts, gives an eightfold classification based partly on feeding patterns and partly on habitat and ignoring anatomical differences: 1) prasaha: those that take their food violently or forcibly. This category includes, as is to be expected, carnivorous animals and birds, but also cows, horses, and camels, because they tear off grass with their teeth. 2) bhümi†aya: animals that live in underground burrows, including snake, frog, porcupine and mongoose. 3) ånüpa: animals inhabiting marshy lands such as boar, rhinoceros, and elephant. 4) våri†aya: animals inhabiting water, including fish, turtles, and crocodiles. 5) våricara: animals operating in water, consisting of various kinds of water birds. 9 This is another significant difference between the vedic and later usages. The term gråmya in the vedic texts refers to domesticated animals, whereas in the later texts on food regulations the term usually refers to animals living in the village, which are forbidden, and not to farm animals, which are permitted. The differing meanings of gråmya in the vedic and post-vedic classficiations of animals may also reflect the changing meaning of the term gråma. Although it refers to a settled village in later Sanskrit, during the vedic period it is likely, as Rau (1957) has argued, that the term refers to a roving band of pastoral people who moved about with their animals. In this sense, the gråmya animals would clearly refer to these domesticated animals. Smith (1994, 254-55) is incorrect in assuming that gråmya is edible and the wild is inedible; this dichotomy is valid for the sacrifice (medhya) but not for food (bhakßya): see note 14 below. 10 For these classifications, see Åpastamba Dharmasütra 1.17.32-36; Gautama Dharmasütra 17.28-35; Baudhåyana Dharmasütra 1.12.1-7; Vasiß™ha Dharmasütra 14.48; Månava Dharma†åtra 5.11-14; Yåj∞avalkya Smr¢ti 1.172-75.

Food for Thought

373

6) jåõgala: animals inhabiting dry land, consisting mostly of different kinds of deer. 7) vißkira (scattering birds) and 8) pratuda (pecking birds), categories we have already encountered. The classificatory systems in the Su†ruta (1.46.28-70) are more complex and better systematized. It gives first a system of six classes, some of which are found also in Caraka: 1) jale†aya, aquatic. 2) ånüpa, inhabitants of marshy lands. 3) gråmya, those living in villages and for Su†ruta probably meaning domestic or farm animals. 4) kravyabhuj, carnivores. 5) eka†apha, animals with single or undivided hoofs. 6) jåõgala, animals inhabiting dry land. This order of enumeration, according to Su†ruta, relates to the health benefits of each class; the ones mentioned later are healthier than those mentioned earlier. Su†ruta then gives a more elaborate classification. All animals, including birds and fish, fall into two broad categories: jåõgala, inhabitants of dry land, and ånüpa, inhabitants of marshy areas. The first group is further classified into eight sub-groups: 1) jaõghåla: fleet-footed animals such as deer. 2) vißkira: scatterers. 3) pratuda: peckers.11 4) guhå†aya: animals living in caves, such as lions, tigers, and wolves. 5) prasaha: unlike Caraka, Su†ruta includes in this class only carnivorous birds. 6) parñamr¢ga: arboreal animals such as the monkey. 7) vile†aya: animals living in burrows. 8) gråmya: village and farm animals. The second group also is divided into five sub-groups: 1) külacara: animals living on banks, such as elephants, rhinos, boars, and some kinds of antelope. 2) plava: birds that dive or swim.12 3) koßastha: animals with shells, such as the conch. 4) pådin: aquatic animals with feet, such as crab and turtle. 5) matsya: fish, which are further subdivided into fresh water fish (nådeya) and sea fish (såmudra).

11 According to Su†ruta [1.46.68], the pratudas live on fruits. This may explain their pecking habit and distinguish them from the ground feeding vißkiras. 12 Note that plava is used elsewhere with reference to a specific bird, a type of crane.

374

LANGUAGE, TEXTS AND SOCIETY

The classifications given in the medical texts relate to the perceived positive and negative health effects of the animals of each class. This criterion contrasts with that of the legal texts, where animals are classified according to dietary regulations. Indeed, many kinds of animals that are prohibited in the legal documents are considered in the medical treatises not only as edible but also as providing health benefits.

2. DIETARY REGULATIONS I turn now to the dietary regulations found in the legal codes, the Dharma†åstras, regulations that relate closely to the classificatory systems we have examined. I want to begin, as any thick description should, with the vocabulary of these regulations. Although there are several terms used to prohibit the eating of a specific food item, two terms stand out both because of their frequency and because they assumed technical meanings within the discourse on diet. They are abhakßya and abhojya, two terms that have very similar connotations in common usage but have acquired specialized meanings within dietary regulations. Because I have discussed these terms in detail in another publication (Olivelle 2002B; above pp.351-66), I will only summarize my conclusions here. A) Abhakßya (“forbidden food”) refers to items of food, both animals and vegetables, that are completely forbidden; they cannot be eaten except under the most dire circumstances.13 Generally, lists of forbidden food contain food sources rather than cooked food served at a meal. Thus, carnivorous animals, village pigs and fowls, web-footed birds, garlic, red resins of trees are all abhakßya.14 B-i) Abhojya (“unfit food”), on the other hand, refers to food that is normally permitted but due to some supervening circumstances has become unfit to be eaten. Lists of unfit food items contain not food sources but food that is actually served at a meal. Thus, food contaminated by hair or insects, food touched by an impure man or woman, food that has turned sour or stale are all abhojya. B-ii) Abhojya takes on a secondary meaning referring not directly to food but to “a person whose food is unfit to be eaten;”15 for exam13 Manu (10.104-108) permits a man facing starvation to eat anything at all, giving the examples of Ajœgarta, Våmadeva, Bharadvåja, and Vi†våmitra. This falls within Manu’s description of åpaddharma, rules for times of adversity. The term abhakßya is also used in the Jain dietary regulations with reference to a standard list of twenty-two forbidden items (Cort 2001: 128-33). 14 The term abhakßya parallels amedhya, which refers to any food that cannot be offered in a sacrifice, and more generally to non-ritual food. The two categories, however, are overlapping but not identical. Unlike in the Jewish prescriptions, sacrificability does not imply edibility and vice versa. The horse, for example, is medhya but abhakßya, whereas a deer is bhakßya but amedhya. Brian Smith (1994: 254) in not quite accurate when he appears to identify sacrificability with edibility. 15 Grammatically, abhojya as “unfit food” is a Karmadhåraya compound, whereas abho-

Food for Thought

375

ple, a person during the period of impurity following a death (å†auca), a physician, or a hunter. Conversely, bhojya in the compound bhojyånna is a person from whom one may accept food. The Gautama Dharmasütra (17.6-7), for example states: “A man who looks after his animals or plows his fields, a friend of the family, his barber, and his personal servant –these are people whose food he may eat (bhojyånnåΔ), as also a merchant who is not an artisan.” When anthropologists working in the field describe food prohibitions and how they relate to caste divisions –that is, from whom can you receive food or water– they are dealing with abhojya. Given the importance of food transactions in modern Indian society, the abhojya kind of food regulation becomes significant. I will not address here issues relating to abhojya, however, both because one can do only so many things in a brief paper and because in the world of the Dharma literature there was little by way of restriction on food transactions between various social groups. Indeed, it was taken as a matter of course that a Brahmin could eat at the house of even a ‡üdra, entertain people of all classes, and even have ‡üdra cooks in his household. My focus today will be on the abhakßya kind of food prohibitions, which alone relate to the animal classifications we have examined. For this study I use the seven extant legal texts. They are, in approximate chronological order, the texts ascribed to Åpastamba, Gautama, Baudhåyana, Vasiß™ha, Manu, Vißñu, and Yåj∞avalkya.16 They span about seven or eight centuries, approximately from the 3rd century BCE to the 4th century CE Legal texts present five categories of permitted and forbidden foods : land animals, birds, fish, milk, and vegetables. My focus will be on the first four and, within them, especially on land animals; prohibited vegetables are few in number and limited to mushroom, garlic, onion, some kinds of leek, and secretions from tree barks, especially those that are red in color.

2.1 Land Animals Land animals, as we have noted, are subject to more complex and extensive classifications than either birds or fish. Given that we are dealing with a pastoral society, these are the animals that humans came into contact with most frequently; further, humans themselves are land animals and are included within these classificatory systems. Among land animals, two broad categories are prohibited (see Chart 2): animals with paws containing five nails (pa∞canakha) and animals with single or unparted hoofs (eka†apha). All our sources contain jya as “a person whose food is unift” is a Bahuvrœhi compound. 16 Åpastamba Dharmasütra 1.17.14-39; Gautama Dharmasütra 17.22-38; Baudhåyana Dharmasütra 1.12.1-15; Vasiß™ha Dharmasütra 14.33-48; Månava Dharma†åtra 5.4-26; Yåj∞avalkya Smr¢ti 1.170-78; Vißñu Smr¢ti 51.21-42.

376

LANGUAGE, TEXTS AND SOCIETY

these prohibited categories, with the exception of Vasiߙha, who is silent with regard to the single-hoofed. The only class of animals, therefore, that is permitted is the cloven-hoofed; these animals constitute the paradigmatic food. Five animals falling within the five-nailed category, however, are explicitly permitted:17 hare, hedgehog, porcupine, tortoise, and monitor lizard. Sometimes the rhinoceros is included in this list, although it has not five-nails but rather a tripartite hoof (see Jamison 1998). Four of our sources contain food prohibitions based on classifications with teeth as the criterion.18 As a class, animals with incisor teeth on both jaws (ubhayatodat) are forbidden. Animals with incisors only on the lower jaw (anyatodat), on the other hand, are permitted with some exceptions, the most notable of which are the camel and the village pig. Note that the permitted category on the basis of teeth, that is, animals with only one row of incisor teeth, coincides with the permitted category on the basis of foot formation, that is, animals with parted hoofs. Chart 2. FOOD REGULATIONS COMPLETELY FORBIDDEN

FORBIDDEN WITH EXCEPTIONS

PERMITTEDWITH EXCEPTIONS

Land Animals Single-hoofed animals [eka†apha/ekakhura]

Five-nailed animals [pa∞canakha]

Double-hoofed animals [dvi†apha/dvikhura]

Incisor teeth in both jaws [ubhayatodat]

Incisor teeth in one jaw [anyatodat]

Carnivores [kravyåd]

Herbivores

Village [gråmya] Unknown [aj∞åta] Solitary [ekacara] Camel Village Pig

Farm/Wild

Birds Carnivores

Village Unknown Solitary Nocturnal Village fowl

Web-footed [jålapåda]

Scatterers [vißkira] Peckers [pratuda] Wild

17 This exception is formulated in the well-known maxim: pa∞ca pa∞canakhåΔ bhakßyåΔ, “The five five-nailed may be eaten.” I will discuss these below under 2.5. 18 Åpastamba, Baudhåyana, and Yåj∞avalkya do not give this classification.

Food for Thought

377

All but Åpastamba and Gautama19 also contain prohibitions based on habitat. In general, animals living in villages (gråmya) are forbidden, whereas those living in the wild as well as farm animals designated as pa†u are permitted. This rule applies equally to land animals and birds, and it overrules all other criteria of permissibility. Thus, the pig or boar, which is permitted because it has both parted hoofs and a single row of incisors, is forbidden when its habitat is the village, whereas its wild counterpart is permitted (see Tambiah 1969, 450: Proposition 1). Likewise, the village fowl is forbidden even though it belongs to the permitted category of vißkira, birds that scratch with their feet. These provisions must have become sufficiently standard by the grammarian Pata∞jali’s time (2nd cent. BCE) for him to use a stock example repeatedly: abhakßyo gråmyakukku™o ’ bhakßyo gråmyasükaraΔ –“ it is forbidden to eat a village cock; it is forbidden to eat a village pig” (on Påñini 1.1.1 [5:16]; 1.1.1 [8:10]; 7.3.14 [320:22]). All but Gautama specify that carnivores are forbidden, even though most authorities apply this to birds rather than to land animals. The reason probably was that carnivorous land animals already fall into the proscribed categories of five-nailed animals and those with incisors on both jaws. Other broad categories of land animals and birds prohibited in some of our sources are the solitary (ekacara), the unknown (aj∞åta), and those that are not specified (anirdiß™a), probably referring to animals and birds not listed specifically as permitted.20

2.2 Birds As we have noted, some forbidden categories, such as carnivores and those living in villages, include both land animals and birds. Of the categories specific to birds, those that scatter their food with their feet,

19 These two authors, however, do prohibit the village pig and the village fowl. With regard to the varying meanings of the term “village” (gråmya), see note 9 above. 20 A†oka’s edicts, the earliest inscriptional evidence we have, confirm at least some of these dietary prescriptions. In his first rock edict, he says that at the time of writing only three animals were killed in the royal kitchen: two peacocks and one deer. The deer falls into the double-hoofed category, and the peacock to the scattering (vißkira). Alberuni’s observations in the early 11th century also broadly confirm the textual rules: “As matters stand thus, it is allowed to kill animals by means of strangulation, but only certain animals, others being excluded. The meat of such animals, the killing of which is allowed, is forbidden in case they die a sudden death. Animals the killing of which is allowed are sheep, goats, gazelles, hares, rhinoceroses, buffaloes, fish, water and land birds, as sparrows, ringdoves, francolins, doves, peacocks, and other animals which are not loathsome to man nor noxious. That which is forbidden are cows, horses, mules, asses, camels, elephants, tame poultry, crows, parrots, nightingales, all kinds of eggs and wine.” Alberuni’s India. tr. E. C. Schau, reprint (Delhi: Low Price Publications, 1989), II, p. 151. Prohibition of eggs is noteworthy because the legal texts are completely silent about eggs, from which we must deduce that eggs were practically never eaten in ancient India; laws generally forbid what is commonly done.

378

LANGUAGE, TEXTS AND SOCIETY

or more accurately, scratch the soil in search of food (vißkira), and birds that peck the food with their beaks (pratuda) are permitted to be eaten, whereas web-footed birds are generally forbidden.21 Gautama (17.19, 34) forbids night birds (nakta∫cara) and those with red feet or beaks (raktapådatuñ∂a), classifications that were not picked up by other authorities. Unlike land animals, however, bird classifications play a less prominent role in the lists. Many sources, such as Manu and Vasiß™ha, simply list individual birds that are forbidden or permitted, a pattern repeated in the case of fish.22 It was harder, I assume, to produce comprehensive classifications of birds and fish. Lawyers found it easier to simply list prohibited species. A comprehensive analysis of bird prohibitions will have to examine closely these forbidden species, some of which are difficult to identify from their Sanskrit names; space does not permit me to do so here. The problems inherent in bird classifications, however, are exemplified in the way later authorities, such as Vasiß™ha, Manu and Yåj∞avalkya, handle the two major categories of birds: those that scatter (vißkira) and those that peck (pratuda). All the more ancient sources declare that these birds are permitted, whereas these later authors state just the opposite.23 The reason for this reversal is unclear, but it may well have to do with a misunderstanding or a fresh interpretation of these classes. At least some commentators, for example, take the “peckers” to be carnivores, such as hawks and woodpeckers.24

2.3 Fish When it comes to fish, our sources, unlike the Hebrew authors of the Deuteronomy and Leviticus, totally give up on broad classifications, resorting instead to listing individual fish. Here also we detect an interesting change from the older to the younger sources. The older sources appear to assume that all fish, except those explicitly forbidden, may be 21 Gautama (17.35) states: “Birds that scratch with the feet or peck with their beaks and that do not have webbed feet may be eaten” (bhakßyåΔ pratudavißkiråjålapådåΔ). There are some exceptions, however; some kinds of cranes, such as the Lakßmaña, are permitted: Åpastamba 1.12.36. 22 Vasiß™ha (14.41-42), for examples, gives this list of forbidden fish: “These are forbidden to be eaten: among fish, the Ce™a, the Gavaya, the porpoise, the alligator, and the crab, as also grotesque fish and snake-head fish.” Vasiß™ha (14.48) has an even longer list of forbidden birds: “Among birds, the following are forbidden: birds that agitate, scratch with their feet or are web-footed; Kalaviõka sparrow, Plava heron, Ha∫sa goose, Cakravåka goose, Bhåsa vulture, crow, Påråvata dove, Kurara osprey, and Såraõga cuckoo; white dove, Krau∞ca crane, Krakara partridge, Kaõka heron, vulture, falcon, Baka egret, and Balåka ibis; Madgu cormorant, ~i™™ibha sandpiper, Måndhåla flying fox, and nocturnal birds; woodpecker, sparrow, Railåtakå bird, Hårœta pigeon, Kha∞jarœ™a wagtail, village cock, parrot, and ‡årika starling; Kokila cuckoo and carnivorous birds; and those living in villages.” 23 See Vasiß™ha 14.48; Manu 5.13; Yåj∞avalkya 1.172-73. 24 Sarvaj∞anåråyaña, for example (on Manu 5.13) takes vißkira to be birds that scratch the earth for worms. Govindaråja, Kullüka, and others take pratuda to be birds such as the woodpecker and nakhavißkira to be carnivores such as hawks.

