The Vedas in Indian Culture and History: Proceedings of the Fourth International Vedic Workshop (Austin, Texas 2007) 9788860323965

Citation preview

The Vedas in Indian Culture and History Edited by Joel P. Brereton

euro 58,00

www.sefeditrice.it

The Vedas in Indian Culture and History Proceedings of the Fourth International Vedic Workshop

Edited by Joel P. Brereton

Società

Editrice Fiorentina

Introduction

THE VEDAS IN INDIAN CULTURE AND HISTORY Proceedings of the Fourth International Vedic Workshop (Austin, Texas 2007) Edited by Joel P. Brereton

Ù  1

2  Ù

The Vedas in Indian Culture and History

‘Alti Studi di Storia intellettuale e delle Religioni’ Series The volumes featured in this Series are the expression of an international community of scholars committed to the reshaping of the field of textual and historical studies of religions and intellectual traditions. The works included in this Series are devoted to investigate practices, rituals, and other textual products, crossing different area studies and time frames. Featuring a vast range of interpretative perspectives, this innovative Series aims to enhance the way we look at religious and intellectual traditions.

Series Editor Federico Squarcini, Ca’ Foscari University of Venice, Italy Editorial Board Piero Capelli, Ca’ Foscari University of Venice, Italy Vincent Eltschinger, École Pratique des Hautes Études, Paris, France Christoph Emmrich, University of Toronto, Canada James Fitzgerald, Brown University, USA Jonardon Ganeri, British Academy and New York University, USA 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 Editorial Coordinator Marianna Ferrara, Sapienza University of Rome, Italy

Introduction

Ù  3

THE VEDAS IN INDIAN CULTURE AND HISTORY Proceedings of the Fourth International Vedic Workshop (Austin, Texas 2007) Edited by Joel P. Brereton

4  Ù

The Vedas in Indian Culture and History

Società Editrice Fiorentina www.sefeditrice.it This edition first published in Italy 2016 by Società Editrice Fiorentina via Aretina, 298 - 50136 Florence, Italy Tel. +39 055 55 32 92 4 | Fax +39 055 55 32 08 5 [email protected]

© 2016 Società Editrice Fiorentina individual chapters © individual contributors The moral right of the authors has been asserted.

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. EBOOK ISBN: 978 88 6032 396 5

Introduction

Ù  5

These studies are dedicated to the memory of our colleague and friend, Frits Staal

6  Ù

The Vedas in Indian Culture and History

Introduction

Ù  7

Contents

Preface 11 Introduction 13 Abbreviations 23 Ludo Rocher The Onset of Vedic Studies: H.T. Colebrooke and the Asiatic Society 25 I. Grammar

and

Text

Eystein Dahl A Note on the Temporal Semantics of the Early Vedic Past Tenses

43

Hans Henrich Hock A Short History of Vedic Prefix-verb Compound Accentuation

61

Tamara Ditrich Historical Development and Typology of dvandva Compounds in the R̥gveda 75 Saraju Rath Observations on Vedic Accents in Grantha Palmleaf Manuscripts 93 II . R e l i g i o n

and

I n t e r p r e tat i o n

Henry John Walker The Birth of the Twin Horse Gods in India and in Greece

107

8  Ù

The Vedas in Indian Culture and History

Madhavi Kolhatkar Lakṣmī: Originally A Marked Animal

121

Madayo Kahle Two Ways to Heaven: R̥ V 10.14 and 10.16

139

Julia Mendoza Tuñón The Path to the Yonder World

147

Joel P. Brereton The Funeral Hymn of Br̥haduktha

167

Mislav Ježić Īśā-Upaniṣad: History of the Text in the Light of the Upaniṣadic Parallels

181

III . R i t u a l , H i s t o r y ,

and

Society

Stephanie W. Jamison R̥ gveda 10.109 “The Brahman’s Wife” and the Ritual Patnī

207

Jarrod L. Whitaker What Makes Indra Indra? On indriyá in the R̥ gveda

221

Shingo Einoo Rites for Rain in the Vedic and Post-Vedic Literature

243

Thennilapuram P. Mahadevan The Institution of Gotra, the R̥ gveda, and the Brahmans

259

Frits Staal† Rathakāro Manasā

289

I V. A t h a r v a v e d a S t u d i e s Elizabeth Tucker The Big-Bellied Heap of Indra

303

Julieta Rotaru “The bráhman that was first born of old….” As It Was Known by the Atharvavedins

319

Michael Witzel A Prosopography of the Śaunakīya Atharvaveda Families of Gujarat As Seen in Their Late Medieval and Early Modern Manuscripts

333

Introduction Contents

V. T h e C o n t i n u i n g

life of the

Ù  9

Veda

Karen Muldoon-Hules Brides of the Buddha, or How Vedic Marital Customs Served Buddhist Ends

385

Shrikant Bahulkar Vedism and Brahmanism in Buddhist Literature

401

Alf Hiltebeitel Epic Aśvamedhas

425

Federico Squarcini To Be Good is To Be vaidika. On the Genesis of a Normative Criterion in the Mānavadharmaśāstra 449 Madhav Deshpande The Yājuṣa Hautra Dispute in Early Modern Maharashtra

467

Laurie L. Patton Notes on Women and Vedic Learning in the 21st Century

477

David M. Knipe Jīrṇa: Reflections of Andhra Āhitāgnis on Old Age and Dying

495

10  Ù

The Vedas in Indian Culture and History

Introduction

Ù  11

Preface

In his keynote address for the Austin Vedic Workshop, Ludo Rocher quoted a passage at the end of Henry Thomas Colebrooke’s 1805 essay “On the Vedas,” in which he wrote, “The ancient dialect in which they [the Vedas] are composed, and especially that of the three first Vedas, is extremely difficult and obscure; and, though curious, as the parent of a more polished and refined language (the classical Sanscrit), its difficulties must long continue to prevent such an examination of the whole Vedas, as would be requisite for extracting all that is remarkable and important in these voluminous works.” Fortunately, succeeding generations of Vedic scholars have taken Colebrooke’s assessment as a challenge rather than a deterrent. Over the last two hundred years, building on scholarship within the Sanskrit tradition, Vedists inside and outside of India have made considerable progress in clarifying Vedic language, history, and religion. As a way of advancing that study and presenting its results, in 1989 Michael Witzel organized the first Vedic Workshop at Harvard. Since then there have been five Vedic Workshops. The second was held in Kyoto in 1999, the third in Leiden in 2002, and the fifth in Bucharest in 2011. The papers in the present collection are from the fourth Vedic Workshop, held in Austin, Texas, in May, 2007. The principal organizers of the Austin Workshop were Joel Brereton, Patrick Olivelle, and Oliver Freiberger, all of The University of Texas. The theme of the workshop was “The Vedas in Culture and History,” from which the current volume takes its title. One of the aims of the committee was to bring together as diverse a group of Vedic scholars as possible, since it saw the Workshop as

12  Ù

The Vedas in Indian Culture and History

an opportunity to promote discussion among scholars from different traditions in different parts of the world. Altogether there were 55 papers presented at the Workshop by scholars from the United States, Japan, Europe, and India. The conference began with a keynote address by Ludo Rocher, who discussed the contributions of H.T. Colebrooke to Vedic scholarship. Two of Colebrooke’s articles had particular significance for the history of Vedic study: his first published essay (1795), which traced the origins of satī to the R̥ gveda, and a later article (1805), which outlined the texts and contents of the Veda. In the first article, the key R̥ g vedic passage that Colebrooke translates as evidence for satī in the R̥ gveda differs from the transmitted text. As Rocher notes, this is a question of the reception of the R̥ gvedic text, not of a deliberate falsification as some early scholars were prone to believe. In his later article Colebrooke illustrated the importance for Vedic study of the anukramaṇīs and other texts of ancient and traditional scholarship. Colebrooke’s essays and Rocher’s address thus raise two central concerns that occupy the contributions in this volume: the text and the interpretation of the Veda and the continuing life of the Veda within a still living tradition. I would like to add a word of thanks to two people who made this conference and this volume possible. The first is my colleague and friend Patrick Olivelle. It was he who suggested at the Third International Vedic Worshop that we at The University of Texas should host the next Workshop in Austin and it was he who arranged support for doing so. Those who know Patrick are well aware of the contributions he has made to this field both on the page and behind it. The second is Federico Squarcini of the Università Ca’ Foscari di Venezia and the Università di Firenze. He has carefully and patiently worked with me to prepare these papers for publication and with the publisher of this volume. I appreciate all that the two of you have done more than I have said here and more than I can say.

Introduction

Ù  13

Introduction

In his introduction to the Proceedings of the first Vedic Workshop (1997), Michael Witzel outlined some of the desiderata of Vedic Studies and discussed the contributions of the volume’s papers to developing new ways of exploring the Veda. His discussion made clear how much needs to be done and how much is being done. Many of the tasks he described remain: we still do not have a full critical edition and translation of the Paippalāda Saṃhitā or a complete translation of the Maitrāyaṇī Saṃhitā, the Kaṭha Saṃhitā, and the Jaiminīya Brāhmaṇa. We still need a comprehensive study of the Vedic verbal system. But progress has been made on a variety of fronts. The work on the Paippalāda Saṃhitā at Leiden has continued, and now we have new editions and translations of eight of its twenty books. Since Witzel wrote, detailed studies of aspects of the Vedic verbal system have appeared, which include, for example, books by Martin Kümmel on the perfect (2000), François Heenen on the desiderative (2006), Eystein Dahl on verbal tense and aspect (2010), and Leonid Kulikov on the ya-presents (2012). These and other significant works are establishing the foundation for a full presentation of the Vedic verbal system. There is a new English translation of the R̥ gveda by Stephanie Jamison and Joel Brereton (2014), and we now have two volumes of a four-volume German translation under the editorship of Michael Witzel and Toshifumi Gotō (2007, 2013). While we still do not have a complete translation of the Maitrāyaṇī Saṃhitā, Kyoko Amano (2012) has given us a German translation of the prose sections of books I and II together with a valuable study of the language of the text. In addition, Thomas Oberlies (1998, 1999, 2012) has published substantial studies of

14  Ù

The Vedas in Indian Culture and History

R̥ g vedic religion that not only consolidate a great deal of scholarship but also offer a distinctive interpretation of the R̥ gveda. Even this very incomplete review of major publications in Vedic studies illustrates the substantial progress that has been made. The papers in this volume show the continuing effort to open up the Vedic corpus and to develop ways of connecting Vedic texts to the study of early Indian history, society, and culture. Foundational to Vedic scholarship are studies of Vedic grammar and Vedic manuscripts. Among the essays in this volume, the four in Part I “Grammar and Text” deal with these basic issues. Eystein Dahl provides a good example of an analysis that both takes account of traditional scholarship and utilizes the insights of contemporary linguistic theory. His paper examines the use of the aorist and imperfect in early Vedic. Pāṇini, followed by many modern scholars, distinguishes between the imperfect indicative, which he understands to denote the remote past, and the aorist indicative, which denotes the recent past. But there are instances in the R̥ gveda, however, which do not show this distinction. Dahl considers verbal aspect as a means of explaining both the normal uses of the aorist and the imperfect and the exceptional cases that do not conform to this usage. The aorist indicative, he argues, has a past perfective character, while the imperfect has a past neutral character. Because it is past perfective, the aorist normally —but not inevitably— signals the immediate past, for the immediate past typically implies that a situation has been completed before the speech time. On the other hand, remote past contexts are semantically less marked because they imply only that there is significant interval between the event described and the speech time. The remote past has therefore generally, but again not necessarily, fallen into the sphere of the more neutral imperfect. The addition of aspect to the interpretation of the aorist and imperfect thus allows a more accurate description of their actual deployment in early Vedic. In his paper, Hans Hock looks at the historical development of the accentuation pattern of verbal prefixes. He shows that the accentuation pattern evolved during the Vedic period, from an early, partly preR̥ g vedic stage, in which every prefix was accented, to a later stage, attested in the Śatapatha Brāhmaṇa and described by Pāṇini, in which the accent alternated between the verb and the immediately preceding prefix. Between these temporal boundaries, the R̥ gveda itself and earlier Yajurvedic texts illustrate varieties of accent patterns that mark the movement from the earlier system to the later. The third paper is by Tamara Ditrich, who investigates the development of dvandva compounds in the R̥ gveda. Ditrich’s analysis focuses on the two earliest types of dvandva compounds: that in which the constituents of the compound are both dual,

Introduction

Ù  15

accented, and declined, and that in which the constituents are dual and accented but only the last constituent is declined. Although there are exceptions, most compounds of these two kinds are devatādvandvas, theonyms invoking two closely related deities. Ditrich supports Stanley Insler’s argument that the constituents of such compounds should be viewed as separate words by pointing out that only these two early types of dvandvas are attested in tmesi and that when in tmesi, the devatādvandvas are not viewed as compounds either by the R̥ gvedapadapāṭha or by Pāṇini. Implied in this analysis is that devatādvandvas are different from other kinds of dvandvas and that their special treatment reflects the power and significance of deities’ names for the Vedic poets. Lastly, Saraju Rath’s paper returns to the issue of Vedic accents, but her concern is the interpretation of the different systems of accentuation found in South Indian Grantha manuscripts. These manuscripts have been ignored by past scholars because their systems of accent notation developed late in the history of South Indian manuscripts (ca. early to mid-16th c.). However, South India has conserved ancient features of Vedic recitation that have disappeared or are rare elsewhere, so that even a late manuscript may provide evidence of early recitation. The essays in Part II “Religion and Interpretation” build on such foundations and offer new interpretations of Vedic texts and new perspectives on Vedic religion. One of the traditions of scholarship on the early Veda has been to study the evidence of other Indo-European cultures in order to clarify the the background and development of R̥ gvedic religious practices and deities. Henry John Walker applies this method, together with a careful reconsideration of Vedic evidence, to understand the nature of the Aśvins in the R̥ gveda. As have earlier scholars, Walker argues that the Aśvins and the Greek Dioskouroi are reflexes of one another. One of the characteristics they share is that there are different traditions involving their status —whether or not they are fully gods— and about their parentage. Although both pairs are referred to as twins, there are a number of traditions regarding their paternity and maternity. In the case of the Aśvins, one R̥ g vedic verse attributes a divided paternity to them, and a mantra quoted by Yāska gives them a divided maternity. As a result of his examination of such traditions, Walker shows that simple formulations that offer a uniform description of the origin and meaning of twins in pre-modern societies have little to recommend them in the face of the complexity of ancient evidence. A second essay that also looks at the origins of a divinity is Madhavi Kolhatkar’s study of the Vedic origins of the goddess Lakṣmī. She argues that Lakṣmī’s name ultimately derives from lákṣman, a “mark,” and more especially, a mark on a cow or other animal

16  Ù

The Vedas in Indian Culture and History

that identifies its owner. The goddess Lakṣmī, she concludes, was originally such a marked animal, which later became anthropomorphized and deified. The papers of Madayo Kahle and Julia Mendoza Tuñón, also are concerned with Vedic origins, but in these, the focus is the development of Vedic concepts of the journey of the dead to heaven. Kahle’s study examines two ways to heaven described in R̥ g vedic funeral hymns. In one, the deceased rejoins his body, transformed and transported by the cremation fire, in heaven. The other way describes the transition of the ásu, the “life,” to heaven. In addition to exploring the differences in these two ways, Kahle’s essay links them to the development of the two “paths” of later Vedic tradition, the devayāna, the “path of the gods,” and pitr̥yāṇa, the “path of the ancestors.” In her essay, Mendoza Tuñón investigates the way to immortality in three passages in the Jaiminīya Brāhmaṇa. The path passes through various units of time and culminates in liberation from time, which is represented by the largest unit of time, the year. Such a journey has parallels in other cultures. A similar path described in the Zoroastrian Hadōxt Nask suggests that the Indian and Iranian eschatologies are descendants of a myth common to the two cultures. A comparable Orphic eschatology attests the influence of Indian descriptions of the path of the dead on the Greek tradition. To these two studies of Indian eschatology, Joel Brereton adds a third, a close study of R̥V 10.56. Although this hymn has often been taken as a eulogy for a dead horse, Brereton argues that it is actually a funeral hymn for a person. The first half of the hymn is dominated by the image of Agni as the horse that transports the body of the deceased to heaven. The second half describes the continued life of the deceased both in heaven and, through his descendants, on earth. Like Brereton, Mislav Ježić also closely examines one Vedic text, but his text is the late Vedic Īśā Upaniṣad. This Upaniṣad has been transmitted in two recensions. By comparing them and parallel passages in other Upaniṣads, Ježić offers a reconstruction of the history of the text, a description of its compositional techniques, and an analysis of its intertextual relations. Such observations on the inner workings of the text also contribute to a clearer exegesis of the Upaniṣad. The papers in Part III “Ritual, History, and Society” show how study of texts can lead to broader observations about the intersections of Vedic ritual and Vedic social roles and about the early history of India. In her essay, Stephanie Jamison builds on her earlier observation that the wife of the sacrificer was brought into the Vedic sacrifice during the late R̥ gvedic period. The focus of her essay is R̥V 10.109, which, she argues, deals with anxieties attendant upon the introduction of the patnī into the solemn

Introduction

Ù  17

ritual. The principal problem for the ritualists was that bringing the wife of the sacrificer into the sacrifice separates her from her normal social identity as a wife and mother. At least during the time of the rite, she becomes the possession of the gods and risks being completely lost to her husband and family. This hymn reflects the attempt to ensure her return into the human world and into her normal social roles. Jamison notes that later exegetes of the hymn miss its point partly because their interpretations reflect the ritual and social conditions of their times. Her essay thus illustrates the need to consider the historical context of both text and commentary in using the latter to understand R̥ gvedic hymns. In his essay, Jarrod Whitaker studies the term indriyá in order to show how R̥ gvedic ritual shaped the ideals R̥ gvedic rulers were expected to embody. The word indriyá denotes the “Indrahood” of Indra, his unique powers and traits as a warrior, which he attained through the ritual. But humans can also aspire to such characteristics and can also gain them by the ritual. The ritual thus defines a warrior ideal and enables men to possess the power to act according to that ideal. In this way, the ritual has a direct influence in defining social roles. In his essay, Shingo Einoo looks at reflections of societal change in the rites for rain in Vedic and post-Vedic literature. While there are many rites for the control of the rain in Vedic literature, there are fewer in the texts compiled after the Veda. Later texts show rather an increasing number of rites for consecrating water reservoirs. This change reflects a technological and economic shift: the advancing ability to build water tanks and the increasing dependence on them. It also reflects a religious one: the development of rites that carried less risk of failure and therefore less risk of compromising the status and reputation of the brahman priests who performed them. The last two essays of the third section also illustrate the effort to understand social history embedded in Vedic texts. T.P. Mahadevan provides a detailed and nuanced study of the development of gotras, brahman lineages that trace lines of descent from r̥ṣi composers of the R̥ gveda. These brahmanical gotras arose from the poets who created the R̥ gvedic hymns and from their descendants who preserved and transmitted the Vedic oral tradition. Indeed, the enactment of these lines of descent in pravara ceremonies replicates the oral transmission of the Veda. The oral tradition was thus instrumental in creating the brahman caste, which is comprised of these brahmanical gotras, and perhaps, Mahadevan suggests, they were instrumental in creating the caste system itself. The last article in this section is by Frits Staal. He looks back at the very earliest history of the Vedic peoples and at the formation of one of the distinctive elements of Vedic culture, its emphasis on knowledge and speech. He approaches this topic

18  Ù

The Vedas in Indian Culture and History

by studying the rathakāra, the “chariot-maker.” Even though the chariot is one of the distinctive cultural elements of the Vedic tradition, there is no evidence that chariot-makers or their ancestors entered South Asia by chariot. Rather, the chariot existed in the minds of chariot-makers. Such specialized and valuable knowledge was one source of the strength of the Vedic peoples, for whom the possession of such knowledge carried social and religious prestige. Staal’s emphasis on the power of knowledge and of words in explaining the cultural dominance of the IndoAryans thus stands in contrast to Whitaker’s analysis, which emphasizes the cult of the warrior. The two articles suggest a tension that further scholarship might usefully explore in understanding how and why the Indo-Aryan peoples came to dominate Indian culture. Partly as a reflection of the on-going preparation of a critical text of the Paippalāda recension of the Atharvaveda, there has been a renewed interest in the Atharvaveda and the Atharvavedic tradition more generally. The three papers in Part IV “Atharvaveda Studies” advance this effort. The essay by Elizabeth Tucker is a close study of two Paippalāda hymns that have no equivalents in the Śaunaka tradition, AVP 11.10 and 11. The two hymns concern a harvest rite, and their focus is the indrarāśi “the heap of Indra,” also called “the big-bellied (mahodara) heap of Indra.” This “heap of Indra” designates a particular portion of grain reserved for brahmans and dangerous for anyone but brahmans. These hymns probably reflect a custom of making an offering to brahmans that removed evil from the harvest and thus brought prosperity to the people. That purpose connects this Atharvavedic rite with later rituals, in which brahmans absorb and thereby eliminate what is inauspicious. The second Atharvavedic study is Julieta Rotaru’s analysis of the application and interpretation of the Atharvavedic pratīka “brahma jajñānam,” “the bráhman that was born.” This phrase serves as the pratīka for two different hymns, AVŚ 4.1 and 5.6, which share the same opening verse. Rotaru’s study examines the Kauśika Sūtra to determine when it uses the pratīka to indicate just the verse and when a whole hymn. And if it indicates a whole hymn, then how —and whether— the Sūtra makes it possible to distinguish which of the two hymns it intends. Her study also examines the various strategies commentaries and later texts use to indicate which hymn they intend by the pratīka. Finally, Michael Witzel reviews the manuscript traditions of the Śaunaka Atharvaveda from the 15th to the 20th centuries to study the copyists, patrons, and families by whom these manuscripts were created and through whom they were transmitted. His study reveals the thin line of transmission of both the oral and the written traditions of the Śaunaka Atharvaveda and the interaction be-

Introduction

Ù  19

tween the two traditions. This fragile transmission helps explain the number of mistakes that occurred and that were perpetuated in both manuscripts and oral recitation of the Atharvaveda. His study alerts scholars to be wary of taking too seriously deviant forms in the Śaunaka traditions, since these may reflect Gujarati peculiarities of script or pronunciation or simply errors in transmission. He further argues that his review shows that not only do we need a critical edition of the Paippalāda recension, but we also need a new edition of the Śaunaka recension, one which more thoroughly documents and examines the oral and manuscript traditions. In the last part of this collection, Part V “The Continuing Life of the Veda,” seven papers carry the discussion of the Veda into the period after its composition. Two of these papers study the relation of the Veda to Buddhism. Karen Muldoon-Hules discusses the reflection of Vedic marital customs in the stories of the Buddhist avadāna literature. The orthodox dharma tradition set marriage as the only desirable option for women, a prescription that left no room for the Buddhist ideal of a celibate life. Two avadāna stories subversively use the Vedic tradition of svayaṃvara, a bride’s “self-choice” of a husband, to justify two women’s entry into monastic life by becoming “brides of the Buddha.” The paper demonstrates both the continuing power of the Vedic tradition and the ability to apply it against brahmanical interpreters who claimed sole authority over it. In his essay, Shrikant Bahulkar demonstrates the continuing influence of Vedic traditions in Buddhist Tantra. Tantric texts show ritual customs, elements, and terminology that derive from Vedic and brahmanical usage. Thus, the Tantras use mantras, including sounds such as om̐, svāhā, and vauṣaṭ, which come directly from Vedic recitation. Similarly, tantric ritual uses sacrificial utensils and ritual practices that derive from Vedic ritual. However, the Tantras employ Vedic terms in a distinctively Buddhist tantric sense, and they adapt Vedic practices to a tantric context. His study demonstrates that Buddhists had knowledge of the Vedic tradition generally, although not of Vedic texts, and it attests to medieval India’s familiarity with Vedic practices, even if the Veda itself remained the possession of ritual specialists. Such general familiarity with the Veda is also apparent in other literature. The next two papers look at the life of the Veda in later Sanskrit literature. In his study, Alf Hiltebeitel examines allusions to the Veda in the Mahābhārata. His focus is the story of Vyāsa’s sexual union with the two princesses, Ambikā and Ambālikā. Hiltebeitel argues that this narrative alludes to the episode in the Aśvamedha in which the mahiṣī or chief queen lies with the sacrificial horse. He supports this interpretation by showing similar oblique ref-

20  Ù

The Vedas in Indian Culture and History

erences to the rites of the Aśvamedha in the Mahābhārata and in the Rāmāyaṇa. While Vedic rites such as the Rājasūya or the Aśvamedha may have been rarely performed in the period of the epics’ composition, these rituals were sufficiently familiar that epic poets could make knowing allusion to them to give greater depth to their narratives. Federico Squarcini examines how the application of the term vaidika in the Mānavadharmaśāstra reflects the historical and cultural context of the text. At the time of the Mānavadharmaśāstra, brahmans’ claim to social and religious primacy was under challenge by other religious groups, especially Buddhists. In order to help brahmans regain their status, Manu urged them to adhere to “Vedic” norms and by thus clothing themselves in the Veda, to rest their claim to social and religious privilege on the Veda. Manu defined “Vedic” to include a broad range of acts and beliefs understood to rest on the authority of the Veda. Thus for him, living according to the Veda meant to trust in the ritual efficacy of Vedic rites, to accept Vedic teachings about the consequences of certain beliefs and actions, to study the Veda, to act according to what Manu understood to be its restrictions and prescriptions, and to identify with a specific brahmanical lineage connected to Vedic study. In that way, Manu envisioned the Veda as defining brahmanical life and with it brahmanical prestige. Bringing us into the early modern period, Madhav Deshpande examines a dispute in 18th- and 19th-century Maharashtra that concerned the right of Yajurvedic priests to perform the functions of a Hotar. The Taittirīya Saṃhitā contains Hautra mantras for nonsoma sacrifices, and therefore the Taittirīyakas claimed that in such rites a Taittirīyaka Yajurvedin could act as Hotar, a role typically belonging to a R̥ gvedin. R̥ gvedins argued that they alone should perform the office of Hotar in all rites. Deshpande documents the arguments, counter-arguments, and even clashes that surrounded this conflict between Yajurvedins and R̥ g vedins. He also notes that this dispute had obvious economic consequences, since exclusion from all but soma rites would have caused severe economic loss for R̥ g vedic communities. The last two papers of this part concern the life of the Veda in contemporary India. In her essay, Laurie Patton discusses the motivations of women to study the Veda and the social acceptance of their study. Based on interviews of women in Pune and Chennai, her paper describes significant differences in attitude and practice concerning Vedic study by women in the two cities. In Pune, most women studying Sanskrit studied Vedic, although few of them participated in priestly recitation. In Chennai, where approximately 50% of the women Sanskrit scholars were nonbrahmans, most of the women studied aesthetics. None studied

Introduction

Ù  21

the Veda itself. Patton points out that Pune has thus preserved a brahman lineage in which women can study the Veda, while Chennai has more thoroughly challenged brahman exclusivism in Sanskrit study but removed Vedic study from the areas pursued by women. Patton’s paper documents the already significant role of women in continuing the traditions of Sanskrit learning, including the traditions of Vedic study. David Knipe’s paper is also based on fieldwork. Since 1980, Knipe has conducted interviews with several generations of pandit families in one area of Andhra Pradesh. One of the topics he has pursued with them has been their reflections on growing old. In this contribution, he discusses their attitudes regarding the inevitablity of illness, their methods of prolonging life, and their performance of the agnihotra and its significance for them. Although his paper concerns old age and death, it is also about how traditional families live the Veda and define their lives according to the Veda. Overall, therefore, these papers show scholars’ efforts to move Vedic studies away from the description of Vedic ritual and Vedic religion as a self-contained system and toward showing how the Veda reflects and forms social and political organization. They also present a more nuanced analysis of the dynamics of the Vedic period. Rather than freezing the early history of Indian religion into discreet categories of early, middle, late, and post-Veda, scholars show an increasing effort to understand the histories within each of these periods and the continuities among them. And finally and most importantly, these papers demonstrate the enduring vitality of both the Vedic tradition and Vedic study. Joel P. Brereton

References Amano, Kyoko. 2009. Maitrāyaṇī Saṃhitā I-II, Übersetzung der Prosapartien mit Kommentar zur Lexik und Syntax der älteren vedischen Prosa. Bremen: Hempen Verlag. Dahl, Eystein. 2010. Time, Tense, and Aspect in Early Vedic Grammar. Leiden–Boston: Brill. Elizarenkova, Tat’jana Ja. 1989-99. Rigveda. Moscow: Nauka. Heenen, François. 2006. Le désidératif en védique. Amersterdam– New York: Rodopi. Jamison, Stephanie W. and Joel P. Brereton. 2014. The Rigveda: The Earliest Religious Poetry of India. 3 Vols. New York: Oxford University Press.

22  Ù

The Vedas in Indian Culture and History

Kulikov, Leonid. 2012. The Vedic -ya-presents: Passives and Intransitivity in Old Indo-Aryan. Leiden Studies in Indo-European, Vol. 19. Amsterdam–New York: Rodopi. Kümmel, Martin. 2000. Das Perfekt im Indoiranischen. Wiesbaden: Richert Verlag. Oberlies, Thomas. 1998. Die Religion des R̥ gveda I: Das Religiöse System des R̥ gveda. Publications of the De Nobili Research Library, Vol. 26. Wien: De Nobili. ______. 1999. Die Religion des R̥gveda II: Kompositionsanalyse der Soma-Hymnen des R̥gveda. Publications of the De Nobili Research Library, Vol. 27. Wien: De Nobili. ______. 2012. Der Rigveda und seine Religion. Berlin: Verlag der Weltreligionen. Witzel, Michael. 1997. Inside the Texts, Beyond the Texts. Harvard Oriental Series, Opera Minora, Vol. 2. Cambridge, MA: Department of Sanskrit and Indian Studies, Harvard University. Witzel, Michael and Toshifumi Gotō. 2007. Rig-Veda. Erster und Zweiter Liederkreis. Frankfurt am Main–Leipzig: Verlag der Weltreligionen. ______. 2013. Rig-Veda. Dritter bis Fünfter Liederkreis. Berlin: Verlag der Weltreligionen.

Introduction

Abbreviations of Vedic and Sanskrit Texts

A ĀGS AiB ĀŚS ĀGS ĀgGS ĀpMP ĀpŚS AV AVP AV-Par AvŚ AVŚ BaudhŚS BĀU BĀU K BĀU M BGS BhG BŚS ChU GDhS GobhGS HirGS HirŚS ĪU ĪU K ĪU M

Aṣṭādhyāyī of Pāṇini Āśvalāyana Gr̥hya Sūtra Aitareyabrāhmaṇa Āśvalāyana Śrauta Sūtra Āśvalāyana Gr̥̥hya Sūtra Āgniveśya Gr̥hya Sūtra Āpastamba Mantrapāṭha Āpastaṃba Śrauta Sūtra Atharvaveda Atharvaveda Paippalāda Atharvaveda Pariśiṣṭa Avadānaśataka Atharvaveda Śaunaka Baudhāyanaśrautasūtra Br̥had-āraṇyaka-upaniṣad Br̥hadāraṇyaka Upaniṣad Kāṇva Br̥hadāraṇyaka Upaniṣad Mādhyaṃdina Baudhāyana Gr̥hya Sūtra Bhagavadgītā Baudāyana Śrauta Sūtra Chāndogya-upaniṣad Gautama Dharmasūtra Gobhila Gr̥hyasūtra Hiraṇyakeśi Gr̥hya Sūtra Hiraṇyakeśiśrautasūtra Īśā-upaniṣad Īśā Upaniṣad Kāṇva Īśā Upaniṣad Mādhyaṃdina

Ù  23

24  Ù

The Vedas in Indian Culture and History

JB Jaiminīya Brāhmaṇa JGS Jaiminīya Gr̥hya Sūtra JUB Jaiminīya Upaniṣad Brāhmaṇa K Kāṇva KāṭhGS Kāṭhaka Gr̥hya Sūtra KaU Kaṭha-upaniṣad KauśS Kauśikasūtra KauB Kauṣītaki Brāhmaṇa KauU Kauṣītaki Upaniṣad KeU Kena-upaniṣad KS Kaṭhasaṃhitā KŚS Kātyāyana-śrautasūtra M Mādhyaṃdina ManB Mantra Brāhmaṇa MānGS Mānava Gr̥hya Sūtra MānŚS Mānavaśrautasūtra Mbh Mahābhārata MDh Mānava Dharmaśāstra MP Matsyapurāṇa MS Maitrāyaṇī saṃhitā MuU Muṇḍaka-upaniṣad P Pāṇini PārGS Pāraskara Gr̥hya Sūtra PB Pañcaviṃśa Brāhmaṇa PS Paippalāda Saṃhitā Rām Rāmāyaṇa R̥ T R̥ ktantra R̥ V R̥ gveda R̥ VPr R̥ gveda Prātiśākhya ŚB Śatapatha-brāhmal̥̀a ŚBK Śatapatha Brāhmaṇa Kāṇva ŚS Śaunaka Saṃhitā ŚŚS Śānkhāyana Śrauta Sūtra ŚvU Śvetāsvatara-upaniṣad TḀ̄ Taittirīya Āraṇyaka TB Taittirīya Brāhmaṇa TrP Taittirīya Prātiśākhya TS Taittirīya Saṃhitā TU Taittirīya Upaniṣad ŚGS Śāṅkhāyana-gr̥hyasūtra ŚvU Śvetāśvatara Upaniṣad ViP Viṣṇu Purāṇa VS Vājasaneyi-saṃhitā VSK Vājasaneyi Saṃhitā Kāṇva VSM Vājasaneyi Saṃhitā Mādhyaṃdina YV Yajurveda

The Onset of Vedic Studies:IH.T. Colebrooke and the Asiatic Society Ù  25 ntroduction

The Onset of Vedic Studies: H.T. Colebrooke and the Asiatic Society Ludo Rocher University of Pennsylvania A few years ago a happy coincidence opened up a treasure of private papers by and about Henry Thomas Colebrooke and the members of Colebrook family that had never been put to use in any scholarly research. The owner, a descendant of Colebrook, gave my wife (and colleague) and me unrestricted access to the documents, with the result that we are engaged jointly in composing a biography of the man whom, in the book1 and henceforth and in this essay, I will call HTC. There is an earlier biography of HTC, by his son Thomas Edward, based on a different set of documents in his possession that can no longer be traced. It is a typical Victorian biography, written by an enthusiast and admiring son, who was not a scholar, at least not in the field of classical Indian studies (Cowell 1873). Even within the field of classical Indian studies HTC’s literary output is vast. In this essay I will concentrate on HTC’s contributions to Vedic studies and some of the long term reactions they provoked. A few words on HTC’s life and career may be in order. He was born in London in 1756, the year in which Shah Alam handed the diwani of Bengal over to Governor Robert Clive. He was the second son of Sir George Colebrook, a prominent London banker, who sometime time after HTC’s birth ran into financial trouble, but who, more importantly, was one of the Directors of the East India Company, and eventually, three times their chairman. HTC left for India at the age of eighteen, in 1783, not to return until thirty years later. Sir George had played an important 1

Rocher and Rocher 2012.