Food for Thought

379

eaten.25 The only broad category of forbidden fish in them is the rather vague one of misshapen or grotesque fish (vikr¢ta). Manu and later writers, on the other hand, do not appear to favor the eating of fish. Indeed, in them the process is reversed; all fish except those listed are forbidden. Indeed, Manu (5.15) says: “A man who eats the meat of some animal is called ‘eater of that animal’s meat’, whereas a fish-eater is an ‘eater of every animal’s meat’. Therefore, he should avoid fish.”26 Yet, possibly bowing to tradition, even Manu (5.16) permits certain kinds of fish, especially in ancestral rituals: “The På™hœna and the Rohita fish may be eaten when they are used in an offering to gods or ancestors; Råjœva, Si∫hatuñ∂a, and Sa†alka fish may be eaten at any time.” One possible reason, besides the geographical locations of these authors, for this shift in attitude with regard to fish may have been the view that fish are by nature carnivores, as exemplified in the well-known maxim of the fish, the bigger fish eating the smaller fish.

2.4 Milk Milk is the most restrictive of all food items. The milk of most animals –women and other five-nailed animals, single-hoofed animals such as mares, and even most animals with parted hoofs, such as sheep and deer– is explicitly forbidden. The only permitted milk is that of cows, buffaloes, and goats.

2.5 Anomaly and Animal Classifications The subject of anomaly was the focus of Mary Douglas’s work on animal categories and purity. Impurity and anomaly imply a classificatory system within which they stand out as exceptions and exceptional; the impure is the “out-of-place”. When an animal breaches the boundaries of its assigned category, it becomes anomalous and the focus of attention and ritual. Birds without feathers, fish without scales, or cloven-footed animals that do not chew the cud were all anomalies to the Hebrew thinkers. Several classes and individual animals singled out for prohibition in the Indian legal texts, likewise, are anomalies within the Indian classificatory systems. Before examining the anomalous species, I want to look at a few animals that fall within the classifications we have examined (see Chart 3). The dog and the lion are representatives of five-nailed ani25 Baudhåyana is the exception. He lists permitted fish species, implying that others are forbidden. The lists contained in Baudhåyana, which are overly systematic, bear the stamp being reworked, however. Baudhåyana, more than any other ancient legal text, has been subject to reworking and additions. 26 yo yasya må∫sam a†nåti sa tanmå∫såda ucyate Ù matsyådaΔ sarvamå∫sådas tasmån matsyån vivarjayet ÙÙ

380

LANGUAGE, TEXTS AND SOCIETY

mals with two rows of teeth. Every feature of the dog makes it inedible, while the only feature of the lion that puts it in the edible category is its habitat, which is the forest. The goat and the deer represent the edible classes of animals, the former living in the farm and pasture land and the latter in the forest; they lack any inedible feature. It is apparent that a single negative feature is sufficient to make an animal inedible, as in the case of the camel; its village habitat alone makes it inedible. Chart 3 Dog

Lion

Goat

One row incisors Two rows incisors



Deer

Horse Camel Pig

Rhino











+

+

+







 +

Hoofed

+

+ 

Single-hoofed 

Double-hoofed



+

Triple-hoofed Five nails





Two horns

+

+

One horn

+

No horns

+

+

Carnivore



 

Herbivore Village



+

+

















Farm Forest



+









 = Prohibited feature  = Permitted features Bold = Prohibited Regular = Permitted Italics = Neutral

Moving beyond the paradigmatically inedible and edible classes, we come to the single-hoofed class of animals. Like the five-nailed carnivores, animals of this class have two rows of incisor teeth, but unlike the former they have hooves rather than nailed paws. Like

Food for Thought

381

their cousins in the farm and the field (pa†u), they have hoofs, but unlike the former, their hooves are whole and unparted and they have two sets of incisor teeth. Interestingly, the edible farm animal has double hoofs (dvi†apha) and a single row of incisors (anyatodat)—2+1, a pattern reversed in the single-hoofed animal, which has single hooves eka†apha) and a double row of incisors (ubhayatodat)— 1+2. Unlike the former, the latter do not have horns, a characteristic they share with the five-nailed animals. Finally, this class includes only two species–the horse and the ass, with the mule as a cross–and has no wild counterparts.27 The camel is another anomalous animal, and it is singled out for explicit prohibition. In the Leviticus it is singled out because, although it is cloven-footed it does not part the hoof. This feature of the camel’s anatomy is passed over in silence by the Indian texts; it is taken to be a full representative of the class with double-hoofs and one row of incisor teeth.28 As such it falls squarely within the edible class. But it is anomalous in that it is not a farm animal but lives in the village, and, like the horse and ass, it is also used exclusively for transport. Again like the horse and the ass, and unlike its double-hoofed cousins, it also has no horns. The camel falls into Tambiah’s (1969, 451) Proposition 3: “An animal that is placed in a class because it shares certain dominant properties of that class may yet be seen as exceptional or anomalous and therefore ambiguous as food or inedible (even if other members of its class are edible) if it shares one or more characteristics with animals of another class which carries strong values and is considered inedible.” As a transport animal living in the village and possibly near the house (see below 3.2.2), the camel shares characteristics with the inedible single-hoofed class of horse and ass. Besides the camel, the village pig and the village fowl are the two other animals singled out for prohibition. The pig fulfills all the requirements of the edible class; it is double-hoofed (dvi†apha) and has a single row of incisors (anyatodat). The fowl belongs to the edible class of scatterers (vißkira). Both species, however, have two varieties, the one living in villages and the other inhabiting forests. The former become anomalous because of habitat and fall under the category of forbidden animals. These animals, when they invade the human habitat of the village, fall into Tambiah’s (1969, 450) Proposition 2: “An unaffiliated animal, if it is seen as capable of leav27 Hunn (1979, 109) makes a similar argument for the anomalous status of the hyrax, hare, camel, and pig within the Hebrew taxonomy. They are all sole representatives of their classes within the Middle East. There is a wild ass in India, but it is limited in range to the deserts of the Rann of Cutch and Baluchistan (Prater 1997, 227). The animal is not noted in the ancient Indian texts. 28 Although camels have two incisor teeth on the upper jaw, the dharma texts regularly include it within the category of animals with one row of incisor teeth (anyatodat). Camels and donkeys, furthermore, were probably associated with lower classes. Brahmins, for example, are forbidden to ride in a cart drawn by a donkey or a camel: Månava Dharma†åtra 11.202.

382

LANGUAGE, TEXTS AND SOCIETY

ing its location or habitat and invading a location or habitat of primary value to man, will be the focus of strong attitudes expressed in the forms of (1) a food taboo and (2) a bad omen or inauspicious sign.” In the Indian case, however, even affiliated animals become the focus of such attention when they invade human space. Conversely, the anomalous is also a feature of animals that are permitted, even though they belong to categories that are forbidden. The five five-nailed animals (pa∞ca pa∞canakhåΔ) offer the classic example. The five animals are: porcupine, hedgehog, hare, tortoise, and monitor lizard (†våvidh, †alyaka, †a†a, kacchapa, godhå).29 Why are these five singled out as edible when other five-nailed animals are forbidden? Even making allowance for the artificiality of the alliterative pa∞ca pa∞canakhåΔ, there must be something that makes this group both distinctive and different from other five-nailed animals. Zimmermann (1987, 174) has given a plausible answer; both Caraka and Su†ruta list all but the tortoise among animals living in burrows (bile†aya, bhümi†aya). The tortoise and the monitor lizard are actually reptiles laying eggs, and distinguished from the mammals that are the most common five-nailed animals. Although I have used the term tortoise to translate the two Sanskrit words kacchapa and kürma used in our sources,30 the Sanskrit can refer to both the land dwelling tortoise and the aquatic turtle. The turtle in particular would have been viewed as anomalous within the five-nailed class of land dwelling carnivores. The tortoise/turtle also lack teeth, having only horny plates; thus it does not fall within the broad category of animals with two rows of incisors (ubhayatodat) to which all other five-nailed animals belong. Four animals of the group, moreover, are herbivores,31 whereas the five-nailed as a class is regarded as carnivorous. In their diet, habitat, and, in the case of some, anatomy, these animals breach the boundaries of the category to which they belong. That the singling out of the anomalous reinforces the categorical boundaries it violates is exemplified in the way legal texts explain the injunctive force of the oft-repeated rule: “The five five-nailed should be eaten” (pa∞ca pa∞canakhåΔ bhakßyåΔ). They consider this rule to fall into the category of parisa∫khyåvidhi developed in the exegetical tradition of Mœmå∫så. Rules of parisa∫khyåvidhi are actually negative or exclusionary rules presented in the garb of a positive injunction. Such an injunction states explicitly: “Do Y among X”, but what it actually means to say is: “Do not do X, excepting Y”, where Y is a sub-category of X. Thus when the rule says “The five five-nailed should be 29 As Zimmerman (1987: 174) has noted, the first two Sanskrit terms, †våvidh and †alyaka, are ambiguous; both can refer to porcupine, hedgehog, or pangolin. 30 Åpastamba, Gautama, Baudhåyana, Vasiß™ha, and Yåj∞avalkya use kacchapa. Only Manu uses kürma. It is unclear what, if any, distinction was conveyed by the two terms. 31 The monitor lizard (Varanus bengalensis) is a carnivore, although it is unclear whether the ancient Indians were familar with its feeding habits. On the monitor, see W. Auffenberg, The Bengal Monitor. Gainsville: University of Florida Press, 1994.

Food for Thought

383

eaten”, what it means to say is “No five-nailed animal, except the five, may be eaten.” The intent behind such an interpretation is that no one is obliged to eat even these five five-nailed animals, which would be the result if the rule was interpreted literally. Singling them out is intended solely to reinforce the prohibition of the five-nailed class. The rhinoceros is even a greater anomaly. It is neither single- nor double-hoofed but triple-hoofed, the only animal in India to have this feature. Although it falls into the category of odd-toed animals (Perissodactyla) just as the forbidden five-nailed class, it has only one set of incisor teeth. It has not two but only a single horn, again a unique feature in India that has drawn a lot of attention in religious writings.32 It also wanders alone and thus falls into the prohibited category of solitary animals. In the legal texts, the rhinoceros is listed frequently with the five five-nailed animals as exceptions to the prohibition of the five-nailed class, even though it does not have five nails.33 I think Zimmerman (1987: 183) is correct in assuming that the rhino’s meat was consumed primarily within the ritual context of an ancestral offering. Clearly, the rhino was not an animal that one could have for dinner on a regular basis. I am not aware of any evidence from ancient India that the rhino or its horn was viewed as an aphrodisiac, although one cannot discount the possibility that this widespread Asian belief may have been present also in India.34 It is evident, however, that symbols of fertility are prominent in ancestral rituals. In most dharma texts rhino meat is recommended as the preeminent food at an ancestral offering. These texts list the length of time that the ancestors will be satisfied with specific kinds of food, and they are unanimous in stating that when rhino meat is offered the satisfaction of the ancestors lasts for ever.35

3. ANALYSIS OF ANIMAL PROHIBITIONS I want to turn now to the social significance of the regulations regarding permitted and forbidden animals. The works of Edmund Leach, Mary Douglas, Stanley Tambiah, and others, even if one takes exception to some aspects of their arguments or conclusions, have demonstrated an unmistakable link between animal prohibitions in a given society and the symbolic worlds relating to the cosmos and soci32 For example, the very ancient Buddhist text Khaggavisåñasutta of the Suttanipåta. For an examination of a folk tale involving the rhinoceros and masculinity, see William S. Sax, “Fathers, Sons, and Rhinoceroses: Masculinity and Violence in the Påñ∂av Lœlå,” Journal of the Americal Oriental Society 117 (1997): 278-93. 33 For possible historical and linguistic reasons for the inclusion of the rhinoceros within this group, see Jamison 1998. 34 Caraka (1.27.84) says that rhino meat is balakr¢t (“giving strength”), where strength may well include sexual potency. Su†ruta (1.46.53) gives longevity as a characteristic of rhino meat. 35 Åpastamba 2.17.1; Gautama 15.15; Manu 3.272; Yåj∞avalkya 1.260; Vißñu 80.14. The skin and horn of rhinos were also used in ancestral rites.

384

LANGUAGE, TEXTS AND SOCIETY

ety created by that society. In the case of the Hebrews and their dietary regulations in the Leviticus, Mary Douglas relates the animal prohibitions to their central cosmogonic narrative which depicts god as setting boundaries to land, water, and air: To grasp this scheme we need to go back to Genesis and the creation. Here a three-fold classification unfolds, divided between the earth, the waters, and the firmament. Leviticus takes up this scheme and allots to each element its proper kind of animal life. In the firmament twolegged fowls fly with wings. In the water scaly fish swim with fins. On the earth four-legged animals hop, jump, or walk. Any class of creatures which is not equipped for the right kind of locomotion in its element is contrary to holiness.36

Animal prohibitions relate in a special way to individuals with whom sexual relationships and marriage alliances are forbidden. LéviStrauss (1966: 104) has formulated a conviction shared by many that “there is an empirical connection between marriage rules and eating prohibitions,” a view reiterated by Leach (1964: 42-43): “it is thus a plausible hypothesis that the way in which animals are categorized with regard to edibility will have some correspondence to the way in which human beings are categorized with regard to sex relations.”

3.1 The Ancient Indian Cosmos Given the complexities of ancient Indian cosmogonies and cosmologies, I want to limit myself here to some of their features that might have a bearing on dietary regulations. A theory that closely relates the cosmic process of rebirth to human diet is the famous doctrine of the five fires (pa∞cågnividyå) expounded in several vedic texts.37 According to this doctrine, a dead person rises up as the smoke of cremation, goes to the heavenly worlds of the sun and the moon, and returns in the form of rain. The rain water is absorbed by 36 Douglas 1966, 55. Hunn (1979, 104), who has attempted to integrate ecological and thus materialist principles into the explanation, supports Douglas’s argument: “Symbolic anthropologists rightly reject piecemeal explanations which seek to explain each animal’s role by reference either to ‘etic’ properties of the particular animal or its material impact on the human society involved. A different explanatory principle might need to be invoked for each animal. The assertion that pigs are prohibited as food because they may pose a threat of trichinosis is a familiar example. Not only are such approaches logically inelegant, they fail to provide an explanation for cultural variation, since immutable characteristics of the organism are cited by way of explanation. They explain too little and they explain too much: too little in their restricted application to only one, or a small set, among the species prohibited, too much in their implied universal applicability.” I agree with Hunn that it is impossible and unnecessary to integrate every food prohibition into an overarching model. Different explanatory models, even materialistic and ecological ones, may be applicable in specific cases, as we see in the Indian examples. 37 The classical formulation is found in Br¢hadårañyaka Upanißad 6.2.9-16; Chåndogya Upanißad 5.4-10. For further details, see Bodewitz 1996b; Schmithausen 1994; Olivelle 1999 (above pp. 13-51).

Food for Thought

385

plants, and when these are eaten by a man they are transformed into semen and deposited in a woman to take a new birth. This rebirth theory takes into account only the eating of vegetables, which is here the paradigmatic “food”; meat eating is outside its purview. In the case of vegetables, we have seen that there are very few dietary restrictions. Restrictions increase as we go up the food chain. We detect what could be called a “bottom line” in these restrictions: all land animals, birds, and fish that eat other animals rather than vegetables are strictly forbidden. Thus, when we go beyond the vegetable kingdom, the paradigmatic food is the meat of vegetarian animals.38 Unlike the Hebrews, then, in India it is not a question of whether a particular species has the physical characteristics specific to its habitat, but whether it participates in the specific characteristics of paradigmatic “food”: vegetables or vegetarian animals.39 The other significant aspect of Indian cosmology is the division of the land into village (gråma) and wilderness (arañya). The distinction of animals along these lines, we have seen, is already noted in the most famous account of cosmogony, the Purußasükta. In the dietary regulations, animals and birds of the village (gråmya) fall under a blanket prohibition. We have noted that “village” animals within these rules are distinguished from “farm” animals, which are the vegetarian animals providing the paradigmatic non-vegetarian food. This geographical divide provides a map of edible animals: they live in the field or the wild, not in the village. Taken together, these two principles are at the root of dietary regulations and underlie the rationale for the classification of animals. Five-nailed animals with two rows of teeth identify all the carnivores, the paradigmatic “eaters” that are “non-food”, whereas the double-hoofed animals with a single row of incisor teeth identify the herbivores, the paradigmatic “food” that are “non-eaters”. The anomalous single-hoofed category comprehend solely animals whose habitat is the village. When such classifications fail to exclude a particular animal, then they are singled out. So we have the double-hoofed camel singled out because its habitat is the village; the double-hoofed pig is also single out, because it is forbidden if its habitat is the village and permitted if it is wild. I think the old classification of birds into scatteres (vißkira) and peckers (pratuda) also points to an attempt at generalizing vegetarian birds. By contrast, the web-footed birds, besides breaching the divide between air and water, are viewed as carnivores by nature. Here also particular birds, 38 The medical texts, however, take a very different stand. Not only do they assume that carnivorous animals are eaten, they also give the medical and dietary properties of each and even call such meat the most nourishing. Zimmermann 1982, 159-79. 39 At least animals perceived as vegetarian, because it would have been difficult for ancient Indians to have an accurate understanding of a particular animal’s diet, especially in the case of wild animals. Indeed, vedic ideology views carnivors as “eaters” and the herbivors as “food”.

386

LANGUAGE, TEXTS AND SOCIETY

such as fowls, are singled out, because although they are edible because of their feeding habits, they become inedible if their habitat happens to be the village.

3.2 Diet and Society The connection between animal prohibitions and marriage regulations, a principle articulated by Lévi-Strauss, is never explicitly drawn in the Indian legal texts or by their commentators. Such connections, if they exist, fall into what Mary Douglas has termed “implicit meanings”. What is explicit in the texts, from the earliest vedic literature down to the legal texts, is the close parallel Indian thinkers have drawn between eating and power. Eating, for them, was a metaphor for the exercise of social power, control, and exploitation. This has been amply demonstrated by Brian Smith (1990). The vedic texts repeatedly call the Brahmins and especially the Kßatriyas –that is, the unified upper crust of society consisting of the priestly and political powers– the eaters, and the Vai†yas and the ‡üdras –that is, the commoner and working classes, including slaves– the food. The Kaußœtaki Upanißad (2.9) contains this petition: The Brahmin is one mouth of yours [i.e., of the creator god Prajåpati]; with that mouth you eat the kings. Make me a food-eater with that mouth. The king is one mouth of yours; with that mouth you eat the Vai†yas. Make me a food-eater with that mouth.40

In the Aitareya Bråhmaña (7.29) we see social hierarchy described from a Kßatriya viewpoint: a Brahmin is a seeker of food, to be moved at will; a Vai†ya is a tributary to another, eaten by another, oppressed at will; and a ‡üdra is the servant of another, removed at will, to be slain at will. The ‡atapatha Bråhmaña (13.2.9.6-8) compares the common people to the vagina and the royal power to the penis, the people to the corn and the royal power to the deer, juxtaposing nicely the acts of sexual and dietary dominance. This social fact of the higher and the powerful eating the lower and the weak is expressed in the well-known maxim of the fish (matsyanyåya), where the big fish eat the small fish. Manu (5.29) articulates this ecological principle well: “The immobile are food for the mobile; the fangless for the fanged; the handless for the handed; and the timid for the brave.”41 Regulations pertaining to marriage reflect, I believe, a similar ideology of hierarchy and power. The ‡atapatha Bråhmaña (1.3.8.5-6) calls 40 bråhmañas ta eka∫ mukha∫ tena mukhena råj∞o ‘tsi tena mukhena måm annåda∫ kuru Ù råjå ta eka∫ mukha∫ tena mukhena vi†o ‘tsi tena mukhena måm annåda∫ kuru Ù 41 caråñåm annam acarå da∫ß™riñåm apy ada∫ß™riñaΔ Ù ahastå† ca sahastånå∫ †üråñå∫ caiva bhœravaΔ ÙÙ This maxim of the fish is also invoked to legitimize the institution of kingship, because without a king “the stronger would grill the weak like fish on a spit” (Manu 7.20).

Food for Thought

387

the husband the eater and the wife the one to be eaten; the wife is the husband’s food. The legal texts permit a man to marry women only from groups hierarchically lower than his own. Here we encounter possibly the most significant principle of Dharma†åstric classification: the rule of anuloma and pratiloma. Literally, “along the hair” and “against the hair”, paralleling the English “with the grain” and “against the grain”, the rule establishes the legitimate and illegitimate directions of sexual relationships. When the male is hierarchically superior to the female, the union is anuloma and legitimate, whereas the opposite is pratiloma and illegitimate. Thus, for example, a Brahmin male may marry, besides a Brahmin woman, also women of the three lower classes, whereas a Kßatriya male may marry only from his own and from the two lower classes; he may not marry a Brahmin woman. Anthropological studies of both north and south India confirm the practice of hypergamy prescribed in the textual tradition. You eat your inferiors both in socio-political relations and in marriage alliances. Giving a girl in marriage is a public statement that the groom’s family is socially superior to one’s own.42 Let me, then, present a hypothesis about food prohibitions within Brahmanical ideology that correlates them first to the power structures of society and second to marriage regulations.

3.2.1 Dietary Regulations and Social Hierarchy The first is the easier to establish. Animals with two rows of incisor teeth and with five nails in each paw are considered to be eaters, being mostly carnivorous, and animals with cloven hoofs are considered to be food. The former are identified with the eating classes of society, namely Brahmins and Kßatriyas, and the latter with the paradigmatic food, namely the Vai†yas and by extension other lower classes. A vedic text makes this connection explicit: “Plants and trees are food and animals are food-eaters, for animals eat plants and trees. Among these [animals], those that have a double set of incisor teeth and are established as belonging to the same class as humans are food-eaters, other animals (pa†u) are food. Therefore, the former somehow surpasses the other animals, for the eater is somehow superior to the food” (Aitareya Årañyaka 2.3.1).43 We can relate this to social reality when we recognize that the paradigmatic pa†u, which in its restrictive sense refers to cloven-footed farm animals, is the goat; in all likelihood in ancient 42 Harmon (1985) has noted with reference to South India that ‡aivites present ‡iva as marrying Vißñu’s sister and thus as his brother-in-law, establishing an implicit superiority of ‡iva, the wife-taker, over Vißñu, the wife-giver. 43 oßadhivanaspatyo ‘nna∫ pråñabhr¢to ‘nnådam oßadhivanaspatœn hi pråñabhr¢to ‘danti Ù teßå∫ ya ubhaytodantåΔ purußasyånuvidhå vihitås te ‘nnådå annam itare pa†avas tasmåt ta itarån pa†ün adhœva caranty adhœva hy anne ‘nnådo bhavati. A similar thought is expressed in the ‡atapatha Bråhmaña (3.7.3.1-6) where the two-footed humans are seen as eaters and the four-footed as food.

388

LANGUAGE, TEXTS AND SOCIETY

India as today, goat was the most common meat. Within vedic classificatory systems, moreover, the goat is closely associated with the Vai†ya (Smith 1994, 270-73). Indeed, farm animals as a class are connected with the Vai†ya in that it is he who is their customary herdsman.44 The horse, a single-hoofed animal with two sets of incisors falling squarely into the forbidden category, on the other hand, is the Kßatriya animal par excellence. More generally, the “eaters” with two rows of incisors and five-nailed paws, are identified with the social “eaters”, the powerful upper classes of society.

3.2.2 Diet, Sex, and Marriage When we come to the marriage regulations, however, the picture becomes more complex and perhaps more interesting and significant. We can identify five broad categories of potential sexual partners who fall within the regulatory system presented in the legal texts, from groups closest to the Self or Ego to those who are farthest; in the columns of Chart 4-A, we move from the closer to the farther as we from left to right. The first group I will call sayoni, those sharing the same womb; but I include here individuals from same family. In an extended family, these would include grandparents, parents, siblings, brothers and sisters-in-law, and the like.45 These form a single unit sharing home, rituals, and property. Marriage or sex among members of this group is strictly forbidden; sex within this group is viewed as the most grievous of sins, one that causes loss of caste (påtaka). Next comes a somewhat ill-defined and elastic group of individuals who are closely related to the father or the mother and who fall under the categories of sapiñ∂a and sagotra.46 Marriage among members of this group is prohibited, but this prohibition is not absolute and there are significant exceptions. In the south, for example, cross-cousin marriage is permitted. The exact extension of these relationships, moreover, has been a matter of controversy. From the third group, occupying the middle and consisting of individuals not closely related yet belonging to one’s 44 Both farm animals and Vai†yas share the characteristic of the “many”. Brian Smith (1994:271) says: “Goats and sheep are animals that herd together, and it would seem to be partly their aggregation and numbers which lead to a connection between them and the Vaishya qualities of earthly abundance (bhüman) and material increase (puß™i).... The two species are said elsewhere to be the beasts which are ‘most manifestly’ like the prolific creator god Prajåpati, in that ‘they bear young three times a year and produce two [offspring] three times [per year]’ (‡atapatha Bråhmaña 4.5.5.6, 9).” Indeed, as Smith (1994: 270) observes, all the double-hoofed animals –cow, goat, and sheep– are connected with Vai†yas. Manu (9.328), morevoer, enjoins the Vai†ya: “ ‘I don’t want to look after farm animals’ –a Vai†ya should never entertain such a wish, and when there is a willing Vai†ya, under no circumstances shall anyone else look after them.” 45 Clearly sex between father and mother is permitted, because before marriage they did not belong to this group of forbidden persons. The prohibition is between siblings, between parents and children, and in an extended household, between father-in-law and daugher-in-law, and between brother-in-law and sister-in-law. 46 For information on these two kinds of kinsmen, see Kane 1962-75, II: 452-500.

Food for Thought

389

own class or caste, come the ideal marriage partners. Beyond this group are those belonging to social classes other than one’s own. Hypergamous marriage and sexual union with such individuals are permitted. The final group constitutes the fringes of society and includes a variety of outcastes, tribals, and foreigners (mleccha). Some legal texts label these groups collectively as antya or lowest; at least in the older legal texts ‡üdras do not fall into this group.47 We thus have total prohibitions at the two extremes of this spectrum –the closest and the farthest from the Ego– with the middle providing the most suitable marriage partners. Chart 4-A. Marriage Regulations Family [sayoni]

Relatives [sapiñ∂a/sagotra]

Group Members [savarña/sajåti]

Non-Group Members [asavarña/asajåti]

Outsiders [antya]

Forbidden Total

Forbidden Exceptions

Permitted Ideal

Permitted

Forbidden Total

Chart 4-B: Food Regulations LAND ANIMALS

Incisor Teeth in Both Jaws

Incisor Teeth in Lower Jaw

Single-hoofed Five-nailed [pa∞canakha] [eka†apha]

Double-hoofed Farm [dvi†apha, pa†u]

Double-hoofed Wild [dvi†apha, mr¢ga]

Unknown /Solitary [aj∞åta/ekacara]

Forbidden Total

Permitted Normal Food

Permitted

Forbidden Total

Forbidden Exceptions

Herd Animals Village [gråmya]

Carnivores [kravyåd]

Camel Village Pig BIRDS

Village [gråmya]

Village Fowl Forbidden Total

Carnivores [kravyåd] Web-footed [jålapåda]

Scatterers/Peckers [vißkira/pratuda]

Unknown /Solitary [aj∞åta/ekacara] Nocturnal [nakta∫cara]

Forbidden Exceptions

Permitted Normal Food

Forbidden Total

MILK

Cow, Goat Buffalo 47 Marriage with ‡üdra women is explicitly permitted. Significantly, the Artha†åstra (3.13.1-3) lists ‡üdra among Åryas, who were not subject to slavery, distinguishing him from Mlecchas, who were.

390

LANGUAGE, TEXTS AND SOCIETY

Animal prohibitions present a much more complex picture. There are multiple and overlapping categories, as we have seen, some based on habitat and diet and others on dental and foot structures. Nevertheless, I think we can detect a correlation between animal categories and social groups within the contexts of diet and marriage (see Chart 4-B). Animals with two rows of incisors, we have seen, are viewed as eaters and not as food. This broad group contains two divisions based on foot structure: those with five-nailed paws and those with undivided hoofs. Of these two, the single-hoofed, we have noted, is an anomalous class. We can draw a parallel between the forbidden single-hoofed class and the immediate family within which sex amounts to incest. Within both groups no exceptions are permitted48 and both are quite restricted groups. All animals of the single-hoofed class live in the village. Further, it is most likely that these animals, used for transportation and carriage, actually resided adjacent to the owner’s house;49 they were members of the family.50 It is this characteristic, we have noted, that makes the authors single out the camel as prohibited. Although it falls under the edible classes because of both foot and dental structures, it nevertheless lives in the village and probably near the home. These animals are also trained in a way other animals are not. They are vinœta, subject to vinaya, or the rules of training and good behavior, parallel to the rules to which humans are subject.51 All animals and birds grouped under “village” (gråmya) also come under a blanket prohibition; they can never be eaten. Eating the village fowl and the village pig, for example, are singled out as particularly abominable; these are anomalous in that, though they fall into the categories of scattering birds and two-hoofed animals that can be eaten, they happen to have the village as their habitat. Broadly then, we have two competing and overlapping criteria in animal classification, the one based on habitat and the other on anatomy. The imagined geography of ancient Indians can be viewed as consisting of concentric and ever widening circles emanating from the center, which is the village. Surrounding the village are the cultivated fields; beyond them are the pasture lands; then the forest. Even the forest, as Zimmerman (1987: 50) has noted, was probably differentiated; the deer and the antelope (mr¢ga), the double-hoofed object 48 The horse is, of course, an exception within the context of the vedic horse sacrifice. But this is never explicitly mentinoned in the dietary regulations. Manu (5.11) appears to allow for some exceptions, but this is a problematic verse and there is no explicit mention of the horse. 49 The well-known Hitopade†a story of the dog and the ass during a robbery illustrates this point (Book2, Story 2). The dog refused to awaken the master because he has not been well fed; so the ass attempts to awaken the master. This story illustrates the fact that the ass was kept near the house. 50 Tambiah (1969) also draws a parallel between animals that live under the house and persons with whom marriage is not permitted. 51 To some degree the draught ox also is trained, but I think their participation in the farm-animal group defines their status with regard to edibility.