26  Ù

The Vedas in Indian Culture and History

role in the appointment Warren Hastings as Governor of Bengal, and he hoped that Hastings would return the favor by taking special care of the second son he was sending out to India. But by then Hastings had problems of his own, and HTC had to start like everyone else, as a simple writer. But from assistant collector and collector, and a post in the judicial sector, he became a diplomat for Wellesley, chairman of the Court of Appeals in Calcutta, and, finally, a member of the Supreme Council. In Calcutta, HTC became a member and later president of the Asiatic Society, and member and President of the College of Fort William; back in London, he became the founder and director of the Asiatic (thereafter Royal Asiatic) Society of Great Britain and Ireland. He died in 1837. Initially there was nothing that pointed to the fact that HTC would pursue a scholarly career, and a Sanskrit scholarly career at that. But he did learn the language when he was posted at Tirhut (1786-89), one of the main centers of Sanskrit learning in North India. He learned it so well, that in 1793 he took the bold step of sending a scholarly paper to the Asiatic Society. He was then collector at Natore, in Rajshahi District. With the modesty becoming a neophyte, he wrote to the secretary of Society: “Should the following authorities from Sanskrit books be thought worthy of a place in the next volume of the Society’s Transactions, I shall be rewarded for the pains taken in collecting them” (Essays 1: 133).2 The paper was read, by the President of the Society, Sir William Jones, at the meeting of 3 April 1794, the last meeting Jones attended before his premature death 24 days later.3 Jones even sent HTC a congratulatory letter (Cannon 1993: 1.lxx), and the paper was published in the fourth volume of the society’s journal, the Asiatic Researches, in 1795.4 It is entitled “On the duties of a faithful Hindu widow.” The fledging scholar did not know 2 I refer to HTC’s works as published in the Miscellaneous Essays (Cowell 1873), rather than to the original publication of the separate articles in the Asiatic(k) Researches. The former are more easily available, and references to the latter are confusing on account of subsequent pirated editions with varying paginations. Equally confusing is the fact that, instead of referring to Cowell’s “new edition,” some scholars continue to quote from an earlier smaller two-volume edition of Miscellaneous Essays, supervised by HTC himself and edited by Friedrich Rosen in 1837, or, more often, to the edition of part of the latter, published in 1858, and, from there, in 1967 and 2001. 3 Jones read the paper himself, not “a young man” (HTC), as Kejariwal 1988: 74-75. 4 In the same year as the essay on the Hindu widow HTC published a volume in collaboration with the merchant Anthony Lambert, entitled Remarks on the Present State of Husbandry and Commerce of Bengal; it was printed for private circulation because of its adversary stand to East India Company policies.

The Onset of Vedic Studies:IH.T. Colebrooke and the Asiatic Society Ù  27 ntroduction

then that one passage in his article was to attract the attention of, and many pages of writing by, a number of very mature scholars, and even by administrators and politicians. For more than one century Charles Rockwell Lanman has taught numerous scholars of Sanskrit that “[t]here is probably no other stanza in the Veda about which so much has been written” (Lanman 1971: 383). HTC’s article on the Hindu widow is essentially just a collection of detached quotations dealing with satī. How he assembled these quotations, at the stage of his acqaintance with Sanskrit, is not clear. It has been said that the source was Raghunandana’s Śuddhitattva (Vidyāsāgara 1895). Much of HTC’s article is to be found in the Śuddhitattva, but not all, and definitely not in the same order.5 On the other hand, it is even less convincing that, at this early stage of his Sanskrit studies, HTC “took them, at secondhand, from some of the sources of treatises in which they are adduced. No one can say that he did not assemble them from volumes as numerous as themselves” (Hall 1868: 198).6 Of one thing we may be assured: HTC’s translation reflects the way in which the pandit helped him read the passage and interpreted it to him. Following a brief paragraph on the prayoga of widow burning, the first quotation in the article is identified as “from the Rigveda.” HTC translated it: Ohm! Let these women, not be widowed, good wives, adorned with Collyrium, holding clarified butter, consign themselves to the fire. Immortal, not childless, not husbandless, well adorned with gems, let them pass into fire, whose original element is water.7 (Essays 1: 135)

Other quotations follow, laying down modifications of and exceptions to this general rule, but none saying that widow burning is prohibited in the Hindu law books. HTC, while appreciating that “the instances of the widow’s sacrifice are now rare,” logically concluded, “All the ceremonies essential to this awful rite are included in the instructions already quoted. But many practices have been introduced though not sanctioned by any 5 HTC’s essay is not a translation of Jagannātha’s Vivāṅgārṇava, as suggested by Max Müller (1856: 34 n.), and, hence, R̥V 10.18.7 is not translated in HTC’s A Digest of Hindu Law on Contracts and Successions (1797-98). 6 Later in the Life he described the essay to his sons as a “task.” T.E. Colebrooke’s interpretation is ambiguous. On the one hand, he acknowledges that, at that time, HTC did not possess the manuscripts necessary for this compilation, while, on the other hand, he takes the words “the pains taken in collecting them” in the article literally and concludes that his father excerpted the quotations from different sources (Life 65-66). 7 Instead of “well-adorned with gems,” the Asiatic Researches read “excellent.”

28  Ù

The Vedas in Indian Culture and History

ritual. A widow, who declares her resolution of burning herself with the corpse, is required to give a token of her fortitude: and it is acknowledged, that one who receded after the ceremony commenced would be compelled by her relations to complete the sacrifice” (Essays 1: 139-140). No criticism was voiced against HTC’s translation until 1854, when Horace Hayman Wilson pointed out that the quotation was based on an “error, if not a willful alteration of the text” (Rost 1862: 2.275). Wilson maintained that the original, i.e., correct reading attested in R̥V 10.18.7d, which must have been the verse HTC translated, was the one adopted and commented on by Sāyaṇa: ā rohantu janayo yonim agre. The “women” were not the widows of deceased husbands, but women who accompanied the corpse to the funeral pile. That was important, because in Wilson’s translation, differently from HTC’s, widow burning did not have a Vedic foundation: May these women, who are not widows, who have good husbands, who are mothers, enter with unguents and clarified butter; without tears, without sorrow, let them first go up to the dwelling (Rost 1862: 2.272).

Wilson’s reaction did not put the matter to rest. As Winternitz pointed out, HTC’s essay made the verse “notorious” (“berühmt”) in the ensuing discussions about satī (Winternitz 1915: 178). It is not surprising that someone like Radhakanta Deva immediately countered Wilson’s interpretation (Rost 1862: 2.293-305). I cannot in this essay go into all the successive arguments and counter arguments, nor can I give an adequate idea of the long list of participants in the debate, including, in recent years, a renewed interest in the subject on the part of historians and anthropologists.8 In the discussion on the correct reading of R̥V 10.18.7d. there is one factor we may not overlook, namely that HTC’s translation “from the Rigveda” dates from a time when Sir William Jones expressed strong negative opinions on the lack of reliability of the pandits attached to law courts, who, he said, when they found no solution to a legal point in the Śāstras, would make up an answer out of thin air (Cannon 1970: 684; cf. 795); the time also when Wilford was handed puranic texts in which the pandit had covered up the original readings and replaced them by readings that would be agreeable to Wilford. Mindful of this tendency 8 One does not have to go as far as Radhakanta Deva, who suggests that the reading agneḥ may be that of any of the unknown śākhās of the R̥g veda (Rost 1862: 299).

The Onset of Vedic Studies:IH.T. Colebrooke and the Asiatic Society Ù  29 ntroduction

of the pandits “to cheat” their foreign interlocutors, Indologists pointed to the fact that, in the case of R̥V 10.18.7d, it was particularly easy to falsify the original reading in agre into agneḥ, because of the similarity of gra and gna in many Indian scripts. The Vedic Variants lists several examples of this alteration, even suggesting that “in most instance the stem agra is evidently original, and the stem agni secondary.” And the authors of the Vedic Variants speak of our case as a “famous falsification” to which “a special historic, as it were, a romantic interest attaches” (Bloomfield and Edgerton 1930-34: 2.402). But not everyone reacted in an equally detached way. The prime example of this was Max Müller, who vented his anger against those Brahmans who “changed” the original Vedic text: “When the question was mooted whether the burning of widows is an essential part of the Hindu religion, the Brahmans were asked [obviously by HTC] to produce an authority for it from the Veda. They did so by garbling a verse, and as the Veda was not yet published, it was impossible at that time to convict them of falsification.” He labled the reading agneḥ “perhaps the most flagrant instance of what can be done by an unscrupulous priesthood,” and concluded that, as a result of HTC’s translation, “here have thousands and thousands of lives been sacrificed, and a fanatical rebellion been threatened on the authority of a passage which was mangled, mistranslated, and misapplied” (Müller 1856: 35).9 Personally I hesitate to believe that HTC’s pandit abused the similarity of the written gra and gna to mislead HTC into translating the stanza from the R̥ gveda differently from the way he knew very well it ought to be read. I rather think that there were indeed real variants in the way brahmans, not wrote, as modern Indologists say, but recited R̥V 10.18.7, depending on how their paramparā felt about satī. All the more so, since the edition of Raghunandana’s Śuddhitattva, which, as I said, may or may not, have been HTC’s source, displays a third reading: ā rohuntu jalayonim agne (Vidyāsāgara 1895: 2.243). Here, again, this may be the reading familiar to the scribe or to the editor of the Śuddhitattva.10 9 According to some, even if R̥V 10.18 did not refer to widow burning, it was customary in India as it was among other ancient Indo-European peoples, and may, therefore, have been a custom of the “Urvolk” (Zimmer 1879: 328-331). Others reacted in the opposite direction, and interpreted the change from agre to agneḥ not as a falsification but as an adaptation of the text to reality, much earlier than HTC’s pandit, namely at the time when widow burning became widespread in India, in the 10th century (Weinbereger-Thomas 1989: 11), or in the 15th century (Fisch 1998: 225). 10 T.E. Colebrooke informed Fitzedward Hall that his father inscribed this reading in the margin of his copy of the Asiatic Researches (Hall 1868: 186).

30  Ù

The Vedas in Indian Culture and History

Anyhow, in the debates about the abolition of widow burning under William Bentnick (in 1892), it was Indians who eagerly pointed to HTC’s essay. The “Petition of the orthodox Hindu community of Calcutta, against the Suttee Regulation” referred to it to demonstrate that “it is objectionable that concremation, being enjoined by the Sruti, which is the most prevalent authority, and original of all the Smrtis, must be performed” (Majumdar 1941: 162). Even Rammohan Roy, one of the leaders in the fight against widow burning, acknowledged HTC’s reading, but he found a way to explain it away with a farfetched allegorical interpretation (Ghose 1982: 2.369).11 After the article on the faithful Hindu widow I will only briefly refer to the fact that in three articles on religious ceremonies of the Hindus (one in 1798 and two in 1801) HTC translated numerous Vedic and texts into English for the first time. One of them was the first complete translation into English of the puruṣasūkta (Essays 1: 183-185). Far more important, however, is another essay that filled 92 pages in the eighth volume of the Asiatic Researches, published in 1805. I am referring to HTC’s essay entitled “On the Védas, or Sacred Writings of the Hindus.” To be sure, the Vedas were known to Europeans before HTC’s 1805 article. Willem Caland starts his survey of the discovery of the Veda in the 16th century (Caland 1918). Before HTC, other Westerners in India even had successfully tried to obtain Vedic manuscripts. Calmette and other French missionaries had sent, or brought, Vedic manuscripts to Paris. Through the intermediary of the maharaja of Jaipur the Swiss Antoine-Louis Polier, who served for thirty years in the army of the East India Company (1759-1789), got hold of the Vedic manuscripts. He sent them for examination to Sir William Jones, who refers to Polier as “my worthy friend” (Cannon 1970: 893), but, after his return home, he deposited them in the British museum in 1789. However, in Paris as well as in London, these manuscripts only continued to gather dust. In Paris, they were consulted for the catalogue, which Alexander Hamilton and Louis Mathieu Langläs prepared in 1807. In London, the first scholar who actually used Polier’s manuscripts was Friedrich Rosen for his edition of a specimen of the R̥gveda in 1830, forty years after their arrival. Meanwhile, much earlier, in 1787, at the Asiatic Society in Calcutta, Jones read a paper: “On the Literature of the Hindus, from the Sanscrit,” a translation, with commentary, of a document prepared by Govardhan Kaul, called Vidyādarśa (Cannon 1993: 4.9311 “Rammohan Ray treated the alleged shastric support with a verbal respect that thinly veiled his contempt for it” (Thomson 1928: 17).

The Onset of Vedic Studies:IH.T. Colebrooke and the Asiatic Society Ù  31 ntroduction

113). Jones’s description of the sources Vedas from this source is extremely sketchy, but thanks to having examined “the complete copy of the four Védas” of Colonial Polier, Jones was at least able to conclude: “On a cursory inspection of those books it appears, that even a learner of Sanscrit may read a considerable part of the Atharva-veda, without a dictionary; but the style of the other three is so obsolete, as to seem almost a different dialect” (Cannon 1993: 4.102). The first hint at HTC’s interest in the Vedas appears in a letter to his father, written from Rajshahi, 6 December 1793. In this letter he seems to have expressed rather freely his opinion about Jones’s papers in the first three volumes of the Asiatic Researches. Unfortunately for us, he explicitly requested Sir George, his father, to reserve these comments for his private perusal. HTC’s son, who wrote his father’s biography, saw these comments, but he, too, complied with the wish of his father, and thus deprived us from the actual text of the letter. As a result, we do not have HTC’s own opinion about Jones, but only his son’s summery as follows: “The Vedas, their nature, contents, and their mythological history, occupy the first place in his [i.e. HTC’s] attention, and he enters into considerable detail on points which he thinks are not sufficiently explicit in Sir William Jones’s paper” (Life 60). From what HTC writes four years later, on 3 February 1797, from Calcutta, it is clear that, by that time, he has actually been studying the Veda with the help of pandits: “I cannot conceive, how it came ever to be asserted that the Brahmins were ever averse to instruct strangers; several gentlemen who have studied the language find, as I do, the greatest readiness in them to give us access to all their sciences. They do not even conceal from us the most sacred text of their Vedas” (Life 89). Finally, at a well attended meeting of the Asiatic Society, on 4 July 1804, he “presented and read short” his paper entitled “On the Vedas or Sacred Writings of the Hindus” (Nair 1995: 2.303-304). The full text appeared in the eighth volume of the Asiatic Researches, in 1805, the same year in which he published his Sanskrit grammar. HTC refers to Polier, Robert Chambers, Claude Martin, and Jones as collectors of Vedic manuscripts, but he adds: “I have been still more fortunate in collecting at Benares, the text and commentary of a large portion of these celebrated books” (Essays 1: 9). Indeed, HTC, more than anyone else, became a collector of Sanskrit and other Indian Manuscripts. After his return to England he presented a collection of 2749 manuscripts to the library of East India House, “the most munificent gift which the library has ever received” (Sutton 1967: 39). It is thanks to his unique collection of manuscripts that HTC was able to compile,

32  Ù

The Vedas in Indian Culture and History

for the first time, a systematic overview of Vedic literature.12 In later publications he was to do the same for Sanskrit grammatical literature, lexicography, belles-lettres, mathematics, astronomy, and, of course, law, for which, he claims, he possessed a copy of every text in existence. The survey of Vedic literature, though systematic, is not complete. His was not a search for manuscripts starting from a prepared list; he had to compile and organize his own list as manuscripts came into his hands. Coming to the content of the essay, there are passages that, from our point of view, look elementary, but which were far from settled at the time. At the beginning, for example, he deals with such items as the apparent incongruity between references to three Vedas on the one hand, and to four Vedas, especially in the purāṇas on the other, a problem that had also occupied Charles Wilkins and Jones. He notes that the difference between the Atharvaveda and the other three is mainly one of “use and purport” (Essays 1: 11). And much later in the text he approaches another topic that still needed discussion at the time: were the Vedic texts truly authentic documents? HTC was convinced they were, but, only three years earlier, in 1802, John Pinkerton still referred to the Vedas as “these forgeries” (Pinkerton 1802: 2.257). After discussing some late Upanishads whose genuineness may be open to question, HTC introduces his eight page long defense of the Vedic texts with a direct mention of Pinkerton: “Entertaining no doubts concerning other works, which have been here described, I think it nevertheless proper to state some of the reasons, on which my belief on their authenticity is founded. It appears necessary since a late author has abruptly pronounced the Vedas to be forgeries” (Essays 1: 88). However, in between such isolated passages, there is the continuous description of the Vedic texts. HTC goes through each of the four saṃhitās and the texts following up on each of them, for some of them more accurately and more completely than others, constantly providing first translations. He warns the reader when his source materials do not allow him to be as precise as he wished. He is well informed on the R̥ gveda, better on the Śuklayajurveda, including the Śatapathabrāhmaṇa: “For want either of a complete index or of an explanatory index, I cannot undertake from a cursory perusal to describe the whole contents of this part of the Veda” (Essays 1: 32). His essay is especially interesting insofar as it reveals two characteristic traits of his scholarly approach to Ve12 He notes, though, “without waiting to examine them more completely than has been yet practicable, I shall here attempt to give a brief explanation of what they contain” (Essays 1: 9).

The Onset of Vedic Studies:IH.T. Colebrooke and the Asiatic Society Ù  33 ntroduction

dic and Sanskrit literature. First, and particularly for the content and internal arrangement of Vedic texts, he makes ample use of anukramaṇīs. And second, and more generally throughout his career, he relies heavily and places great trust in the interpretation provided by the commentators. It is fair to say that, in general, his knowledge of the sūtras is minimal, whereas, for each śākhā, he is extremely well informed on the Upaniṣads. He translates the entire Aitareya Upaniṣad, large extracts from the Br̥hadāraṇyaka and Chāndogya, and from several Upaniṣads belonging to the Atharvaveda. HTC was, rightly, proud of his essay on the Vedas. After the eighth volume of the Asiatic Researches has been shipped to England, he wrote to his father, on 14 September 1807: “I am curious to know what judgment is passed on the 8th volume of the Researches. On my former contributions to that publication I was pretty much indifferent, but I cannot say that I am so much so to my treatise on the Védas. If it be as little noticed as the rest, it will occasion on me some disappointment, for it is the result of very laborious study” (Life 228). To be sure, the Essay on the Veda did not become part of popular reading in Britain. Nearly a full century later Max Müller still lamented that “people went on dreaming about the Veda, instead of reading Colebrook’s essays” (Müller 1900: 193).13 But the article was read in the scholarly world, and not only in England and in English. In 1840, Guillaume Pauthier published an abridged version of it in his Livres sacrés de l’Orient (Pauthier 1840). In 1847 Ludwig Poley come out with a rather free translation into German (Poley 1847). And widely read authors such as William Ward copied HTC unabashedly, tens and tens of pages at a time, in their own works on Indian religion —which, like Ward’s, were popular enough to go into multiple editions.14 In the Edinburgh Review, Alexander Hamilton, who did not hesitate occasionally to disagree with HTC’s writings, complained that HTC “has now supplied the most important desideratum in Indian literature” (Hamilton 1808: 36). In fact, he went much farther than what he said: “The treatise of which we now have to speak, is, from its subject, the most curious; and from the ability, candour, and research displayed by its author, the most entitled to approbation, of any that have hitherto appeared in the Asiatic Researches” (ibid.: 47). When Rosen edited his specimen of the Cf. anon.: “Vedic India,” Calcutta Review 32 (1859): 400-436 at 401-402. Ward’s four-volume Account of the Writings, Religion, and Manners of the Hindoos (1811), became more popular, from its two-volume abridgement onward, under the title A View of History, Literature and Mythology of the Hindoos (1815-18). 13 14

34  Ù

The Vedas in Indian Culture and History

R̥ gveda, he listed among the tools that made his work possible, next to works of Pāṇini, Yāska, and Sāyaṇa, “the learned essay on the Vedas by the most famous Colebrooke, which throws so much light on the most recondite subjects, that anyone whose studies lead him in that direction, must follow it like the light of a glowing star” (Rosen 1830: 4). And, a few years later, Wilson acknowledged: “It must have been a work of great labour, and could have been done by no one but himself…. The essay is still the only authority available for information respecting the oldest and most important writings of the Hindus” (Life 231). However, others were less enthusiastic about the essay. Rudolph Roth, for one, did acknowledge that HTC was first to throw light on the darkness of Vedic literature, but, “even then,” he says, “the true meaning of the hymns partly escaped him” (Meisig 1994: 41-42). Roth points out, for example, that, although it is “a thing of rare occurrence in his thorough and careful researches,” HTC totally misunderstood the Yama-Yamī hymn (Meisig 1994: n. 319). After studying the Vedic manuscripts, i.e. HTC’s manuscripts, at the India House, Roth proclaims that he himself finally established the interrelation of the Vedic texts, “for Colebrooke’s well known essay of 1805 had not thrown any light on this” (Meisig 1994: 514). One paragraph, the very last one, of HTC’s essay in particular created a major commotion. It reads: The preceding description may serve to convey some notion of the Vedas. They are too voluminous for a complete translation of the whole; and what they contain would hardly reward the labour of the reader; much less that of the translator. The ancient dialect in which they are composed, and especially that of the three first Vedas, is extremely difficult and obscure; and, though curious, as the parent of a more polished and refined language (the classical Sanscrit), its difficulties must long continue to prevent such an examination of the whole Vedas, as would be requisite for extracting all that is remarkable and important in these voluminous works. But they well deserve to be occasionally consulted by the oriental scholar (Essays 1: 102).

Not all reactions were negative. In fact, Böhtlingk quoted HTC’s words in the preface to the Petersburg Dictionary, adding: “What Colebrooke said fifty years ago turned out to be true: the difficulty of the Vedic language will, for a long time to come, be an obstacle to the kind of investigation that will extract everything that is of interest in this literature” (Böhtlingk and Roth 1855: iv). When Max Müller began thinking of editing the R̥ gveda with Sāyaṇa’s commentary, he was convinced that mere extracts would do. He adds: “And I had Colebrooke’s authority for hold-

The Onset of Vedic Studies:IH.T. Colebrooke and the Asiatic Society Ù  35 ntroduction

ing that opinion.” And he quotes HTC’s entire final paragraph. We might never have has Müller’s complete edition of Sāyaṇa had not the “strongest remonstrations” of Burnouf make him change his mind (Müller 1874: vi). But, to some, HTC’s final paragraph was nothing less than blasphemous. Roth nearly went that far, when he said that, whatever the difficulties, the mass of Vedic writings has to be explored to the hilt: “Not only those interested in Indian literature, but also the historian and the archeologist, as well as the linguist, all of these are counting on the spoils that will become theirs” (Meisig 1994: 666). And, even more than to Roth, the paragraph was offensive to Franz Bopp. In 1815 (Paris, 24 February), still very much under the influence of his teacher and guide, the philosopher Karl Joseph Windischmann, he wrote to the latter in connection with HTC’s essay: “With foolish prejudice (“mit solchem albernen Vorurteil”), he cannot for himself create any excitement for a task that indeed requires a lot of labor. A large section, probably one half, of the Vedas, may be ceremonial and of less interest to us. But the other half includes the loftiest philosophy. The Vedas are my goal. They always have been. Anything I have done so far, and anything I am doing now, is only preparation, initiation into the mysteries” (Lefmann 1891: Anhang, 17-19). And a few weeks later (5 April 1815), he added that Windischmann ought not to expect much from HTC’s essay (Lefmann 1891: Anhang, 23-24). Yet, at the very moment Bopp wrote this, he was busy copying for Windischmann all the Vedic passages HTC translated in the essay. And some of these passages he himself —or Windischman who edited the book— inserted, in a German translation by J. Merkel, in his Conjugationsystem der Sanskritsprache (Bopp 1816: 271-312). As Windisch puts it in his final paragraph, HTC had irremediably spoiled his credibility with Bopp (“Colebrooke hatte es bei Bopp verdorben”) (Windisch 1919: 70). Bopp was never to forgive him, even after HTC received him hospitably, and, to Bopp’s astonishment (“gegen meine Erwartung”), helped him with manuscripts, right from the first day, during his visit to London in 1818 (Lefmann 1891: Anhang, 58). In closing, I wish to draw attention to a misinterpretation of HTC that is not uncommon among classic Indologists. A typical example is Willem Caland, when he said that “from 1794 onward Colebrooke devoted himself exclusively (‘uitsluitend’) to the study of Indian languages and literature” (Caland 1918: 299). Caland would not have been correct even if he had nuanced his text and said that, from 1794 onward, HTC exclusively devoted his spare time to the study of Sanskrit language and literature. Philologists often tend to overlook something that others have recently reminded us of rather emphatically, namely that, throughout his

36  Ù

The Vedas in Indian Culture and History

thirty years in India, HTC was a full time, very busy, and not unambitious servant of the East India Company. As such, postcolonial scholars say, he was an integral part of that early group of orientalists who gathered knowledge about India, not for the sake of knowledge itself, but in order better to rule, i.e., more thoroughly to dominate, their Indian possessions and those inhabitants. I submit that both evaluations, the philological and the postcolonial, of a man like HTC are one-sided and needlessly exclusive. There is little doubt that searching for Vedic texts, deciphering manuscripts well enough to be able to organize them and describing their contents, had little to do with being a servant of the East India Company —except that his daily official duties forced him to burn his midnight lamp in order to pursue his research (Life 12 and 86). But the situation is different for many other pieces HTC wrote, his volume on the husbandry and commerce of Bengal and his endlessly long and detailed minutes on subjects such as slavery in India or on the permanent settlement of the revenue of the Ceded and Conquered Provinces. Except for the fact that these, too, were executed in a scholarly fashion, they were written with a view to improve British rule in India. But, in between the extremes of pure scholarship and pragmatic administration, there are many publications that partake of both, and his maiden essay on widow burning belongs to this class. His motivation to select this “awful rite” for his first “task” in ancient Indian literature may be sought in the long history of western fascination with the Indian satī. Yet, in HTC’s case there was more to it than that. We tend to forget that, as soon as it became known that a widow was about to die with or without her husband he, the local Collector in a faraway corner of Bengal, faced the unenviable responsibility of allowing the event to go on or prohibiting it. A decision that was not pleasing to the authorities of the East India Company could have a serious impact on his career. As a fervent adherent to the policy of noninterference with Hindu religious practices proclaimed under Warren Hastings in 1772, HTC turned for advice to his pandit who provided him with relevant ancient texts.15 15 In the case of widow burning, HTC reiterated his opinion in a lecture on mīmāṃsā before the Royal Asiatic Society in London, in 1826 (Essays 1: 346). Notwithstanding Wilson’s disagreement with HTC’s translation of R̥V 19.18.7d, he, too, remained a partisan of noninterference. See, e.g., his letter to the Military Secretary to Government of 25 November 1828, at the time of the widow burning debate: “My opinions are adverse to any authoritative interfering with the practice” (Majumdar 1941: 133). As Lata Mani and others pointed out, the debate was less about whether or not widow burning was “a cruel or barbaric act,” than whether or not “the prohibition of widow immolation was consonant with enforcing the truest principles of ‘Hindu’ religion” (Mani 1998: 15).

The Onset of Vedic Studies:IH.T. Colebrooke and the Asiatic Society Ù  37 ntroduction

Later in his Indian career he reacted similarly to the question of slavery, which he found attested in the ancient dharmaśāstras, and which, therefore, was an integral part of the Indian Tradition. And, of course, it was he who, nearly single-handedly, determined the course of Anglo-Hindu law by his translation, first, of Jagannātha’s Vivādabhaṅgārṇava (1798), and, even more so, of Vijñāneśvara’s Mitākṣarā and Jīmūtavāhana’s Dāyabhāga (1810). I cannot expand on the question of HTC’s “orientalism” in this essay. We deal with it in greater detail in our biography entitled The Making of Western Indology: Henry Thomas Colebrooke and the East India Company, situating HTC’s name in between scholarship and Company service. References Bloomfield, Maurice, Edgerton, Franklin and Murray Barnson Emeneau. 1930-34. Vedic Variants. Philadelphia: University of Pennsylvania. Böhtlingk, Otto and Rudolf Roth. 1855. Sanskrit-Wörterbuch. Erster Theil. St. Petersburg: Kaiserliche Akademie der Wissenschaften. Bopp, Franz. 1816. Über das Conjugationssystem der Sanskritsprache ... Nebst ... einigen Abschitten aus den Veda’s. Ed. by K.J. Windischmann. Frankfurt am Main: Andräische Buchhandlung. Caland, Willem. 1918. De ontdekkingsgeschiedenis van den Veda. Amsterdam: Verslagen an Mededelingen van de Koninklijke Akademie van Wetenchappen, Afdeling Letterkunde, 5e Reeks, 3e Deel, 261-334. Cannon, Garland Hampton. (ed.). 1970. The Letters of Sir William Jones. 2 Vols. Oxford: Clarendon Press. ———. (ed.). 1993. The Collected Works of Sir William Jones. 13 Vols. New York: New York University Press. Cowell, Edward Byles. (ed.). 1873. Miscellaneous Essays by H.T. Colebrooke, with a Life of the Author by his Son, Sir T.E. Colebrooke. 3 Vols. London: Karl J. Trübner. Essays 1, 2 = Cowell 1873, Vols. 2-3. Fisch, Jörg. 1998. Tödliche Rituale. Die indische Witwenverbrennung und andere Formen der Totenfolge. Frankfurt: Campus Verlag. Ghose, Jogendra Chunder. (ed). 1982. The English Works of Raja Rammohun Roy. 4 Vols. Delhi: Cosmo Publications. Hall, Fitzedward. 1868. “The Source of Colebrooke’s Essay ‘On the Duties of a Faithful Hindu Widow.’“ JRAS n.s. 3, 183-198. Hamilton, Alexander. 1808. “Review of Asiatic Researches, Vol. 8.” Edinburg Review 12, 36-50.

38  Ù

The Vedas in Indian Culture and History

Kejariwal, O.P. 1988. The Asiatic Society of Bengal and the Discovery of India’s Past 1784-1818. Delhi: Oxford University Press. Lanman, Charles Rockwell. 1971. A Sanskrit Reader. Cambridge, MA: Harvard University Press. Lefmann, Salomon. 1891-1895. Franz Bopp, sein Leben und seine Wissenschaft. 2 Vols. Berlin: Georg Reimer. Life = Cowell 1873, Vol. 1. Majumdar, Jatindra Kumar. (ed.). 1941. Raja Rammohun Roy and Progressive Movements in India. A Selection from Records (17751845). Calcutta: Art Press. Mani, Lata. 1998. Contentious Traditions. The Debate of Sati in Colonial India. Berkeley, CA: University of California Press. Meiseg, Konrad. (ed.). 1994. Rudolph von Roth. Kleine Schriften. Glasenapp-Stiftung, Vol. 36. Stuttgart: Franz Steiner. Müller, Frederich Max. 1856. “Comparative Mythology.” In: Chips from a German Workshop. Vol. 2, 1-43. London: Longmans, Green, 1867. ———. (ed.). 1874. Rig -Veda-Sanhita. Vol. 6. London: Allen. ———. 1900. My Auobiography. A Fragment. London: Longmans, Green. Nair, P. Thankappan. (ed.). 1995. Proceedings of the Asiatic Society. Vol. 2. Calcutta: Asiatic Society. Pauthier, J.-P. Guillaume. 1840. “Notice sur les Vedas, ou livres sacrés des Hindous, par H.T. Colebrooke.” In: Les livres sacrés de l’Orient, traduits ou revus et publiés par G. Pauthier. Paris: Didot, 307-329. Pinkerton, John. 1802. Modern Geography. A Descritption of the Empires, States, Kingdoms and Colonies. 2 Vols. London: Caddell. Poley, Ludwig. 1847. H.Th. Colebrooke’s Abhandlung über die heiligen Schriften der Inder. Leipzig: Karl J. Trübner. Radhakanta, Deva. 1859. “Remarks by Raja Radhakanta Deva on the preceding article [‘On the supposed Vaidik Authority for Burning of Hindu Widows, and on the Funeral Ceremonies of the Hindus’]; with observations.” In: Rost 1862. Vol. 2, 293-309. Rocher, Ludo and Rosane Rocher. 2012. The Making of Western Indology: Henry Thomas Colebrooke and The East India Company. London: Routledge. Rosen, Friedrich August. 1830. Rigvedae Specimen. London: Joannes Taylor. Rost, Reinhold. (ed.). 1862. Essays and Lectures Chiefly on the Religion of the Hindus by the Late H.H. Wilson. 2 Vols. London: Karl J. Trübner. Roy, Rammohun. 1830. “Abstract of the Arguments Regarding the Burning of Widows considered as a Religious Rite.” In:

Introduction

Ù  39

Ghose 1982. Vol. 2, 365-372. Sutton, Stanley Cecil. 1967. A Guide to the India Office Library. London: Her Majesty’s Stationary Office. Thompson, Edward John. 1928. Suttee. A Historical and Philosophical Enquiry into the Hindu Rite of Widow-burning. London: Allen & Unwin. Vidyāsāgara, J. (ed.). 1895. Śuddhitattva. In: Raghunandana’s Smr̥titattva. Calcutta: no publisher. Vol. 2, 233-412. Weinberger-Thomas, Catherine. 1989. “Cendres d’immortalité. La crémation des veuves en Inde.” Archives de sciences socials des religiones 67, No. 1, 9-51. Winternitz, Moriz. 1915. “Die Witwe im Veda.” WZKM 29, 172203. Zimmer, Heinrich. 1879. Altindisches Leben. Berlin: Weidmann.

40  Ù

The Vedas in Indian Culture and History

Introduction

Ù  41

I. Grammar and Text

42  Ù

The Vedas in Indian Culture and History

A Note on the Temporal SIemantics of the Early Vedic Past Tenses ntroduction

Ù  43

A Note on the Temporal Semantics of the Early Vedic Past Tenses Eystein Dahl University of Bergen, Norway 1. Introduction It is well known that the Early Vedic aorist indicative is often found in immediate or recent past contexts and that the imperfect is typically found in non-immediate or remote past contexts as illustrated by the examples under point (1). (1) The temporal remoteness readings of the Early Vedic past tenses. ́ a) asmābhir ū nú praticákṣiyā abhūd “She (the goddess of dawn) has just now become visible to us.” (R̥ g veda 1.113.11) b) devā ́nāṃ yugé prathamé ásataḥ sád ajāyata “In the (time of the) first generation of gods, being was born from not-being.” (R̥ g veda 10.72.3)

In the first example, the adverb ū nú “just now” unequivocally locates the situation denoted by the aorist indicative form abhūd “became, has become” in the immediate past, whereas the locative phrase devā ́nāṃ yugé prathamé “in the (time of the) first generation of the gods” unequivocally locates the situation denoted by the imperfect form ajāyata “was born” in the non-immediate or remote past. Although it is uncontroversial that the Early Vedic past tenses are strongly associated with these temporal remoteness readings at this stage, they still have other readings which apparently are at odds with their remoteness readings and the semantic properties of the Early Vedic aorist indicative and imperfect therefore

44  Ù

The Vedas in Indian Culture and History

remain disputed. One line of research seems to presuppose that the two past tenses may be characterized as temporal remoteness categories, that their other attested readings can be derived from their temporal remoteness meanings in a principled manner and that there is little or no evidence for an aspectual difference between these two categories (cf. e.g. Tichy 1997: 592-593). Other scholars, like Gonda (1962) and Hoffmann (1967: 151-160) recognize the possibility of explaining at least some of the syntactic differences between the aorist indicative and the imperfect in aspectual terms (cf. also Delbrück 1897, Mumm 2002). Most scholars seem to agree, however, that notions pertaining to tense and, in particular, temporal remoteness are more central in the Early Vedic past tense system than notions pertaining to aspect. The claim that the two Early Vedic past tenses represent temporal remoteness categories is analogous to Pāṇini’s distinction between bhūte adyatane “in the past referring to the present day” and bhūte anadyatane “in the past not referring to the present day” which according to his definition represents the temporal reference of the aorist indicative and the imperfect respectively. As the distinction between what has happened on the present day, i.e. in the hodiernal past, and what has happened before the present day, i.e. in the prehodiernal past, is the most common grammaticalized remoteness distinction in natural language, Pāṇini’s definitions seem to be typologically significant (cf. Dahl 1984, Comrie 1985). There are, however, some prima facie problems with the view that the aorist indicative and imperfect obligatorily denote the hodiernal-immediate and prehodiernal-remote past respectively. More specifically, the aorist indicative is occasionally used in contexts where a hodiernal-immediate past interpretation seems to be precluded or at least highly unlikely and we sometimes, albeit very rarely, find the imperfect in explicitly hodiernal past contexts, apparently excluding a prehodiernal-remote past interpretation. Consider for instance the examples under point (2). ́ iva ít tr̥ṣṇájo nāthitāso ́ ádīdhayur dāśarājñé vr̥t āsaḥ ́ (2)a. úd dyām vásiṣṭhasya stuvatá índro aśrod urúṃ tŕ̥t subhyo akr̥ṇod ulokám “The suppressed ones, who were surrounded in the battle of the ten kings looked towards the sky as if they were thirsty. Indra listened to the praising Vasiṣṭha. He made the land wide for the Tr̥t sus.” (R̥ g veda 7.33.5) b. yád adyá tvā prayatí yajñé asmín hótaś cikitvo ávr̥ṇīmahi ihá “For we chose you as hotar at this ongoing sacrifice here today, o wise one.” (R̥ g veda 3.29.16)

In the first example, a hodiernal-immediate past interpretation seems unlikely because of the locative phrase dāśarājñé “in

A Note on the Temporal SIemantics of the Early Vedic Past Tenses ntroduction

Ù  45

the battle of the ten kings” which apparently specifies the time during which the situations denoted by the verb phrases ádīdhayur “looked, were looking,” aśrod “listened to,” and akr̥ṇod “made” took place. In the second example, on the other hand, the adverb adyá “today” unambiguously locates the imperfect form ávr̥ṇīmahi “chose” in the hodiernal-immediate past. These considerations suggest that Pāṇini’s definitions of the two main past tenses do not capture the full range of interpretations associated with the aorist indicative and the imperfect in Early Vedic and that the Early Vedic past tense system differs typologically from the system described by Pāṇini. This is not surprising as there are several other differences between Early Vedic and the language described by Pāṇini, who not infrequently notes special rules for the language of the Vedic hymns. In this paper I will take a closer look at the temporal semantics of the Early Vedic aorist indicative and imperfect, trying to establish whether their remoteness readings belong to their semantic properties, i.e. whether these categories systematically convey the immediate and non-immediate past, or whether these readings should rather be seen as pragmatic, i.e. context-dependent, arising from a more general semantic difference between these two categories. 2. Tense, aspect, and temporal remoteness: a time-relational approach Notions pertaining to tense, aspect, and temporal remoteness are notoriously difficult to define and a time-relational framework provides a suggestive way of distinguishing such notions (cf. Reichenbach 1947, Kamp and Reyle 1993, Klein 1995 and Kiparsky 1998). I assume that sentences in natural language implicitly refer to four temporal parameters, namely speech time or the time of the utterance, event time or the run time of the situation denoted by the verbal predicate, reference time or the time of the discourse and perspective time or the time from which something is seen as past, present and future. The reference time parameter may either be left implicit or be made explicit by various types of adverbial expressions. For instance, in the examples under point (1) and (2) above the reference time parameter is specified by the adverbial expressions ū nú “just now,” devā́nāṃ yugé prathamé “in the (time of the) first generation of the gods,” dāśarājñé “in the battle of the ten kings,” and adyá “today” respectively. The notion of perspective time is closely related to the notion of speech time and in the unmarked case, these two parameters coincide. However, I will assume that the perspective time parameter is independent of the speech time parameter and that it can be shifted by various morphosyntactic processes. For instance, the relative past relation expressed by categories like the English

46  Ù

The Vedas in Indian Culture and History

pluperfect is considered to represent a grammaticalized shift in temporal perspective, expressing that the reference time of the sentence is past relative to a contextually given time which constitutes its perspective time parameter and which is not identical with speech time. Moreover, I assume that the basic values of the temporal parameters are intervals and that that various relations may hold between the parameters. For our present purposes, the most important relations are inclusion, general overlap, precedence, and immediate precedence. Within this framework tense may be defined as a relation between perspective time/speech time and reference time, whereas aspect may be defined as a relation between reference time and event time. Furthermore, simple past may be defined as the relation “reference time prior to perspective time/ speech time,” immediate/recent past may be defined as “reference time immediately prior to perspective time/speech time,” the imperfective aspect as “reference time included in event time,” the perfective aspect as “event time included in reference time,” and the neutral aspect as “reference time overlapping with event time.” These tense and aspect categories are underspecified in several respects, hence being dependent on the immediate discourse context for their full interpretation. For instance, the temporal relation “reference time prior to perspective time,” i.e. past, denotes the maximal interval prior to perspective time/ speech time, that is the maximal past. This interval has an indefinite initial point and a definite endpoint, namely perspective time/speech time. The basically unspecified character of the semantic past is reflected in its truth-conditions. For a sentence like A man killed a woman to be true it is necessary and sufficient that there is some indefinite time in the past at which the situation expressed by that sentence holds. Likewise, the sentence in the following passage may be taken to refer to some indefinite time in the past, although the situation alluded to suggests that the reference time is not immediately prior to perspective time/ speech time, but rather is located in the mythical, remote past. (3) áhann áhim pariśáyānam árṇaḥ “He (Indra) killed the dragon that surrounded the waters.” (R̥ g veda 6.30.4)

The fact that the Early Vedic imperfect is compatible with a hodiernal past interpretation, as illustrated by example (2b) above suggests that the mythical/remote past interpretation of passages like the one just cited is induced by world knowledge and not because the imperfect has an inherent remote past time reference.

A Note on the Temporal SIemantics of the Early Vedic Past Tenses ntroduction

Ù  47

Similarly, the immediate or recent past relation may be assumed to refer to denote the maximal interval immediately prior to perspective time/speech time, with an indefinite initial point and a definite endpoint. For a sentence like She just got married to be true it is necessary and sufficient that there is some indefinite time in the immediate past at which the situation denoted by that sentence holds. Again, it is reasonable to assume that the immediate discourse context and, more generally, world knowledge play an important role in the exact interpretation of a statement concerning the immediate past. Consider for instance the following sentence: ́ (4) dadhikrā vṇo akāriṣaṃ jiṣṇór áśvasya vājínaḥ “I have praised Dadhikrāvan, the victorious swift horse.” (R̥ g veda 4.39.6)

This sentence is in the last verse of a hymn where the poet praises the steed Dadhikrāvan and the aorist indicative form akāriṣam “I praised, have praised” clearly refers to a situation which has just been or is about to be terminated at speech time and hence seems to presuppose that the aorist indicative at the very least is compatible with a perfective-like interpretation. In this case, the immediate past reference time of the sentence is implied by the immediate discourse context and, on the assumption that the aorist indicative has an immediate past meaning, is signalled by the choice of grammatical form. The assumption that the initial point of the reference time interval is indefinite also in the case of immediate past categories is corroborated by the fact that a sentence like She just got married two years ago seems fine. In cases where the reference time interval of a sentence with immediate past reference is prolonged a current relevance interpretation is often invited, consider for instance She just got married two years ago, but now she will get a divorce. On the assumption that the Early Vedic aorist indicative is associated with an immediate/recent past interpretation, this could be the reason why it is used in contexts like the following: ́ r̥tasāpa ́ āsan ́ ́ (4) yé cid dhí pūrva sākáṃ devébhir ávadann r̥tāni té cid ávāsur nahí ántam āpúḥ “For even the previous (sages), who were performing pious work and were discussing righteously with the gods, even they gave up, for they did not reach the end.” (R̥ g veda 1.179.2)

In this hymn, Lopāmudrā attempts to make her husband Agastya give up his penance. The passage clearly seems to invite the inference that Agastya should not be ashamed to give up, as even the far more powerful sages of old gave up. This might be

48  Ù

The Vedas in Indian Culture and History

the reason why the aorist indicative form avāsur “they gave up” is used in this and similar contexts. The aspectual categories are likewise taken to be semantically underspecified in various respects. The perfective aspect expresses that there is some time at which the relation “event time included in reference time” is true, whereas the imperfective aspect expresses that there is some time at which the relation “reference time included in event time” is true. Consider for instance the sentences Yesterday he read the book and Yesterday he was reading the book. In the first case, the run time of the situation is represented as being contained within the reference time interval, specified by the temporal adverb yesterday. In the second case, the run time of the situation may be taken to be represented as lasting longer than the reference time interval, hence denoting the relation “reference time included in event time.” These basic topological configurations apparently give rise to a tension between the two temporal parameters involved, in the sense that in the absence of any contextually given information about the reference time the perfective aspect tends to select the largest possible reference time interval and the smallest possible event time interval, whereas the imperfective aspect tends to select the largest possible event time interval and the smallest possible reference time interval. This assumption is in line with the traditional descriptions of the perfective aspect as “punctual” and the imperfective aspect as “durative,” in the sense that the smallest possible event time interval will be conceived of as a punctual event and the largest possible event time interval will be conceived of as an indefinitely large event. Among other things, the time relations specified by the perfective and imperfective aspect may be taken to explain their different use in narrative texts, where perfective categories tend to represent two or more situations as sequential or non-overlapping, whereas imperfective categories tend to represent two or more situations as overlapping, as illustrated by the two discourse fragments He entered the room. She laughed. and He entered the room. She was laughing. In the first example, the second sentence apparently denotes a situation following the situation expressed by the first sentence. In the second example, on the other hand, the situation denoted by the second sentence is represented as being overlapping with the situation denoted by the first sentence. The so-called neutral aspect, on the other hand, has received less attention in the discussion on aspect than the former two aspectual categories. It seems to be more pragmatically flexible than the imperfective and perfective aspects, being compatible both with a perfective-like and an imperfective-like interpretation. This may be illustrated by the two possible readings of the Norwegian sentences Han bygget huset. Hun besøkte ham, which

A Note on the Temporal SIemantics of the Early Vedic Past Tenses ntroduction

Ù  49

can be used with the meaning “He built the house. She visited him” or with the meaning “He was building the house. She visited him.” The difference between the two available interpretations may be described in terms of full vs. partial overlap between reference time and event time. A sequential interpretation arises when reference time and event time are interpreted as being coextensive, i.e. fully overlapping. A progressive-like interpretation, on the other hand, arises when the relationship is one of partial overlap between reference time and event time. Thus the neutral aspect seems to have a broader range of interpretations than the two other aspectual categories discussed here. Before concluding this section, a brief note on the lexical dimension of aspectual interpretation is in order. I distinguish three aspectually relevant classes of predicates, namely states, activities, and events. Event predicates comprise verbs like eat an apple, win a race and die, that is, verbs inherently specifying an endpoint, and may be characterized as telic. State predicates like be sick or know French and activity predicates like run or eat apples, on the other hand, do not inherently specify an endpoint and may accordingly be characterized as atelic. Moreover, activity predicates and state predicates differ in at least one important respect. Whereas activity predicates like run or eat apples are typically conceived of as consisting of several identifiable stages and may be characterized as dynamic, state predicates like be sick or know French do not have this property. Finally, I assume that the temporal and aspectual information in a given sentence is compositional and in principle predictable from the meaning of the individual lexical and grammatical constituents. For instance, the combination of a grammatical category with past time reference, a basic perfective meaning and a telic verbal predicate yield the interpretation that the situation was completed prior to speech time, due to the endpoint specified by the verbal predicate and to the time-relations expressed by the semantic past and the perfective aspect respectively. There is a considerable cross-linguistic correspondence with regard to how grammatical categories with certain temporal and aspectual properties interact with different types of lexemes and these patterns of interaction provide important information about the aspectual properties of a given grammatical category (cf. also Smith 1997[1991]). 3. The aorist indicative and the imperfect in Early Vedic 3.1 Introduction In the present section I will consider a recent account of the temporal properties of the Early Vedic past tenses given by Tichy

50  Ù

The Vedas in Indian Culture and History

(1997), who claims that the Early Vedic aorist indicative either has an immediate/recent past interpretation or a relative past interpretation, whereas the Early Vedic imperfect always has a nonimmediate past interpretation. Unlike previous accounts like for instance Gonda (1962) and Hoffmann (1967), Tichy provides an unambiguous definition of immediate and non-immediate past couched in terms of the absence or presence of a relevant interval between the time of the situation and speech time (cf. Tichy 1997: 592 with fn. 8). This definition is fairly similar to the one adopted in the present work and can easily be incorporated in the time-relational framework developed here. Tichy (1997: 592-593) notes that the immediate and non-immediate past is systematically conveyed by the aorist indicative and the imperfect respectively, but that there is no evidence for an aspectual opposition in the Early Vedic past tense system. Within the present framework her claim may be restated in the following way. The aorist indicative and the imperfect have exactly the same aspectual properties, but differ in their temporal reference in that the aorist indicative invariably denotes the immediate past and that the imperfect invariably denotes the non-immediate past. Although Tichy is not explicit on this point, it seems that she assumes that the basic aspectual value of the Early Vedic past tenses would correspond to what I defined above as the neutral aspect (cf. Tichy 1997: 593-594). If I have understood her correctly, then, her claim is that the Early Vedic aorist indicative is a neutral immediate past category, whereas the imperfect is a neutral non-immediate past category. I will examine the claim that the aorist indicative invariably denotes the immediate past in the above sense and that the imperfect invariably denotes the non-immediate past in section 3.2. In section 3.3 I will take a closer look upon the interaction between the two Early Vedic past tenses and various types of verbal predicates to establish whether they actually have the same aspectual characteristics. 3.2 The temporal reference of the Early Vedic past tenses In the previous sections we have already noted that the aorist indicative in many cases is used with an immediate or recent past reference time which is either made explicit by various types of adverbial expressions or is implicitly suggested by the discourse context. I cite two more examples of this use of the aorist indicative for convenience. ́ ́ (5) a. ábhūd deváḥ savitā ́ vándiyo nú na idānīm áhna upavāciyo nŕ̥bhiḥ “God Savitar, who is praiseworthy for us, has just now come into being.” (R̥ g veda 4.54.1)

A Note on the Temporal SIemantics of the Early Vedic Past Tenses ntroduction

Ù  51

b. ápāḥ sómam ástam indra prá yāhi “You have drunk the soma. Go home, Indra.” (R̥ g veda 3.53.6)

Moreover, the assumption that the aorist indicative has an immediate past time reference provides an elegant account of the use of the aorist indicative in passages like the following, in which the immediate discourse context seems to presuppose a non-immediate or recent past reference time. (6) ní āvidhyad ilībíśasya dṝḷhā ́ ví śr̥ṅgíṇam abhinac chúṣṇam ́ ́ táro maghavan yā vad ójo vájreṇa śátrum avadhīḥ índraḥ yā vat pr̥tanyúm “He struck down Ilībiśa’s strongholds, Indra split the horned Śuṣṇa apart. With full strength, with full power, o bountiful one, you killed Pr̥tanyu, the enemy, with the mace.” (R̥ g veda 1.33.12)

This verse seems to refer to the non-immediate mythical past, something which is clear from the situations described by the sentences. Interestingly, we find two imperfect forms, namely āvidhyat “struck down” and abhinat “split” in the third person and one aorist indicative form, namely avadhīḥ “you killed” in the second person. A second person form indicates that the poet addresses the god directly and could be taken as evidence that the situation is somewhat more closely related to the speech situation than the situations denoted by the previous two predicates. Although the aorist indicative in many cases is used with an immediate/recent past time reference, this interpretation is not available in all cases. Consider for instance the following example: (7) śúnaś cic chépaṃ níditaṃ sahásrād yūpād ́ amuñco áśamiṣṭa hí ṣáḥ “You (Agni) even released the fettered Śunaḥśepa from the thousand bounds, for he had been prepared for sacrifice.” (R̥ g veda 5.2.7)

Again, the immediate discourse context seems to suggest that the reference time is not in the immediate past, but rather in the non-immediate mythical past. The situation denoted by the aorist indicative form áśamiṣṭa “was prepared” rather seems to be conceived of as temporally located prior to the situation denoted by the imperfect form amuñcas “you released.” On the assumption that the aorist indicative is an immediate past category, one could analyze this use of the aorist indicative as a syntactic way of shifting the perspective time parameter from speech time to another contextually salient time, in this case the time of the event denoted by the preceding imperfect form.

52  Ù

The Vedas in Indian Culture and History

Tichy (1997: 599) generalizes this observation, claiming that aorist indicative forms in remote past contexts as a rule have a “resultative or relative past” interpretation, in this respect differing from the imperfect. This claim, however, remains somewhat vague, as it is not substantiated by any explicit criteria. In some cases, the aorist indicative may be taken to be syntactically coordinated with other grammatical categories having a relative past interpretation, as for instance the so-called gerund, as illustrated by the following example. In such cases a relative past interpretation is not unlikely, but nevertheless only remains a possibility, unlike the example just cited, where it seems to be the only available interpretation. ́ víśvam adhāg ā yudham ́ (8) sá pravoḷhrn iddhé ̥ ̄ ̣́ parigátyā dabhī ter agnaú sáṃ góbhir áśvair asr̥ jad ráthebhiḥ a. “Having encircled Dabhīti’s abductors and burned the whole weaponry in the ignited fire, he presented (him) with cows, horses and wagons.” (R̥ g veda 2.15.4) b. “Having encircled Dabhīti’s abductors he burned the whole weaponry in the ignited fire. He presented (him) with cows, horses and wagons.” (R̥ g veda 2.15.4)

As far as I can see, both these interpretations are in principle available and there is no theory-neutral reason to prefer one to the other, except perhaps the fact that the aorist indicative in some cases is used with an unequivocal relative past interpretation. On the other hand, we sometimes find aorist indicative forms in passages where it is not obvious that a relative past interpretation is likely. Consider for instance the following examples: (9) a. abhí sidhmó ajigād asya śátrūn ví tigména vr̥ṣabhéṇā púro ́ matím atirac ’bhet sáṃ vájreṇa asr̥ jad vr̥trám índraḥ prá svām ́ chāśadānaḥ “Going straight towards his enemies, he split the fortresses with his sharp bull. Indra made Vr̥tra collide with the mace. Having become superior he advanced his own thought.” (R̥ g veda 1.33.13)

́ ivét tr̥ṣṇájo nāthitāso ́ ádīdhayur dāśarājñé vr̥tāsaḥ ́ b. úd dyām vásiṣṭhasya stuvatá índro aśrod urúṃ tŕ̥t subhyo akr̥ṇod ulokám “The suppressed ones, who were surrounded in the battle of the ten kings were looking towards the sky as if they were thirsty. Indra listened /? had listened to the praising Vasiṣṭha. He made the land wide for the Tr̥tsus.” (R̥ g veda 7.33.5)

The most natural interpretation of these passages seems to be that the situations denoted by the aorist indicative forms abhed

A Note on the Temporal SIemantics of the Early Vedic Past Tenses ntroduction

Ù  53

“split” and aśrot “heard, listened” are temporally located after the situations denoted by the immediately preceding predicates and before the situations denoted by the immediately following predicates. There is no indication that the aorist indicative forms signal a shift in the perspective time, nor that they indicate that the situations denoted by the predicates are more currently relevant than the situations denoted by the imperfect forms in the same immediate discourse context. Interestingly, there is a close parallelism between the verse in the second example and the immediately following verse in the same hymn: ́ (10) daṇḍā ́ ivéd goájanāsa āsan párichinnā bharatā ́ arbhakāsaḥ ábhavac ca puraetā ́ vásiṣṭha ād́ ít tŕ ̥tsūnāṃ víśo aprathanta “The weak Bharatas were cut on all sides like sticks for driving cattle. When Vasiṣṭha became their leader, then the clans of the Tr̥ t sus spread out.” (R̥ g veda 7.33.6)

The two relevant passages are construed in an almost identical manner and, in fact, seem to describe one and the same set of circumstances and thus to provide an exceptionally clear basis for comparison. The first two pādas describe introduce the narrative background, the third pāda describes the first contextually salient situation and the fourth pāda describes an immediately following contextually salient situation. The fact that these two verses are so closely related both thematically and structurally in my opinion indicates that the aorist indicative form aśrot “heard, listened” and the imperfect form ábhavat “became” have identical properties as regards temporal reference, something which in turn may be taken to show that the Early Vedic aorist indicative in some cases is used in non-immediate past contexts with a simple past value. These considerations suggest that we may distinguish three types of temporal reference for the Early Vedic aorist indicative, namely immediate past (including hodiernal past), simple non-immediate past, and relative past. As regards temporal reference, then, the Early Vedic aorist indicative seems to be fairly flexible. The Early Vedic imperfect, on the other hand, generally seems to occur in non-immediate or remote past contexts as noted in the previous sections, but is also occasionally found in hodiernal past contexts, as illustrated by example (2b) above. Again, we may distinguish two main types of evidence for the non-immediate/remote past interpretation, namely contexts where we find an explicit adverbial expression and contexts where the situations alluded to may be considered to have taken place in the non-immediate historical or mythical past. Consider for instance the following examples:

54  Ù

The Vedas in Indian Culture and History a. táma āsīt támasā gūḷhám ágre “In the beginning, darkness was hidden in darkness.” (R̥ g veda 10. 129.3) b. vr̥ṣāyámāṇo avr̥ṇīta sómaṃ tríkadrukeṣu apibat sutásya ā ́ ́ ́ áhīnām sā yakam maghávādatta vájram áhann enam prathamajām “Eager he chose the soma. He drank of the juice at the place of the Trikadrukas. The munificent one took the resounding mace. He killed that firstborn of the dragons.” (R̥ g veda 1.32.3)

In the first example, a non-immediate past reference time is indicated by the adverb ágre “in the beginning.” In the second example, on the other hand, a non-immediate past reference time is suggested by the immediate discourse context which apparently alludes to the well-known myth of Indra’s first draught of soma, which enabled the god to overcome the dragon. The fact that the imperfect is used in these cases clearly shows that it is compatible with a non-immediate past interpretation in the above sense. The question remains, however, as to whether the Early Vedic imperfect always implies that there is a relevant temporal interval between the reference time of the sentence and speech time/ perspective time. It has been noted previously that the aorist indicative is sometimes used in remote past contexts implying that the situation in some sense is currently relevant at speech time/perspective time. In some cases, the imperfect clearly seems to be used in a similar manner. Consider for instance the following passage: ́ (12) a. utó ghā té puruṣíyā íd āsan yéṣām pūrveṣām áśr̥ṇor ŕ̥ṣīṇām ádhā aháṃ tvā maghavañ johavīmi “Indeed, also the former R̥ ṣis, whom you listened to, were human. Therefore I call loudly upon you, o generous one.” (R̥ g veda 7.29.4) ́ ̐ aradad vájrabāhur ápāhan vr̥trám paridhíṃ b. índro asmām ́ nadī nām devó anayat savitā ́ supāṇís tásya vayám prasavé yāma urvī ḥ́ “Mace-armed Indra dug us out, he warded off Vr̥tra, the enclosure of the rivers. God Savitar guided us with his beautiful hands. On his impetus we are flowing widely.” (R̥ g veda 3.33.6)

In the first passage, the situation denoted by the imperfect form áśr̥ṇor “listened to, used to listen to” seems to be located in the non-immediate past because of the adjective pū́r veṣām “earlier, former.” The following discourse context clearly indicates that the poet, recognizing that he is human, hopes that Indra will listen to his invocation, like the god has listened to human r̥ṣis previously. As far as I can see, this case is more or less parallel to the use of the aorist indicative form avāsur “they gave up” in the

A Note on the Temporal SIemantics of the Early Vedic Past Tenses ntroduction

Ù  55

passage cited under point (4) above and hence it is questionable whether there is any systematic difference between the aorist indicative and the imperfect in this respect. In the second passage, however, the non-immediate mythical past reference time is suggested by the immediate discourse context, which may be taken as an account of the historical origin of the rivers. It is not immediately clear that there one might assume a relevant interval between the situation denoted by the imperfect form anayat “he guided” and that denoted by the following present form yāmas “we are going/flowing,” as both these forms seem to refer to one and the same situation, which is represented as having lasted from some time in the mythical past down to the time of speech. These considerations suggest that the hypothesis that there is a systematic opposition in Early Vedic between the immediate past aorist indicative and the non-immediate imperfect might be too strong. Although there indeed seems to be a strong tendency that the aorist indicative is used in hodiernal/immediate past contexts, whereas the imperfect is used in prehodiernal/nonimmediate past contexts both these categories have uses which are at odds with their temporal remoteness interpretations. We therefore need to consider the second part of the claim made by Tichy (1997), namely that the Early Vedic aorist indicative and imperfect have the same aspectual properties. This will be the topic of the next section. 3.2 The aspectual properties of the Early Vedic past tenses In this section I will take a look at how the Early Vedic aorist indicative and imperfect interact with various types of verbal predicates to establish whether it is correct that they have the same aspectual properties. First, we may note that the aorist indicative and the imperfect tend to induce a sequential interpretation when two or more aorist indicative or imperfect forms of telic verbal predicates are found in one and the same discourse context. Consider for instance the following examples: (13) a. avadyám ’va~ mányamānā gúhākar índram mātā ́ vīríyeṇā nír̥ṣṭam átha úd asthāt svayám átkaṃ vásāna ā ́ ródasī apr̥ṇāj ́ jā yamānaḥ “As if she regarded him as a shame, the mother hid Indra, who was bristling with manliness. But he rose up, wearing his armour; (just) coming into being he was filling the two worlds.” (R̥ g veda 4.18.5) ̥ ́ dhúnim étor aramṇāt só asnātṝ ń apārayat suastí tá b. sá īm mahī ṃ ́ rayím abhí prá tasthuḥ utsnā ya

56  Ù

The Vedas in Indian Culture and History “He (Indra) calmed the great river in its course. He brought those who could not swim over in safety. Having come out of the water they reached prosperity.” (R̥ g veda 2.15.5)

It is noteworthy, however, that imperfect forms of telic verbal predicates sometimes seem to be used to express that the situation denoted by the verbal predicate is simultaneous with or is temporally overlapping with another situation in the same immediate discourse context. Compare the following examples. (14) a. ágūhat támo ví acakṣayat súvaḥ “He concealed the darkness, letting the sun shine” (R̥ g veda 2.24.3) b. índro vr̥trám avr̥ṇoc chárdhanītiḥ prá māyínām aminād várpaṇītiḥ “Courageously Indra warded off Vr̥tra, artfully destroying (the sorcery) of the sorcerers.” (R̥ g veda 3.34.3)

It is noteworthy that there are no unambiguous examples of aorist indicative forms of telic verbal predicates with a temporally overlapping interpretation in Early Vedic. Such forms apparently always induce a sequential interpretation, as illustrated by example (13a) above. Aorist indicative forms of atelic verbal predicates, on the other hand, may assume either an inchoative-ingressive interpretation, i.e. focus the entry into the situation denoted by the verbal predicate or an egressive interpretation, i.e. focus the exit of the situation denoted by the verbal predicate. In the former case, the aorist indicative is compatible with a temporally overlapping interpretation. Compare the following examples: ́ (15) a. yé rejáyanti ródasī cid urvī ́ (…) yád áyāsur ugrā ḥ “They (the Maruts), who cause even the wide heaven and earth to shake (…), when they, the strong ones have set out on their journey.” (R̥ g veda 7.57.1)