Food for Thought

391

of the hunt, occupied “humanized” areas, sometimes even game reserves (mr¢gavana) for the royal hunt, whereas the deep dreaded forest was inhabited by vicious animals. We can see these circles emanating from the center, the Self/family in terms of marriage and the village in terms of food, where eating and marrying are totally forbidden. In the pasture lands and the deer fields live the edible animals, corresponding to the close and distant groups with which marriage alliances are permitted. The five-nailed animals with two rows of incisor teeth in a special way represent “us”; humans also belong to this group (see Aitareya Årañyaka 2.3.1 cited above). Significantly, however, some animals belonging to this group are permitted, such as the group of five. Here we have a prohibition with limitations, paralleling the marriage prohibition within the somewhat elastic boundaries of sapiñ∂a and sagotra, which also represents “us” in the sense of a kin group.52 Within Bråhmañical ideology, the paradigmatic food is the farm animal, technically called pa†u. It has double-hoofs and a single row of incisor teeth. These animals live neither in the wilderness nor in the village; they occupy the ideal not-too-distant and not-too-far geography of the farm and pasture land, paralleling individuals of one’s own class or caste who are not closely related. These represent the “other”, best for eating and for marrying. The double-hoofed farm animals have their wild counterparts, which are also permitted. Just as there is a blanket prohibition of village animals and birds, so there is a generic permission to eat wild animals and birds, so long as they fall within the permitted categories in terms of teeth, hoofs, and diet. When we eliminate the fivenailed and carnivorous animals, then, we are left with the Artiodactyla family encompassing wild oxen, sheep, goats, deer, antelopes, and hogs. These wild counterparts of farm animals are permitted, but it is clear that they do not constitute the normal or the normative meat; such wild animals, for example, are not medhya or fit to be sacrificed. Here we have, I think, a parallel to the third category of marriage partners, those from other classes or castes. They are not ideal, normative, or normal; but they are (sometimes grudgingly) permitted.53 Finally, we come to the fringe groups of society, the antya, with whom marriage alliances are forbidden. These groups find parallels in several categories of prohibited animals that belong to the fringes of 52

One common feature of prohibited categories, whether single-hoofed or fivenailed, is that they do not have horns, whereas all the permitted categories, both farm and wild, have a pair of horns. This is a feature, however, that is passed over in silence in our sources. 53 Indeed, a Brahmin husband is forbidden to employ a wife from another class as his ritual partner in sacrifices. Such wives are also, in a sense, amedhya, unfit to participate in the ritual.

392

LANGUAGE, TEXTS AND SOCIETY

the classificatory systems.54 First, there are the unknown (aj∞åta) animals, which are forbidden because their classification is unknown or dubious. Clearly, such animals belong to the wild category; it is unlikely that village or farm animals could have been categorized as unknown. Then there are the animals and birds that wander alone (ekacara) rather than in herds; these are also forbidden. It has been noted repeatedly that in traditional societies, and especially in India, marriage is not a union between two individuals but an alliance between two families or clans. One does not marry a persoan or eat an animal that lives alone and unattached to a larger group or herd.55

CONCLUSIONS The anomalous, when singled out for special attention, affirms and reinforces the normative categories. Animals normatively and normally fall into two broad categories: the eaters having incisor teeth in both jaws and five nails in each paw, and the food having a single row of incisor teeth and double hoofs in each foot; the former are meat eaters, the latter are grass eaters; the former eat the latter. As the many tragic endings of double-hoofed animals in the Pa∞catantra, the Indian book of animal tales, exemplify, there can be no friendship, no close social intercourse between grass-eaters and meat-eaters, between the eaters and the food. Marriage, however, appears to be quite a different matter; it is not so much a friendship as a consumption. At the broadest level, then, it is across these dividing lines that eating and marrying alike take place.

54 Although the dog is not singled out in the lists of prohibited animals, it is viewed with particular abhorence in Brahmanical literature. Its close association with the Cåñ∂åla in lists of diverse kinds appear to connect them also with the fringe groups of society. 55 The parallels to marriage restrictions in the case of birds, and especially fish, are more difficult to draw. By and large, I think, several of the criteria for the prohibition of birds and perhaps fish are drawn from those pertaining to land animals. Given the scope of this paper, however, I will not deal with regulations with respect to birds, fish, or milk. I agree with Hunn, however, in his criticism of Mary Douglas, that a comprehensive theory with regard to all food restrictions in a given society is not only impossible but perhaps undesirable. We must allow for the probability, that several criteria, including medical and ecological, may play a part in these regulations.

References

Aiyangar, K. V. Rangaswami. ed. 1941. Br¢haspatismr¢ti (reconstructed). Gaekwad’s Oriental Series, 75. Baroda: Oriental Institute. Alsdorf, L. 1950. “Contributions to the Textual Criticism of the Ka™hopanißad,” Zeitschrift der Deutschen Morgenländischen Gesellschaft 100: 621-37. Alter, J. 1992. “The Sannyasi and the Indian Wrestler: The Anatomy of a Relationship.” American Ethnologist 19: 317-36. Åpastamba Dharmasütra. Ed. with Haradatta’s commentary Ujjvalå by U. C. Pandeya. Kashi Sanskrit Series, 93. Varanasi: Chowkhamba Sanskrit Series Office, 1969. See also Bhler 1868, 1879-82, 1892. Archak, K. B. (Ed.) 1998. Manusmr¢ti and Women: A Collection of Research Papers. Dharwad: Shrutigeeta Prakashana. Bartlett, R. 1994. “Symbolic Meanings of Hair in the Middle Ages.” Transactions of the Royal Historical Society 4: 43-60. Berg, C. 1951. The Unconscious Significance of Hair. London: George Allen and Unwin. Berger, P. L. 1969. The Sacred Canopy. Garden City, N.Y.: Doubleday. Berger, P. and Luckmann, T. 1967. The Social Construction of Reality. Garden City, NY: Doubleday. Berreman, G. D. 1971. “The Brahmanical View of Caste.” Contributions to Indian Sociology (n.s.), 5: 16-23. Bhåskara. Commentary on the VeS. Ed. by Pandit Vindhyesvarœ Prasåda Dvivedin. The Chowkhambå Sanskrit Series, 70, 185, 209. Benares: 1915. Bodewitz, H. 1996. “Redeath and its Relation to Rebirth and Release,” Studien zur Indologie und Iranistik 20: 27-46. _____. 1996a. “Redeath and its Relation to Rebirth and Release,” Studien zur Indologie und Iranistik 20: 27—46.

396

LANGUAGE, TEXTS AND SOCIETY

_____. 1996b. “The Pa?cågnividyå and the Pit®yåna/Devayåna.” Studies on Indology: Professor Mukunda Madhava Sharma Feliciation Volume, ed. A. K. Goswami and D. Chutia, pp. 51—57. Delhi: Sri Satguru Publications. Böhtlingk, O. 1885a. “Bemerkungen zu Fhrer’s Ausgabe und zu Bhler’s Uebersetzung des Vâsiß™hadharmaçâstra.” Zeitschrift der Deutschen Morgenländischen Gesellschaft 39: 481-88. _____. 1885b. “Bemerkungen zu Bhler’s Ausgabe und Uebersetzung des Åpastambîjadharmasûtra.” Ibid. 39: 517-27. _____. 1885c. “Einige Bemerkungen zu Baudhâjana’s Dharmaçâstra.” Ibid. 39: 539-42. _____. 1885d. “Die Verbalwurzeln sku und skubh.” Ibid. 39: 328. _____. 1886a. “Bemerkungen zu Bhler’s Artikel im 39. Bande dieser Zeitschrift, S. 704 fgg.” Ibid. 40: 144-47. _____. 1886b. “Nachtägliches zu Vasish™ha.” Ibid. 40: 526. _____. 1889a (Ed. and trans.). Br¢hadârañjakopanishad in der Mâdhjaμdina-Recension. St. Petersburg: Kaiserliche Akademie der Wissenschaften. _____. 1889b (Ed. and trans.). Khândogjopanishad. Leipzig: Verlag von H. Haessel. _____. 1890. “Drei kritisch gesichtete und bersetzte Upanishad mit erklärenden Anmerkungen,” Berichte ber die Verhandlungen der königlich sächsischen Gesellschaft der Wissenschaften zu Leipzig, Philologisch-historische Classe. 42: 127-97. Contains Ka™ha, Aitareya, and Pra†na Upanißads. _____. 1897a. “Bemerkungen zu einigen Upanishaden.” Ibid., 49: 78100. _____. 1897b. “Kritische Beiträge.” Ibid., 49: 127-38. _____. 1898. “Kritische Beiträge.” Ibid., 50: 76—86. _____. 1901. “Kritische Beiträge.” Ibid., 53: 7-17. Breckenridge, C.A. 1986. “Food, Politics and Pilgrimage in South India, 1350-1650 A.D.” In Khare and Rao 1986. Brereton, J. 1981. The R¢gvedic Adityas. American Oriental Society, Vol. 63. New Haven: American Oriental Society. _____. 1986. “‘Tat Tvam Asi’ in Context,” Zeitschrift der Deutschen Morgenländischen Gesellschaft 136: 98-109. _____. 2004. “Dhárman in the R¢gveda.” In Olivelle 2004b, pp. 449-89. Bronkhorst, J. and Deshpande M. M. (Eds.) 1999. Aryan and NonAryan in South Asia: Evidence, Interpretation and Ideology. Harvard Oriental Series, Opera Minora Vol. 3. Cambridge, Mass.: Department of Sanskrit and Indian Studies, Harvard University. Bhler, G. (Ed.) 1868. Åpastamba Dharmasütram. 1st ed. Bombay: Education Society’s Press. 2nd ed. Bombay: Government Central Book Depot 1892. A third edition was published in Pune by the Bhandarkar Oriental Research Institute in 1932, but this is merely a reprint of the 2nd edition with numerous numerous additional printing errors.

References

397

_____. 1879-82 (Trans.). Sacred Laws of the Åryas [containing the Dharmasütras of Åpastamba, Gautama, Vasiß™ha, and Baudhåyana]. 2 vols. SBE 2, 14. Oxford: Oxford University Press. _____. 1885. “Einige Noten zu Böhtlingk’s Bemerkungen ber Fhrer’s Ausgabe und meine Uebersetzung des Vâsish™hadharmaçâstra.” Zeitschrift der Deutschen Morgenländischen Gesellschaft 39: 704-9. _____. 1886a. (Tr.). The Laws of Manu. The Sacred Books of the East, 25. Oxford: Clarendon Press. _____. 1886b. “Einige Bemerkungen zu Böhtlingk’s Artikel ber Åpastamba.” Zeitschrift der Deutschen Morgenländischen Gesellschaft 40: 527-48. Buitenen, J. A. B. van. 1962. The Maitråyañœya Upanißad: A Critical Essay, with Text, Translation and Commentary. Disputationes Rheno-Trajectinae, VI. The Hague: Mouton. _____. 1979. “Ånanda, or All Desires Fulfilled.” History of Religions 19: 27-36. Reprint: Studies in Indian Literature and Philosophy: Collected Articles of J. A. B. van Buitenen, ed. Ludo Rocher (Delhi: Motilal Banarsidass, 1988), pp. 323-32. Burnell, A. C. (Tr.) 1995. The Ordinances of Manu. Completed and edited by E. W. Hopkins. 1884. Reprint. Delhi: Mushiram Manoharlal. Cardona, G. 1997. “Vedic Tradition and Descriptions of Grammarians,” in Witzel 1997, pp. 33-38. Carman, J. and Marglin, F. A., eds. 1985. Auspiciousness and Purity. Leiden: Brill. Carrithers, M. 1985. “An Alternative Social History of the Self,” in The Category of the Person, ed. Michael Carrithers, Steven Collins, and Steven Lukes (Cambridge: Cambridge University Press), pp. 234-56. Charpentier, J. 1928-29. “Kå™haka Upanißad, Translated with an Introduction and Notes.” Indian Antiquary 58: 201-7, 221-29; 59: 1-5. Collins, S. 1982. Selfless Persons: Imagery and Thought in Theravåda Buddhism. Cambridge: Cambridge University Press. _____. 1998. Nirvana and Other Buddhist Felicities: Utopias of the Pali imaginaire. Cambridge: Cambridge University Press. Cooper, W. 1971. Hair: Sex Society Symbolism. New York: Stein and Day. Cort, J. 2001. Jains in the World: Religious Values and Ideology in India. New York: Oxford University Press. Daalen, L. A. van. 1988. “The Particle kila/kira in Sanskrit, Prakrit, and the Pali Jåtakas.” Indo-Iranian Journal, 31: 111—37. Daniel, V. 1984. Fluid Signs: Being a Person the Tamil Way. Berkeley: University of California Press. Das, R. M. 1962. Women in Manu and His Seven Commentators. Varanasi: Kanchana Publications.

398

LANGUAGE, TEXTS AND SOCIETY

Das, V. and Uberoi, J. S. “The Elementary Structure of Caste.” Contributions to Indian Sociology (n.s.), 5: 33-43. Dasgupta, S. 1922. A History of Indian Philosophy. 4 volumes. Cambridge: Cambridge University Press. Derrett, J. D. M. 1973. “Religious Hair.” Man (n.s.) 8: 100-03. _____. (Ed. and Tr.) 1975. Bhåruci’s Commentary on the Manusmr¢ti (The Manu-‡åstra-Vivaraña, Books 6-12). 2 vols. Wiesbaden: Franz Steiner. Deshpande, M. 1985. “Historical Change and the Theology of Eternal Sanskrit,” Zeitschrift fr vergleichende Sprachforschung 98: 122-49. _____. 1993a. Sanskrit & Prakrit: Sociolinguistic Issues. Delhi: Motilal Banarsidass. _____. 1993b. “The Changing Notion of ‡iß™a from Pata∞jali to Bhartr¢hari.” Asiatsche Studien 47.1: 95-115. _____. 1999. “What to do with the Anåryas? Dharmic Discourses of Inclusion and Exclusion.” In Bronkhorst and Deshpande, 1999, pp. 107-127. _____. Forthcoming-a. “Revisiting the Notion of ‡iß™a in Bhartr¢hari.” Proceedings of the International Seminar on Bhartr¢hari (Delhi 2003). Delhi: Motilal Banarsidass. _____. Forthcoming-b. “Source of Validation: Linguistic Usage or Received Rules of Grammar: Perspectives from the Late Sanskrit Grammarians.” In Boundaries, Dynamics and Construction of Traditions in South Asia, ed. Federico Squarcini. Firenze: Firenze University Press, 2005. _____. Forthcoming-c. “Changing Perspectives in the Sanskrit Grammatical Tradition and the Changing Political Configurations in Ancient India,” in Olivelle, forthcoming-c. Deussen, P. (Trans.) 1897. Sechzig Upanishad’s des Veda. Leipzig: F. A. Brockhaus. English tr. by V. M. Bedekar and G. B. Palsule. Sixty Upanißads of the Veda. 2 vols. Delhi: Motilal Banarsidass, 1980. _____. 1912. The System of Vedånta. Tr. C. Johnston. Reprint 1973 (New York: Dover). Dirks, N. B. 1990. “The Original Caste: Power, History and Hierarchy in South Asia,” in Marriott 1990, pp. 59-77. Doniger, W. tr. 1991. The Laws of Manu. London: Penguin. Douglas, M. 1984 [1966]. Purity and Danger: An Analysis of the Concepts of Pollution and Taboo. London: ARK Paperbacks. _____. 1982. Natural Symbols: Explorations in Cosmology, reprint of 1970. New York: Pantheon. _____. 2000. Leviticus as Literature. Oxford: Oxford University Press. Dumont, L. 1960. “World Renunciation in Indian Religions,” Contributions to Indian Sociology 4: 33-62. _____. 1980 [1966]. Homo Hierarchicus: The Caste System and Its Implications. Revised Edition (trans. M. Sainsbury, L. Dumont, and B. Gulati). Chicago: University of Chicago Press.

References

399

Dundas, P. 1985. “Food and Freedom: The Jaina Sectarian Debate on the Nature of the Kevalin,” Religion 15: 161-98, Dundes, A. 1997. Two Tales of Crow and Sparrow: A Freudien Folkloristic Essay on Caste and Untouchability. Lanham, Maryland: Rowman & Littlefield. Edgerton, F. 1953. Buddhist Hybrid Grammar and Dictionary. 2 volumes. Reprint. Delhi: Motilal Banarsidass, 1972. Eilberg-Schwartz, H. 1990. The Savage in Judaism: An Anthropology of Israelite Religion and Ancient Judaism. Bloomington: Indiana University Press. Emeneau, M. B. 1968—69. “Sanskrit Syntactic Particles–kila, khalu, nünam.” Indo-Iranian Journal 11: 241—68. Erdosy, G. ed. 1997. The Indo-Aryans of Ancient South Asia: Language, Material Culture and Ethnicity. Reprint of 1995. Delhi: Munshiram Manoharlal. Falk, H. 1986. “Vedisch upanißád.” Zeitschrift der Deutschen Morgenländischen Gesellschaft 136: 80-97. Fezas, J. 1996. “†u†rüßå dans l’Artha†åstra: obéissance, devoir conjugal ou maison du beau-père?” In N. Balbir and G-J. Pinault (eds.), Langue, style et structure dans le monde indien: Centenaire de Louis Renou (pp. 385-412). Paris: Librairie Honoré Champion. Firth, R. 1973. Symbols: Public and Private. London: George Allen and Unwin. Fitzgerald, J. L. 2004. The Mahåbhårata: 11. The Book of the Women, 12. The Book of Peace, Part One. Volume 7. Chicago: Chicago University Press. _____. Forthcoming. “Bhœßma’s Instruction of Yudhiß™hira in the Mahåbhårata: New Perspectives on the Development and Growth of the Epic Between the Empires,” in Olivelle, forthcoming-c. Frauwallner, E. 1953. Geschichte der indischen Philosophie, Vol. 1. Salzburg: Otto Mller. Tr. V. M. Bedekar. Delhi: Motilal Banarsidass, 1973. Frazer, J. G. 1913. The Golden Bough: A Study in Magic and Religion. 3rd Ed. London: Macmillan. Part II, pp. 252-87. Freed, R. S. & Freed, S. A. 1980. Rites of Passage in Shanti Nagar. Anthropological Papers of the American Museum of Natural History, Vol. 56, Pt. 3. New York. Frenz, A. (Ed. and trans.) 1968-69. “Kaußœtaki Upanißad.” IndoIranian Journal 11: 79-129. Fhrer-Haimendorf, C. von. 1971. “Tribes and Hindu Society.” Contributions to Indian Sociology (n.s.), 5: 24-27. Geiger, M. & W. 1920. Påli Dhamma vornehmlich in der kanonischen Literature. Abhandlungen der Bayerischen Akademie der Wissenschaften, Phil-philologische und historische Klasse, XXXI Band, 1. Munich.

400

LANGUAGE, TEXTS AND SOCIETY

Geldner, K. F. 1928 (1911) . “Die Lehre der Upanißaden.” In K. F. Geldner, Vedismus und Brahmanismus, pp. 109—69. Reprint. Tbingen: Verlag von J. C. B. Mohr. _____. 1951-57. Der Rig Veda. 4 vols. Harvard Oriental Series 33-36. Cambridge, Mass.: Harvard University Press. Gennep, Arnold van. 1960. The Rites of Passage. Tr. M. B. Vizedom and G. L. Caffee. Chicago: University of Chicago Press. Gomez, L. 1976. “Proto-Mådhyamika in the Påli Canon.” Philosophy East and West 26: 137-165. Gonda, J. 1955. “Reflections on Sarva- in Vedic Texts,” Indian Linguistics 16: 53-71. _____. 1956. “The Sœmantonnayana as Described in the Gr¢hyasütras.” East and West 7: 12-31. _____. 1959. Epithets in the R¢gveda. The Hague: Mouton. _____. 1960-61. “Prayata.” Bhåratœya Vidyå: Munshi Indological Felicitation Volume: 20-21: 45-51. _____. 1965a. “Bandhu in the Bråhmañas,” Adyar Library Bulletin 29: 1-29. _____. 1965b. Change and Continuity in Indian Religion. Disputationes Rheno-Trajectinae 9. The Hague: Mouton. _____. 1974. The Dual Deities in the Religion of the Veda. Amsterdam: North-Holland. _____. 1976. Triads in the Veda. Amsterdam: North-Holland. _____. 1981. “The Treatment of Hair Cuttings in the Gr¢hyasütras.” R¢tam (Akhila Bhåratœya Sa∫skr¢ta Parißad, Lucknow) 10 (1-2): 37-40. _____. 1985. The Ritual Functions and Significance of Grasses in the Religion of the Veda. Amsterdam: North-Holland Publishing Company. Gren-Eklund, G. 1978. A Study of the Nominal Sentences in the Oldest Upanißads. Studia Indoeuropaea Upsaliensia, 3. Uppsala: Almquist & Wiksell. Griffith, R. T. H. 1916—17. The Hymns of the Atharva-Veda. 2 vols. Benares: Lazarus. Hacker, P. 1965. “Dharma im Hinduismus.” Zeitschrift fr Missionswissenschaft und Religionswissenschaft 49: 93-106. _____. 1978 (1947). “‡aõkaråcårya and ‡aõkarabhagavatpåda: Preliminary Remarks about the Authorship Problem.” Kleiner Schriften, ed. L. Schmithausen, pp. 41—58. Wiesbaden: Steiner. Halbfass, W. 1988. India and Europe: An Essay in Understanding. Albany: State University of New York Press. Hallpike, Christopher R. 1969. “Social Hair.” Man (n.s.) 4: 256-64. _____. 1987. “Hair.” The Encyclopedia of Religion. Ed. M. Eliade et al., 6: 154-57. Hara, M. 1986. “The Holding of the Hair (Ke†a-grahaña).” Acta Orientalia 47: 67-92. Harman, W. 1985. “Kinship Metaphors in the Hindu Pantheon: ‡iva

References

401

as Brother-in-Law and Son-in-Law,” Journal of the American Academy of Religion 53: 411-30. Hauschild, R. (Ed.) 1927. Die ‡vetå†vatara-Upanißad: Eine kritische Ausgabe mit einer Übersetzung und einer Übersicht ber ihre Lehren. Abhandlungen fr die Kunde des Morgenlandes, XVII.3. Leipzig. Heesterman, J. C. 1957. The Ancient Indian Royal Consecration: The råjasüya Described According to the Yajus Texts. ’s-Gravenhage: Mouton. _____. 1968. “The Return of the Veda Scholar.” In Pratidånam, ed. J. C. Heesterman et al., pp. 436-47. The Hague: Mouton. _____. 1978. “Veda and Dharma.” In The Concept of Duty in South Asia, ed. W. D. O’Flaherty and J. D. M. Derrett, pp. 80-95. New Delhi: Vikas. Hershman, P. 1974. “Hair, Sex and Dirt.” Man (n.s.) 9: 274-98. Hertel, J. (Ed.) 1924. Muñ∂aka-Upanißad. Kritische Ausgabe mit Rodarneudruck der Erstausgabe (Text und Kommentare) und Einleitung. Indo-Iranische Quellen und Forschungen, III. Leipzig: H. Haessel Verlag. Hiltebeitel, A. 1981. “Draupadœ’s hair.” In Autour de la déesse hindoue, ed. M. Biardeau, pp. 179-214. Paris: Éditions de l’École des Hautes Études en Sciences Sociales. _____. 2001. Rethinking the Mahåbhårata: A Reader’s Guide to the Education of the Dharma King. Chicago: University of Chicago Press. _____. 2004. “The Nåråyañœya and the Early Reading Communities of the Mahåbhårata.” In Olivelle, Forthcoming-c. Horsch, P. 1967. “Vom Schoefpungsmythos zum Weltgesetz”, Asiatische Studien 21: 31-61. Translated as “From Creation Myth to World Law: The Early History of dharma.” In Olivelle 2004b, pp. 423-48. Hume, R. E. (Trans.) 1931. The Thirteen Principal Upanishads. 2nd ed.; Oxford: Oxford University Press. Hunn, E. 1979. “The Abominations of Leviticus Revisited: A Commentary on Anomaly in Symbolic Anthropology,” in Classifications in their Social Context, ed. R. F. Ellen and D. Reason, pp. 103-16. London: Academic Press. Insler, S. 1989—90. “The Shattered Head Split and the Epic Tale of ‡akuntalå.” Bulletin d’Études Indiennes, 7—8: 97—139. Jaini, P. S. 1979. The Jaina Path of Purufication. Berkeley: University of California Press. _____. 1991. Gender and Salvation: Jaina Debates on the Spiritual Liberation of Women. Berkeley: University of California Press. _____. 1993. “Fear of Food? Jaina Attitudes on Eating,” in Jain Studies in Honour of Jozef Deleu, ed. R. Smet and K. Watanabe. Tokyo: Hon-no-Tomosha, pp. 339-53. Jamison, S. W. 1988. “Våyav Indra† ca Revisited.” Mnchener Studien zur Sprachwissenschaft 49: 13-59.

402

LANGUAGE, TEXTS AND SOCIETY

_____. 1991. The Ravenous Hyenas and the Wounded Sun: Myth and Ritual in Ancient India. Ithaca: Cornell University Press. _____. 1996a. Sacrificed Wife/Sacrificer’s Wife: Women, Ritual, and Hospitality in Ancient India. New York: Oxford University Press. _____. 1996b. “Vedic menì, Avestan maëni, and the power of thwarted exchange,” Studien zur Indologie und Iranistic 20: 187-203. _____. 1997. “Sanskrit påriñåhya ‘household goods’: Semantic Evolution in Cultural Context.” Festschrift Eric Hamp, ed. D. Q. Adams. Vol. I. Journal of Indo-European Studies Monograph 23: 139-45. _____. 1998. “Rhinoceros Toes: Manu 5.17-18 and the Development of the Dharma System.” Journal of the American Oriental Society 118: 249-56. _____. 2000. “A Sanskrit Maxim and its Ritual and Legal Applications.” In Anusantatyai: Festschrift fr Johanna Narten zum 70. Geburtstag, ed. A. Hintze and E. Tichy. Mnchener Studien zur Sprachwissenschaft, Beiheft 19, pp. 111-125. Jolly, J. 1880 (Trans.). The Institutes of Vishnu. Sacred Books of the East, vol. 7. Oxford: Oxford University Press. _____. 1885. Outlines of an History of the Hindu Law of Partition, Inheritance, and Adoption. Tagore Law Lectures 1883. Calcutta: Thacker, Spink, and Co. _____. 1887 (Ed.). Månava Dharma-†åstra: The Code of Manu. Original Sanskrit Text Critically Edited According to the Standard Sanskrit Commentaries. London: Trbner. _____. 1913. “Artha†åstra und Dharma†åstra,” Zeitschrift der Deutschen Morganländischen Gessellschaft 67: 49-96. _____. 1923. (Ed. and trans.) Artha†åstra of Kau™ilya. 2 vols. Lahore: Motilal Banarsidass. Kane, P. V. 1933. Kåtyåyanasmr¢tisåroddhåra. Poona: Oriental Book Agency. _____. 1962-75. History of Dharma†åstra. I.1 (1968), 1.2 (2nd ed., 1975), II. 1-2 (2nd ed., 1974), III (1973), V.1 (1974), V.2 (1962). Kangle, R. P. (Ed. and trans.) 1960-65. The Kau™ilya Artha†åstra. Three Parts. Bombay: University of Bombay. _____. 1964. “Manu and Kau™ilya,” Indian Antiquary, 3rd series, 1: 48-54. Keith, A. B. 1914. The Veda of the Black Yajus School. Harvard Oriental Series, 18-19. Cambridge, Mass: Harvard University Press. Khare, R. S. 1976. Culture and Reality: Essays on the Hindu System of Managing Foods. Simla: Indian Institute of Advanced Study. _____. 1992.(Ed.) The Eternal Food: Gastronomic Ideas and Experiences of Hindus and Buddhists. Albany, NY: State University of New York Press. Khare, R. S. and Rao, M. S. A. 1986. (Eds.) Food, Society and Culture. Durham: Carolina Academic Press.