́ b. hárī nú ta indra (…) svārám asvārṣṭām ví samanā ́ bhūmir aprathiṣṭa áraṃsta párvataś cit sariṣyán “Your two bay horses (…) have now sounded their sound, the earth has spread out evenly, even the mountain that wanted to flee has become calm.” (R̥ g veda 2.11.7)

Imperfect forms of atelic verbal predicates, on the other hand, have a more restricted range of interpretations, apparently always denoting temporally overlapping situations. Consider for instance the following examples: ́ abhíto anavanta ihéha vatsaír víyutā yád ā ́san (16) a. sám átra gā vo

A Note on the Temporal SIemantics of the Early Vedic Past Tenses ntroduction

Ù  57

“The cows bellowed simultaneously on all sides, for they were separated from their calves here and there.” (R̥ g veda 5.30.10)

́ āyan paraśūm ́ ̐ r abibhran vánā vr̥ścánto abhí viḍbhír āyan b. devāsa “The gods were coming, they were carrying axes, (and) cutting wood they were approaching with their attendants.” (R̥ g veda 10.28.8)

Moreover, it is noteworthy that aorist indicative forms of performative verbs are occasionally used to express that the situation takes place at speech time. The imperfect, on the other hand, apparently never occurs in performative sentences. Consider for instance the following example: (17) agním astoṣi r̥ g míyam “I praise Agni who is to be celebrated with verses.” (R̥ g veda 8.39.1)

The fact that the Early Vedic aorist indicative and imperfect shows a different range of interpretations when combined with various types of predicates is a strong indication that they differ in their aspectual properties. More specifically, the various lexically determined readings of the aorist indicative suggest that it has a basically past perfective character, whereas the various lexically determined readings of the imperfect rather suggest that it has a basically past neutral character. Now the question arises as to whether the temporal remoteness readings of the aorist indicative and imperfect can be derived from a past perfective and past neutral meaning respectively. This will be the topic of the concluding section. 3.3 Semantics or pragmatics? Aspect and temporal remoteness in Early Vedic In the previous sections it was noted that the Early Vedic aorist indicative and imperfect seem to convey a systematic aspectual opposition between the perfective and the neutral aspect. Why, then, is the aorist indicative almost exclusively used in immediate past contexts, whereas the imperfect is almost exclusively used in non-immediate past contexts? In this section I will suggest a tentative explanation of this distribution. Immediate past contexts are semantically relatively marked as they not only presuppose that the situation is located prior to perspective time/speech time, but typically also that it has been terminated prior to speech time/perspective time and in many cases implies that a state resulting from the termination of the situation denoted by the verbal predicate holds at perspective time/speech time (resultative interpretation). Remote past con-

The Vedas in Indian Culture and History

58  Ù

texts, on the other hand, may be taken to represent a semantically less marked type of context among other things because such contexts are not typically associated with a resultative implicature, but rather with the implicature that there is an interval of some duration between the reference time interval of the discourse context and perspective time/speech time. On the assumption that the aorist indicative has a past perfective character in Early Vedic and hence expresses that the situation is included within the reference time interval which is located prior to perspective time/speech time, whereas the imperfect has a past neutral character and hence expresses that the run time of the event overlaps with the reference time interval which is located prior to perspective time/speech time, it is reasonable to ask which of these two productive formations may be regarded as the most optimal expression of immediate past, given that the notion of immediate past is semantically marked in the sense just suggested. A past perfective category would in most cases unambiguously express that the situation has been terminated prior to speech time, whereas a past neutral category would be ambiguous between a simple past and a past continuous reading. Therefore a past perfective category may be expected to block a semantically less specific past neutral category from immediate past contexts. In semantically relatively unmarked non-immediate past contexts, on the other hand, the aspectual difference between a past perfective and a past neutral category in general might be expected to play a less important role. The choice between a perfective or imperfective category might in such cases be assumed to be determined by other factors, e.g. stylistic or metrical considerations. It is therefore interesting to note that most of the aorist indicative forms used in non-immediate past tense contexts as a rule have a morphologically simpler form than the corresponding imperfect forms. In this paper I have argued that aspect seems to play a role in the Early Vedic past tense system and that the temporal remoteness readings of the aorist indicative and imperfect should be regarded as contextually conditioned readings and hence pragmatic rather than semantic in nature. This paper is a revised version of previous papers presented at the UCLA Program in Indo-European Studies April 25th 2007, at the Fourth International Vedic Workshop, UT Austin May 24th 2007, at the 26th East Coast Indo-European Conference, Yale University June 17th 2007 and at the 2nd Conference on Greek and Latin from an Indo-European Perspective, University of Oslo July 5th 2007. I thank the audiences for comments and critical discussions after the presentations, above all Stephanie Jamison,

*

A Note on the Temporal SIemantics of the Early Vedic Past Tenses ntroduction

Ù  59

Calvert Watkins, Stanley Insler, Hans Hock, Frank Köhler, Julia M. Mendoza Tuñón, Craig Melchert, Paul Kiparsky, Mark Hale, Michael Weiss, Andrew Garrett, Jay Jasanoff and Dag Haug. I am particularly grateful to Stephanie Jamison for inviting me to give a lecture at UCLA and to her and Calvert Watkins for their generous hospitality during my stay in Los Angeles, to Joel Brereton for inviting me to the Vedic Workshop at UT Austin and to Stanley Insler for inviting me to the ECIEC conference and presenting his counterarguments to my views in his paper presented immediately after my paper, hence laying the ground for a fruitful and inspiring discussion afterwards. Finally, I have discussed most of the issues elaborated on in this paper with Paul Kiparsky during my stay at Stanford January to June 2007 and his critical and insightful comments have substantially improved my understanding of issues concerning tense and aspect in Early Vedic and beyond. Needless to say, I resume the full responsibility for the remaining errors and idiosyncrasies in this paper.

References Comrie, Bernard. 1985. Tense. Cambridge, MA: Cambridge University Press. Dahl, Eystein. (2008). Studies in the Vedic tense and aspect system. PhD Thesis, University of Oslo. Dahl, Östen. 1984. “Temporal distance: remoteness distinctions in tense-aspect systems.” In: B. Butterworth, B. Comrie and Ö. Dahl (eds.), Explanations for language universals. Berlin: Mouton, 105-122. Delbrück, Berthold. 1897. Vergleichende Syntax der indogermanischen Sprachen. Zweiter Theil. Strassburg: Karl J. Trübner. Gonda, Jan. 1962. The Aspectual Function of the R̥ g vedic Present and Aorist. S-Gravenhage: Mouton. Hoffmann, Karl. 1967. Der Injunktiv im Veda. Heidelberg: C. Winter. Kamp, Hans and Uwe Reyle 1993. From Discourse to Logic. Introduction to Modeltheoretic Semantics of Natural Language, Formal Logic and Discourse Representation Theory. Dordrecht: Kluwer. Kiparsky, Paul. 1998. “Aspect and Event Structure in Vedic.” In: R. Singh (ed.), The Yearbook of South Asian Languages and Linguistics 1998. New Delhi–London: Thousand Oaks–Sage Publications, 29-61. Klein, Wolfgang. 1995. “A time-relational analysis of Russian aspect.” Language 71, No. 4, 669-695. Mumm, Peter-Arnold. 2002.“Retrospektivität im R̥ gveda: Aorist und Perfekt.” In: H. Hettrich (ed.), Indogermanische Syntax – Fragen und Perspektiven. Wiesbaden: Reichert, 157-188. Reichenbach, Hans. 1947. Elements of Symbolic Logic. London: Macmillan.

60  Ù

The Vedas in Indian Culture and History

Smith, Carlota S. 1997[1991]. The Parameter of Aspect. Dordrecht: Kluwer. Tichy, Eva. 1997 “Vom indogermanischen Tempus / Aspekt-System zum vedischen Zeitstufensystem.” In: Emilio Crespo and José-Luis García Ramón (eds.), Berthold Delbrück y la sintaxis indoeuropea hoy. Actas del Coloquio de la indogermanische Gesellschaft 21-24 septiembre de 1994. Madrid–Wiesbaden: Reichert, 589-609.

A Short History of VedicIntroduction Prefix-verb Compound Accentuation

Ù  61

A Short History of Vedic Prefix-verb Compound Accentuation* H ans Henrich Hock University of Illinois at Champaign-Urbana

1. Introduction The two standard accounts of the accentuation of Vedic finite verbs with multiple prefixes are Pāṇini (ca. 400 BC) and Delbrück (1888), the latter apparently informed by Pāṇini and the evidence of the Śatapatha Brāhmaṇa. Both varieties of Late Vedic offer the patterns in [1], with accent alternation between the verb and the immediately preceding prefix, and non-accentuation of any earlier prefixes. By contrast, the R̥ gveda has structures such as [2], with accent on more than one prefix or on the leftmost prefix and the verb. Delbrück notes some of these complications. However, his account is incomplete and fails to address the historical developments. I argue that at the earliest stage, accent variation must have affected only finite verbs (main clause vs. dependent clause, etc.1), while prefixes retained their accent. The R̥ gveda presents an initial step toward the Late Vedic situation, with a great amount of variation, and constructions with prefix ā tending to be more innovative. The situation is similar in the Taittirīya Saṃhitā and Brāhmaṇa. The latter, however, offers a triple-prefix structure [3], with alternating accent. Combined with other evidence, this suggests that the development toward Late Vedic was initiated by a tendency toward alternating accent within the same domain. I * I am grateful to Stephanie Jamison for having started me off on this topic. I also have benefited from comments by George Cardona who has prodded me to look more deeply and by Stanley Insler who reminded me of my own earlier observations on accent alternation in Vedic. 1 See Klein 1992, Hock 2006 for recent discussions.

62  Ù

The Vedas in Indian Culture and History

discuss possible problems, map out further developments, and conclude by pointing out the implications, as well as the surprising complication that the early Maitrāyaṇī Saṃhitā as well as (optionally) the Atharvaveda present the same situation as Late Vedic. [1] a. abhyúddharati (P 8.1.28 + 8.1.70) — Main Clause accentuation b. abhyuddhárati (P 8.1.66 + 8.1.70 + 8.1.71) — Dependent Clause accentuation [2] a. úpa prá yantu … (R̥V 1.40.1c) b. yátrābhí saṃnávāmahe (R̥V 8.69.5c) [3]

́ (TB 1.1.5.4) táṃ … prajā ́ abhí samā vartante

2. Prefix or adverb/adposition? Before turning to the main argument of this paper it is necessary to discuss a possible problem in dealing with the development of prefix-verb combinations in the Veda. As is well known, the process of univerbation of verbs and original adverbials such as ní or ā́ had begun even during the pre-Vedic period. It has, after all, parallels in all the other Indo-European languages, including Hittite. However, in the traditional form of the earliest Vedic texts, such adverbials commonly have their own accent and, if they do, are generally not treated as compounded with following verbs in the Padapāṭhas and Prātiśākhyas.2 As a consequence they are presented as separate words in the written versions of the Vedic texts. This raises the question whether in examples such as [2a] and [2b], the elements úpa and abhí (and perhaps even prá) are adverbials (or even adpositions), rather than prefixes. This alternative interpretation could be considered supported by the fact that Pāṇini (1.4.58-60 and 1.4.83-98) recognizes that a subset of the elements that can be used as verbal prefixes (upasargas) can also be used adverbially or as adpositions (karmapravacanīyas). 2 In fact, the Prātiśākhyas, at least selectively, recognize certain problems. Consider the case of the Atharvaveda Prātiśākhya. Sūtras 4.1 and 4.2 suggest that only one prefix (upasarga) can be compounded with an accented verb, but 4.4 states that in certain cases something recognized as a prefix can be non-compounded and hence, by implication, accented: 4.1: upasarga ākhyātenodāttena samasyate 4.2: aneko ’nudāttenāpi (Example: anu-sampráyāhi) 4.4: pūrveṇāhivipaśyāmyādiṣu (Example: yā ́vat te abhí vi-páśyāmi The situation is similar in the Vājasaneyi Prātiśākhya (see 6.2 and 6.4 vs. 6.5, 6.8-10).

A Short History of VedicIntroduction Prefix-verb Compound Accentuation

Ù  63

In fact, this uncertainty of interpretation may be considered evidence that univerbation has not been completed in early Vedic. While it is indeed difficult to be certain about the interpretation of every single example in early Vedic, a number of factors suggest that the elements in question are indeed prefixes in the majority of cases where they occur in chains of such elements preceding the verb, as in [1] and [2]. First, it should be obvious that separability of prefixes from verbs does not constitute evidence for lexical non-univerbation. Consider the case of ā́ plus dā with middle voice inflection, as in [4a] and [4b], the latter with tmesis. In both cases we find the same noncompositional meaning “take,” irrespective of whether ā́ is adjacent to the verb or not. This is no doubt why the extant text of Pāṇini3 has the sūtras 1.4.81-82 which permit separation of prefixes in Vedic. The case is similar for English particle-verb combinations such as blow up “explode, make explode” and German separable prefix verbs such as (sich) vorsehen : sieht sich vor “be careful : is careful.” ́ [4] a. ādatta vajram abhí yád áhiṃ hánn (R̥V 5.29.2c) ́ b. ā ́ sā yakaṃ maghávā adatta vájram ́ áhīnām (R̥V 1.32.3ab) áhann enaṃ prathamajām

Second, while nominal derivatives of verbs with multiple prefixes are rare in the R̥ gveda, a fair number can be found; see [5]. Significantly, these include not only elements that Pāṇini recognizes as possible karmapravacanīyas (abhi, upa, pari, pra, and ā) but also an element not recognized as such (vi). [5] Vedic nominal derivatives from verbs with multiple prefixes pariprayát (pres. pple. R̥V 9.68.8a)

upaprayát (pres. pple.; R̥V 1.74.1a, 1.103.4c, 4.39.5b) viprayát (pres. pple.; R̥V 9.22.5b) abhipracákṣe (inf.; R̥V 1.113.6c)4 ́ samākr̥ta (ta-pple.; 10.84.7a) abhiprabhaṅgín (R̥V 8.45.35b) abhyāvartín (R̥V 6.27.5b) abhyāyaṃsenya (R̥V 1.34.1d) ́ upābhr̥ti (1.128.2d) ́ upā yana (2.28.2c) upārúh (9.68.2b)

There is some question as to whether these sūtras are original or not. This word lacks a corresponding finite form in the R̥ gveda, which goes to show (if proof of this were needed) that the attested forms in the Vedic corpus do not provide a complete picture of Vedic grammar. 3 4

The Vedas in Indian Culture and History

64  Ù

Third, we can find quasi-minimal pairs within the R̥ g veda, such as the ones in [6] and [7]. Example [6] shows that in (quasi-) repetitions, the same elements may sometimes be all unaccented [6a], suggesting univerbation, sometimes be all accented [6b], and sometimes have accent on one and no accent on the other [6c]. The examples in [7], which could easily be multiplied, show that in parallel combinations, those with ā́ occurring closest to the verb tend to lack accent on the preceding element [7a], while those with other elements in the same position, such as prá have accent on the preceding element [7b]. This evidence strongly suggests that the accented preverbal elements are in fact in most cases prefixes. [6] a. yáṃ víprā … b. dyumnaír c. yátra ⁐

(a)bhi- abhí abhí

pra- mandúr … (R̥V 8.12.13ab) prá ṇonumaḥ (R̥V 1.78.1c) saṃ- návāmahe (R̥V 8.69.5c)

́ [7] a. imáṃ yajñám idáṃ váco jujuṣāṇá upāgahi (R̥V 1.91.10ab) ́ sumán me ’dhā yi ́ mánma (R̥V 1.162.7a) b. úpa prāgāt

Fourth, there are numerous instances where the R̥ g veda accents both preverbal elements, while the Atharvaveda and Maitrāyaṇī Saṃhitā do not,5 thereby following the pattern of accentuation in [7a], which began to be introduced in the R̥ gveda. See [8a] vs. [8b]. Note, moreover, that here again the first element, ví, is not recognized by Pāṇini as a karmapravacanīya. ́ saméta páśyata [8] a. sumaṅgalī ŕ iyáṃ vadhūŕ imāṃ ́ (á)thāstaṃ ́ saúbhāgyam asyaí dattvā ya ví páretana (R̥V 10.85.33)

́ saméta páśyata b. sumaṅgalī ŕ iyáṃ vadhūŕ imāṃ saúbhāgyam asyaí dattvā ́ daúrbhāgyair vipáretana (AV 14.2.28) ́ vipáretana (AV 14.2.29) … (á)thāstaṃ

Finally, in one case nominalization supports the view that finite-verb combinations with multiple accented preverbal elements are in fact univerbated and that the accented elements in these structures are indeed prefixes; see [9] [9] a. Finite:

́ úpa prá yantu marútaḥ sudānava(ḥ) (R̥V 1.40.1c)

b. Non-Finite: upaprayát (pres. pple.; R̥V 1.74.1a, 1.103.4c, 4.39.5b)

5 As we will see, the AV and MS pattern is explainable as an extension of the R̥ gvedic one.

A Short History of VedicIntroduction Prefix-verb Compound Accentuation

Ù  65

3. The R̥gvedic situation Although, as we have just seen, there is strong evidence for univerbation of multiple prefixes with verbs —even in the R̥ gveda, it is generally agreed that prehistorically the prefixes in question were independent adverbials and therefore had their own accent, and that the only accent variation occurred on the finite verb, depending on clause type. Prehistorically, therefore, we expect the accentuation corresponding to [9a] = [3a] to have been as in [10a], and the corresponding dependent-clause pattern to have been as in [10b]. Notice the use of the symbols P and V + acute or grave to indicate accented and unaccented prefixes and verbs. [10] a. úpa Ṕ b. úpa Ṕ

prá Ṕ

yantu 

(main clause)

prá Ṕ

yántu 

(dependent clause)

In the R̥ gveda, by contrast, dependent-clause patterns of the type [10b] coexist with two other accentual patterns; see [11a-c]. [11] a. yát sīm

ánu Ṕ

prá Ṕ

b. yátra ⁐ abhí saṃ- ̀ Ṕ P c. yáṃ víprā… (a)bhi- pra- P ̀ P ̀

mucó badbadhānā ́ (R̥V 4.22.7c)  návāmahe (R̥V 8.69.5c)  mandúr… (R̥V 8.12.13ab) 

Of these, [11c] conforms to the Late Vedic norm of the Śatapatha Brāhmaṇa and Pāṇini, by limiting the accent to the verb and leaving both prefixes unaccented. [11b] appears to represent some kind of intermediate stage in the development, since it is neither the most original one (which must have been the type [10b]/[11a]) nor the ultimately victorious one [11c]. Relevant for present purposes is that this pattern has an alternating accentuation, between accented first prefix (abhí), unaccented second prefix (pra), and accented verb (návāmahe). Main clauses, by contrast, generally show precisely the accent pattern of [10a], as in [9a], repeated for convenience in modified form. However, if the second prefix is ā́, the first prefix is usually unaccented, accentuation occurring only sporadically; see [12].

The Vedas in Indian Culture and History

66  Ù

́ [9] a. úpa prá yantu marútaḥ sudānava(ḥ) (R̥V 1.40.1c) Ṕ Ṕ  ́ no adyá sumánā up(a) -ā -́ gahi (R̥V 2.32.5c) [12] a. tābhir Ṕ Ṕ  ́ vyā kaḥ (R̥V 2.38.8d) b. sthaśó jánmāni savitā ́ ví- ā ́ akaḥ (according to the Padapāṭha)



ṔṔ 

The dominant type with ā́-prefix is, of course, identical with the pattern found in Late Vedic, but its limitation to this prefix6 in the R̥ gveda suggests that at this stage it is only beginning to come in. As in the case of [11b], we can look at this incipient pattern as a case of alternating accentuation, in fact, the mirror-image of [11b]. However, it would also be possible to compare it to [11c] by arguing that both cases share the fact that prefixes that do not immediately precede the verb lose their accent. A choice between these two analyses will become possible once we consider the later Vedic evidence of the Taittirīya Saṃhitā and Brāhmaṇa, which appears to be supported by an intriguing formation in the Atharvaveda. 4. The later Vedic evidence of the Taittirīya Saṃhitā and Brāhmaṇa In dependent clauses, the Taittirīya Saṃhitā and Brāhmaṇa generally follow the most advanced pattern of the R̥ gveda [11c], the pattern that is also found in Late Vedic. Compare the examples in [13]. In a few cases, however, we can observe the intermediate pattern of [11b]; see [14]. [13] a. yá

abhi- prati- gr̥ṇyād́ (TS 3.2.9.5) ̀ P P ̀ 

6 Plus one occurrence with áva in a late hymn, R̥V 10.110.10a. —The reasons for why ā ́ should have been in the vanguard of the change are not entirely clear. Delbrück (1888) claims that the more innovative behavior of structures with ā ́ is attributable to the fact that ā ́ is almost always closest to the verb, does not substantially change the verb’s meaning, and is never used as an adposition. There are evident problems with the latter two parts of this claim. First, as is well known, ā ́ tends to change the meaning of the verb, as in [4ab] above. Second, as is also known, ā ́ is in fact used as an adposition. Delbrück is, however, correct in his first claim, namely that ā ́ is almost always closest to the verb. Combined with the fact that it consists of a single sound and, moreover, tends to fuse with initial vowels of following verbs to the point of virtual loss of identity, this might account for the special behavior of ā ́. But this is speculation at best and, as we will see, does not account for the overall developments from the R̥gveda to later Vedic texts.

A Short History of VedicIntroduction Prefix-verb Compound Accentuation

́ b. yáthāhitasyāgnér áṅgārā

abhy-ava-várteran (TB 1.1.5.9) P̀ P̀ 

́ lokāń [14] yád imāṃ

́ (TS 2.3.6.1) abhyàtiricyā tai ́ abhí- ati- ricyā tai Ṕ P̀ 



Ù  67

In main clauses, the R̥ gvedic ā́-prefix pattern [12a] still prevails and has been extended to structures with áva7 as the immediately preverbal prefix; see [15]. Elsewhere, however, these texts are remarkably conservative, as shown by the examples in [16].8 Note that these examples come from prose, rather than mantra and yajus (where their multiple prefix accentuation could be explained as a carry-over from an earlier period). sam- ā ́ P ̀ Ṕ

[15] a. táṃ

rohati (TS 2.5.6.2) 

b. púnar evaínaṃ vāmáṃ vásu⁐ p(a)- ā ́ P ̀ Ṕ

vartate (TB 1.1.2.3) 

́ evá díśi rudráṃ c. svā yām

dayate (TS 2.6.6.6) 

nir- áva- P̀ Ṕ

[16] a. té devā ́ abruvann étemaú ví páry- ūhāméty (TS 2.5.8.2) Ṕ Ṕ  ́ ̐ saṃ mṝdho b. índraṃ vr̥tráṃ jaghnivām ̥ ’bhí prá –avepanta (TS 2.5.3.1) Ṕ Ṕ 

́ c. sá … pr̥thivī m

ánu sám Ṕ Ṕ

acarat (TB 1.1.5.4) 

In addition to such passages with two prefixes I have been able to locate several passages with three prefixes; see [17].9 What is remarkable about these passages is that here we find robust evidence for an alternating accent pattern, which is parallel to the transitional accentuation in dependent clauses, as in [11b], repeated for convenience. Moreover, note that the ví of [17a] is not included in Pāṇini’s inventory of karmapravacanīyas. [17] a. īdŕṅ ̥ vaí rāṣṭráṃ

ví Ṕ

pary- ā -́ P ̀ Ṕ

vartayat(i) (TS 2.5.1.1) 

Which, as noted earlier, follows this pattern once in the R̥ gveda. Note that in [16a] the first prefix is ví, which Pāṇini does not include among the karmapravacanīyas. 9 I owe example [17a] to Insler, May 2007. 7 8

The Vedas in Indian Culture and History

68  Ù

b. táṃ … prajā ́ abhí sam- Ṕ P ̀ c. sāyáṃ paśáva [11] b. yátra ⁐

ā́

(a)vartanta (TB 1.1.5.4; sim. ibid.) Ṕ 

úpa Ṕ

sam P ̀

ā ́ Ṕ

abhí Ṕ

saṃ- P ̀

návāmahe (R̥V 8.69.5c) V́

vartante (TB 3.2.1.5) 

While such passages with three prefixes are hard to find, I have been able to locate three such passages in the Atharvaveda as well.10 Two of them, [18ab], conform to the Late Vedic pattern of accentuation. The third, [18c], however, is quite anomalous, with accent on the second prefix and no accent on the preceding and following prefixes, or on the verb. While not identical to what we find in [17], [18c] shares with [17] the principle of accent alternation, except that here the domain of alternation appears to be restricted to the prefixes, excluding the verb. [18 ]a. anu- P ̀ b. abhi P ̀



saṃ- prá- yāhi (AV 11.1.36a) P ̀ Ṕ  pary ā ́ P Ṕ

c. pitŕ̥ bhya

(a)vartanta (AV 15.7.4; prose) ... .

upa- sám- P ̀ Ṕ

parā- P ̀

(a)ṇayād (AV 18.4.50d) 

In fact, the principle of an alternating pattern of linguistic features, especially suprasegmental ones, in a given domain is a widespread phenomenon and has been recognized as the Obligatory Contour Principle (OCP); see Leben 1973, Goldsmith 1976, and the recent discussion in Prince and Smolensky 2004. For Vedic I have argued for a similar accent alternation in ClauseInitial Strings (Hock 1882, 1996).11 The latter, to be sure, shows OCP effects at the abstract level of the Initial-String template, while in the case of prefix-verb accentuation we are dealing with a more surfacey phenomenon. Still, there is ample independent evidence for the OCP as a motivating force in accentual and other suprasegmental phenomena. It is therefore legitimate to invoke it here. See below for triple-prefix structures in the Maitrāyaṇī Saṃhitā. As I realized only recently, my Initial-String account has been anticipated by Delbrück (1900: 56). —Under Hale’s syntactic approach (1996), the accent alternation in Initial Strings would be epiphenomenal. 10 11

A Short History of VedicIntroduction Prefix-verb Compound Accentuation

Ù  69

Before turning to the evidence of the Atharvaveda and Maitrāyaṇī Saṃhitā, let me briefly summarize the historical developments that will account for the early Vedic situation and its relation to Late Vedic. Let us start with the hypothetical antecedents in [19a], without univerbation and with accent variation only between main clause and dependent clause verbs. Univerbation places the prefix or prefixes and the verb into the same prosodic domain, making it possible for the OCP to apply. [19b] shows what appears to be the “mainstream” development, ignoring the idiosyncratic Atharvaveda pattern in [18c]. For the main-clause version of [19b] see [17]; for the dependent-clause version, compare [11b]. The dependent-clause development appears to have taken place earlier, being attested already in the R̥ gveda, while the main-clause development is only incipient in the R̥ gveda, being restricted to combinations of prefix + ā ́ + verb. [19] M ain Clause Dependent Clause a. ((abhí)sám) ā ́ vartate ((abhí)sám) ā ́ vártate ((Ṕ) Ṕ) Ṕ  ((Ṕ) Ṕ) Ṕ 

b. ((abhí)sam) ā ́ vartate ((abhi)sám) ā vártate ((Ṕ) P̀ ) Ṕ  ((P̀ ) Ṕ) P 

The further development to Late Vedic and Pāṇini (see [1] above) can be attributed to the following factors. First, as we have seen earlier, the main-clause pattern with two prefixes in [12a], which becomes dominant in later Vedic, could be interpreted as either obeying the OCP or a principle of accent alternation restricted to the verb and its immediately preceding prefix. In the R̥ gveda, the OCP interpretation prevailed. Later Vedic, by contrast, reinterpretated the situation and opted for the second interpretation —accent alternation between the verb and its immediately preceding prefix. This reinterpretation may have been helped by the fact that the pattern of accent alternation between the verb and its immediately preceding prefix occurred naturally in the most common type of prefix-verb combination, containing just one prefix. As shown in [20], even under the older OCP account, the accent in main clauses would be on the prefix and in dependent clauses on the verb. [20] a. Main Clause: b. Dependent Clause:

[Ṕ P̀ ] Ṕ  [P̀ Ṕ] P 

Finally, in Late Vedic prefixes generally can occur only before the verb; tmesis is rare; and postposition just about impossible. Under the circumstances, prefix-verb combinations are

70  Ù

The Vedas in Indian Culture and History

acting more and more like single words, rather than multi-word lexemes, so that the common rule will apply that words (normally) only have one accent. 5. A quick glance at the Atharvaveda and Maitrāyaṇī Saṃhitā Example [8b], repeated for convenience, suggests that unlike the Taittirīya tradition, the Atharvaveda has a general main-clause accent loss of first prefixes in double-prefix structures, extending the R̥ gvedic pattern that restricted accent loss to the context before ā́ (and áva). Similar examples can be found in the Maitrāyaṇī Saṃhitā, as in [21a]. In fact, these patterns are identical to what was characterized as “Late Vedic” up to this point. Notice in this regard the prevailing dependent-clause pattern in [21b], with no accent on any of the prefixes, but only on the verb. ́ saméta páśyata [8] a. sumaṅgalī ŕ iyáṃ vadhūŕ imāṃ ́ (á)thāstaṃ ́ ví páretana (R̥V 10.85.33) saúbhāgyam asyaí dattvā ya ́ saméta páśyata b. sumaṅgalī ŕ iyáṃ vadhūŕ imāṃ saúbhāgyam asyaí dattvā ́ daúrbhāgyair vipáretana (AV 14.2.28) ́ vipáretana (AV 14.2.29) … (á)thāstaṃ

́ [21] a. gā trāṇi devā ́ abhisám̐ viśantu (MS 2.5.6) ́ b. yád dhanyámāno hástau pratiprasārayati (MS 1.8.2)

In fact, however, the situation is more complicated. While in many cases where the Atharvaveda (near-) repeats R̥ g vedic passages we find non-accentuation of the first prefix (as in [8b]), in others we find the R̥ g vedic pattern retained, as in [22].12 [22] a. vayám indra tvāyávo ’bhí prá ṇonumo vr̥ṣan (AV 20.18.4ab = R̥V 7.31.4ab) b. úpa prá yantu náro agnírūpāḥ (AV 4.31.1d) ≈ abhí prá … (R̥V 10.84.1d)

Even in what seems to be original Atharvaveda material, we find variation between accentuation and lack of accent, as in [23]. The tendency is for prefixes to be accent verse-initially and at the beginning of a- and c-lines, but unaccented elsewhere. [23] a. ā ́ prá cyavethām … (AV 18.4.49a; quasi-prose) b. ā práyaccha dákṣiṇād ótá savyā t́ (AV 7.26.8d) 12 In the following I have drawn on data from Bloomfield 1906, Lubotsky 1997, and Whitney 1881.

A Short History of VedicIntroduction Prefix-verb Compound Accentuation

Ù  71

For the Atharvaveda, then, we can say that the pattern for double prefix structures varies, in both “native” and “borrowed” material, with maybe some conditioning by position within the verse. For the Maitrāyaṇī Saṃhitā, the situation is different. Here the pattern in [21] is normal in “native” passages. In mantra or yajus passages, however, we find that the normal pattern has accent on all prefixes [24ab], even in triple-prefix combinations [24cd]. Non-accentuation of the first prefix, as in [24e], appears to be rare. ́ sumán me ’dhā yi ́ mánma (MS 3.16.15a = R̥V 1.162.7a [24] a. úpa prāgāt ́ (MS 4.9.1.4 ≈ R̥V 1.40.1c b. úpa práyantu marútaḥ sudānavā(ḥ) c. úpa sáṃpráyāt(a) (MS 2.12.4.21a) d. ágne cyávasva sámánupráyāhi (MS 2.12.4.20a) e. saṃprácyavadhvam (MS 2.12.4.21a —from the same passage as [24c])

The highly archaic patterns in [24a-d], especially [24cd], are not necessarily surprising, given the generally highly archaic nature of the Maitrāyāṇī Saṃhitā. But precisely because of this generally archaic nature, the “Late Vedic” prose pattern of [21] is quite surprising —as is the fact that the Atharvaveda, commonly considered to be closest in antiquity to the R̥ gveda, exhibits the same “Late Vedic” pattern, at least optionally. The “Late Vedic” behavior of the Atharvaveda and the Maitrāyaṇī Saṃhitā as regards multiple prefix accentuation, thus, raises interesting challenges to our understanding of the historical development and chronology of the Vedic tradition. At this point, I don’t even have a hunch as to what might account for this highly innovative behavior. 6. Conclusions As I hope to have shown, the accentuation of verbal prefixes changed over the course of the Vedic period, from an early, partly prehistoric stage with accentuation of every prefix (and with verb accentuation dependent on clause type) to the Late Vedic stage of the Śatapatha Brāhmaṇa and Pāṇini with accent alternation restricted to the verb and the immediately preceding prefix. Moreover, there is good evidence that the initial phase of the change was driven by the OCP, leading to an intermediate stage with alternating accent. Interestingly, the post-R̥ gvedic texts of the Atharvaveda, Maitrāyaṇī Saṃhitā, and the Taittirīya tradition do not agree on the pace of the change, and the two otherwise very archaic texts of the Atharvaveda and Maitrāyaṇī Saṃhitā show more innovative,

72  Ù

The Vedas in Indian Culture and History

“Late Vedic” behavior. The explanation for this difference still awaits explanation. Beyond that, I hope that my paper, in combination with the work of Jared Klein (1992) and my recent paper at the World Sanskrit Conference in Edinburgh (2006), has shown that closer investigations of Vedic accentuation can lead to interesting insights and thus are a fruitful focus for future research. Abbreviations inf. Infinitive P Particle pple. Participle pres. Present V Verb References

Indigenous Indian Grammatical Treatises Atharvaveda Prātiśākhya. 1862. Ed. by William Dwight Whitney. New Haven. Repr. Varanasi: Chowkhamba Sanskrit Series, 1962. Pāṇini’s Aṣṭādhyāyī. 1887. Ed. by Otto von Böhtlingk. 2nd ed., repr. Hildesheim–New York: Georg Olms Verlag, 1971. Vājasaneyi Prātiśākhya. 1858. Ed. by Albrecht Weber. Indische Studien 4. Berlin: F.A. Brockhaus, 65-171; 177-331.

Other References: Bloomfield, Maurice. 1906. A Vedic Concordance. Repr. Delhi: Motilal Banarsidass, 1964. Cardona, George. 1993. The Bhāṣika Accentuation System. Studien zur Indologie und Iranistik 18, 1-40. Delbrück, Berthold. 1888. Altindische Syntax. Halle: Waisenhaus. ———. 1900. Vergleichende Syntax der indogermanischen Sprachen, 3. (= Grundriss der vergleichenden Grammatik der indogermanischen Sprachen … by Karl Brugmann and Berthold Delbrück, 5.) Strassburg: Karl J. Trübner. Goldsmith, John. 1976. Autosegmental Phonology. PhD Diss., MIT. Hale, Mark. 1995. “Deriving Wackernagel’s Law: Prosodic and Syntactic Factors Determining Clitic Placement in the Language of the Rigveda.” In: A. Halpern and A. Zwicky (eds.),

A Short History of VedicIntroduction Prefix-verb Compound Accentuation

Ù  73

Approaching Second: Second Position Clitics and Related Phenomena. Stanford: CSLI Publications, 165-198. Hock, Hans Henrich. 1982. “Clitic Verbs in PIE or DiscourseBased Verb Fronting? Sanskrit sá hovāca gā́rgyaḥ and Congeners in Avestan and Homeric Greek.” Studies in the Linguistic Sciences 12, No. 2, 1-38. ———. 1996. “Who’s On First: Toward a Prosodic Account of P2 Clitics.” In: A. Halpern and A. Zwicky (eds.), Approaching Second: Second Position Clitics and Related Phenomena. Stanford: CSLI Publications, 199-270. ———. 2006. “Vedic Verb Accent Revisited.” Paper read at the World Sanskrit Conference, Edinburgh, Scotland (revised). Klein, Jared S. 1992. “On Verbal Accentuation in the Rigveda. American Oriental Society Essay Number 11. New Haven, CT: American Oriental Society. Leben, William. 1973. Suprasegmental Phonology. PhD Diss., MIT. Distributed by Indiana University Linguistics Club. Lubotsky, Alexander. 1997. A R̥gvedic Concordance, 2 parts. New Haven, CT: American Oriental Society. Prince, Alan, and Paul Smolensky. 2004. Optimality Theory: Constraint Interaction in Generative Grammar. Malden, MA–Oxford, UK: Blackwell. Whitney, William Dwight. 1881. Index Verborum to the Atharva-Veda = Journal of the American Oriental Society 12.

74  Ù

The Vedas in Indian Culture and History

Historical DevelopmentIntroduction and Typology of dvandva Compounds

Ù  75

Historical Development and Typology of dvandva Compounds in the R̥ g veda Tamara Ditrich

I. Introduction Several types of dvandva compounds, consisting of two coordinate constituents, can be identified in the R̥ gveda as those with: 1. both constituents in the dual and declinable, each having its own accent [e.g., R̥ V 1.2.9a mitrā́váruṇā (N. A. Du.) “Mitra and Varuṇa”; R̥ V 7.66.1a mitráyorváruṇayoḥ (G. L. Du.)]; 2. both constituents in the dual, the first constituent being indeclinable and each constituent having its own accent [e.g., R̥ V 1.23.5c mitrā́váruṇā (N. A. Du.) “Mitra and Varuṇa”; R̥ V 10.130.5a mitrā́váruṇāyoḥ (G. L. Du.)]; 3. both constituents in the dual, the first constituent being indeclinable and the whole compound having only one accent —on the last syllable of the final constituent [e.g., R̥ V 1.162.2d indrāpūṣṇóḥ (G. L. Du.) “of Indra and Pūṣan”]; 4. the first constituent in stem form and the whole compound having only one accent [e.g., R̥ V 1.14.3a indravāyū́ (N. A. Du.) “Indra and Vāyu”]; 5. the first constituent in stem form, the whole compound having one accent and plural endings [e.g., R̥ V 10.190.2c ahorātrā́ṇi (N. A. Pl.) “days and nights”]; 6. the compound having only one accent and appearing in the collective neuter singular [e.g., R̥ V 10.14.8b iṣṭāpūrtám “what was offered and given”]. This typology was already established in the early linguistic studies of Vedic and has been followed more or less ever since

76  Ù

The Vedas in Indian Culture and History

(e.g., Wackernagel 1957: 149–173; Burrow 1969: 216–217). It is generally accepted that different types of dvandva compounds represent different stages in their historical development, beginning with types 1 and 2, the most archaic and frequently used types in the R̥ gveda, followed by types 3 and 4, where the whole compound received one accent and gradually the first constituent occurred in stem form. Types 5 and 6 represent the last stages in the development: they are very rare in the R̥ gveda and no theonyms occur in these dvandvas. This paper examines two interrelated aspects of the typology and historical development of dvandvas. First, it examines the typological scheme as presented above, focusing particulary on the first two types, and revisits Insler’s interpretation of types 1 and 2 as two separate words rather than compounds (Insler 1998). Second, it argues that dvandva compounds cannot be viewed as a single category but rather a distinction must be made between those comprising theonyms and those comprising non-theonyms, as each displays different linguistic features. II. Historical development and typology of dvandvas Most dvandva compounds in the R̥ gveda consist of two theonyms. I have examined all attestations of the most prominent pairs of R̥ g vedic dual deities: Indra and Vāyu, Mitra and Varuṇa, Indra and Varuṇa, Indra and Agni, Dyaus and Pr̥thivī, Uṣas and Nakta. Their names occur in the R̥ gveda in dvandva compounds (with the components in juxtaposition or in tmesi), elliptic duals, asyndeta, and syntagms constructed with coordinative conjunctions. For each pair, the number of attestations for these constructions in the R̥ gveda is the following: Table 1 Dual deities

Dvandva

Dvandva in tm.

Elliptic Dual

Asyndeton

Conjunctions

Mitra and Varuṇa

95

3

8

80

15

Indra and Varuṇa

45

7

0

13

1

Dyaus and Pr̥thivī

83

20

5

57

59

Uṣas and Nakta

15

2

4

0

1

Indra and Vāyu

23

0

0

18

7

Indra and Agni

92

3

0

10

5

The first four pairs occur in dvandvas of types 1 and 2 (mitrā́váruṇa-; uṣā́sānáktā; índrāváruṇa-; dyā́vāpr̥thivī́ ), the last

Historical DevelopmentIntroduction and Typology of dvandva Compounds

Ù  77

two (indravāyú-; indrāgní-) are usually considered to be dvandvas of type 4 (e.g. Wackernagel 1957: 154), although it is uncertain which type of dvandva is formed by Indra and Agni. In my past research on the distribution of dvandva compounds and other coordinative constructions in the ten books of the R̥ gveda (Ditrich 2006), I have shown that there is no indication that dvandvas of types 1 and 2 occur more frequently in the older layers of the R̥ gveda and later types in the younger layers. For example, Dyaus and Pr̥thivī occur in the dvandva of the oldest type (dyā́vāpr̥thivī́ ) yet they have considerably more attestations (in juxtaposition as well as in tmesi) in the later maṇḍalas, especially in the latest maṇḍala 10. On the other hand, Indra and Vāyu are attested in dvandvas of type 4 (indravāyú-) which is considered to be a later type; however, there are considerably more attestations of this dvandva in the oldest maṇḍalas, i.e., the family books (Ditrich 2006: 128–129; 133–134). This investigation has indicated that the historical development and typology of dvandvas needs revisiting. Dvandvas of types 1 and 2 The oldest types 1 and 2 are very common in the R̥ gveda; they are probably of Indo-European origin and certainly existed in the Indo-Iranian period.1 These dvandvas usually occur in juxtaposition (e.g. dyā́vākṣā́mā) though they are also attested in tmesi (e.g., R̥ V 10.12.1 dyā́vā ha kṣā́mā).2 Tmesis of dvandva compounds is quite common in the R̥ gveda: the largest number of instances is found with the theonyms dyāvā and pr̥thivī (e.g., R̥ V 1.143.2 dyā́vā śocíḥ pr̥thivī́ ).3 I have argued (Ditrich forthcoming) that the tmesis of devatādvandvas seems to be one of the stylistic devices of the R̥ g vedic poetic language: constituents of dvandvas in tmesi occur elsewhere (i.e., in the same hymn or in other hymns) in dvandvas having their components in juxtaposition, as well as in 1 There are examples of dvandvas that appear both in type 1 and 2 in the R̥gveda, e.g. the older type 1 (e.g., in G. L. Du., R̥V 7.66.1a mitráyorváruṇayoḥ) and the younger type 2 (e.g., in G. L. Du., R̥V 10.130.5a mitrā ́váruṇayoḥ). Similar expressions are attested in Av., e.g. ahurami θra “Ahura and Mitra”; ahuraēibyami θraēibya. In the R̥gveda most attestations of types 1 and 2 are in N. A. V. Du., some are also in G. L. Du., but —unlike in Avestan— none in I. D. Ab. Du. 2 Dyā ́vā ha kṣā ́mā (R̥V 10.12.1) is attested only once in tmesi whereas the more common dyā ́vākṣā ́mā has 8 attestations. Dvandvas in tmesi seem to be of IndoIranian origin; they occur also in Avestan (e.g., ašičā ārmaitī) though not as frequently as in Vedic (Duchesne-Guillemin 1936: 45). 3 The pair has 20 attestations in tmesi (in N. A. Du. (dyā ́vā … pr̥thivī )́ : 1.63.1; 1.143.2; 1.159.1; 2.12.13; 2.41.20; 5.43.2; 6.11.1; 7.43.1; 7.53.1; 8.97.14; 10.35.3; 10.46.9; 10.91.3; and in V. Du. (dyā ́vā … pr̥thivī): 1.185.2–8).

78  Ù

The Vedas in Indian Culture and History

the elliptic duals, asyndeta, and constructions with coordinative particles. No semantic difference between these constructions is indicated; they are often found alternating in the same hymn, especially in hymns dedicated to the pair in question. When components of dvandvas of types 1 and 2 are in tmesi, the constituents are freely distributed within the syntax of the sentence which is one of the main reasons for Insler’s argument that dvandvas of the two oldest types should be interpreted not as compounds but rather as two separate words bound by semantic content yet nonetheless grammatically independent (Insler 1998). As pointed out by Insler, other scholars simply follow the R̥ gveda-padapāṭha which reads those dvandvas, with the components in juxtaposition, as single words. Compounds are usually analysed in the R̥ gveda-padapāṭha (i.e., the components are separated by an avagraha) when the first member is in a stem form, and the compound has only one accent. The compounds in tmesi are not restored in the R̥ gveda-padapāṭha unless they are proper names; e.g. R̥ V 5.2.7 Sp śúnaścicchépam, Pp śúnaḥ-śepam cit. However, devatādvandvas in tmesi are never restored; e.g., R̥ V 10.12.1 dyā́vā ha kṣā́mā is not restored in the padapāṭha although there are occurrences of dyā́vākṣā́mā in the R̥ gveda (e.g., R̥ V 1.96.5). Jha (1992: 168–169) believes that the constituent of compounds in tmesi expressing proper names lose their meaning and semantic focus unless they are together, whereas in the case of dvandva compounds having constituents in the dual, both constituents have a direct semantic relationship in the sentence, and therefore it is not necessary to restore them in the padapāṭha. His explanation reflects the views of traditional Sanskrit grammarians, particularly of Kātyāyana and Patañjali. Thus the padapāṭha, by not restoring the compounds comprising theonyms in tmesi, indicates that they were perceived as two independent words that occur in several alternative constructions, which seem to be expressions of the specific poetic style of the R̥ gveda. The R̥ gveda-padapāṭha also does not restore constructions having the first member in the dual and the second in the singular which is considered to be the oldest stage in the development of Vedic dvandvas (e.g., R̥ V 6.68.5: Sp indrā … varuṇa, Pp indrā … varuṇā),4 nor does it restore those constructions with the second member in dual and the first in singular (e.g., divás … pr̥thivyā́s).5 Therefore, it seems that dvandva compounds of types 4 This construction is regarded by Edgerton (1902: 117–118) as an intermeditate stage in the development of dvandvas from elliptic duals. 5 Some scholars believe that the transition from the syntactical collocation

Historical DevelopmentIntroduction and Typology of dvandva Compounds

Ù  79

1 and 2 with components in juxtaposition are not viewed in the R̥ gveda-padapāṭha as compounds but as single words, and when in tmesi, as two independent single words. Similar interpretation of devatādvandvas is given by traditional Sanskrit grammarians. It seems that Pāṇini does not conceive dvandvas in tmesi as compounds. Although he does not explicitly state that the constituents of a compound should be directly connected, this could be implied by his treatment of compounds as prātipadikas with a fixed order of the constituents. Each component of a dvandva compound in tmesi may be derived within the system of the Aṣṭādhyāyī as a case of the elliptic dual; however, most of the R̥ g vedic elliptic duals are not explicitly discussed by Pāṇini.6 Kātyāyana, in his rephrase of Pāṇini’s definition of dvandva compounds, says that each constituent of a dvandva denotes the sense of both members.7 Furthermore, in Vārttika 6,8 he points out that this applies also to non-compounded expressions, e.g., in the case of dvandvas in tmesi such as dyā́vā śocíḥ pr̥thivī́ (R̥ V 1.143.2) each dual, dyāvā or pr̥thivī, would denote, on its own, “Heaven and Earth.” According to Kātyāyana, each single constituent of both the elliptic duals and dvandvas represents the other constituent as well. This view is followed by Patañjali: he interprets each component of a dvandva compound in tmesi to be an independent word and comments that if simultaneous reference may occur in a non-compounded construction (giving as an example two duals in tmesi, dyāvā ha kṣāmā), how much more so when it takes place in a compound.9 to a dual compound is seen in a few examples of coordinative constructions where the second of two coordinate words is in the dual to indicate that an associated pair is meant; see Wackernagel (1957: 149-152) and Macdonell (1910: 156). 6 In the Aṣṭādhyāyī, the section on ekaśeṣa (P 1.2.64–1.2.73) deals with forms conveying more than one lexical meaning. The sūtra P 1.2.64 may account for the derivation of elliptic duals, but the interpretation of this rule raises numerous difficulties within the system of the Aṣṭādhyāyī and has therefore received much attention from the traditional commentaries as well as from modern scholarship; e.g., Joshi and Roodbergen (1993: 110), in their survey of commentaries on P 1.2.64, refute the view that ekaśeṣa is supposed to account for elliptic duals: they point out that the elliptic duals are a case of “particular meaningusages which, as such, lie outside the domain of grammar.” 7 Vārttika 2 on P 2.2.29: yugapadadhikaraṇavacane dvandvaḥ “a dvandva [is optionally formed] when the items denoted are simultaneously referred to [by each single constituent]”; translated by Joshi and Roodbergen (1974: xxx). It seems that Kātyāyana wants, by using yugapad, to include elliptic duals where one word simultaneously stands for the meaning of another word too and therefore the term yugapadadhikaraṇavacana applies to both elliptic duals and dvandvas; for discussion see Joshi and Roodbergen (1974: xxx). 8 Vārttika 6 on P 2.2.29: vigrahe ca yugapadvacanaṃ jñāpakaṃ yugapadvacanasya. 9 Bhāṣya 18 on Vārttika 6 on P 2.2.29.

80  Ù

The Vedas in Indian Culture and History

Dual deities that are attested in dvandva compounds of types 1 and 2 are alternatively expressed also in the elliptic dual. There is a close link between the older types of dvandva compounds and the elliptic duals: some scholars believe that an abbreviation of dvandva compounds of type 1 and 2 may be a possible origin of elliptic duals (Wackernagel 1957: 150–151), whereas others take the elliptic dual as the starting point from which dvandvas developed (Delbrück 1900: 138–139). Edgerton (1910: 110–120) proposes four stages for the development of dvandva compounds in Sanskrit, starting with the elliptic dual and ending with dvandva.10 Three of the six pairs of deities examined here occur in the elliptic dual (see Table 1), i.e., Dyaus and Pr̥thivī (dyā́vā in 5 instances), Mitra and Varuṇa (both theonyms occur in the elliptic dual, mítrā in 5 instances, and varuṇā in 3 instances), Uṣas and Nakta (uṣā́sā in 3 instances, and uṣā́sau in 1 instance). These dual deities are also attested in dvandvas of types 1 and 2, having their constituents in juxtaposition as well as in tmesi (see Table 1). The exception is the pair Indra and Varuṇa: though expressed in dvandvas of the oldest type, having most attestations in the family books and no occurrence in the latest maṇḍala 10 (Ditrich 2006: 131–132), the pair is not attested in the elliptic dual. However, these deities occur more frequently than other pairs in dvandvas in tmesi (in terms of their having the highest proportion of attestations, see Table 1) which probably represent an earlier stage in the development of the elliptic dual. It is evident that only those dual deities that occur in dvandvas of types 1 and 2 are also expressed, apart from tmesis, in the elliptic dual which indicates their status as independent words. On the other hand, the theonyms which occur in later types of dvandvas (e.g., Indra and Vāyu) are attested neither in the elliptic dual nor in dvandvas in tmesi. Furthermore, differences between older and later types of dvandvas can be observed in the occurrences of their components in syntagms constructed with the particle ca. As discussed by Klein (1981) and further developed by Jamison (1988), when dual theonyms occur in syntagms Vocative+Nominative+ca, they are 10 The four stages proposed by Edgerton (1910: 110–120) are: 1. elliptic duals: the more important member of a pair of nouns is used alone in the dual to express the pair, e.g. mitrā ́ (Indo-European stage, attested in all major language groups); 2. elliptic dual together with the singular of the other member: this construction is used to convey the idea of the pair more precisely, e.g. mitrā ́ … váruṇo … ca (Indo-European stage, attested in all major language groups); 3. dvandva compounds of type 1: the complementary singular is replaced by the dual, resulting in two duals in order to keep the pair-notion alive, e.g. mitrā ́váruṇā (Indo-Iranian stage, attested in Vedic and Avestan); 4. dvandva compounds with the first member in stem form, e.g. mitravaruṇau (Indo-Aryan stage, attested rarely in Vedic but most common in Classical Sanskrit).

Historical DevelopmentIntroduction and Typology of dvandva Compounds

Ù  81

found elsewhere in the R̥ gveda expressed in dvandvas, usually in the same hymn. Jamison (1988: 19–20) points out that the pair Indra and Vāyu is disproportionately represented in syntagms Vocative+Nominative+ca (vā́yav índraś ca or índraś ca vā́yav) because indravāyú is the only true morphological compound among R̥ g vedic devatādvandvas. All dual theonyms that occur in dvandvas of type 1 and 2 have also attestations in various syntagms constructed with coordinative particles. However, the components of the older types of dvandvas, when forming syntagms with the particle ca, appear mostly in hymns that do not address the pair in question and —unlike those of type 4— they are usually not attested in the same hymns in dvandva.11 Furthermore, the components of dvandvas of types 1 and 2 examined here occur rarely in a syntagm Vocative+Nominative+ca.12 These differences between older and later types of dvandvas indicate that the components of the older types could be viewed as two separate words rather than members of true morphological compounds. R̥ g vedic dvandvas of types 1 and 2 comprise mostly theonyms. Those consisting of two non-theonyms are very rare in the R̥ g veda; e.g., the nouns pitŕ̥- and mātŕ̥- although often used as epithets of dual deities and having over 100 attestations (most frequently referring to Heaven and Earth), occur, unlike theonyms, only once in dvandva compound (R̥ V 4.6.7 mātárāpitárā).13 There are only a few more examples of dvandvas of type 1 and 2 comprising 11 E.g., Dyaus and Pr̥ t hivī occur in syntagms constructed with the particle ca only in hymns dedicated to another deity/deities (i.e., in seven hymns dedicated to another deity and in one dedicated to all deities); the pair does not recur in the same hymn in dvandva compound (Ditrich forthcoming). 12 Mitra and Varuṇa occur, out of a total of 201 attestations in various coordinative constructions, in 11 syntagms with the particle ca (see Table 1); only three of these attestations are in Vocative+Nominative+ca (R̥V 5.64.5; 7.66.17, 18). Uṣas and Nakta are attested in only one construction with the particle ca, in N. A. Du. (1.73.7) and have no occurrence in Vocative+Nominative+ca. Indra and Varuṇa have no attestations in syntagms with ca, but appear once in a syntagm constructed with two particles ca (R̥V 7.83.6). Dyaus and Pr̥t hivī occur altogether in 8 syntagms constructed with ca but have only one attestation in the syntagm Vocative+Nominative+ca; the theonyms occur most frequently connected with the particle utá. 13 Pitŕ̥- and mātŕ̥- are attested 108 times in various coordinative constructions in the R̥ gveda: in dvandva only once; in the elliptic dual pitárā 49 times and mātárā 34 times; in asyndeta 15 times; in syntagms with one particle ca 5 times; with two particles ca 3 times; and with utá once. The pair refers unambiguously to Heaven and Earth in 73 instances; to parents (usually metaphorically comparing human parents with various deities) in 17 instances, to fire-sticks in 4 instances; to the Aśvins in 3 instances; and to Nakta and Uṣas once. The meaning of 14 attested coordinative constructions formed by pitŕ̥- and mātŕ̥- is uncertain and has been interpreted in various ways both by traditional commentaries and by modern scholarship (Ditrich 2005: 227–229).

The Vedas in Indian Culture and History

82  Ù

non-theonyms in the R̥ g veda (Collitz 1882: 287–298): turváśāyádū “Turváśa and Yádu” (R̥V 4.30.17); and dhúnīcúmurī “Dhúni and Cúmuri” (R̥V 6.20.13); both are proper names and have only one attestation each. Thus, furthering Insler’s argument, R̥ gvedic dvandvas of types 1 and 2 can be viewed as two independent words for the following reasons: only these types alternatively occur in the elliptic dual; only these types have their components attested in tmesi; when in tmesi these types are not viewed as compounds, neither by the R̥ gvedapadapāṭha nor by Pāṇini; when their components are expressed in syntactic constructions with coordinative particles they indicate different linguistic and stylistic features from younger types of dvandvas. Furthermore, R̥ gvedic dvandvas of types 1 and 2 consist almost exclusively of theonyms; those comprising non-theonyms, though too scarce to allow any general conclusions, never occur in tmesi. The evidence listed above suggests that it is only devatādvandvas that can be interpreted as two separate words. Dvandvas of types 3 and 4 Type 3 is very rare in the R̥ gveda: it is regarded as the third stage in the development of Vedic dvandvas where the first constituent loses its accent and the second constituent is accented on the last syllable. There is some disagreement as to whether dvandva compounds formed by Indra and Agni (indrāgnī́ ) belong to type 3 or 4 owing to sandhi of the final vowel of indrawith the initial vowel of agni-: some scholars (e.g. Grassmann 1873: 216; Reuter 1892: 178) interpret this compound as type 3 (i.e., the first component has a dual ending, indrā-agní), while others (e.g. Wackernagel 1957: 154) follow Sāyaṇa’s14 analysis and classify it as type 4 (i.e., the first component is in a stem form, indra-agní). My investigation indicates that indrāgnī́ could be placed among the older types for the following reasons: its distribution in the ten maṇḍalas shows that there are more attestations in the oldest layers of the R̥ gveda (Ditrich 2006: 132); the pair occurs, like all dvandvas of type 1 and 2, alternatively in tmesi, whereas Indra and Vāyu, the only other pair of theonyms that belongs without ambiguity to type 4, has no attestations in tmesi (see Table 1); the pair has only one attestation in a syntagm Vocative+Nominative+ca (R̥ V 3.25.4 ágna índraśca), and the theonyms do not occur in the same hymn in dvandva —unlike Indra and Vāyu which have, of all dual deities, by far the larg14

See Sāyaṇa on R̥V 1.21.1.

Historical DevelopmentIntroduction and Typology of dvandva Compounds

Ù  83

est number of attestations in this type of syntagm and mostly occur in the same hymn in dvandva;15 Indra and Agni —unlike Indra and Vāyu— are never expressed in the inverted syntagm Vocative+Nominative+ca. These features suggest that dvandvas comprising Indra and Agni follow the patterns similar to those of dual deities that occur in older types of dvandvas rather than those of type 4. Type 4 is very rare in the R̥ gveda but in post-Vedic Sanskrit it is the most common type of dvandva compound. Indravāyū́ seems to be the only example of a devatādvandva that unambigously belongs to this type: it has the first constituent in stem form and the whole compound has only one accent.16 This dvandva has no attestations in tmesi nor do the components occur in elliptic dual. There is a close relationship between the syntactic constructions with the particle ca and dvandva compounds of type 4: Indra and Vāyu occur more frequently than other dual deities in the syntagm Vocative+Nominative+ca, both in regular and inverted order (vā́yav índraś ca or índraś ca vā́yav)17 and, as shown by Jamison (1988), it is the only example among devatādvandvas of the R̥ gveda that, having only one accent and the first constituent in stem form, operates like a true morphological compound. There are only a few examples of dvandva compounds of type 4 comprising non-theonyms in the R̥ gveda:18 they have only one or two attestations each, mostly in the latest maṇḍala 10.

15 Indra and Vāyu occur in five of a total of seven constructions in Vocative+Nominative+ca: vā ́yavī ́ndraśca (R̥V 1.2.5, 1.2.6, 4.47.3), vāyo … índraśca (R̥V 1.135.4), índraśca vāyo (R̥V 4.47.2); the pair occurs in all these hymns also in dvandva compound. 16 There has been some discussion among scholars as to whether the compound indravāyū ́ is a dvandva with a single or a double accent: Grassmann (1873: 215) and Reuter (1892: 179) interpret it as a compound with two accents, i.e. índravāyú, without giving any evidence or developing an argument for this claim. On the other hand, Wackernagel (1957: 156–157) proposes the accent to be on the last syllable of the compound; similarly also Macdonell (1910: 156), Geldner (1907: 29), and Oldenberg (1888: 2). Arnold (1905: 290) suggests, on metrical grounds, that the first component of this compound should be índrā-; in this case the compound would belong to type 2; his interpretation is convincingly refuted by Oldenberg (1909-12: 2). 17 Indra and Vāyu occur in seven instances in syntagms constructed with the particle ca: five constructions are in Vocative+Nominative+ca (1.2.5; 1.135.4; 1.2.6; 4.47.2; 4.47.3). The components of all these constructions also occur in the same hymn in a dvandva compound. 18 The following dvandva compounds of type 4, consisting of non-theonyms, are attested in the R̥ gveda: r̥ksāmā b́ hyām “for R̥ c and Sāman” (R̥V 10.85.11, 10.114.6); satyānr̥té “truth and untruth” (R̥V 7.49.3); sāśanānaśané “which eats and which does not eat“ (R̥V 10.90.4) (Ditrich 2005: 454).

84  Ù

The Vedas in Indian Culture and History

Dvandvas of types 5 and 6 Dvandva compounds of type 5 (the first constituent is in stem form, the whole compound has only one accent and a plural ending) and type 6 (the whole compound has only one accent and appears in the collective neuter singular) are very rare in the language of the R̥ gveda but much more common in the later Vedic literature. There are only a few attestations of these two types:19 they all comprise non-theonyms and occur only in maṇḍala 10. Although the number of dvandva compounds comprising nontheonyms in the R̥ gveda is too small to allow conclusions, it seems that their distribution, unlike the distribution of those comprising theonyms, does reflect the generally accepted internal chronology of the R̥ gveda: the older types are attested in older maṇḍalas and the later types in the later maṇḍalas. Origin of dvandvas The origin of dvandvas is usually explained from two frequent grammatical practices in the Vedas: the use of asyndeton and the elliptic dual,20 both originating in the Indo-European period. Some scholars (e.g., Wackernagel 1957: 149–152; Edgerton 1910: 117–118) believe that the transition from the syntactical collocation to a dual compound is seen in a few examples of coordinative constructions where the second of two coordinate words is in the dual to indicate that an associated pair is meant (e.g. divás … pr̥thivyā́s). However, there is no clear indication of the antiquity of this construction in the R̥ gveda: among pairs that I have examined, these constructions (e.g., R̥V 8.25.2 mitrā́ … váruṇaḥ; R̥V 8.27.2 uṣā́sā náktam) are attested in maṇḍala 8 which is of an uncertain relative age and often considered a later maṇḍala of the R̥ gveda. Another example, often given by scholars in order to demonstrate the transition from the syntactic collocation to a dual compound, is diváṣpr̥thivyóḥ which, according to Wackernagel (1957: 149–152) and others, is seen to originate from two coordinate words the second of which is in the dual to in19 The following dvandva compounds of types 5 and 6 are attested in the R̥ gveda: ahorātrā ́ṇī “days and nights” (type 5, one attestation, R̥V 10.190.2); ajāváyaú “goats and sheep” (type 5, one attestation, 10.90.10); dhanabhakṣéṣu “gifts and enjoyments” (type 5, one attestation, 10.102.1); iṣṭāpūrténa “by what was offered and given” (type 6, one attestation, 10.14.8) (Collitz 1882: 287–298). 20 Modern scholars usually explain the origin of dvandvas either from asyndeta (Reuter 1892: 172–173) or from elliptic duals (Brugmann 1922: 301; Delbrück 1900: 138–139; Edgerton 1910: 110–120; Renou: 1952: 124) or from both asyndeta and elliptic duals (Wackernagel 1957: 150–151; Macdonell 1910: 156).

Historical DevelopmentIntroduction and Typology of dvandva Compounds

Ù  85

dicate that an associated pair is meant; however, the compound diváṣpr̥thivyóḥ has two attestations in the family books (maṇḍalas 2 and 5) and two in maṇḍala 10, thus giving no clear evidence of its antiquity.21 All dvandvas that I have examined are attested in Vedic only and have no parallels in other old Indo-European languages, although the constituents of most of them are of Indo-European or at least Indo-Iranian origin.22 However, the expression of those pairs in dvandva compounds and in a variety of other coordinative constructions seems to have developed within Indo-Aryan and to reflect the specific style of Vedic. In my past research (Ditrich 2006) I have shown that the distribution of coordinative constructions comprised of theonyms —unlike those comprised of non-theonyms— does not reflect the usually accepted internal chronology of the R̥ gveda which situates the family books (maṇḍalas 2–7) and maṇḍala 10 at opposite ends of the chronological spectrum: the older and younger types of devatādvandvas are used throughout the R̥ gveda without any marked differences among the ten maṇḍalas, whereas dvandvas comprising non-theonyms— though too scarce to allow any general conclusion— do reflect the suggested historical development of dvandvas, having the older types attested in older maṇḍalas and the later types in the later maṇḍalas. I have argued (Ditrich forthcoming) that various types of dvandvas and other coordinative constructions expressing dual theonyms seem to be stylistic variants: different stylistic patterns are found depending on whether the dual theonyms are attested in hymns which address the very same pair or in hymns addressing all deities, or a different deity. III. Dvandvas in relation to other coordinative constructions The components of devatādvandvas are frequently alternatively expressed in other coordinative constructions in the R̥ gveda, i.e., asyndeta, elliptic duals, and syntagms constructed with copulative conjunctions. For the dual deities examined here, the distribution of occurances in different coordinative constructions is as follows: 21 Furthermore, although the padapāṭha restores proper names in tmesi it does not restore the tmesis of two theonyms having the second member in the dual and the first in the singular (e.g. R̥V 5.62.3: Sp mítrarājānā varuṇā, Pp mítra-rājānā varuṇā), nor the constructions having the first member in the dual and the second in the singular (e.g. R̥V 6.68.5: Sp indrā … varuṇa, Pp indrā … varuṇā); only the singular ending -a is replaced with the dual ending -ā. 22 Indra, Vāyu, Mitra, Uṣas, Nakta, Dyaus and Pr̥t hivī are of IE origin, having parallels in several IE languages. Only Varuṇa has uncertain etymology and no Iranian parallels.