References

403

Kinsley, D. 1975. The Sword and the Flute: Kålœ & Kr¢ßña, Dark Visions of the Terrible and the Sublime in Hindu Mythology. Berkeley: University of California Press. Kubie, L. S. 1937. “The Fantasy of Dirt.” The Psychoanalytic Quarterly 6: 388-425. Lamotte, E. 1958. Histoire du Bouddhisme Indien: Des origines à l’ère ‡aka. Bibliothèque du Muséon 43; Louvain: Université de Louvain Institut Orientaliste. Lang, K. 1995. “Shaven Heads and Loose Hair: Buddhist Attitudes toward Hair and Sexuality,” in Doniger, W. and EilbergSchwartz, H. ed., Off with Her Head! The Denial of Women’s Identity in Myth, Religion, and Culture (Berkeley: University of California Press), pp. 32-53. Lariviere, R. W. (Ed. and trans.) 1989. The Nåradasmr¢ti. Two Parts. Philadelphia: Department of South Asia Regional Studies. _____. 2004. “Dharma†åstra, Custom, ‘Real Law’ and ‘Apocryphal’ Smr¢tis.” In Recht, Staat und Verwaltung im klassischen Indien, ed. B. Kölver, pp. 97-110. Munich: Oldenbourg, 1997. Reprinted in Olivelle 2004b, pp. 611-27. Laszlo, F. 1971. Die Parallelversion der Manusmr¢ti im Bhavißyapuråña. Abhandlungen fr die Kunde des Morgenlandes, XL, 2. Wiesbaden: Franz Steiner. Leach, E. R. 1958. “Magical Hair.” Journal of the Royal Anthropological Institute 88: 147-64. _____. 1964. “Anthropological Aspects of Language: Animal Categories and Verbal Abuse,” in New Directions in the Study of Language, ed. E. H. Lenneberg, Cambridge, Mass.: M.I.T Press, pp. 23-63. _____. 1971. “ ‘Esprit’ in Homo Hierarchicus.” Contributions to Indian Sociology (n.s.), 5: 13-16. Leslie, I. J. 1991. “A Problem of Choice: The Heroic Satœ or the WidowAscetic.” In Problems of Dharma: Rules and Remedies in Classical Indian Law, ed. I. J. Leslie, pp. 46-61. Panels of the VIIth World Sanskrit Conference, ed. J. Bronkhorst. Vol. IX. Leiden: E. J. Brill. Lévi-Strauss, C. 1963. Totemism. Boston: Beacon Press. _____. 1966. The Savage Mind. Chicago: University of Chicago Press. Limaye, V. P. and Vadekar, R. D. (Eds.) 1958. Eighteen Principal Upanißads. Poona: Vaidika Sa∫†odhana Mañ∂ala. Lincoln, B. 1977. “Treatment of Hair and Fingernails Among the Indo Europeans.” History of Religions 16: 351-62. Lingat, R. 1973. The Classical Law of India. Tr. J. D. M. Derrett. Berkeley: University of California Press. Lders, H. 1914. “Setaketu.” Festschrift Windisch. Reprinted in Philologica Indica Göttingen: Vandenhoeck & Ruprecht, 1940, pp. 346—61. Madan, T. N. 1971. “On the Nature of Caste in India: A Review

404

LANGUAGE, TEXTS AND SOCIETY

Symposium on Louis Dumont’s Homo Hierarchicus.” Contributions to Indian Sociology (n.s.), 5: 1-13. _____. 1988. Non-Renunciation: Themes and Interpretations of Hindu Culture. Delhi: Oxford University Press. _____. 1989. “Caste and the Ordering of Hindu Society,” in The Cambridge Encyclopedia of India, pp. 364-66. Cambridge: Cambridge University Press. Malamoud, C. 1972. “Observations sur la notion de ‘reste’ dans le bråhmanisme.” Wiener Zeitschrift fr die Kune Sdasiens 16: 5-16. _____. 1982. “On the Rhetoric and Semantics of Purußårthas.” In Way of Life: King, Householder, Renouncer. Essays in Honour of Louis Dumont, ed. T. N. Madan, pp. 33-54. Delhi: Oxford University Press. Mandlik, V. N. (Ed.) 1886. Månava-Dharma-‡åstra (with seven commentaries). Bombay: Ganpat Krishnaji’s Press. Marglin, F. A. 1985. “Female Sexuality in the Hindu World.” In Immaculate & Powerful: The Female in Sacred Image and Social Reality, ed. C. W. Atkinson, C. H. Buchanan, and M. R. Miles, pp. 39-59. Boston: Beacon Press. Marriott, M. (Ed.) 1968. “Caste Ranking and Food Transactions: A Matrix Analysis.” In Structure and Change in Indian Society, ed. M. Singer and B.S. Cohn, pp. 133-71. Chicago: Aldine. _____. 1990. India through Hindu Categories. New Delhi: Sage Publications. Maue, D. 1976. (Ed.) Br¢hadårañyakopnißad I: Versuch einer kritischen Ausgabe nach akzentuierten Handschriften der Kåñva-Rezension mit einer Einleitung und Anmerkungen. Doctoral Dissertaion, University of Gießen (unpublished). Maurer, W. H. 1986. Pinnacles of India’s Past: Selections from the R¢gveda. University of Pennsylvania Studies on South Asia, 2. Amsterdam/Philadelphia: John Benjamins Publishing Company. Mauss, Marcel. 1973 [1935]. “Techniques of the Body,” tr. Ben Brewster, Economy and Society 2.1: 70-88. Mayrhofer, M. 1953-76. KurzgefaBtes etymologisches Wörterbuch des Altindischen — A Concise Etymological Sanskrit Dictionary. Heidelberg: Carl Winter Universitätsverlag. Mines, D. P. 1990. “Hindu Periods of Death ‘Impurity’,” in Marriott 1990, pp. 103-30. Morgenroth, W. (Ed.) 1958. Chåndogya-Upanißad. Versuch einer kritischen Ausgabe mit einer Übersetzung und einer Übersicht ber ihre Lehren. Inaugural-Dissertation, Friedrich Schiller-Universität, Jena (unpublished). Mller, F. Max. (Trans.) 1879-84. Reprint 1981. The Upanishads. 2 parts. SBE 1, 15. Delhi: Motilal Banarsidass. Muller-Ortega, P. E. 1989. The Triadic Heart of ‡iva: Kaula Tantricism of Abhinavagupta in the Non-Dual Shaivism of Kashmir. Albany: State University of New York Press.

References

405

Nakamura, H. 1983. A History of Early Vedånta Philosophy. Tr. T. Legget, S. Mayeda, T. Unno, et al. Delhi: Motilal Banarsidass. Norman, K. R. 1997. “Dialect Variation in Old and Middle IndoAryan,” in Erdosy 1997, pp. 278-92. Oberlies, T. 1995. “Die ‡vetå†vatara-Upanißad: Einleitung-Edition und Übersetzung von Adhyåya I.” Wiener Zeitschrift fr die Kunde Sdasiens 39: 61-102. Obeyesekere, G. 1981. Medusa’s Hair: An Essay on Personal Symbols and Religious Experience. Chicago: The University of Chicago Press. _____. 1990. The Work of Culture: Symbolic Transformation in Psychoanalysis and Anthropology. Chicago: The University of Chicago Press. Oertel H. 1937. Zum altindischen Ausdrucksverstärkungstypus satyasya satyam ‘das Wahre des Wahren’ = ‘die Quintessenz des Wahren’. Sitzungsberichte der Bayerischen Akademie der Wissenschaften, Philosophisch-historische Abteilung, Jahrgang 1937, Heft 3. Munich: Verlag der Bayerischen Akademie der Wissenschaften (KS-O, II, 815-62). O’Flaherty, W. D. 1973. Asceticism and Eroticism in the Mythology of ‡iva. London: Oxford University Press. Olivelle, P. 1984. “Renouncer and Renunciation in the Dharma†åstras,” in Studies in Dharma†åstra, ed. Richard Lariviere (Calcutta: Firma KLM), pp. 81-152. _____. 1991. “From Feast to Fast: Food and the Indian Ascetic. In Problems of Dharma: Rules and Remedies in Classical Indian Law, ed. I. J. Leslie, pp. 17-36. Panels of the VIIth World Sanskrit Conference, ed. J. Bronkhorst. Vol. IX. Leiden: E. J. Brill. _____. 1993. The ņrama System: The History and Hermeneutics of a Religious Institution. New York: Oxford University Press. _____. 1995a. “Deconstruction of the Body in Indian Asceticism,” in Asceticism, ed. Vincent L. Wimbush and Richard Valantasis (New York: Oxford University Press), pp. 188-210. _____. 1995b. “Food in India.” Journal of Indian Philosophy 23: 367-80. _____. 1996a. (Trans.) Upanißads. World’s Classics Series. Oxford: Oxford University Press. _____. 1996b. “DharmaskandhåΔ and Brahmasa∫sthaΔ: A Study of Chåndogya Upanißad 2.23.1,” Journal of the American Oriental Society 116: 205-19. _____. 1997. “Orgasmic Rapture and Divine Ecstasy: The Semantic History of ånanda,” Journal of Indian Philosophy 25: 153-80. _____. 1998a. (Ed. and trans.) The Early Upanißads: Scholar’s Edition. New York: Oxford University Press. _____. 1998b. “Caste and Purity: A Study in the Language of the Dharma Literature.” Contributions to Indian Sociology 32: 189-216. _____. 1998c. “Unfaithful Transmitters: Philological Criticism and Critical Editions of the Upanißads,” Journal of Indian Philosophy 26: 173-87.

406

LANGUAGE, TEXTS AND SOCIETY

_____. 1999. (Trans.) Dharmasütras: The Law Codes of Ancient India. Oxford: Oxford University Press. _____. 2000. (Ed. and trans.) Dharmasütras: The Law Codes of Åpastamba, Gautama, Baudhåyana, and Vasiß™ha. Delhi: Motilal Banarsidass. _____. 2002a. “Structure and Composition of the Månava Dharma†åstra.” Journal of Indian Philosophy 30: 535-74. _____. 2002b. “Abhakßya and abhojya: An Exploration in Dietary Language,” Journal of the American Oriental Society , 122: 345-54. _____. 2004a. (Ed. and trans.) Manu’s Code of Law: A Critical Edition and Translation of the Månava-Dharma†åstra. New York: Oxford University Press. _____. 2004b. (Ed.) Dharma: Studies in Its Semantic, Cultural, and Religious History. Special issue of Journal of Indian Philosophy 32: 421-873 _____. 2004c. “Manu and the Artha†åstra: A Study in ‡åstric Intertextuality.” Journal of Indian Philosophy. 32: 281-91 _____. 2004d. “Semantic History of dharma: The Middle and Late Vedic Periods.” In Olivelle 2004b, pp. 491-511. _____. forthcoming-a. “Power of Words: The Ascetic Appropriation and the Semantic Evolution of dharma.” In Peter Fluegel and Gustaaf Houtman (eds.), Asceticism and Power in the Asian Context. London: Curzon. _____. Forthcoming-b. “Manu and Gautama: A Study in ‡åstric Intertextuality.” In Wilhelm Halbfass Commemoration Volume, ed. K. Preisendanz and E. Franco. Vienna: Verlag der Österreichische Akademie der Wissenschaften. _____. Forthcoming-c. (Ed.)Between the Empires: Society in India 300 BCE to 400 CE. New York: Oxford University Press. Orenstein, H. 1970. “Death and Kinship in Hinduism.” American Anthropologist 72: 1357-77. Orr, L. 2000. Donors, Devotees, and Daughters of God: Temple Women in Medieval Tamilnadu. New York: Oxford University Press. Padma, B. 1991. Costume, Coiffure and Ornaments in the Temple Sculpture of Northern Andhra. Delhi: Agam Kala Prakashan. Pandey, R. B. 1969. Hindu Sa∫skåras (Socio-Religious Study of the Hindu Sacraments). 2nd ed; Delhi: Motilal Banarsidass. Pandey, U. C. (Ed.) 1967. Yåj∞avalkyasmr¢ti (with the commentary of Vij∞åne†vara). Kashi Sanskrit Series 178. Varanasi: Chowkhamba Sanskrit Series Office. Parpola, A. 1981. “On the Primary Meaning and Etymology of the Sacred Syllable Om.” Proceedings of the Nordic South Asia Conference. Studia Orientalia 50: 195—213. Paul, D. Y. 1985. Women in Buddhism: Images of the Feminine in the Mahåyåna Tradition, 2nd ed. Berkeley: University of California Press. Pérez Coffie, C. A. 1994. Br¢hadårañyakopanißad II: Critical edition of the

References

407

second chapter of the Kåñva recension according to accented manuscripts with a critical-exegetical commentary. Harvard Ph.D. dissertation (unpublished). Pollock, S. 1985. “The Theory of Practice and the Practice of Theory in Indian Intellectual History.” Journal of the American Oriental Society 105: 499-519. _____. 1989. “The Idea of ‡åstra in Traditional India.” In The Sastric Tradition in the Indian Arts, ed. A. L. Dallapiccola and S. Z. Lallemant, pp. 17-26. Beiträge zur Sdasienforschung. Wiesbaden: Steiner. _____. 1990. “From Discourse of Ritual to Discourse of Power in Sanskrit Culture.” Journal of Ritual Studies 4: 316-45. _____. 1997. “The ‘Revelation’ of ‘Tradition’: †ruti, smr¢ti, and the Sanskrit Discourse of Power.” In Lex et Litterae: Studies in Honour of Professor Oscar Botto, ed. S. Lienhard and I. Piovano, pp. 395417. Alessandria: Edizioni dell’Orso. Prater, S. H. 1997. The Book of Indian Animals. Tenth impression of 1948. Bombay: Bombay Natural History Society. Preus, J. S. 1987. Explaining Religion: Criticism and Theory from Bodin to Freud. New Haven: Yale University Press. Quigley, D. 1993. The Interpretation of Caste. Oxford: Clarendon Press. Radhakrishnan, S. 1953. The Principal Upanißads. London: George Allen & Unwin. Råmånuja. Commentary on the VeS. Ed. and tr. by R. D. Karmarkar. 3 parts; University of Poona Sanskrit and Prakrit Series, 1. Poona: 1959-64. Tr. by G. Thibaut. SBE 68, 1904. Ramanujan, A. K. 1990. “Is There an Indian Way of Thinking? An Informal Essay,” in Marriott 1990, pp. 41-58. Rau, W. 1957. Staat und Gessellschaft im alten Indien nach den BråhmañaTexten dargestellt. Wiesbaden: Otto Harrassowitz. _____. 1960. “Bemerkungen zu ‡aõkaras Br¢hadårañyakopanißadbhåßya.” Paideuma: Mitteilungen zur Kulturkunde 7: 115-21. _____. 1964. “Versuch einer deutschen Übersetzung der ‡vetå†vataraUpanißad,” Asiatische Studien 17: 25-46. _____. 1965. “Versuch einer deutschen Übersetzung der Muñ∂akaUpanißad.” Asiatische Studien 18-19: 216-26. _____. 1970. Weben und Flechten im vedischen Indien. Akademie der Wissenschaften und der Literature. Abhandlungen der Geisten- und Sozialwissenschaftlichen Klasse, Nr. 11. Wiesbaden: Steiner Verlag. _____. 1971. “Versuch einer deutschen Übersetzung der Kå™hakaUpanißad.” Asiatische Studien 25: 158-74. _____. 1981. “Versuch einer deutschen Übersetzung der TaittrœyaUpanißad.” In Festschrift der wissenschaftlichen Gesellschaft an der Johann Wolfgang Goethe-Univerversität Frankfurt am Main (Wiesbaden: Franz Steiner), 349-73.