86  Ù

The Vedas in Indian Culture and History

Table 2 Dual deities

Dvandva

Elliptic Dual

Asyndeton

Conjunctions

Indra and Vāyu

47.9%

0.0%

37.5%

14.6%

Mitra and Varuṇa

48.8%

4%

39.7%

7.5%

Indra and Varuṇa

78.8%

0.0%

19.7%

1.5%

Indra and Agni

86.4%

0.0%

9.1%

4.5%

Dyaus and Pr̥t hivī

46.0%

2.2%

25.4%

26.4%

Uṣas and Nakta

77.3%

18.2%

0.0%

4.5%

For all the pairs examined the preferred coordinative construction is dvandva, followed, for most pairs, by asyndeton. I have shown (Ditrich forthcoming) that the theonyms occur in dvandva compounds most frequently in hymns that address the pair in question, whereas in hymns addressing another deity the most common expression is asyndeton. Syntagms constructed with coordinative conjunctions are significantly less preferred and elliptic duals are the most marginal option. The high percentage of constructions with conjunctions for Dyaus and Pr̥thivī23 is due to the frequent usage of the particle utá in hymns R̥ V 1.94–1.115, attributed to Kutsa Āṅgirasa, which seem to follow specific stylistic patterns —Dyaus and Pr̥thivī always appear connected by the particle utá in the final stanza of each hymn.24 The relatively high usage of conjunctions for Indra and Vāyu is due to the frequent occurrences of syntagms Vocative+Nominative+ca, attested mainly in maṇḍala 1; the components of these syntagms nearly always occur in the same hymn also in dvandva. Only those theonyms that occur in dvandvas of types 1 and 2 appear elsewhere in the R̥ V also in tmesi and elliptic dual, and follow different linguistic patterns when attested in syntagms with coordinative particles. As noted above, dvandvas comprising non-theonyms are very rare in the R̥ gveda; each compound has one or, only rarely, two attestations. The components of these dvandvas usually do not occur in other coordinative constructions. Non-theonyms which are used to refer to dual deities usually do not occur in dvandva neither do they display such a variety of coordinative expressions 23 Dyaus and Pr̥thivī occur in 59 syntagms constructed with coordinative particles: 37 are constructed with the particle utá; 8 with one particle ca; 7 with two particles ca; two with the particle ná; and 5 with the particle ā ́. 24 E.g., R̥V 1.94.16cd: tán no mitró váruṇo māmahantām áditiḥ síndhuḥ pr̥thivī ́ utá dyaúḥ.

Historical DevelopmentIntroduction and Typology of dvandva Compounds

Ù  87

as the theonyms do (Ditrich 2005: 369–370; 389–397). This can be demonstrated in the example of the coordinative constructions comprising nouns pitŕ̥- and mātŕ̥-: although most coordinative constructions of these two nouns refer to dual deities (at least 98 out of a total of 108 attestations),25 the pair occurs only once in dvandva compound (R̥ V 4.6.7 mātárāpitárā) and does not follow expressive paradigms comprising the usual variations between different types of nominal constructions for dual theonyms. By far the most common construction expressing coordinative relation between non-theonyms is asyndeton, followed by much less common syntagms with conjunctions. I have shown in my analysis of all coordinative constructions in R̥ V 1.1–1.50 (Ditrich 2005: 389–397) that only rarely are two non-theonyms that are attested in asyndeton also expressed elsewhere in a syntagm with coordinative conjunctions, and they do not occur, apart from the nouns pitŕ̥- and mātŕ̥-, in the elliptic dual or dvandva compound.26 Two non-theonyms in syntagms with the particles ca and utá seem to express a casual relation between two nouns; usually they have only a few attestations in the whole R̥ gveda. On the other hand, dual theonyms clearly follow different typological and stylistic patterns: they appear in a variety of coordinative constructions, usually within the same hymn, having dvandvas as the preferred expression (see Table 2). These idiosyncrasies of dual theonyms seem to reflect Vedic ideas about the extreme importance and the magical power of divine names: they follow special, well-established stylistic paradigms and do not reflect the historical development of the language that must have taken place during the composition of the R̥ gveda. IV. Numerals As discussed above, when studying dvandvas and other coordinative constructions, a distinction has to be made between those having theonyms as components, and those consisiting of nontheonyms. Numeral dvandvas, though comprising non-theonyms, display specific features that require them to be discussed as a separate category. The cardinal numerals from eleven to nineteen 25 Pitŕ-̥ and mātŕ-̥ appear, apart from one attestation in dvandva compound (mātárāpitárā), in the following constructions: 83 times in the elliptic dual (with 49 occurences of pitárā, and 34 of mātárā); 15 times in asyndeton; and 9 times in syntagms constructed with conjunctions ca and utá; for the complete list see Ditrich 2005: 279–280. 26 E.g., gó- and áśva- occur 42 times in asyndeton, but never in any other coordinative construction; dvipád- and cátuṣpad- have 20 attestations in asyndeton and only one occurrence in a syntagm with the particle ca and one with utá.

88  Ù

The Vedas in Indian Culture and History

and those between the decades form dvandva compounds which are of Indo-European origin. In Vedic as well as in later Sanskrit the cardinal numerals from eleven to nineteen consist of two numerals: the first numeral is in the dual (ékā-daśa and dvā́-daśa), plural (tráyo-daśa) or singular (from cátur-daśa to náva-daśa), the second numeral is just in the stem form daśa and —like numerals from five to ten— has neither gender distinction nor case ending. The cardinal numerals between the decades (and hundreds) are dvandva compounds comprising the numbers from one to nine for the first component and decades for the second component.27 Numeral dvandvas always have one accent on the first component of the compound,28 and the order of constituents follows the rule that the constituent signifying the lower number is placed first.29 The components of compounded cardinal numerals can be expressed in a dvandva compound as well as in an asyndeton or syntagms constructed with a connective particle;30 this variety of coordinative constructions for numerals is probably of IndoEuropean origin.31 Although the alternative ways of expressing compounded numerals occur already in the R̥ gveda, they are much more common in later Sanskrit. I have shown in my analysis of all coordinative constructions in R̥ V 1.1–1.50 (Ditrich 2005: 455–457) that numeral dvandva compounds signifying numbers higher than twenty alternate considerably more often with syntactic strings than lower numbers do; e.g., the numerals ékādaśa and dvā́daśa occur only in dvandva,32 no alternative coordinative construction is attested in the R̥ gveda; the numeral tráyastriṃśatis attested in dvandva compound as well as in asyndeton, constructions with one particle ca and with two particles ca.33 27 The numerals from one to nineteen are usually used as adjectives and those above nineteen as nouns. 28 According to P 6.2.35, the numeral, standing as the first member of a dvandva compound, preserves its accent. The attestations in Old Greek give the evidence that the accent on the first member is inherited from Indo-European (Wackernagel 1957: 379). 29 According to Vārttika 8 on P 2.2.34 the constituent signifying the lower number is placed first; see also Wackernagel (1957: 379). 30 E.g., constructions with one particle ca: R̥V 3.6.9c triṃśátaṃ trī ́ṃś ca “thirty three”; with two particles ca: R̥V 10.93.15a saptatíṃ ca saptá ca “seventy-seven”; in asyndeton: (R̥V 2.19.6c navatím … náva “ninety-nine”; or both in asyndeton and syntagm with two particles ca: R̥V 10.52.6ab trī ́ni śatā ́ trī ́ sahásrāṇi … triṃśác ca devā ́ náva ca “three thousand three hundred thirty-nine gods.” 31 Similar alternations are attested in Avestan, where a compounded numeral can be expressed also in a syntagm using one or two coordinative particles ča; in Old Greek with the particle καί; and in Latin with the particles et or atque. 32 The numeral ékādaśa has three attestation (R̥V 1.139.11abc), and dvā ́daśa four (R̥V 1.164.48; 1.25.8; 4.337; 10.114.5). 33 The numeral tráyastriṃśat- appears in the following constructions: once in

Historical DevelopmentIntroduction and Typology of dvandva Compounds

Ù  89

Numeral dvandvas follow special patterns that are different from dvandvas comprised of theonyms and non-theonyms: they have one accent on the first component, inherited from IndoEuropean; the constituent signifying the lower number is placed first; their derivation is completely accounted for by Pāṇini;34 and they are analysed as compounds in the padapāṭha.35 Moreover, the coordination of two cardinal numerals is significantly more frequently expressed in syntagms constructed with the particle ca than the coordination of two theonyms or non-theonyms;36 the least frequent coordinative construction for numerals —unlike theonyms— seems to be dvandva, especially for numbers higher than twenty. V. Conclusion Firstly, this paper has examined the typology of dvandva compounds, focusing particularly on the first two types, and revisited Insler’s interpretation of types 1 and 2 as two separate words rather than compounds (Insler 1998). It was shown that only the components of the oldest types of dvandvas are attested in tmesi and alternatively occur in the elliptic dual; that the oldest types of dvandvas in tmesi are not viewed as compounds, neither in the R̥gvedapadapāṭha nor by Pāṇini; when the components of the oldest types of dvandvas are expressed in syntactic constructions with coordinative particles there is a marked difference between older and younger types; that constructions comprisdvandva compound (R̥V 1.45.2); once in asyndeton (R̥V 8.28.1), once in a syntagm with one particle ca (R̥V 3.6.9); and once with two particles ca (R̥V 8.30.2). 34 The numerals 11 to 19 and those between the decades are treated by Pāṇini as dvandvas, the rules for their derivation are provided. According to Vārttika 8 on P 2.2.34 the constituent signifying the lower number is placed first. In contrast, the derivation of dvandvas comprising theonyms is not completely accounted for by Pāṇini (Ditrich 2005: 429–435). 35 Most numerals fulfill both criteria for the analysis of dvandva compounds in the padapāṭha, i.e., the compound has a single accent, the first member is in stem form, and its components are separated by an avagraha; e.g. R̥V 10.114.7: Sp cáturdaśa, Pp cátuḥ-daśa. The numerals ekā ́daśa and dvā ́daśa, having a single accent and the first members in the dual, are not analysed as compounds; however, the compounds formed with trayas- as the first member are analysed in the padapāṭha even though trayas is not a stem form; e.g. R̥V 1.45.2 Sp tráyastriṃśat, Pp tráyaḥ-triṃśat. 36 E.g., the numeral náva- is attested in the R̥gveda 24 times, having 21 attestations in coordinative constructions with navatí-; the numeral navatí- is attested 26 times, having 19 attestations in coordinative constructions with náva. Návaand navatí- occur together in various coordinative constructions in the R̥ gveda: 9 times in asyndeton; 9 times in syntagms constructed with one particle ca; and once with two particles ca; the numerals are not attested in dvandva.

90  Ù

The Vedas in Indian Culture and History

ing one constituent in the dual and the other in the singular (which, some scholars believe, represent the most archaic stage of development of dvandvas) do not occur in the oldest layers of the R̥gveda. Similarly, there is no evidence that the oldest types of dvandvas comprising theonyms appear in older maṇḍalas, and later types in later maṇḍalas, and thus reflect stages in the historical development of dvandva compounds. The question as to whether the oldest types of dvandva compounds should be interpreted as two separate words can be answered affirmatively. The examination of different types of dvandvas and the variant usage of their components in other coordinative constructions shows that these compounds are stylistic variants expressing coordination rather than stages in the historical development of dvandvas in Indo-Aryan. A prominent stylistic feature of the R̥gveda is the repetition of theonyms, usually in hymns addressing the deity/deities in question: these are either repetitions of the identical form (e.g., in nearly every stanza the deity’s name is repeated in the vocative) or repetitions of theonyms in a variety of expressions (in various cases, or, with dual deities, in various coordinative expressions). The theonyms display specific grammatical and stylistic features, the reason for which lies in Vedic ideas about the extreme importance of divine names. It seems that dvandva compounds of types 1 and 2 can be interpreted as stylistic variants rather than as reflections of the historical development of dvandvas in Indo-Aryan; this is also indicated by their distribution in the ten maṇḍalas which does not reflect the generally accepted internal chronology of the R̥gveda. Secondly, it is argued in this paper that dvandva compounds cannot be viewed as a single category but rather a distinction has to be made between those comprising theonyms and those comprising non-theonyms, each displaying different linguistic features. It has shown that dual theonyms display specific grammatical and linguistic features: several types of constructions are used to express a copulative relation between pairs of theonyms (dvandvas being preferred), whereas non-theonyms display hardly any such alternations, as illustrated by the example of pitŕ̥and mātŕ̥-. There is also a marked difference between the distribution of dual theonyms and dual non-theonyms in the R̥gveda: the older types of dvandvas comprising non-theonyms —unlike those comprised of theonyms— are attested in older maṇḍalas and the later types in the later maṇḍalas. Furthermore, numeral dvandvas are different from those consisting of theonyms as well as of non-theonyms, especially in the following aspects: they have one accent on the first component; the coordination of two cardinal numerals is significantly more frequently expressed in syntagms constructed with the particle ca than coordination

Historical DevelopmentIntroduction and Typology of dvandva Compounds

Ù  91

of two theonyms or non-theonyms; the least frequent coordinative construction for numerals —unlike theonyms— seems to be dvandva, especially for numbers higher than twenty. Therefore, the examination of the typology and historical development of dvandvas, and of their relationship with other coordinative constructions (asyndeta, elliptic duals, and syntagms constructed with copulative conjunctions) indicates that dvandvas cannot be viewed as a single category but rather a distinction has to be made among those comprising theonyms, non-theonyms or numerals, each displaying different linguistic and stylistic features. References Arnold, Edward Vernon. 1905. Vedic Metre: In its Historical Development. Cambridge: Cambridge University Press. Repr. 1967. Brugmann, Karl. 1922. Kurze vergleichende Grammatik der indogermanischen Sprachen. Berlin–Leipzig: W. de Gruyter. Burrow, Thomas. 1969. The Sanskrit Language. London: Faber and Faber. Collitz, Hermann. 1882. “Ueber eine besondere Art vedischer Composita.” In: Verhandlungen des fünften internationalen Orientalisten-Congresses: gehalten zu Berlin im September 1881, Vol. II, 2. Berlin: A. Asher & Co. Weidmannsche Buchhandlung, 287–298. Delbrück, Berthold. 1900. Vergleichende Syntax der indogermanischen Sprachen. Theil 3: Grundriss der vergleichenden Grammatik der indogermanischen Sprachen. Ed. by Karl Brugmann and Berthold Delbrück, Band 3-5. Strassburg: Karl J. Trübner. Ditrich, Tamara. 2005. “Dvandva Compounds in the R̥ gveda: A Stylistic and Typological Analysis of Coordinative Nominal Constructions in R̥ gveda 1.1–1.50.” PhD Thesis. Brisbane: University of Queensland. ———. 2006. “Dvandva Compounds in the R̥ gveda: The Chronology of the Ten Maṇḍalas Revisited”. Poligrafi 41/42, 123-148. ———. (forthcoming). “Stylistic Analysis of Coordinative Nominal Constructions for Dual Deities in the R̥gveda”. In Proceedings of the 13th World Sanskrit Conference, Edinburgh, July 2006. Duchesne-Guillemin, Jacques. 1936. Les composés de l’Avesta. Études de morphologie iranienne. Vol. I. Paris: Librairie E. Droz. Edgeron, Franklin. 1910. “Origin and Development of the Elliptic Dual and of Dvandva Compounds.” Zeitschrift für vergleichende Sprachforschung auf dem Gebiete der indogermanischen Sprachen 43, 110-120. Geldner, Karl F. 1907-1909. Der Rigveda in Auswahl. 2 Vols. Stutt-

92  Ù

The Vedas in Indian Culture and History

gart: W. Kohlhammer. Grassmann, Hermann. 1873. Wörterbuch zum Rig-Veda. Leipzig: F.A. Brockhaus. Insler, Stanley. 1998. “mitrā́váruṇā or mitrā́ váruṇā?” In: J. Jasanoff et al. (eds.), Mír Curad: Studies in Honor of Calvert Watkinspp. Innsbruck: Institut für Sprachwissenschaft der Universität Innsbruck, 285-290. Jamison, Stephanie W. 1988. “Vāyav Indraś ca Revisited.” Münchener Studien zur Sprachwissenschaft 49, 13-59. Jha, V.N. 1992. A Linguistic Analysis of the R̥ gveda-padapāṭha. Sri Garib Dass Oriental Series No. 142, Pre-Pāṇinian Grammatical Traditions, Part 1. Delhi: Sri Satguru Publications. Joshi, Shivram D. and Jouthe A.F. Roodbergen. 1993. The Aṣṭādhyāyī of Pāṇini: With Translation and Explanatory Notes. Vol. 2. New Delhi: Sahitya Akademi. ———. 1974. Vyākaraṇa-Mahābhāṣya: Bahuvrīhidvandvāhnika (P. 2.2.23–2.2.38). Publications of the Centre of Advanced Study in Sanskrit, Class C, No. 9. Poona: University of Poona. Klein, Jared S. 1981. “The Origin of the Rigvedic vā́yav índraś ca Construction.” Münchener Studien zur Sprachwissenschaft 40, 73-91. Macdonell, Arthur Anthony. 1910. Vedic Grammar. Strassburg: Karl Trübner. Oldenberg, Hermann. 1888. Die Hymnen des Rigveda. Band I: Metrische und textgeschichtliche Prolegomena. Berlin: Wilhelm Hertz. ———. 1909-1912. R̥ gveda. Textkritische und exegetische Noten. 2 Vols. Abhandlungen der Königlichen Gesellschaft der Wissenschaften zu Göttingen, Philologisch-historische Klasse. Berlin: Weidmannsche Buchhandlung. Repr. Nendeln, Liechtenstein: Kraus Reprint, 1970. Renou, Louis. 1952. Grammaire de la langue védique. Collection “Les langues du monde,” Vol. 9. Paris: IAC. Reuter, J. 1892. “Die altindischen Nominalkomposita, ihrer Betonung nach untersucht.” Zeitschrift für vergleichende Sprachforschung auf dem Gebiete der indogermanischen Sprachen 31, 157–232; 485–612. Wackernagel, Jacob. 1957. Altindische Grammatik: Einleitung zur Wortlehre, Nominalkomposition. Band II,1. 2nd ed. Göttingen: Vandenhoeck & Ruprecht.

Introduction Observations on Vedic Accents in Grantha Palmleaf Manuscripts

Ù  93

Observations on Vedic Accents in Grantha Palmleaf Manuscripts Saraju R ath International Institute for Asian Studies, Leiden, The Netherlands 1. Introduction 1.1 According to A.C. Burnell, in his Elements of South-Indian Palaeography (1878: 82): “The different methods used for the different Vedas are all of very recent origin, comparatively; and have arisen in different parts of India much about the same time, and in consequence of the decay of the old way of learning the Vedas by heart. In S[outh] India there is no pretence of a complete or even uniform system, and mss. with accents do not appear to occur before the middle of the sixteenth century.” Hence, in the view of Burnell: “Palaeographically, the notation of the Vedic accents is a subject almost devoid of interest.” Burnell’s negative judgment of the importance of accents in South Indian Vedic manuscripts is perhaps the reason why this subject has subsequently been largely neglected in handbooks and studies of the South Indian scripts such as Grantha. Witzel, in his important and stimulating 1974 article devoted to “Systems of Marking Vedic Accents,” almost entirely neglects South Indian systems of marking accents except for a brief paragraph, which does not go beyond the observations Burnell made in 1878.1 By 1 Witzel 1974 : 485 (par. 2.2.3): “A similar system is also met with in some old South Indian W [sic; meant is probably: Y V] and RV mss, i.e. those in Grantha script mark the udātta by a superscribed u, and those in Telugu by superscribed circle.” In the accompanying note 49 Witzel refers to Renou’s École Védiques et la Formation du Veda (Paris 1947), 16 note 1, to Burnell South Ind. Paleography, p. 60 (probably to the first edition, not to the second, enlarged and improved edition used by me, where the relevant passage is found on p. 81-82), and to von

94  Ù

The Vedas in Indian Culture and History

now we know that South India has conserved very ancient features of the recitation of Vedic texts which have disappeared elsewhere (cf. Staal 1961 and 1983). This in itself is a sufficient reason to have a closer look at the notation of Vedic accents in this area. We will here focus on systems of accent notation used in Grantha manuscripts of Vedic texts. This is necessary (a) in order to have a more complete grasp of the Grantha script so that misreadings and misinterpretations are avoided; and (b) in order to make better use of available sources that can provide information on the textual form of the Vedas and the tradition of their recitation. 1.2 In Vedic texts and in the oldest form of Sanskrit, an accent system is found in which each word has in principle a single main accent which it maintains or gives up depending on its role in the sentence. Pāṇini and the Prātiśākhyas give quite detailed descriptions of the accent system. A tone accent is still there in the traditional recitation of Vedic texts. Neither Pāṇini’s and the Prātiśākhyas’ descriptions nor the current actual recitation in traditional schools will be the main subject of this paper although they are both relevant and important. Instead of that my focus is here on how vedic accents are noted in Grantha palmleaf manuscripts. A description of the notation of Vedic accents is required in order to fill an important lacuna in available handbooks on South Indian scripts from A.C. Burnell onwards. It is true that a system of accent notation developed relatively late in the South Indian scripts, as was already pointed out by Burnell in 1878. It is also true that in several respects the Prātiśākhyas give more detailed information on accents and their pronunciation than what is found in accented manuscripts. However, it is well known that South India has conserved ancient features in the recitation of Vedic texts which have disappeared or are very rare elsewhere.2 Therefore, an effort should be made to analyse, read, and interpret the systems of accent notation found in South Indian Grantha manuscripts, so that they can be used for the history of recitation and for the establishment of critical editions. To this the present paper seeks to provide a contribution. We will study a small numSchroeder’s introduction to his edition of the Maitrāyaṇī Saṃhitā, p. XXXII. For the accentuation of southern manuscripts Renou and von Schroeder again refer only to p. 60 of Burnell’s book. 2 Staal 1983: 172 : “The Nambudiri Vedic tradition makes an archaic impression when compared to other Vedic traditions in present-day India.… In the domain of recitation, there is a unique survival of a very ancient feature: in a special form of R̥ gveda recitation, called jaṭāpāṭha, the originally raised udātta accent, which has elsewhere disappeared, is still recited at the highest pitch ...” Cf. also Staal 1961: 43, 47.

Introduction Observations on Vedic Accents in Grantha Palmleaf Manuscripts

Ù  95

ber of selected examples which will lead to several conclusions regarding Vedic accentuation in Grantha palmleaf manuscripts. 2. Generalities on systems of accentuation 2.1 Before turning to the peculiarities in notating vedic accents in South Indian Grantha palmleaf manuscripts, a brief general overview of Vedic accents and their notation is necessary. From the observations and descriptions given in Pāṇini’s grammar and in the various phonetic and grammatical treatises for specific Vedic schools (śākhās), the Prātiśākhyas, we know that the accent was mainly a musical or a tone accent. The high pitch tone which is regarded as the main accent is called udātta.3 It contrasts with the low pitch tone, which is called anudātta.4 The tradition distinguishes a third accent which is a combination of udātta and anudātta, and which is known as svarita.5 To explain the nature of the svarita, it is said that its first part, in case it immediately follows an udātta, is in a high pitch, to the extent of half a short vowel.6 Along with these texts, R̥ gvedaprātiśākhya,7 R̥ ktantra,8 and Patañjali’s Mahābhāṣya9 and other grammatical treatises also give similar indications about the nature and pronunciation of accents in recitation, from which it becomes clear, however, that the part of the svarita immediately following udātta came to be pronounced in a higher pitch in some traditions, which is still the common practice in modern Veda recitation.10 2.2 With regard to the notation of accent in Vedic manuscripts in Devanāgarī, it is to be noted that the earlier recitation system was completely based on oral tradition. After the writing system was introduced, the need for preservation of the oral pattern was felt, and so far as the use of the vedic mantras (to be used in Śrauta and Gr̥hya ritual performances) in the later, medieval period is concerned, it is expected that the recitation system followed the principles and rules mentioned in the phonetical and grammatical texts. Different śākhās in different Vedic schools show a numA 1.2.29 uccair udāttaḥ | TPr 1. 38. A 1.2.30 nīcair anudāttaḥ | TPr 1. 39. 5 A 1.2.31 samāhāraḥ svaritaḥ | TPr 1. 40. 6 A 1.2.32 tasyādita udāttam ardhahrasvam | TPr 1.41, tasyādir uccaistarām udāttād anantare yāvad ardhaṁ hrasvasya. 7 R̥V Pr 3.1ab udāttaś cānudāttaś ca svaritaś ca trayaḥ svarāḥ āyāma-viśrambhākṣepais ta ucyante | 8 R̥T 2.6.3 ādyārdhamātrā svaritam | 9 Patañjali’s Mahābhāṣya on A 1.2.33 sapta svarā bhavanti | udāttaḥ, udāttataraḥ, anudāttaḥ, anudāttataraḥ, svaritaḥ, svarite ya udāttaḥ so ’nyena viśiṣṭaḥ | 10 See discussion in Macdonell 1910: 77 and Staal 1961: 22-24. 3 4

The Vedas in Indian Culture and History

96  Ù

ber of styles of ritual performance and recitation. Only in the past few decades the various types of Vedic recitation have been well studied, recorded, and published in a descriptive way. Apart from the difference in Vedic schools or śākhās, more detailed information has become available on regional variation. This research progresses in two ways: (i) theoretical, i.e., based on written texts and manuscripts, preserved so far in the tradition; (ii) based on the direct recitation, i.e., the musical notes as used by the traditional reciter/chanter during their performances on various occasions. 2.3 The pattern of accent markings followed in Devanāgarī manuscripts11 is generally well known and is described in handbooks. Its main characteristics are as follows: i. No mark is used to indicate udātta (although linguistically it is the major accent) in the R̥ g - and Atharvaveda and in the Taittirīya tradition (not in the Maitrāyaṇī and Kāṭhaka Saṃhitā and, depending on the interpretation of the accent marks, the Śatapatha Brāhmaṇa). ii. In the Maitrāyaṇī and Kāṭhaka Saṃhitā the udātta is marked with a vertical stroke above the character (similar and usual way for the svarita mark seen in the R̥ gveda and texts that follow its accentuation). iii. The only accent mark found in the Śatapatha Brāhmaṇa is a horizontal bar (often not a straight but a slightly bent one) below the syllable which gets udātta value in other vedic texts. iv. A small horizontal bar is used below the syllable to show the low pitch accent, i.e., anudātta (in R̥ g vedic texts as well as the Maitrāyaṇī and Kāṭhaka Saṃhitā). v. A vertical stroke above the syllable (on the right side top) indicates the svarita accent (in R̥ gvedic texts). vi. To mark a dependent svarita, Maitrāyaṇī manuscripts give a horizontal stroke or three similar strokes crossing the middle of the syllable and Kāṭhaka manuscripts put just a dot below the accented syllable. vii. The mark used in Maitrāyaṇī manuscripts to show an independent svarita is indicated by a curve below while in Kāṭhaka manuscripts it is shown by a curve below if an unaccented syllable follows, and by a hook below if the following syllable is accented. 2.4 The system of accentuation marks found in South Indian Vedic manuscripts show a much greater variation and are less consistent in following a fixed pattern, as already observed by 11

See Macdonell 1916: 448-451.

introduction oBserVations on Vedic accents in grantha Palmleaf manuscriPts

Ù  97

Burnell cited above. The systems of marking accents in Southern manuscripts vary not only because of different vedic recitation practices prescribed and followed in different śākhās, but also in their notation from one script to another. The marks used in Vedic manuscripts written in Malayalam script do not appear in the same way in Telugu or Grantha manuscripts. The most interesting but also challenging part in this is that the no specific marks have been prescribed or recorded anywhere for this. A thorough study of all the South Indian Vedic manuscripts preserved in different libraries in India and abroad is therefore required on the basis of which we have to infer the various systems followed by the scribes. My present study12 on the Vedic Grantha manuscripts focuses on the marks distinct from the usual systems. 3. Examples and observations 3.1 Accent marks in Vedic manuscripts developed quite late (cf. now also Falk 1993: 266f.), and are attested only occasionally in relatively recent manuscripts (cf. Burnell 1878: 82), and this applies to both North and South Indian scripts. Burnell, according to the passage cited above, thinks that in the South manuscripts with accents “do not appear to occur before the middle of the sixteenth century.” On the basis of my own reading experience I would conclude that the transition from non-accented to accented Vedic manuscripts was indeed around the middle of the sixteenth century or a little earlier. By way of illustration we may consider the manuscripts used for the R̥ gveda Saṃhitā- and Pada-pāṭha Vaidika Samshodhana Mandala edition with Sāyaṇa’s commentary (Poona, 1933-1951). Some old palmleaf manuscripts used for this edition are in Grantha and in old Grantha-Malayalam character and they are entirely without any trace of accentuation. Neither do we find indications of anunāsika, varieties of anusvāra, or samastapada.

Ms. illustration I. 12 This study on South Indian Vedic manuscripts is still continuing. Here I present some of my findings regarding manuscripts written in Grantha script because most of the Vedic manuscripts found in South India are in this script.

the Vedas in indian culture and history

98 Ù

This is one of the oldest palmleaf manuscripts out of few others used for the above R̥ V edition. The date of this manuscript is not mentioned in the colophon. However, on the basis of paleography and material, it must be from ca. 15th–(early)16th century C.E. Observations: in old Grantha manuscripts (as in old manuscripts generally) we find: • •

no mark used for udātta, anudātta, svarita varieties no mark used for anunāsika, varieties of anusvāra.

3.2 We find specific svara marks in some Grantha manuscripts (in irregular way) from the early 16th century onwards, for instance in the following manuscript of the R̥ gveda samhitā text (preserved in the Adyar Library, Chennai), with R̥ V 1.1 on the first folio:

Ms. illustration II.

On the basis of paleography and material, the manuscript can be assigned to the early 16th century A.D. Observations: • • •

the udātta is marked with a small horizontal bar (above) there is no mark for anudātta or svarita no mark for anunāsika

3.3 Yajurveda: a Taittirīya Padapāṭha manuscript (1.6.1):

Ms. illustration III.

introduction oBserVations on Vedic accents in grantha Palmleaf manuscriPts

Ù  99

In this Padapāṭha manuscript some special marks are used to indicate the split or division of padas. Observations: • udātta:

mark (above the syllable)

• anunāsika:

mark in between characters,

• samastapada (to split members of compound etc.):

,

, See the following clips from Ms. illustration III: (i) (ii) Read (i) as: ap-su, (ii) as: sā-akṣāt. • a special mark is placed under the last syllable of the word that is sarvānudātta. Compare the manuscript clips with my Grantha and Devanāgarī transcription:

Elsewhere the padapāṭha represents an accented finite verb karoti (udātta sign on -ro-) as follows:

3.4 Sāmaveda: for instance a passage giving the Pravatbhārgavasāman, to be chanted by the Udgātar in the Prāyaṇīyeṣṭi.

the Vedas in indian culture and history

100 Ù

Ms. illustration IVa, folio 1a, left-hand half

Ms. illustration IVb, folio 1a, right-hand half

Compare the accented Grantha version of proāyāsā-it ... in the first line with the same passage in Devanāgarī (which has here proayāsā-it ...) with svara marks:

Observations: • svarita mark ‘2’ (Devanāgarī) corresponds with: the letter ‘ra’ (in between two characters) • ra (Devanāgarī) used for adding two mātrās corresponds with: ° (above the character) • ‘2’ between characters indicating preṅkha13 which adds two 13

Preṅkha is one out of seven vikr̥ti notes prescribed in Sāmagāna texts. They

introduction oBserVations on Vedic accents in grantha Palmleaf manuscriPts

Ù  101

mātrās to the preceeding syllable corresponds with: pre • horizontal bar above the preṅkha ‘2’ corresponds with the syllables: kai, khai, gai, rai etc., thus, for instance, rai pre

3.5 Sāmaveda saṁhitā (Pūrvārcika) (clip).

Ms. illustration V.