408

LANGUAGE, TEXTS AND SOCIETY

_____. 1997. “The Earliest Literary Evidence for Permanent Vedic Settlements”, in Inside the Texts Beyond the Texts: New Approaches to the Study of the Vedas, ed. M. Witzel, pp. 203-206. Harvard Oriental Series. Opera Minora 2. Cambridge: Harvard University Department of Sanskrit and Indian Studies. Renou, L. 1946. ‘“Connexion” en védique, “cause” en bouddhique.’ In Dr. C. Kunhan Raja Presentation Volume. Madras: The Adyar Library. _____. 1955. “Remarques sur la Chåndogya-Upanißad.” In Études védiques et påñininéennes. 17 vols. Paris: Publication de l’ICI. 1: 91-102. _____. 1964. “Sur deux mots du R¢gveda. 1. dharman. 2. jar-”, Journal Asiatique. 252: 159-167. _____. 1978. “La vie et le droit dans l’Inde: le dharma,” in L’Inde fondamentale, pp. 175-198. Paris: Hermann. Renou, L. and Silburn, L. 1949. “Sur la notion de ‘Brahman’,” Journal Asiatique. 1949, 7-46. Reynolds, R. 1949. Beards: Their Social Standing, Religious Involvements, and Decorative Possibilities, and Value in Offence and Defence Through the Ages. New York: Harcourt Brace Jovanovich. Rivière, P. G. 1969. “Myth and Material Culture: Some Symbolic Interrelations.” In Forms of Symbolic Action. Ed. R. F. Spencer, pp. 151-66. Seattle: American Ethnological Society. Salomon, R. 1981. “A Linguistic Analysis of the Muñ∂aka Upanißad.” Wiener Zeitschrift fr die Kunde Sdasiens. 25: 91-105. _____. 1991. “A Linguistic Analysis of the Pra†na Upanißad,” Wiener Zeitschrift fr die Kunde Sdasiens. 35: 47-74. _____. 1997. “On Drawing Socio-Linguistic Distinction in Old IndoAryan: The Question of Kßatriya Sanskrit and Related Problems,” in Erdosy 1997, pp. 258-77. ‡a∫kara. Commentary on the Vedånta Sütra. Ed. by Nåråyan Råm Åcårya. 3rd edition; Bombay: Nirñaya Sågar Press, 1948. Tr. by G. Thibaut. 2 pts. SBE 34, 1890; SBE 38, 1896. ‡a∫kara. Commentary on the Chåndogya Upanißad. ‡rœ†å∫karagranthåvaliΔ. Vol. 5. ‡rœraõgam: ‡rœvåñœvilåsamudråyantrålaya, n.d. Scharfe, H. 1993. Investigations in Kau™alya’s Manual of Political Science. Wiesbaden: Harrassowitz. Schmidt, H-P. 1980. “Ancient Iranian Animal Classification.” Studien zur Indologie und Iranistik 5-6: 209-44. _____. 1987. Some Women’s Rites and Rights in the Veda. Poona: Bhandarkar Oriental Research Institute. Schopen, G. 1997. Bones, Stones, and Buddhist Monks: Collected Papers on the Archeology, Epigraphy, and Texts of Monastic Buddhism in India. Honolulu: University of Hawaii Press. Schwarz, W. 1974-78. Rckläufiges Wörterbuch des Altindischen: Reverse Index of Old Indian. 4 parts; Wiesbaden: Otto Harrassowitz.

References

409

Segal, R. A. 1983. “In Defense of Reductionism.” Journal of the American Academy of Religion 51: 97-124. Senart, E., 1909. “Upås-upanißad” in Florilegium à Melchior de Vogé (Paris: Imprimerie Nationale, 1909), pp. 575-87. _____. 1930.(Ed. and trans.) Chåndogya-Upanißad. Paris: Société d’Édition «Les Belles Lettres». _____. 1934. (Ed. and trans.) Br¢had-Årañyaka-Upanißad. Paris: Société d’Édition «Les Belles Lettres». Seneviratne, H. L. 1992. “Food Essence and the Essence of Experience.” In Khare 1992, pp. 179-200. Albany. Shamasastry, R. (Tr.) 1961. Kau™ilya’s Artha†åstra. 7th ed. of 1915 original. Mysore: Mysore Printing and Publishing House. Shastri, S. R. 1959. Women in the Sacred Laws. Bombay: Bharatiya Vidya Bhavan. Smith, B. K. 1989. Reflections on Resemblance, Ritual, and Religion. New York: Oxford University Press. _____. 1990. “Eaters, Food, and Social Hierarchy in Ancient India: A Dietary Guide to a Revolution in Values.” Journal of the American Academy of Religion 58: 177-205. _____. 1994. Classifying the Universe: The Ancient Indian Varña System and the Origins of Caste. New York: Oxford University Press. Söhnen, R. 1981. “Die Einleitungsgeschichte der Belehrung des Uddålaka Åruñi: Ein Vergleich der drei Fassungen KaußU 1.1, ChU 5.3 und B®U 6.2.1—8.” Studien zur Indologie und Iranistik, 7: 177—213. Sprockhoff, J. F. 1979. “Die Alten im alten Indien: Ein Versuch nach brahmanischen Quellen,” Saeculum 30: 374-433. _____. 1981. “Årañyaka und Vånaprastha in der vedischen Literatur: Neue Erwägungen zu einer alten Legende und ihren Problemen.” pt.1. Wiener Zeitschrift fr die Kunde Sdasiens 25: 19-90. _____. 1984. “Årañyaka und Vånaprastha in der vedischen Literatur: Neue Erwägungen zu einer alten Legende und ihren Problemen.” pt. 2. Wiener Zeitschrift fr die Kunde Sdasiens 28: 5-43. _____. 1991. “Årañyaka und Vånaprastha in der vedischen Literatur: Neue Erwägungen zu einer alten Legende und ihren Problemen.” pt. 3. Wiener Zeitschrift fr die Kunde Sdasiens 35: 5-46. Sternback, L. 1974. The Månava Dharma†åstra I-III and the Bhavißya Puråña. Varanasi: All India Kashi Raj Trust. Talbot, C. 2001. Precolonial India in Practice: Society, Region, and Identity in Medieval Andhra. New York: Oxford University Press. Tambiah, S. J. 1969. “Animals are Good to Think and Good to Prohibit.” Ethnology 8: 423-59. _____. 1973. “From Varna to Caste through Mixed Unions,” in The Character of Kinship, ed. J. Goody, pp. 191-229. Cambridge: Cambridge University Press.

410

LANGUAGE, TEXTS AND SOCIETY

Thapar, R. 1990 [1984]. From Lineage to State: Social Formation in the Mid-First Millennium B. C. in the Ganga Valley. Bombay: Oxford University Press. _____. 1997 [1961]. A†oka and the Decline of the Mauryas. 2nd. edn. Delhi: Oxford University Press. Thieme, P. 1952. “Brahman,” Zeitschrift der Deutschen Morgenländischen Gesellschaft 102: 91-129. Reprint: Paul Thieme, Kleiner Schriften. Vol. 2, pp. 100-38. _____. 1963. “‘Jungfrauengatte’: Sanskrit kaumåraΔ patiΔ.” Zeitschrift fr vergleichende Sprachforschung auf dem Gebiete der indogermanischen Sprachen, 78:161-248. Also in Kleine Schriften (Wiesbaden: Franz Steiner, 1984), I: 426-512. _____. 1968a. “Åde†a.” In Mélanges d’Indianisme a la Mémoire de Louis Renou, pp. 715-23. Paris: Éditions E. de Boccard. _____. 1968b. “Ambrosia,” in Indogermanische Dichtersprache, ed. R. Schmitt (Darmstadt: Wissenschaftliche Buchgesellschaft), pp. 113-32. Originally published in P. Thieme, Studien zur indogermanischen Wortkunde und Religionsgeschichte. Berichte ber die Verhandlungen der Sächsischen Akademie der Wissenschaften zu Leipzig; Phil.-Hist. Klasse, Band 98, Heft 5 (Berlin: 1952), pp. 15-34. Trautmann, T. R. 1971. Kau™ilya and the Artha†åstra: A Statistical Investigation of the Authorship and Evolution of the Text. Leiden: Brill. Turner, B. S. 1984. The Body and Society. Oxford: Basil Blackwell. Turner, R. L. 1966. A Comparative Dictionary of the Indo-Aryan Languages. London: Oxford University Press. Uberoi, J. P. S. 1967. “On Being Unshorn.” In Sikhism and Indian Society pp. 87-100. Transactions of the Institute of Advanced Study, Vol. IV. Simla. Vajracharya, G. 1988. “Apala Versus Romasa: The Vedic View About the Hair and Fertility.” Paper presented at the American Oriental Society Meeting, Chicago, March 1988. Unpublished draft. Varenne, J. 1960. La Mahå-Nåråyaña Upanißad. 2 vols. Paris: Édition E. de Boccard. Vetter, T. 1990. “Some Remarks on Older Parts of the Suttanipåta,” in Earliest Buddhism and Madhyamaka, ed. D. S. Ruegg and L. Schmithausen, pp. 36-56. Leiden: Brill. Wezler, A. 1982. “Zum Verständnis von Chåndogya-Upanißad 5.1.12.” Studien zur Indologie und Iranistik, 8/9: 147—68. _____. 1993. “On a Prose Passage in the Yuktidœpikå of Some Significance for the History of Indian Medicine.” Journal of the European Åyurvedic Society 3: 282—304. _____. 2004. “Dharma in the Veda and the Dharma†åstras.” In Olivelle 2004b, pp. 629-54. This is revised translation of “Über den sakramentalen Charakter des Dharma nachsinnend.” In Raum-zeitli-

References

411

che Vermittlung der Transzendenz: Zur ‘sakramentalen’’ Dimension religiöser Tradition, ed. G. Oberhammer and M. Schmcker (Vienna: Österreichische Akademie der Wissenschaften, 1999), pp. 63-113. Whitaker, J. L. 2000. “Divine Weapons and tejas in the two Indian Epics,” Indo-Iranian Journal 43: 87-113. White, D. G. 1992. “You Are What You Eat: The Anomalous Status of Dog-Cookers in Hindu Mythology.” In Khare 1992, pp. 53-93. Whitney, D. 1890. “Böhtlingk’s Upanishads,” American Journal of Philology 11: 405-39 (review of Böhtlingk 1889a, 1889b). Wijewardene, G. 1968. “Address, Abuse and Animal Categories in Northern Thailand.” Man n.s., 3: 76-93. Willman-Grabowska, H. 1934. “Evolution sémantique du mot ‘dharma’,” Rocznik Orientalistyczny, 10: 38-50. Winternitz, M. 1926-28. “Dharma†åstra and Artha†åstra.” Sir Asutosh Memorial Volume. Patna. Vol. 1, 25-48. Witzel, M. 1984. “Sur le chemin du ciel,” Bulletin D’Études Indiennes 2: 213-79. _____. 1987. “On the Localisation of Vedic Texts and Schools.” In India and the Ancient World: History, Trade and Culture before A.D. 650, ed. G. Pollet, pp. 174-213. Orientalia Lovaniensia Analecta, 25. Leuven: Department OriÒntalistiek. _____. 1989. “Tracing the Vedic Dialects.” in Dialectes dans les littératures Indo-aryennes, ed. C. Caillat, pp. 97-265. Publications de L’Institut de Civilisation Indienne 55. Paris: de Boccard. _____. 1997. (Ed.) Inside the Texts Beyond the Texts: New Approaches to the Study of the Vedas. Harvard Oriental Series, Opera Minora 2. Cambridge, Mas.: Department of Sanskrit and Indian Studies. Yagi, T. 1994. “A Note on bhojya- and bhakßya.” In A Study of the Nœlamata: Aspects of Hinduism in Ancient Kashmir, ed. Y. Ikari, pp. 377-397. Kyoto: Institute for Research in Humanities, Kyto University. Zimmer, H. 1962. Myths and Symbols in Indian Art and Civilization. Ed. J. Campbell. New York: Harper. Zimmermann, F. 1987. The Jungle and the Aroma of Meats: An Ecological Theme in Hindu Medicine. Berkeley: University of California Press.

Index

abhakßya 351—66, 374—75 abhojya 351—66, 374—75 abortion 203, 259—60, 354 åcåra 153, 159—60, 164, 168—69, 191, 197 ade†a 68—69 adharma 140, 143—44, 146, 149, 150, 153 Åditya 86 Advaita Vedånta 70, 75—76, 215 Agni 90, 139, 142 agnihotra rite 25, 50 agniß™oma rite 161 åhavanœya fire 25 Ajåta†atru 43, 45—46, 116 Alberuni 377 All-gods 80 Alsdorf 290, 294—95 Alter 14, 323 amala 221, 239 Ambedkar 247 amedhya 221, 228, 237—40, 242, 391 am®ta/å: see immortality anadhyåya 242—43 anådya 351—52 ånanda 75—99 Ånanda 119 Ånandagiri 34 animals, classification of 369—74 anu†iß™a 34—36, 40 åpaddharma 211—12 Apålå 338 apeya 351, 354, 357 apüta 221, 235

aprayata 221, 234 Apsaras 80—81 arghya water 37 artha 174—76, 275 Årya 148, 163 åryåvarta 134, 158, 168—70 å†auca 220, 225, 229, 242 asceticism 53—74, 105, 120, 121—35, 151, 157—58, 166—67, 325, 331—37 ascetics 54—55, 108, 114, 147, 258, 340, 342—43, 346, 348—49 A†oka 127—29, 131—33, 137, 151—52, 157—58, 177, 192, 377 å†rama 53—57, 60, 73—74, 166—67, 185, 187—88, 335 Aß™åvakra 49 a†uci 220, 223, 225, 227—29, 233, 240, 242 a†uddha 220, 228, 232—33 a†vamedha 139, 141 A†vapati 25—26, 32, 41—42, 47 åtman 33, 76—77, 84—85, 90, 93—94, 99, 118, 144 Atri 199 Auffenberg, 382 Avadhüta 336—37 åyus 103, 105, 109—11 Bådaråyaña 76, 99 Baka 47 Bandin 49 Bålåki (D®pta Bålåki) 43, 45—46, 116 bee 68—69

414

LANGUAGE, TEXTS AND SOCIETY

begging 333—34 Behagel 85 Berg 342 Berger 321, 365 Berreman 218 bhakßya: see abhakßya Bharata 96 Bhåskara 55, 64, 66—67, 74 bhojya: see abhojya Bh®gu 48, 188, 195—97, 199, 207, 214— 15, 272 Biardeau 180 Bible 14 birth 118—19, 264—65, 332, 338, 385 Bodewitz 14, 16, 18—19, 25, 47—48, 108, 384 body 58, 61—62, 238, 321, 336 five 93—94 Bodhisattva 61—62 Böhtlingk 42, 44, 56, 63—65, 287—98, 301—20 Brahmå 46, 69, 101 brahmacårin 57, 60, 64—65, 243 brahmacarya 37, 47, 54—55, 64—65, 147, 167 Brahman 18, 44—45, 53—54, 60—61, 65— 74, 75—77, 87—89, 93—94, 98, 122, 125, 144 brahmastha 53—74 Brahmin 21, 23, 25, 37, 42—47, 49—50, 64, 75, 95, 121, 128, 131—34, 140—41, 143, 147, 151—52, 155—57, 161—62, 164—70, 176—77, 185, 187—188, 219, 224, 240, 244, 281—82, 323, 327, 331—33, 338, 340, 346, 353, 363, 365, 369, 386—87, 391 breath 18—20, 24—26, 146, 162, 180 see also faculties, pråñågnihotra Brereton 123, 137—38, 157, 297 B®haspati 139 Bronkhorst 14, 38, 41 Buddha(s) 46, 49, 61, 63, 119—20, 121— 22, 128—29, 133, 150—52, 164, 196, 333, 335, 337 Buddhism 59, 61—63, 95—98, 113, 117, 119—20, 122, 127, 129—32, 134, 137, 143—44, 149—53, 157—58, 164, 170, 173, 177, 195—97, 248—49, 257, 331, 335, 349—50 Bhler 50, 180—81, 198—99, 205—207, 261, 268, 275, 278, 282—283, 302— 20, 357—58 Burnell 278, 282 cakravartin 122, 150

Cåñakya 327 Candragupta Maurya 327 Cardona 320 Carman 245 Carrithers 103—04 caste 217—19, 245, 247—48, 331, 333—34, 365 castration 342 caula 328—29 Charpentier 294 Citra Gåõgyåyani 13—51 Collins 105 commentaries 287—99, 300—20 Cort 374 critical editions 216, 287—99, Cüla Bhågavitti 23 Daalen 36 Da¬hanemi 335 Damayantœ 258 Daniel 330 dar†apürñamåsa 48 daughter 23, 153, 251—52, 255 Dasgupta 97 death 101—20, 144, 166, 254, 264—65, 330, 348 paths of the dead 13—51 personified 33, 37, 50 see also funeral rites Derrett 278 de†a (evidence) 276——79 Deshpande 34—35, 134, 169—70, 243, 294, 297, 312, 320 Deussen 40, 42, 55—56, 65, 76, 292—93 Devåpi 142, 153 dharma 54, 56—57, 60—61, 63, 65, 74, 121—35, 137—54, 155—77, 182—85, 187—93 195—97, 206—207, 215, 275 dharman 123, 138 dharmaskandha 53—74, 147 Dhaumya Åyoda 13 dhruva 138—39 dietary regulations 351—66, 367—92 Dirks 218 Doniger 278, 282 Douglas 105, 238, 241—42, 244—45, 321, 337, 342, 351, 365, 368, 379, 383— 84, 386, 392 Draupadœ 255, 327 drinking (and liquor) 78, 81, 86, 98, 161—62, 253 DuΔßanta 107 Dumont 118, 217—19, 229, 241—42, 244, 365

Index

Dundas 333 Dundes 218 Dußyanta 258 Dyutåna Marüta 138 Edgerton 62 Eggeling 82, 84, 110, 143, 338—40 Eilberg—Schwartz 331 Emenau 36 excrement 226—28, 243, 337—38, 346 execution, simple and colorful 283 faculties 19 contest among 15—20, 24 Falk 69, 154 fathers, see manes Fezas 276 fire fire doctrine 15—51, five fires 13—51, 92, 94 Firth 325, 327, 331, 337 Fitzgerald 152, 158, 177, 180, 272 food 18—21, 24—25, 33, 83—87, 91, 94— 95, 102, 105, 108, 114, 126, 333—34, 351—66, 367—92 prohibitions 351—66, 367—69, 374— 83 see also dietary regulations Fhrer 309 funeral rites 329 Frazer 341—42 Frenz 43, 288 Frauwallner 18 Freed and Freed 329 Frer—Haimendorf, von 218 Gådhi 197 gambling 80—81, 98, 191, 210 Garbe 294 Gårgœ Våcaknavœ 116—17 Geldner 42, 59, 77—78 gender 102, 104, 247, 321—50 Gerow 202 gods 21, 25, 58, 64, 77, 79—81, 83—84, 108—12, 119, 139—40, 143, 145—46, 149, 338, 341 Gombrich 368 Gomez 130 Gonda 43, 69, 71—72, 102, 124, 234, 328, 339 Gramsci 105 Gren—Eklund 84 Griffith 79 guests 37

415

Halbfass 123 Hallpike 324—31, 342—43, 345, 347 hair 321—50, 353—54, 374 shaving 327—34, 348 Hara 340 Haradatta 301—20 Harmon 387 Hartman 41 Hauschild 289, 294, 296—97 heaven 103, 105, 111—13, 115 Heesterman 124, 160 Hertel 289, 294, 296 Hershman 236, 329, 336—38, 342—43 Hiltebeitel 152, 158, 177, 180, 196, 272, 326—27 Hinber, von 153 Hirañyagarbha 197 homosexuality 259 Hopkina 179 Horsch 122—23, 137, 157, 194 Hume 42, 53, 55—56, 65, 296 Hunn 381, 384, 392 husband 112, 116, 191, 249—60, 332, 387 immortality 73, 78, 101—20 am®ta 102, 106, 109—10 am®tå 115 am®tatva 71, 106 impurity 219—45, 264—66, 270, 337—38, 353, 365 Indra 37, 39, 47, 59, 77, 82, 86, 139—43, 196, 338 Indråñi 82 initiation 328—29 reintegration from 329 upanayana 328—29 inheritance 251—52, 271 Insler 46, 142 Jåbåla, Satyakåma 19—20, 23—24, 33, 37—38, 47—48 Jaimini 99 Jaini 120 Jainism 95, 97—98, 113, 117, 119—20, 248, 257, 331—33 Jala 49 Jamison 35, 78, 83—86, 102, 112, 181, 248, 250, 253, 261, 352, 383 Janaka 17, 37, 45, 49—50, 89, 116 Janaka Åyasthüna 23 Jåna†ruti 37, 39, 47 jati: see caste Jåtukarña 48—49 Jolly 214, 252, 275—76, 278, 361—62

416

LANGUAGE, TEXTS AND SOCIETY

Kaho∂a 49 Kaliyuga 34 Kalmåßapada 363 Kane 55—57, 65—66, 109, 156—57, 221, 261, 272—73, 275, 302, 330—31, 333, 339, 357, 388 Kangle 276, 278—79, 282 kañ™aka†odhana 191, 210—11 karma 185, 193—94, 196, 214 Kashikar 148 Kåtyåyana 153, 155 Kåtyåyanœ 115 Kau™ilya 121 Keith 80, 141 Khare 220 king 13—14, 36—37, 40—42, 45—51, 116, 122, 124—29, 131, 137, 139—45, 149— 50, 163, 172, 174—75, 185, 188—92, 209—11, 223, 275—85, 325, 329, 335, 386 Kinsley 326 Kishwar 247 K®ßña 122, 201 kßatra 125—26, 131, 140—41, 144—45, 150—51 Kßatriyas 23, 25, 37, 42—43, 45—48, 125, 141, 189, 258, 386 Kubera 197 Kubie 241 Lakoff 368 Lakßmana 96 Laszlo 208 Lamotte 62—63 Lariviere 152—53, 159—50, 171, 179—80, 205, 248 laukika injunctions 161—62 law 56, 123—24, 142—45, 150, 162, 174— 76, 180—81, 189—92, 209—10, 223, 248—85, 330—31, 333, 374—83 Leach 218, 335, 337, 342—43, 367—68, 383—84 lectio difficilior 292, 295—97 Leslie 333 Levi—Strauss 322, 365, 368, 384, 386 liberation (mokßa) 103, 105, 113—16, 119—20 Limaye 288, 293 Lincoln 328, 339 Lingat 156—57 Luckman 321, 365 Lders 49 Madan 217—20, 243, 245

Mådhava 51 Madhuka Paiõgya 23 mahåbhißeka 140 Mahåvœra 122 Makhådeva 335 Maitreyœ 66, 115—17 mala 221, 238 Malamoud 72, 236 manas 89—90 manes 21, 64, 104—09, 114, 118 mantha rite 15, 19—24 Manu 46, 113, 117, 152—53, 159, 181— 216, 272 Marglin 245 Marœci 196—97, 207 marriage 105—109, 111—12, 114, 116, 120, 127, 149, 166, 191, 247—60, 324, 329—31, 336, 368—69, 384, 387—92 Marriot 218, 220 Maue 288 Maurer 58 Mauss 103, 321 Mayrhofer 58 medhya 221, 237—38, 370, 372, 391 Mehendale 92 menstruation 259—60, 326—27, 332, 348 Mœmå∫så 157—60, 171, 176 Mines 229 Minkowski 196 Mitra 123—24, 138—39, 145 Mitra-Varuña 123—24, 138 moi 103—105, 110—11, 113, 115, 118—19 moon 21 Mller 42, 55—56, 65, 202, 261, 275, 289, 294 Muller-Ortega 76 Nåciketas 33, 37, 50, 294 Nakamura 76 Nala 258 Nårada 39, 47, 196 nirmala 221, 239 niyoga 201—02 Norman 320 numerology 71—73 Oberlies 288 Obeyesekere 322, 334, 337, 340, 342— 46, 349 Oertel 78, 126, 150 O‘Flaherty 83 Oguibénin 38 Orenstein 229 orgasm 76—77, 82, 87, 92, 98, 107

Index

Orr 250 Padma 324 Pa∞cålas (Kuru—Pa∞cåla) 39, 45—46 Påñ∂ava 335 Pandey 329 Parameß™hin 142 Parpola 35 parßñœgraha 283 Paul 120 Paußya 337 pavitra 221, 230, 235—36 påvana 221, 230, 235—36 penance 192—193, 212—14, 271, 330, 361 penis 75, 79, 81, 85—89, 95, 97, 107, 331, 342, 345—46, 386 Pérez—Coffie 288 personne 103—106, 108, 110, 113—14, 118 pit®medha 87 Pollock 114, 160, 164, 172, 176, 248 pollution 217, 330—31 see also impurity power 121—35, 219 prak®ta (official) 282—83 Prajåpati 19—20, 37, 39, 46—47, 70—71, 79—80, 82—83, 85, 97—98, 112, 141— 42, 196, 255, 339, 388 pramåña 162, 164, 168, 197 pråñågnihotra 17—18, 24—25 prañava (o∫) 35, 65—68, 70—72, 74 Pra†åsas 81 Prater 381 Pratyåsas 81 Pravåhaña Jaivali 13—51, pravåsayet 281—82 prayata 220, 233—34 pregnancy 23 Preus 76 Priyavrata Rauhiñåyana 84 procreation 19, 166, 254—55, 258, 385 prostitution 254 P®thu 197 punarm®tyu (re—death) 108 purity 217—45, 266, 365 see also impurity Purußa 58, 79, 144 purußamedha sacrifice 81, 126, 140 püta 221, 235 putrikå 252 püyate 221, 230, 234—36 Quigley 217—18, 241, 365 Radhakrishnan 56, 65

417

Raikva 37, 39, 47 rain 20—21, 80, 86, 142, 384—85 råjasüya 124, 139—42, 150, 157 Råma 96, 102, 122, 335 Råmånuja 46, 55—56, 64, 66—67, 74 Ramanujan 220 rasa 22, 91, 107 Rau 90, 92, 117, 139, 289, 294, 296—97, 372 rebirth (sa∫såra) 103—104, 106—109, 113—16, 119—20, 384—85 Renou 14, 40, 69, 123 Reynolds 322 Rivière 347 Rudra 139 ‡abara 99 sacrifice 19, 25—26, 43—44, 49—51, 53, 55, 58, 63—64, 72, 74, 78, 81—82, 83— 87, 90, 92, 105, 107—109, 111—13, 116, 118—19, 126—27, 131—32, 138—40, 147—49, 157—58, 161, 325, 327—29, 337, 352, 370 sacrificial animal 139, 143 sacrificial post 138 Sådhyas 68 ‡åkalya 35, 117 ‡akuntalå 107—108, 258 Salomon 289, 293, 295—96, 320 såman 56, 66—69, 71—73, 146 sa∫dhyå 262—63 ‡a∫kara 287—99 Sanatkumåra 39, 47, 196 †aõkå (anxiety) 243, 245 ‡antanu 142, 153 †apatha 283—84 †åstra 155—77, 180—81, 206—207, 272— 74, 275 satya 145—47, 151 †auca 220, 223, 225—27, 229, 232, 236 ‡aunaka 199 sautråmañœ sacrifice 81 Savit® 139, 263 Såyaña 58, 79—80, 85, 143 Sax 383 Scharfe 182, 281 Schmidt 258, 338 Schmithausen 14, 16, 384 Schopen 196, 249—50 Segal 76 self, conception of 103—20 semen 15—16, 19—24, 78, 80, 82—83, 85— 87, 90—92, 102, 105, 107, 114, 118, 227, 337—40, 342, 385

418

LANGUAGE, TEXTS AND SOCIETY

Senart 56, 64—65, 290—293 Seneviratne 337, 350 sex 17—23, 33, 75—83, 85—96, 98, 105— 109, 118—20, 248, 254—60, 325, 331— 33, 337—38, 340, 342, 345—48, 384, 388—90 Shamasastry 278 Siddhårtha Gautama 122 Sikhism 323, 336, 340, 343, 347, 349 sin 123—124, 192—93, 212—14, 228 †iß™a 34—35, 153, 158, 168—71 ‡itœman 81 ‡iva 387 sleep 82, 91—92, 94—95 Smith 69, 71, 219, 368, 372, 374, 386, 388 sm®ti 153, 159—60, 171, 180, 197 social hierarchy 369, 386—89, 392 †odhana 220, 230—31 †odhayati 220, 230—31 Söhnen 16, 26, 34, 36—44 soma 19—23, 68, 77—78, 83—85, 102, 112—13, 139, 263, 352 son 23, 33—34, 36, 41, 44, 50, 77, 88—89, 102, 104—109, 111, 113—15, 118—19, 153, 249, 251—52, 254, 258 space 89—91, 95 Sprockhoff 55—56, 65 †raddha 108—109, 263 †ramaña 131—32 152 Sternback 207 strœdhåna 250—52 †ubha 245 †uci 220, 222—29, 232, 237, 240, 245 †uddha 220, 232—33 †uddhi 220, 224, 226, 229, 231—33, 240, 245 †udhyati 220, 229—35 Sujåta 49 †ulka 255—56 sun 67—69, 82, 146 ‡una-‡œra 86 sütra 155—77 svadharma 148 svådhyåya 143, 147 ‡vetaketu 13—51, 257 taboo 341, 351 Talbot 250 Tambiah 229, 368, 377, 381, 383, 390 tapas 121 tejas 121 Thapar 113 Thieme 44, 69—70, 102—103, 109—10

Toru Yagi 352, 355 Trautmann 276 Turner 59, 321 Uberoi 323, 336 ucchiß™a 221, 236—37 Uddåla tree 13 Uddålaka Åruñi 13—51, 69 Uddålaka Jåtaka 13 udgœtha 56, 66—67, 70, 72—73 Udumbara 21 Upakosala 37, 39, 47—48 Ußan Våja†ravasa 50 Ußasti 39 Utaõka 337 Vajracharya 338 Vålmiki 101 vagina 19—20, 85—87, 92, 386 van Buitenen 34, 75—78, 94—95, 335 van Gennep 328 Varenne 18 varña 182—85, 187—93, 195, 199—200, 206—207, 211—212, 218—19, 229, 240—41, 245, 270—71 Varuña 123—25, 137—43, 145, 149, 151 Våyu 86 Vayu Niyutvat 82 Vedakar 288, 293 Vetter 130 Vidyårañya see Mådhava Virocana 33, 37, 39, 47, 196 virtues, five 55 Vißñu 61, 339, 387 Vi†vedevas 140 vivåsa 282 vows 327 V®tra 59 vyavahårapada 175—76, 186, 189—91, 209—10, 284 Warren 333 water 142, 144 wealth 113—15, 143, 248—54, 256 Wezler 123, 133, 152—53, 157—59, 162— 65, 171, 248 Whitney 78—80, 124, 287, 290, 292 widow 330—31, 333 wife 23, 106, 112, 116, 141, 191, 249—54, 332, 387 Wijewardene 367 wind 146 Winternitz

Index

Witzel 45, 112, 320 woman 19—23, 33, 101—109, 112—13, 115— 20, 247—60 and hair 321—50 economic agency of 249—54 property of 249—54 sexual agency of 254—60

Yajurveda 19 Yama see death Yamœ 85 Yåska 86, 153, 155 yautaka 251 Yudhiß™hira 255

Yådava Prakå†a 46 Yåj∞avalkya 17, 23, 37, 45—46, 66, 89, 115—17

Zelliot, 247 Zimmer 333 Zimmerman 372, 382—83, 385, 390

419