Observations: • udātta : ° (above the character) • anudātta: ட mark (above character) • svarita: left unmarked 3.6 Sāmaveda-saṁhitā manuscript: This is a Sāmaveda Uttarārcika text which indicates all accents (udātta, anudātta, svarita). From the style of the script and the quality of the material, it can be tentatively dated to 18th–(early) 19th century A.D.

Ms. illustration VI.

are: Preṅkha, Namana, Karṣaṇa, Vinata, Atyutkrama, Samprasāraṇa, Abhigīta. Cf. Burnell 1876: xliii, Tarlekar 1985: 1ff.

The Vedas in Indian Culture and History

102  Ù

Observations: • udātta : ° (above the character) • anudātta: ட mark (above character) • svarita: mark (above character) 4. Conclusions On the basis of the examples reviewed above and the observations based on them we can formulate the following conclusions regarding systems of notating Vedic accents in South Indian palmleaf manuscripts in Grantha script, the script in which the largest number of Vedic manuscripts are available. The points observed are not found in the available handbooks but are important for a proper reading of the accented texts. (1) By preference the udātta is marked in R̥ g -, Sāma-, and Yajurvedic mss.; sometimes also the anudātta and the svarita. This suggests that the South Indian Vedic schools had till relatively recent centuries an accent system still close to that described by Pāṇini where the udātta has the highest pitch, not the svarita as is suggested by accented Devanāgarī-manuscripts of the R̥ g vedic texts and which corresponds with the current practice of recitation. (2) Usually specific marks are used to show accents, no letters or numerals (except sometimes in Sāmaveda gāna mss.). (3) In case digital numerals are used to indicate accents they are (till ca. 19th century) not put above the syllable, but they are inserted on the same line between the characters. (4) Accent marks are generally placed along with the writing (with the same pen and ink), not afterwards in a different (red) ink as is usual in Devanāgarī manuscripts. (5) Later Sāmavedic mss. (from the 19th century onwards) containing the gāna part only show digital numerals and mātrās (in the form of letters) above the syllable. Finally, what Prof. V.N. Jha once pointed out with regard to Pada-texts applies in particular also to South Indian Pada-manuscripts: there is a “need for the study of the different Pada-texts from [a] grammatical point of view. This will enable us to reconstruct the grammars known to the respective Padakāras and their chronology and thereby … know the history of development of grammatical thought in India from pre-Pāṇinian times onwards.” (Jha 1987: 101)

Introduction Observations on Vedic Accents in Grantha Palmleaf Manuscripts

Ù  103

Manuscript illustrations Ms. illustration I: courtesy Vaidika Saṁśodhana Maṇḍala, Pune. Ms. illustration II: courtesy Adyar Library, Chennai. Ms. illustration III: courtesy Adyar Library, Chennai. Ms. illustation IVa and b: courtesy Library of Kern Institute, Leiden. Ms. illustration V: clip from Pūrvārcika ms., courtesy Adyar Library, Chennai. Ms. illustration VI: courtesy Library of Kern Institute, Leiden. References Burnell, Arthur Coke. 1876. Ārsheyabrāhmaṇa, being the Fourth Brāhmaṇa of the Sāma Veda. Mangalore: Stolz & Hirner, Basel Mission Press. ———. 1878. Elements of South-Indian palaeography from the 4th to the 17th century. Second enl. and improved edition. London: Karl J. Trübner. Falk, Harry. 1993. Schrift im alten Indien: Ein Forschungsbericht mit Anmerkungen. Tübingen: Gunter Narr Verlag. Jha, V.N. 1987. Studies in the Padapāṭhas and Vedic Philology. Delhi: Pratibha Prakashan. Macdonell, Arthur Anthony. 1910. Vedic Grammar. Strassburg: Karl J. Trübner. Repr. New Delhi: Munshiram Manoharlal Publ., 2000. ———. 1916. A Vedic Grammar for Students. London. Repr. New Delhi: D.K. Printworld, 1999. Staal, Frits. 1961. Nambudiri Veda Recitation. Leiden: Brill. ———. 1983. Agni: The Vedic Ritual of the Fire Altar. Vol. I. Berkeley, CA: Berkeley University Press. Tarlekar, Ganesh Hari. 1985. The Sāman Chants: A Review of Research. Bombay, Baroda: Indian Musicological Society. Witzel, Michael. 1974. “On some unknown systems of marking the Vedic accents.” Vishveshvaranand Indological Journal 12, 472-502.

104  Ù

The Vedas in Indian Culture and History

Introduction

Ù  105

II. Religion and Interpretation

106  Ù

The Vedas in Indian Culture and History

The Birth of the TwinIntroduction Horse Gods in India and in Greece

Ù  107

The Birth of the Twin Horse Gods in India and in Greece Henry John Walker Bates College

“Dioscurism” For over a hundred years now, the mythology of twins has been dominated by the strange theory of “Dioscurism,”1 which was created by the famous biblical scholar James Rendel Harris at the beginning of the twentieth century.2 According to this theory, all pre-modern peoples are terrified of twins; they believe that twins are born when two fathers sleep with one mother (“dual paternity”), and that one father and one twin are human, the other father and the other twin are divine. The Classical scholar Wiseman, who demonstrates that “Dioscurism” will not work in the case of Romulus and Remus, states rather optimistically that “Rendel Harris is a forgotten man.”3 Yet the work of Harris on twins is still cited in the 2005 edition of the Encyclopedia of Religions,4 and in the only twentieth-century monograph devoted to the Aśvins, Zeller laments that Indologists have not appreciated “the significance of Harris’ comprehensive material and his 1 Harris 1906 and 1913, Sternberg 1929: 164-169, Krappe 1930: 53-99, Ward 1968: 3-8, Zeller 1990: 16-17 and 18-22. 2 The Hollis Online Catalog at Harvard has no less than 148 entries under his name. 3 He qualifies this, however, by adding “But Harris’s ‘twin-lore’ lives on, in ever more exotic manifestations” (Wiseman 1995: 30). 4 Harris 1906 appears in the bibliography to Bianchi’s article on twins (Bianchi 2005: 9418), while in Chemery’s article on meteorogical beings, Harris 1906 and 1913 are praised as “two old but still fascinating studies” (Chemery 2005: 5996).

108  Ù

The Vedas in Indian Culture and History

conclusions.”5 Since Harris ultimately came to the extraordinary conclusion that there had once been an ancient and universal cult of twins, and that this cult was the origin of all subsequent religions,6 except for his own brand of Christianity, it might be advisable to question the validity of “Dioscurism” itself. Dual Paternity in Greece In Greek myth, the union between a god and a woman is not considered disturbing in any way. When a poet of Hesiod’s school wrote a poem honouring famous women (the Catalogue of Women), he started it with the following words: Sing now about the tribe of women, o sweet-voiced Muses of Olympus, o daughters of aegis-bearing Zeus, the women who were the best and most beautiful on earth, who untied their belts and, because of golden Aphrodite, slept with gods.7

The catalogue is an extremely long one, but these good and beautiful women are not being criticized, and no stigma attaches to their offspring. The phrase “good and beautiful” (kalos kai agathos) is regularly used to describe members of the aristocracy, so these “best and most beautiful” women are at the very pinnacle of Greek society, and their “splendid children”8 are the great heroes and heroines of Greece. Such unions between immortal gods and mortal women are remarkably common in myth, but by the time Hesiod was telling such tales, the happy days when earthly women could sleep with gods had gone by.9 Nevertheless, these ancient myths were very important to the Greeks, because all the best families of Greece originated in this manner. The Greek genealogist Hekataios, who was of course an expert on these matters, claimed that a god had started his family 16 generations ago.10 He was a rational, Ionian thinker and the forerunner of Herodotus, 5 Zeller 1990: 17. Zeller regrets that she was unable to consult Boanerges (Harris 1913), the work in which Harris reached this conclusion (Zeller 1990: 16 note 115). I doubt she would have valued his theory so highly if she had done so. 6 The cult of the heavenly twins is “the oldest religion of the world” (Harris 1906: 28 and 152; see Harris 1913: 30 and 266). In Boanerges, Harris declares that there had also been a world-wide belief in a thunder-god, who was worshipped as the father of the twin-gods. All peoples had believed in the “twin-cult and thunder-cult” (Harris 1913: 354); “we might almost say that on these two dreads hang nine-tenths of subsequent religion” (Harris 1913: 30). 7 Hesiodic Catalogue of Women, fr 1 MW, 1-5. 8 “Aglaa tekna,” Hesiodic Catalogue of Women, fr 1 MW, 14. 9 Hesiodic Catalogue of Women, fr 1 MW, 6-7. 10 Herodotus 2: 143.

The Birth of the TwinIntroduction Horse Gods in India and in Greece

Ù  109

but he readily accepted the idea that his family descended from a hero ancestor, and that this hero was the child of a god and a woman. There was, moreover, no doubt or ambiguity about the status of these heroic ancestors. They may have performed extraordinary deeds, but they were quite definitely mortal, and the cult of the heroes centered on their graves. Heroes, both twins and those born alone, often have two fathers (one divine and one human), or to be more precise, their paternity is ambiguous. The Greeks want to think of their hero-kings both as rightful kings and as legitimate heirs to their human fathers, but also as exceptional heroes, and therefore of divine origin. Theseus, for example, has a human father, Aegeus the King of Athens, and a divine father, Poseidon the god of the sea. Even some historical characters enjoy such double or ambiguous paternity: Alexander the Great had a human father, Philip II, the king of Macedonia, and a divine father, Zeus, the king of the gods. This claim, already advanced in his life-time, led a later Roman historian to complain that Alexander in asserting it “had betrayed his father, his country, and his humanity!”11 To claim a divine origin for the Macedonian royal family (from Zeus via Herakles) was, of course, perfectly acceptable,12 but gods had stopped sleeping with princesses long before the 4th century B.C., and no historical king of Macedonia had ever claimed that he had a divine father. For the Greeks, there was nothing wrong or sinister or ambiguous in being the child of a sexual relationship between a god and a woman, even when a human father was there to cast doubt on your claims. On the contrary, it made you a very fine human being, a hero or a heroine. Having two fathers certainly did not make you a twin, and conversely, being a twin did not require two fathers. The most important twins in the Greek pantheon are not the Dioskouroi, but Apollo and Artemis, and they are the children of Zeus and Leto.13 Since both their parents are divine, there is no ambiguity about the status of these twins: Apollo and Artemis are gods too. In the Iliad, there are three pairs of men who are explicity called twins (didumoi, or didumaone paide), and in two of these cases Homer specifies that there is one human Valerius Maximus, Memorable Deeds and Sayings, 9:5, ext.1. The genealogy of the royal family is recorded by Herodotus (8: 137-139). Some anti-Macedonian Greeks did deny this story, but they also denied that the Macedonians were Greeks! (Borza 1982: 10-11) 13 Homer tells us that they are brother and sister (Iliad 20:71) and that Leto is their mother (Iliad 24:606). The first explicit description of them as twins (didumoi, a relatively rare word in early and classical Greek and nowhere used of the Dioscuri!) appears in Pindar, Olympian Odes 3:35, which dates from 476 B.C. 11

12

110  Ù

The Vedas in Indian Culture and History

father alone.14 In the third case, they are the sons of the god Poseidon and a mortal woman Molione, though she is actually married to Aktorion.15 All three sets of twins are, of course, mortal.16 As a final example, the Spartans believed that their unusual dual monarchy was created when the twins Procles and Eurysthenes inherited the throne, but these twins had only one father, Aristodemus, and one mother, Argeia.17 Since both their parents were human (though ultimately descended from gods, like all aristocratic Greeks), these kings were human too. Dual paternity is, therefore, a common and unremarkable phenomenon in Greek myth, and it has nothing to do with twins or terrifying ambiguity. Divided Paternity in Greece There are, in fact, only two real cases of the phenomenon that “Dioscurism” upholds as a general rule: where the double paternity is not ambiguous, but is divided clearly between two twins.18 Both these cases are highly unusual, because the offspring of the divine father and the human mother is not a hero, but a god.19 The first is the famous case of Herakles and his twin brother, Iphikles. Zeus sleeps with their mother; on the same night, their human father also sleeps with her. Herakles the great hero is the son of Zeus, the insignificant Iphikles is the offspring of the hu14 Krethon and Orsilokhos (Iliad 5:548-549) are the twin sons of Diokles; Aisepos and Pedasos are the twin sons of Boukolion (Iliad 6:21-26). 15 They are called twins (didumoi) at Iliad 23:641; sons of Aktorion and Molione (Aktoriōne Molīone paide) at Iliad 11:750. They are the sons of Poseidon implicitly at Iliad 11:751, explicitly in the Hesiodic Catalogue of Women, fr 17B MW. The Moliones are depicted as conjoined twins in Greek art and in later Greek literature (Catalogue of Women, fr 18 MW). 16 Krethon and Orsilokhos are killed by Aeneas (Iliad 5:541-542); Aisepos and Pedasos are killed by Euryalos (Iliad 6:27-28); the Moliones are killed by Herakles (Ibycus fr. 285 PMG, Pindar, Olympian Odes 10: 26-34). 17 Herodotus 6:52. 18 The third case cited by Harris (1906: 5) is that of Amphion and Zethos, the twin heroes who built Thebes. Both of them are the sons of Zeus and Antiope at Odyssey 11:260-265. The fifth-century poet, Asios of Samos, wrote that Antiope bore “Amphion and noble (dios) Zethos” to Zeus and Epopeus (Asius fr 1 W). Harris mistranslates the adjective dios as “divine” in an attempt to demonstrate that one twin is be a god and the other a man. Though its etymological meaning is “heavenly,” dios is commonly applied to heroes and heroines, and is normally translated as “noble.” Both Zethos and Amphion were mortal and perished shortly after the deaths of their own children. 19 The only comparable case is that of Zeus and Semele, the mother of Dionysus. In this case too “a mortal woman gave birth to an immortal,” as Hesiod says (Theogony, 942).

The Birth of the TwinIntroduction Horse Gods in India and in Greece

Ù  111

man father. Herakles is unique among the Greek heroes, in that he becomes a god after his death. As Pindar puts it, he is a “hero god” (hērōs theos)20 —quite a contradiction, as far as Greek religion is concerned. Herakles is, therefore, the only Greek hero who lives on Olympus, and the only Greek god who possesses a ghost.21 Herakles is obviously too exceptional a character to be regarded as a typical Greek twin. Homer and Hesiod know that Herakles is the son of Zeus, that he is both a god and a man, but when they speak of his birth, they do not mention his brother.22 Iphikles does not appear as his twin brother until about 600 B.C.23 in the Hesiodic Shield of Herakles. Even in this work, however, Iphikles plays no role in his brother’s adventures; it is in fact Iolaos, the son of Iphikles, who acts as the charioteer and constant companion of Herakles.24 This is strange, because in Greek myth, twins are always very close to each other and work together as a team. So Herakles and Iphikles are very unusual twins: they have different fathers; they do not behave towards each other as twin brothers normally do; and even though both of them are mortal heroes, Herakles becomes a god after his death. This apotheosis is highly unusual, but it has nothing to do with his being a twin. The other case of two fathers divided between twins is that of the Dioskouroi. According to the well-known version of this myth, Zeus appears to their mother Leda in the form of a swan. She lays an egg from which the twins emerge. Castor is mortal, and he is the son of their human father Tyndareos; Pollux is a god, and he is the son of their divine father Zeus. This version, however, took a while to establish itself. There is nothing about swans or eggs until the 3rd century B.C., and even the story of the human Castor and the divine Pollux is just one of the stories the Greeks told about the Dioskouroi. From the earliest period of Greek literature, we find four versions: 1. In Homer’s Iliad, both of the twins are human and quite dead.25 2. In Homer’s Odyssey, both of them are the sons of the human father Tyndareos, and both of them are part-time gods. They Pindar, Nemean Odes 3:22. Homer is already aware of this paradox, and he tells us that the ghost of Herakles is down in the Underworld with all the other dead mortals, but Herakles himself, the god Herakles, is up with the other gods on Mount Olympus. 22 Homer, Iliad 14:323-324 and Iliad 19:97-105; Hesiod, Theogony 943-944. Hesiod does, however, mention Iolaus as the companion of Herakles at Theogony 317. 23 Hesiodic Shield of Herakles 53-56. 24 Hesiodic Shield of Herakles 70-121, especially line 74. 25 Homer, Iliad 3:236-244. 20 21

The Vedas in Indian Culture and History

112  Ù

spend one day as dead corpses in the earth, and the next day as gods in heaven!26 Such a complicated arrangement is unheard of anywhere else in Greek thought. 3. The early sixth-century Homeric Hymn to the Dioskouroi declares in its first line that both of them are the sons of Zeus, in the second line that both of them are the sons of Tyndareos, and throughout the hymn regards both of them as gods, as does the contemporary Hesiodic Catalogue of Women.27 4. Finally, the later sixth-century Kypria makes Castor human and Pollux divine, but (as in version 2) Zeus grants them parttime immortality.28 So we have four versions altogether: both twins are fully human, both twins are fully divine, both twins are half-divine, and finally, one twin is human, one twin is divine. Even in this last version, it is always stated that on the death of his brother Castor, the divine Pollux shares his divinity with his dead brother. So the two brothers end up being half-divine. The final status of each of the Dioskouroi is, therefore: dead, divine, or both. Obviously, the Dioskouroi were regarded as straddling the line between gods and men, and these versions were various attempts at defining their status. The Dioskouroi were torn between their divine identity as Indo-European horse-gods and their human identity as local Spartan heroes, who were simply the mortal sons of the mortal king Tyndareos. It would be ill-advised to choose one of these stories as the “real” version that defines the Greek concept of the Dioskouroi, even if that version eventually becomes the most popular one. It would be very wrong-headed to declare that this version reveals the Greek attitude to twins. Although they are exceptional in the Greek world, the Dioskouroi are very important for our purposes because they are, in effect, the Greek Aśvins. This is clear from the following points: 1. The Dioskouroi and the Aśvins are the only twin brothers among the major gods of Greece and India. 2. They are the horse-gods in both countries. Some scholars are disturbed by the discrepancy that the Dioskouroi ride horses, whereas the Aśvins drive them in a chariot. This arises simply because the image of the Aśvins was established in the bronze age, when driving chariots was fashionable, and that of the Di-

Homer, Odyssey 11:297-304. Hesiodic Catalogue of Women fr 24 MW. 28 Stasinus Cypria fr 9 W. 26 27

The Birth of the TwinIntroduction Horse Gods in India and in Greece

Ù  113

oskouroi in the archaic age, when riding became respectable.29 3. The Dioskouroi and the Aśvins are described as young men (kouroi in Greek, yuvānā in Sanskrit) and they are the sons or grandsons of the Indo-European sky-god, Dyēus (Dios kouroi in Greek, Divo napātā in Sanskrit). 4. They rescue people from danger, and in myth their status is lower than that of the other gods —the Dioskouroi initially lived on earth and became gods later, the Aśvins were not allowed to join the other gods at first. In cult, however, they are treated like the other gods —the Dioskouroi receive the white victims of Olympian gods, the Aśvins receive soma. 5. Finally, both the Dioskouroi and the Aśvins are the sons or brothers of the same goddess, known as Helen to the Greeks, and Saraṇyū to the Indians.30 This goddess lives among mortals, but later abandons them, leaving a substitute behind (the eidōlon of Helen, the savarṇā of Saraṇyū). She is also involved, as wife or daughter, in an animal metamorphosis: in India, Saraṇyū herself and her mortal husband take the form of horses and produce the Aśvins; in Greece, her divine parents take the form of birds and produce Helen. There can be little doubt that the Dioskouroi and the Aśvins are different avatars of the same Indo-European gods, but this does not mean that we can take one detail from one version of the story of the Dioskouroi and impose it mercilessly on their Indian cousins. Divided Paternity in India Indian myth was very familiar with heroes born from gods. Most royal families were ultimately descended from Vivasvant (in his role as sun-god) or Soma, and thereby belonged to the solar or lunar dynasties. It was equally happy with double paternity. The Pāṇḍavas are the legal sons of Pāṇḍu, but the biological sons of various gods. As in ancient Greece, this enables the sons to be the legal heirs of their father, and it likewise guarantees that they will be great heroes. Divided paternity is quite a different matter, and in the case of the Aśvins, there is only one place in the R̥ gveda where each of them has a different father, R̥ V 1.181.4. 29 Even though Homer composed his epics in the archaic age, he preserves the bronze age prejudice against riding. If Homer had depicted the Dioskouroi, he would, of course, have presented them as chariot-drivers, just like all his other heroes. No Homeric hero ever disgraces himself by riding a horse. 30 Greek hel and Sanskrit sar derive from Proto-Indo-European *sel, “to move quickly” (Skutsch 1987: 189-190).

The Vedas in Indian Culture and History

114  Ù

iheha jātā sam avāvaśītām arepasā tan(u)vā nāmabhiḥ svaiḥ jiṣṇur vām anyaḥ sumakhasya sūrir divo anyaḥ subhagaḥ putra ūhe Born here and there, the flawless ones harmonize in body and in their names. One of you is a victorious patron, (a son) of Sumakha, the other is called a fortunate son of Dyaus.

Even if we translate sumakhasya sūrir as “patron of the sacrifice,” the very fact that the “other” Aśvin is called the son of Dyaus, implies that the first Aśvin is the son of somebody else, so we clearly have two fathers here. In another hymn, we find a similar account of their separate births, but nothing about two different fathers (R̥ V 5.73.4cd): nānā jātāv arepasā sam asme bandhum eyathuḥ Born separately, you two flawless ones enter into friendship with us.

The phrase “born here and there” occurs one more time in the R̥ gveda, at R̥ V 5.47.5cd, where it applies to a pair of twin sisters. d(u)ve yad īm bibhr̥to mātur anye iheha jāte yam(i)yā sabandhū Two women other than his mother carry him, born here and there, twin sisters of one family.

The identity of the child is unclear, it is either Agni or Sūrya, and the two twin sisters who babysit him are either Dyāvāpthivī or Naktoṣasā.31 In either case, they do not have two fathers, and iheha jāte means that Earth and Sky were born at opposite levels of the universe, or that Night and Dawn were born at opposite ends of the sky. I think it more likely that we are dealing with the twin sisters Night and Dawn, because if we turn to another passage where twins are described as behaving “separately,” just as the Aśvins were born “separately,” we shall find the same twin sisters there (R̥ V 3.55.11ab). nānā cakrāte yam(i)yā vapūṃṣi tayor anyad rocate kr̥ṣṇam anyat

31

Geldner 1951: 2.53 (commentary on 5cd).

The Birth of the TwinIntroduction Horse Gods in India and in Greece

Ù  115

The twin sisters have produced separate beauties, one of them is shining, the other is dark.

Here there is no doubt about their identity, they are obviously Night and Dawn. The last two passages show that two sisters can be born in different places (iheha jāte), have very different appearances (nānā vapūṃṣi), and yet still be twins (yam(i)yā). By themselves, therefore, the phrases “born here and there” (iheha jātā, R̥ V 1.181.4a) or “born separately” (nānā jātāv, R̥ V 5.73.4c) do not disqualify the Aśvins from being twins; the real problem lies with the two fathers. Night and Dawn have only one father, Dyaus, and we are explicitly told that they are twins in these hymns. The Aśvins have two fathers in R̥ V 1.181; they are twins everywhere else, but are they twins in this hymn? Zeller believes that they are the twin sons of one mother by two fathers, a mortal father, Sumakha (whom she convincingly identifies as Vivasvant),32 and a divine father, Dyaus: “So the Aśvins have two fathers, and can, therefore, be distinguished as a human and a divine partner. This very fact, however, proves their identity as twins beyond a doubt.”33 Zeller produces no Vedic evidence for a human Aśvin, or for the notion that a woman who sleeps with a god and a man will give birth to twins. The only parallel she can find is the story of Nakula and Sahadeva from the later epic tradition.34 In their case, however, we do not have a human father and a divine one, but three fathers —their legal human father, Pāṇḍu, and their two divine biogical fathers, the Aśvins. Furthermore, each twin is the offspring and incarnation of his divine fathers alone. Finally, Nakula and Sahadeva are both equally mortal. Their story is an embarrassing refutation of “Dioscurism.” There is, however, one example of two fathers sleeping with one mother in the R̥ gveda itself, and this is the case of the testtube babies, Vasiṣṭha and Agastya (R̥ V 7.33). The gods Mitra and Varuṇa see the immensely attractive Apsaras Urvaśī and cannot restrain themselves from ejaculating into a pot. This pot acts as a surrogate womb,35 and from it the two babies are born. kumbhe retaḥ siṣicatuḥ samānam tato ha māna ud iyāya madhyāt tato jātam r̥ṣim āhur vasiṣṭham Zeller 1990: 33 and 35. Zeller 1990: 34. 34 Zeller 1990: 33 note 210 and 35. 35 Pots often act as surrogate wombs in Vedic thought (Jamison 1991: 231 and 236-239). 32

33

The Vedas in Indian Culture and History

116  Ù

They poured their combined semen into the pot; from the middle of it Māna came up, from it they say that the r̥ṣi Vasiṣṭha was born. (R̥V 7.33.13bd)

Vasiṣṭha is “the son of Mitra and Varuṇa”36 and he is “born from the Apsaras,”37 so we finally have a case of two sons being the biological offspring of two fathers and one mother. They are also born separately, as were the Aśvins and the twins Night and Dawn. Māna Agastya is born from the middle of the pot, and Vasiṣṭha from an undisclosed location in the pot.38 This might seem like a triumphant vindication of “Dioscurism,” but its proponents did not choose to exploit this evidence, because (unfortunately for their theory) all of the parents are divine, both of the sons are r̥ṣis, and each son is the child of both fathers!39 The birth of these two r̥ṣis is a highly unusual one, but this story, the only case of “dual paternity” in the R̥ gveda, makes it clear that in India, as in ancient Greece, there is not much trace of “Dioscurism.” Instead of forcing Indian myths to conform to a rigid formula of dubious validity, we should look to the Vedic tradition itself if we want to understand R̥ V 1.181. A half-mantra (ardharc) quoted by Yāska provides the simple answer:40 vāsātyo anya ucyate uṣaḥ putras tavānyaḥ One is called the son of Vasāti, the other your son, o Uṣas.

If we add this information to what we know from R̥ V 1.181, we discover that one Aśvin is the son of an unidentified father41 and utāsi maitrāvaruṇo vasiṣṭha (R̥V 7.33.11a). apsarasaḥ pari jajñe vasiṣṭhaḥ (R̥V 7.33.12d). R̥V 7.33.11b specifies that Vasiṣṭha was “born from the mind of Urvaśī” (urvaśyā ... manaso ’dhi jātaḥ). This could also be translated as “born from (Mitra and Varuṇa’s) desire for Urvaśī,” but then Vasiṣṭha would be the child of his fathers alone and Urvaśī would play no role is his birth. The statement that he was born from the Apsaras excludes such an interpretation. See Geldner 1951: 2.213 (commentary on 11b) and Dandekar 1974: 228. 38 Later interpreters will specify the exact location of their births: Vasiṣṭha is born on the ground, Agastya in the jar. There is also a third son, Matsya, who is born in the water (Br̥haddevatā 5.151-152). 39 Agastya and Vasiṣṭha are both called Maitrāvaruṇi (son of Mitra-Varuṇa). 40 Yāska, Nirukta 12, 2. Zeller says that “this explanation goes back to Yāska,” but he is citing a Vedic verse (r̥c), not inventing a new explanation. See Zeller 1990: 32 n. 201, and (supporting Yāska) Geldner 1951: 1.261 (commentary on 4 and note 1). 41 If sumakha is the name of a man at R̥V 1.181.4c, he is probably Vivasvant, the first sacrificer, as Zeller suggests (Zeller 1990: 33). 36 37

The Birth of the TwinIntroduction Horse Gods in India and in Greece

Ù  117

Vasāti (Night), the other Aśvin is the son of Dyaus (Sky) and Uṣas (Dawn). This is an alternative to the famous story of their birth in R̥ V 10.17, and the seer who composed it was obviously speculating about the role of the Aśvins as morning gods. It is an unusual story, for it denies that the Aśvins are twins, but it is readily understandable and does not oblige us to introduce an extraneous theory about the birth of twins in order to explain it. Where twins come from Greek and Indian myths indicate that women from India and Greece were perfectly capable of sleeping with one man and giving birth to twins, and that they also had a remarkable gift for sleeping with several partners (gods and men) and yet producing only one child at a time. The notion that twins are the result of a man and a god sleeping with one mortal woman finds little support from the ancient Greeks and Indians. Let us see, instead, what they themselves actually said about twins. The first Greek author to discuss the birth of twins is Aristotle. He believes that a woman supplies the matter for a future child, and this matter is made of menstrual blood; the man supplies the form through his semen.42 Aristotle uses two images to explain this. In the first one, the father is a carpenter, the mother supplies both the wood and the workshop in which the baby is manufactured, the child is the piece of furniture they produce.43 The second more accurately reflects his views: the woman supplies milk, the father supplies rennet, and the child coagulates like cheese from this altered milk.44 The role of women in providing the raw matter for embryos is important, because Aristotle believes that twins are born when there is too much matter for one embryo alone.45 When he speaks of matter, he is referring to the mother’s contribution.46 In humans, as in all large animals that normally produce one offspring, the mother will produce enough matter for one embryo, or in exceptional cases, she will produce enough for two. Admittedly, the father must provide 42 Aristotle, Generation of Animals 729 A 10-12. He does not believe that semen is always necessary for coagulating the female’s matter; in some animals, the male may coagulate her matter merely by his bodily heat (Aristotle, Generation of Animals 729 B 22-27). 43 Aristotle, Generation of Animals 729 B 16, 730 B 1-8. 44 Aristotle, Generation of Animals 729 A 13-14, 771 B 23-24. 45 Aristotle, Generation of Animals 772 A 33-36. Compare 772 A 2-8 (animals) and 772 B 13-15 (deformities). 46 He defines a male as the efficient cause, and a female as the material cause of future offspring (Aristotle, Generation of Animals 716 A 5-7).

The Vedas in Indian Culture and History

118  Ù

enough semen to coagulate this matter,47 but it is the mother who supplies the matter in the first place. So the number of men she sleeps with is quite irrelevant to the production of twins. To borrow an analogy used by Aristotle in a related context,48 it is like trying to heat water; no matter how much heat you apply, the water will never go beyond its boiling point.49 Extra fathers would be wasting their time. We do not know how earlier Greek thinkers explained the birth of twins. Aristotle mentions a theory that both men and women produced semen, and that the new embryo resulted from the combination of the two.50 Such a view would hardly require two separate fathers to produce twins. From Aristotle again and also from independent fifth-century evidence, we know that other Greeks thought that a woman played almost no part in reproduction. She merely supplied a womb for the embryo, but the offspring was produced from the man’s semen alone.51 This theory might require double fertilization for twins, but once again would hardly require two fathers. Two biological fathers are never required for twins per se; they only crop up when one twin is human and the other is divine, a very rare occurrence even in Greek myth.52 When, finally, we return to India, we shall find that the same thing is true there. Indian thinkers associated the birth of twins with the problem of the mule. They realized that mules could not reproduce, and that the ability to produce mules was given to horses and donkeys. The reproductive power of the mule has, therefore, been distributed among other animals. The male donkey is dviretas (“with double semen”), meaning that he can produce either a donkey or a mule.53 The Taittirīya Saṃhitā explains it as follows: Aristotle, Generation of Animals 729 A 17-21, 772 A 17-22. He is explaining why the size of an embryo is limited rather than why the number of embryos is limited. 49 Aristotle, Generation of Animals 772 A 13-16. 50 Aristotle, Generation of Animals 727 B 34 - 729 A 33. 51 Aristotle merely mentions this view (Generation of Animals 726 A 33-34); he does not discuss it as a serious possibility. It first appears at Aeschylus, Eumenides 658-661; Diodorus Siculus claims it is an Egyptian theory (Library of History 1:80). 52 Divided paternity occurs only in the case of the Dioskouroi and Herakles. When the fifth-century poet Asius says the same about the Theban twins, Amphion and Zethus (Asius fr 1 W; cited at Harris 1906: 5), he is obviously changing their story (they were both the sons of one father, Zeus, in Homer; see Odyssey 11:260-265) to match that of the Dioskouroi. By the time Asius wrote his epic, the Dioskouroi were already thought of as the sons of two different fathers. 53 Taittirīya Saṃhitā 5.1.5.5; Taittirīya Saṃhitā 7.1.1.1; Aitareya Brāhmaṇa 4.9.3; Śatapatha Brāhmaṇa 6.3.1.23. 47

48

The Birth of the TwinIntroduction Horse Gods in India and in Greece

Ù  119

He (Prajāpati) followed the mule, took its seed, and placed it in the ass. Therefore the ass has double seed.

The Taittirīya Saṃhitā went further, however, and declared that mares were also dviretas, since they could produce either a horse or a mule: He placed it in the mare. Therefore the mare has double seed.

This makes it clear that the word dviretas, in spite of its obvious etymology, is not restricted to males alone and means something like “doubly reproductive.” Finally, the Taittirīya Saṃhitā declares that the reproductive capacity (retas, semen) of the mule is placed in creatures (or humans, prajāsu) and this is why they can produce twins: He placed it in creatures/humans. Therefore twins are born.

Instead of having the ability to produce one offspring from either of two types, creatures/humans can produce two offspring of one type. The capacity to produce the additional offspring (a mule or a second child) lies in the parents themselves, not in adding to their number. Twins are born because of “double semen” (dviretas), not because of double paternity. In the case of India, however, we do not have to rely on the intellectualized explanations of the Taittirīya Saṃhitā. We will find the very same outlook in the R̥ gveda itself. In R̥ V 10.17 we are told that the goddess Saraṇyū sleeps with one man, Vivasvant, and gives birth to twins, Yama and Yamī. Later in the story, she once again gives birth to twins, the Aśvins. This is not just a coincidence, because during Saraṇyū’s absence, Vivasvant had slept with her substitute, Savarṇā. The outcome was quite different this time; Savarṇā could only give birth to one child, Manu.54 In each case, we have one man sleeping with one woman, but what is most striking about this story is that it is the woman, Saraṇyū, who has the genetic capacity to bear twins. She does not need any help from a second man, and Indologists do not need any help 54 His birth is not described in R̥V 10.17, but it is clear from other parts of the R̥g veda that he is the son of Vivasvant (manau vivasvati, R̥V 8.52.1) and of the substitute, Savarṇā. He is referred to as Manu and Sāvarṇya at R̥V 10.62.8b and 9c, and as Manu and Sāvarṇi at R̥V 10.62.11a and 11c. See Bloomfield 1893: 178-180; Geldner 1951: 2.374 (commentary on 1a); Geldner 1951: 3.232 (introduction to hymn 10.62). It does not matter greatly whether the poet was thinking of the original Manu, or a particular descendant of Savarṇā; the important point is that the original Manu was her son. See Bloomfield 1893: 180.

120  Ù

The Vedas in Indian Culture and History

from “Dioscurism.” They can safely ignore “Harris’s comprehensive material and his conclusions.” References Bianchi, Ugo. 2005. “Twins.” In: Jones 2005, Vol. 14, 9411-9419. Bloomfield, Maurice. 1893. “Contributions to the Interpretation of the Veda. III. The Marriage of Saraṇyū, Tvaṣṭar’s Daughter.” Journal of the American Oriental Society 15, 172-188. Borza, Eugene N. 1982. “Athenians, Macedonians, and the Origins of the Macedonian Royal House.” In: Hesperia Supplements 19 (Studies in Attic Epigraphy, History and Topography. Presented to Eugene Vanderpool), 7-13. Chemery, Peter C. 2005. “Meteorological Beings.” In: Jones 2005, Vol. 9, 5992-5996. Dandekar, Ramchandra N. 1974. “The two births of Vasiṣṭha.” In: Manfred Mayrhofer and Wolfgang Meid (eds.), Antiquitates Indogermanicae. Studien zur indogermanishcen Altertumskunde und zur Sprach- und Kulturgeschichte der indogermanischen Völker. Gedenkschrift für Hermann Güntert. Innsbruck: Institut für Sprachwissenschaft der Universität Innsbruck, 223-232. Geldner, Karl F. 1951. Der Rig-Veda, aus dem Sanskrit ins Deutsche übers. und mit einem laufenden Kommentar versehen von Karl Friedrich Geldner. Cambridge, MA: Harvard University Press. Harris, James Rendel. 1906. The Cult of the Heavenly Twins. Cambridge: University Press. ———. 1913. Boanerges. Cambridge: University Press. Jamison, Stephanie. 1991. The Ravenous Hyenas and the Wounded Sun. Myth and Ritual in Ancient India. Ithaca, NY: Cornell University Press. Jones, Lindsay. (ed.). 2005. The Encyclopedia of Religion. Detroit: Macmillan Reference USA. Krappe, Alexandre Haggerty. 1930. Mythologie universelle. Paris: Payot. Skutsch, Otto. 1987. “Helen, Her Name and Nature.” Journal of Hellenic Studies 107, 188-193. Sternberg, Leo. 1929. “Der antike Zwillingskult im Lichte der Ethnologie.” Zeitschrift für Ethnologie 61, 152-200. Ward, Donald. 1968. The Divine Twins. Berkeley, CA: University of California Press. Wiseman, Timothy Peter. 1995. Remus. A Roman Myth. Cambridge: Cambridge University Press. Zeller, Gabriele. 1990. Die vedischen Zwillingsgötter. Untersuchungen zur Genese ihres Kultes. Wiesbaden: Otto Harrassowitz.

Introduction Lakṣmī: Originally A Marked Animal

Ù  121

Lakṣmī: Originally A Marked Animal M adhavi Kolhatkar Deccan College, Pune

Introduction and the problem Lakṣmī, the goddess associated with wealth, beauty, and auspiciousness is very important in the pantheon of Indian deities, even though it is true that there are very few temples separately dedicated to her and that she is generally and mostly found in Viṣṇu temples accompanying him. Her popularity is seen increasing day by day, especially in the modern industrialized society. The Dhanalakṣmī Vrata, “the vow of worshipping the goddess Dhanalakṣmī,” becoming more and more popular among the middle class people, both higher and lower, is an obvious example of it. However, that does not mean that she is a modern deity. She is in fact one of the oldest Indian deities. The references to her are found in the Vedic literature, right from the R̥ V . Regarding the meaning and etymology of the word Lakṣmī it is always said that Lakṣmī is the one who has lakṣma “mark,” i.e. she is the goddess who possesses marks, the marks of beauty or luck or fortune. Herein an attempt is made to find if they really are the marks of beauty, if so then what they are, and whether they are mentioned in the Vedic literature; and if they are not recorded there, then what is the lakṣma after which the Great Goddess has acquired her epithet from the Vedic times. The meaning of lakṣman and lakṣmī in the dictionaries Before going to the actual texts it would be better to see what meaning the dictionaries assign to the words lakṣman and lakṣmī. While commenting upon the word lakṣmī occurring at

122  Ù

The Vedas in Indian Culture and History

R̥ V . 10.71.2, the Nirukta derives the word as lakṣmīr lābhād vā | lakṣanād vā [lapsyanād vā] | lāñchanād vā | laṣater vā syāt prepsākarmaṇaḥ | lagyater vā syād āśleṣakarmaṇaḥ | lajjater vā syād aślāghākarmaṇaḥ | “Lakṣmī (mark) is (so called) from obtaining, or from indicating, [or from a desire to obtain], or from marking, or it may be derived from (the root) laṣ, meaning to desire, or from lag, meaning to cling, or from lajj, meaning not to praise,” (Transl. Sarup).

Böhtlingk and Roth have given the meanings of lakṣman: “(von lag wie lakṣa or lakṣmī) 1)Mal, Merkmal, Marke, Zeichen.” The meanings of lakṣmī are given: “1) Merkmal, Zeichen ... 2) mit oder ohne pāpī ein schlimmen Zeichen, bevorstehendes Unglück, Unglück ... 3) ein gutes Zeichen ... 4) Schönheit, Anmuth, Pracht,” etc. In the entry lakṣma, Monier Williams (1899 s.v.) has directed to look for the word devalakṣma and has stated the meaning: “characteristic, the brahmanical cord.” In the entry lakṣmaṇa are the meanings “having marks or signs or characteristics, endowed with auspicious signs or marks, lucky, fortunate.” Lakṣman according to him is “n. mark, sign, token, characteristic, a good or lucky mark, excellence, a bad mark, stain, blemish” and also “definition.” The word Lakṣmī (f.), “nom. īs, rarely ī; also ifc. as m.f., but n. i cf. lakṣmīka, a mark, sign, token, (with or without pāpī) a bad sign, impending misfortune, (but in the older language more usually with puṇya) a good sign, good fortune, prosperity, success, happiness, wealth, riches, beauty, loveliness, grace, charm, splendour, luster; N. of the goddess of fortune and beauty (frequently in the later mythology identified with Śrī and regarded as the wife of Viṣṇu or Nārāyaṇa; accord. to R[āmāyaṇa] 1.45.40-43, she sprang with other precious things from the foam of the ocean when churned by the gods and demons for the recovery of the Amr̥ta ... she is also variously regarded as a wife of Sūrya … of Prajāpati … of Dharma and mother of Kāma, sister of mother of Dhātr̥ and Vidhātr̥, as wife of Dattātreya; ... the Good Genius or Fortune of a king personified ..., royal power, dominion majesty.” Apte’s dictionary states regarding lakṣman: “1. a mark, sign, token, characteristic; 2. A speck; 3. definition; 4. the chief, principal; 5. a good or bad mark; 6. a pearl.” Then under lakṣmī f. [lakṣ-i muṭ ca Uṇ. 3.158, 160] “1. fortune, prosperity, wealth, 2. Good fortune, good luck; 3.success, accomplishment, 4. Beauty, loveliness, grace, charm, splendour, lustere; 5. The goddess of fortune, prosperity and beauty, regarded as the wife of Viṣṇu, (She is said to have sprung from the ocean along with the other precious things or ‘jewels’ when it was churned for nectar by the gods and demons) etc.” Regarding the etymology of the word lakṣma, Mayrhofer (195680 s.v.) has suggested “Zeichen, gutes Zeichen,” i.e. “mark, good

Introduction Lakṣmī: Originally A Marked Animal

Ù  123

mark,” R̥ V 10.71.2; “Glück, glückliches Zeichen,” i.e. “luck, lucky mark” (AV ff.), personified as “die Glücksgöttin, Lakṣmī,” i.e. “the goddess of luck, Lakṣmī.” The opinions of previous scholars Scholars have naturally been interested in the origin, nature, development and also in the etymology of the name of Lakṣmī. Various scholars have presented various opinions in this matter. E.g. Oldenberg (1967: 46) comments, “Die Verwandtshaft von lakṣmī mit lakṣman ‘Zeichen, Merkmal’ liegt auf der Hand. … lakṣman ist das äussere Merkmal glücklicher oder unglücklicher Disposition : lakṣmī ist diese Disposition selbst, die durch ein lakṣman angezeigt wird oder werden kann.” However in the footnote to this, he expresses his doubt: “Dass lakṣmī geradezu im Sinne von lakṣman stehen kann, bezweifle ich. Das Petersburg Wörterbuch nimmt das für den ältesten Beleg des Worts an, den einzigen r̥gvedischen (aus jünger Gegend des R̥V) bhadraiṣā lakṣmīr nihitādhi vāci 10.71.2 zu welcher Auffassung ich keinen Anlass finde. Man konnte etwa noch an AV 1.18.1, TS 2.1.5.2, TB 2.1.2.2 denken; dass aus diesen Stellen ein wirkliches Zusammenfliessen der beiden einander immerhin nachstehenden Vorstellungen folge, glaube ich doch nicht.” Further he continues, “Diese Doppelseitigkeit der lakṣmī-vorstellung die stehend in den Beiworten bhadra, śiva, besonders puṇya, und anderseits pāpa zum Ausdruck kommt, tritt dann weiterhin zurück, und es bleibt ausschliesslich oder fast ausschliesslich der günstige Sinn des Wortes. Damit nun ist dieses der Vorstellung der Śrī, wie wir sie sich entwickeln gesehen haben, ganz nah gerückt. ... So werden sie beide ārdrā beide padmamālinī genannt. Die entgegensetzen feindlichen Mächte heissen nicht mehr ‘böse lakṣmī’ sondern alakṣmī. … Die lakṣmī ist also jetzt etwas an sich Gutes … Göttin des Glückes und der Schönheit.” The summary in English would be, “It is obvious that there is connection between the word lakṣma, ‘mark’ and lakṣmī.” Then quoting the passage from ŚB 8.4.4.11; 5.4.3 etc. he concludes, “lakṣman is an outer mark either of a lucky or fortunate or of an unfortunate disposition; and Lakṣmī herself is such disposition, which is or which can be shown through a lakṣman.” In the footnote to this he adds, “I doubt whether Lakṣmī can directly mean lakṣman. The double-sided representation of Lakṣmī, i.e. with the epithets bhadra, śiva, puṇya etc. on one hand and pāpa on the other, gradually recedes and only the good or auspicious meaning of the word remains.” Further he states (47), “Both Śrī and Lakṣmī are ārdrā and padmamālinī. Later the other ‘hostile’ or ‘malignant’ part is no more called ‘böse,’ i.e., ‘angry’ Lakṣmī, but as Alakṣmī (cf.

124  Ù

The Vedas in Indian Culture and History

also GobhGS 4.6.3; ĀpMP 1.1.5). Thus Lakṣmī basically is the goodness incarnate … and also the Goddess of luck or fortune and beauty.” However, as is clear, he has not explained what the marks are, whether they be of luck or of beauty. Gonda (1954: 215), hesitating to subscribe to Oldenberg’s view expresses his opinion that “Lakṣmī is an object or a being, the very existence or presence of which means something (auspicious - puṇya l. or evil : pāpī l.); a lakṣman- is, more vaguely, a token or mark, a fact connected with the external form of beings or objects as perceived by the senses, which may induce man to infer that there is something auspicious, favourable etc. It seems worthy to notice that the feminine polysyllabic -ī- stems in Vedic denote, as a rule, beings or objects; part of them refer to powersubstances to lead an independent existence: tandrī- ‘weariness’; nāndī- ‘refreshment, bliss.’” He further states that AV 6.141 “has the word in connection with marks made in the ears of cows: the text, which is included among the puṣṭika-mantras is to accompany the ceremony of earmarking cattle for the sake of prosperity. In st.3 the Aśvins are requested to make the lakṣma in order to thousand-fold thriving (poṣa-) …. For Varahamihira and other post-Vedic authors in general lakṣmī denoted ‘luck, fortune.’… This sense is ‘developing’ in the Vedic texts: TS 2.1.5.2 … AiB 2.40.8 … ‘a prospering destiny, a condition of prosperity.’” Thus having dealt with the topic in details, Gonda (217) concludes, “So Lakṣmī may ‘originally’ have been the divinity representing the signs, evidence, or prognostications (of luck and prosperity).” Thieme (1951: 211 n.1) has derived lakṣman from the root rakṣ. According to Moti Chandra (1949: 21) “lakṣman is the outward visible sign of a happy or unhappy disposition; Lakṣmī is the disposition itself which is proclaimed or even can be proclaimed by a lakṣman. According to AV (8.115) the human being is born with hundred and one lakṣman … The noble conception of Lakṣmī finds expression in such terms as bhadra, śiva, and puṇya, while its evil epithet pāpa recedes into background, and finally only the auspicious meaning of the term holds the field.” Hartmann1 and Meyer share views concerning lakṣmī totally opposite to those above. According to Hartmann (1933: 13ff.), lakṣmī originally expressed the sense of alakṣmī alone, and represented a sign, which was evil and unholy without any trace of auspiciousness in it. Lakṣmī first was the goddess of poverty. Similarly Meyer (1937: 2,90) also opines that originally lakṣmī is nothing auspicious but is a disease of the crops that stuck to the plants 1 I was not able to see the original work, hence have given her opinion from Gonda’s and Dhal’s work.

Introduction Lakṣmī: Originally A Marked Animal

Ù  125

like a scar or a mark. He derives the word from the root lag. This lakṣmī and the other lakṣmī “der romischen Robigo, der Genie des Getreiderostes,” —“the spirit of dry corn” were originally related to each other. Afterwards, the attributes as disease were transferred to the new divinity called Alakṣmī and the “auspicious corn- spirit” merged with Śrī and developed into Lakṣmī, the goddess of wealth and fortune. Dhal (1978: 42) in his lengthy dissertation Goddess Lakṣmī: Origin and Development concluded: Thus the development of the concept of lakṣmī in the sense of prosperity, well-being can be connected with lakṣma meaning — sign or symbol. Kalidasa’s use -lakṣma lakṣmīm tanoti is not accidental. … That lakṣma which was found to bring good luck, prosperity could easily be turned as lakṣmī. … TS recommends offering of a hornless goat to Brahmaṇaspati; and this hornless goat is designated as Lakṣmī, because the particular lakṣma of having no horns leads to prosperity and well-beingkṣurapavir vā eṣā lakṣmī yat tūparaḥ samr̥ddhyai [TS 2.1.5.7] Those particular signs or symbols which are indicative of prosperity, well-being, are called lakṣma and gradually the term lakṣmī itself came to denote prosperity, well-being, opulence etc. The humped back bull which is taken as a sign of prosperity, and which is to bring prosperity to the person, became identified with lakṣmī. Similarly the hornless goat which is a sign of prosperity, when sacrificed to Brahmaṇaspati destroys one’s enemies, and brings about the prosperity of the sacrificer; the object which served as a symbol or sign (lakṣmī) of prosperity subsequently was identified with prosperity.

In AV 7.115 which is used in a propitiatory rite against Nirr̥ti, the goddess of death and destruction occur the words puṇyā lakṣmī and pāpī lakṣmī. Referring to that Dhal concludes, “In the above Suktas we mark the development of the concept in two ways, viz. the pāpī lakṣmī and puṇyā, śivā, bhadrā lakṣmī. The signs or symbols which bring fortune, good luck are designated as puṇyā lakṣmī, but the signs which bring bad luck, misfortune and misery are designated as pāpī lakṣmī. … This pāpī lakṣmī ultimately became Alakṣmī of the later age, and carried all the evil traits which bring misfortune, misery, bad luck in one’s life.” Thus it can be seen that all the scholars agree on one point that lakṣma and lakṣmī are connected with each other, though the connection and also the nature of the marks do not become clear. Most of them have taken lakṣma itself as lakṣmī, and to denote one and the same thing. As is mentioned previously, the word Lakṣmī occurs already in R̥ V 10.71.2 bhadraiṣām lakṣmīr nihitādhi vāci, which Palsule (1969: 29), translates as “Lakṣmī (is so called) from (her) lakṣaṇa i.e. bhāsana (shining) (The word

The Vedas in Indian Culture and History

126  Ù

adhi means she) becomes capable.” Also Lakṣmī is referred to in the Śrīsūkta. However, both these occurrences are comparatively late. At AiB 2.40 it is said puṇyā eva lakṣmīḥ puṇyām eva tat lakṣmīm sambhāvayati puṇyām lakṣmīm saṃskurute. In ŚB 8.4.4.8, there is a general statement that ye te paśavaḥ puṇyā tā lakṣmayaḥ. And at TS 2.1.5.7 it is said, kṣurapavir vā eṣā lakṣmīḥ yat tūparaḥ samr̥ddhyai. As noted above, Dhal has stated that “the particular lakṣma of having no horns leads to prosperity and well-being,” however, without stating any convincing reason. There are some important mythological passages in the Vedic texts, viz. MS, TS, etc., which are important regarding the etymology of the word Lakṣmī. Let us look into the myths occurring there. Lakṣmī in the Vedic texts The myth in the TS 2.1.5.1, 2 reads: indro valasya bilam apaurṇot | sa ya uttamaḥ paśur āsīt tam pr̥ṣṭham prati saṃgr̥hyodakkhidat | taṃ sahasram paśavo ’nūdāyan | sa unnato ’bhavad | yaḥ paśukāmaḥ syāt sa etam aindram unnatam ālabheta | indram eva svena bhāgadheyenopadhāvati sa evāsmai paśūn prayacchati paśumān eva bhavati unnataḥ bhavati | sāhasrī vā eṣā lakṣmī yad unnato | lakṣmiyaiva paśūn avarundhe | yadā sahasram paśūn prāpnuyād atha vaiṣṇavaṃ vāmanam ālabheta | etasmin vai tat sahasram adhyatiṣṭhat tasmād eṣa vāmanaḥ samīṣitaḥ paśubhya eva prajātebhyaḥ pratiṣṭhāṃ dadhāti “Indra opened the hole of Vala.2 He held the back of the one, which was the topmost animal, and pulled it out. Following it came out thousand animals. It became tall. He (i.e. the sacrificer) should offer this tall one to Indra. By that he approaches Indra with his own share. Indra bestows cattle upon him. He becomes the one possessing the cattle. It is tall. This tall one is the thousand-fold Lakṣmī. With this Lakṣmī does he obtain the animals or cattle. When one obtains thousand animals he should offer an animal to Viṣṇu. Upon it did that thousand rested. Hence this dwarf one stretched. It gives support to animals when born.”

Similar is the myth at MS 2.5.3. Thus it reads: indro vai valam apaurṇot | tataḥ sahasram udait | tasya sahasrasyāgrataḥ kubhra udait | tasmād etaṃ sāhasrī lakṣmīr ity āhuḥ yaś ca veda yaś ca na “Indra uncovered (the hole or the hiding place of) Vala. Thousand came up from it. A humped one was the first one to come up among those thousand. Hence that was called thousand-fold Lakṣmī by one who knows and one who does not. ” 2

Keith gives Vr̥t ra outside and Vala into bracket.

Introduction Lakṣmī: Originally A Marked Animal

Ù  127

The same story occurs in KS also, but there is no reference to Lakṣmī in it. In MS the word lakṣmī occurs once again at 4.2.1 as goṣṭho vai nāma eṣā lakṣmīḥ “The cowpen really is this Lakṣmī.” From the above passages it is clear that lakṣmī here is a collective noun. The animal which was followed by thousand animals, that male animal is called here as “thousand-fold lakṣmī.” Then what is this lakṣmī? Also, if the word is connected with lakṣman, then what can it be? One point, however is certain that almost all the scholars already quoted have stated that lakṣma itself is lakṣmī and that lakṣmī is the one having auspicious (and sometimes also inauspicious) marks. In the above passages, however there is no reference at all to any kind of auspicious or inauspicious signs or marks. Hence the question remains unsolved as to what these marks are. Some important passages in MS There are some important passages in this regard in the MS, which previous scholars dealing with lakṣman and lakṣmī have not taken into consideration. It might throw some light on our present problem. The fourth kāṇḍa in MS is a khila kāṇḍa, “an appendix.” The second chapter in it is called Gonāmika “regarding the names of the cow,” and is unique in the whole of Vedic literature. Gandhe (1976-77: 19) has concluded regarding it, “The analysis of the data ... shows us that Gonāmika was essentially a sort of manual of cattle-keeping … Its conclusion [inclusion?] in the Maitrāyaṇī-Saṃhitā represents the period of transition from cattle-keeping to a mixed economy based on agriculture and cattle-keeping.” Regarding agriculture at 4.2.2 it is said, yadā susasyam bhavati atha pratitiṣṭhati yadā na sasyam bhavati atha na pratitiṣṭhati “When there is good crop, there is stability and when not, there is instability.” It also speaks about some aspects of the life of cowherds and about some of their rites, beliefs, and practices. It is in one of such passages that the word lakṣma occurs in the context of one such practice, which is to mark the cattle in one’s own herd with some specific sign. According to MS 4.2.9, “If the cattle are not so marked then they would not come back home.” It was a sign or mark of identification as well as of ownership. Thus it reads, yāvatīnām idaṃ karomi bhūyasīnām evottarāṃ samāṃ kriyāsam iti gavāṃ lakṣma kuryāt bhūyasīnām eva uttarāṃ samāṃ karoti “‘May I do this to more next year than I am doing this year,’ saying so he marks the cows (or/and bulls). Thus, next year he does this to more cattle.” The word lakṣma occurs six times in this passage, both separately as well as in compounds. It tells us that the marks were variously made. Each family had a different mark. The cows of Vasiṣṭha were sthūṇākarṇī, those

128  Ù

The Vedas in Indian Culture and History

of Agastya were viṣṭyakarṇī, whereas those of Jamadagni were karkarikarṇī. It is difficult to understand how exactly the marks were. It is said that no hole should be made on cow’s ear, lest they go under the control of Nirr̥ti. It is necessary to take into consideration in this context what Burrow has suggested regarding the etymology of the word lakṣma, which is quoted by Mayrhofer under the entry lakṣmī. It is very significant and noteworthy. However scholars have not taken cognizance of it. Burrow has derived it from the root dah and has commented, “The change of d to l was a particular characteristic of Eastern Iranian, that is to say, of those Iranians who were most immediately in contact with the Indo-Aryans. In the case of these words there are two possibilities: either the Sanskrit words are due to early borrowing of Eastern Iranian, or possibly the isogloss l > d may have extended at some time into the IndoAryan area, in which case the words are dialect words which have been adopted into the standard language (cf. Lat. Lacrima etc.). Ultimately the words are derived from dah “to burn,” the original meaning being a mark branded on the horns of cattle, sheep etc. Avestan daxsa —Skt. lakṣa— still maintains the etymological sense of “branding.” The relation of Avestan daxsta- “sign” to the Skt. participle lakṣita- is exactly the same as that between Skt. participle bhakta “food” (kṣatríya, rudrá>rudríya, vīrá>vīryà ( kh, and the replacement of Skt. nāsti by ‘modern’ nahī. Such deviations are not marked in the sequel.

344  Ù

The Vedas in Indian Culture and History

This was copied by one Rudra Ojhā,52 commissioned by Śiva Pañcolī for the study by his sons and grandsons. He was a son of Nārāyaṇa P., son of Vāsaṇa P., son of Lahūā P., son of Vasiṣṭha P., son of Viśrupa P., and was known as (with the by-name) Ābhyantara-Nāgara Pañcolī living in Ahanala-(+Anahila)-purapattana (Patan); copied on Sunday, the second day of the bright half of the month of Āṣāḍha, Vikrama Saṃvat 1651, that is in the summer of 1594 CE. —The names in the colophon agree with those of AV-Par Ms. R of 1584 CE and with those in the earlier part of ŚPP’s B of 1673 CE, allowing to establish a lineage from the mid 15th down to the late 17th century. And, colophon after book 20: saṃvat 1652 varṣe āṣāḍha vadi paḍavo [śa]ni-vāsare Ojhā-Rudrallikhitaṃ | | adya ahanalapurapattana-madhya-vāstavyaṃ | ābhyaṃttara-nāgara-jñātīya-paṃculī śrī Viśru[pa]-suta paṃ˚ | Vasīṣṭa | suta paṃ˚ | Lahūā | suta paṃ˚ | Vāsaṇa | suta paṃ˚ | Nārāyaṇa | suta paṃculī  Śivā liṣāvitaṃ.53

This gives the same information, but it was written very soon after the first part of the Ms., on the 6th (?) day of the dark half of the month of Āsāḍha VS 1652 (1595 CE), a Saturday, —thus merely three (to maximum four) weeks after the first half of the AVŚ Ms. (that is, if the calendar used is amānta). 1598-(BORI 386) An Atharvaveda Kramapāṭha Ms. (Deccan College 133/1879-80); 1598 CE svasti saṃvat 1598 varṣe āśvana va di 1 budhe paṃcolī … likhitam idam.

Written in the early autumn of 1598, apparently by a Pañcolī Brahmin. The name seems to have been erased, as is fairly common upon sale of a Ms.; see above 1421-(SPr). 1607-(BORI 410) A Gopatha Brāhmaṇa Ms. (Deccan College 88, 1880-1) 1607 CE

52 Ojhā is a typical north Indian family name, derived from upādhyāya. See Turner 1969: 2301. The modern meanings vary from “Brahman teacher” to “sorcerer”; in Gujarat even “potter.” 53 Followed by the interesting note similar to the one just mentioned: putrapautrādikapaṭhanārthe || śubhaṃ bhavatu || aparaṃ  pustakaṃ  vīkṣya | śodhanīyaṃ sadā budhaiḥ || hīnādhikyai svarair varṇaiḥ-|| -r asmākaṃ dūṣaṇam [nahi] ||. It stresses that one should compare other Mss. and that it is not the scribe’s fault if accents marks (svara) have disappeared.

A Prosopography of the Śaunakīya Atharvaveda Families Introduction

Ù  345

svasti śrī saṃvat 1654 varṣe pauṣamāse | śuklapakṣe | paṃcamyāṃ tithau | ravivāsare |

No further details. 1611-(II.APrM) An AV-Prātiśākhya Ms. (ed. Surya Kanta), 1611 CE saṃvat 1668 varṣe dvitīya āṣādha vadi 6 ṣaṣṭhī bhr̥guṇe vāsare

The Ms. was copied in the summer of 1611 CE at an unknown location, but certainly in Gujarat as it follows local writing style, see the facsimile in Surya Kanta 1613-(Ch) A Kauśika Sūtra Ms. (ed. Bloomfield); 1613/14 CE saṃvat 1670 | kāśyāṃ gaṅgāsamipaṃ  viśveśvarasaṃnidhau | BhaṭaGaṅgādhareṇa likhitam

This Ms. was written at Kāśī, near the Viśveśvara temple, by Bhaṭṭa Gaṅgādhara. 1613-(Ch) A Kauśikasūtra Ms. with Paddhati, 1613/4 CE Chambers (Berlin Libr. no. 119), already written at Kāśī, saṃvat 1670 (1613/4 CE). 1614-(9.V) An AV-Prātiśākhya Ms. (ed. Surya Kanta), 1614/1619 CE (a) saṃvat 1676 varṣe jyeṣṭha vadi (1619 CE) (b) saṃvat 1671 varṣe āśvina vadi 12 some adyeha śrī vācāle nagaravastavya-nāgara-jñātīya-Rāäla-śrī puṇyakaputra Mādhavajī-bhrātr̥Viśvanātha adhyāyanārtham

The Ms. consists of six AV texts; last section (Prāyaścitta) in S. 1671 (b). The Ms. was copied in the spring of 1619 CE at Vācāla by Mādhava, son of the Nāgara Brahmin Rāala, for study purposes of his brother Viśvanātha. (Surya Kanta thought that this and other Mss. were “from the same district in Deccan”).

1619-(N.5) An AV-Prātiśākhya Ms. (ed. Surya Kanta), 1619 CE saṃvat 1676 varṣe āśvina vadi 13 some adyeha śrī-staṃbhatīrthavāstavyaṃ śrīmad-ābhyaṃtaranāgara-jñātīya paṃca[ka]lpī śrī Murāri

346  Ù

The Vedas in Indian Culture and History tadanuja-Nr̥siṃhena svayaṃ likhitam idam. śrī gopījanavallabho jayati

The Ms. was copied in the summer of 1619 CE by Nr̥siṃha, son of the Pañcolī Murāri, living at Stambhatīrtha (Cambay, Khambat). Note the Kr̥ṣṇaite benediction at the end, for which see the Excursus, below. See now BORI 397 (Deccan College 175/1880/1. 1624-(15) (bha2) A Kauśikasūtra MS with Keśava-Paddhati, (ed. Bloomfield), 1624 CE saṃvat 1680 pravartamāne māgha vadi paṃcamī śukre svasya nirīkṇaṇārthaṃ tathātmaja-Gaṇeśasyāvalokanārthaṃ idaṃ saṃhitāvidheḥ kauśikenoktaṃ bhāṣyāṃ yāhyanagare sāraṃgapuravāsinā tathā śaunakīśākhādhyāyinā Gaṇeśenāṃlokhiyā-davena ca

The Ms. was written by a Gaṇeśa Dave at YahyanagaraSaraṃgapura for his son Gaṇeśa, in January/February of 1624 CE. For other Daves see below, 1684-(P), however at Broach in Gujarat. Saraṃgapura probably is not the one near Chaka, Allahabad, but one in Gujarat. A second colophon has: saṃvat 1883 (1826/7 CE) nāmāgasara-māse śuklapakṣe aṣṭamyāṃ ravivāsare likhitaṃ granthasaṃpūrṇaḥ. This Ms. apparently is a copy of a Ms. written in 1623/4 CE. 1627-(P) An AV-Prātiśākhya Ms. (ed. Deshpande, Caturādhyāyī), 1627/1677 CE (a) saṃvat 1685 varṣe māgaśara vādi 5 gurau likhitam (b) saṃvat +1734 varṣa āśvana śudi 15 dine paṃcolī śrī Īśvara lakṣaṇagraṃtha bhaṇe che. pañcolī śrī Govindajī  kanera bhaṇe cha (by a second hand; saṃvat 1637 in the edition is a misprint, as confirmed by the author).

The Ms. was copied in the winter of 1627 CE, and then added to by Pañcolī Īśvara in the autumn of 1677 CE. 1638-(Bü) Atharvaveda Saṃhitā Ms. (Bühler, Gujarat Cat. 4) Two Mss.: “Atharvaveda Saṃhitā Padāni,” 500 fol. and Kāṇḍa 20. The Mss. belonged to Dave Narabherāma at Bharūch (Broach). 1641-(BORI 412) A Gopatha Brāhmaṇa Ms. (Deccan College 89, 1880-81); 1641 CE svasti śrī  saṃvat 1718 varṣe śāke 1584 pravartamāne jyeṣṭha vadya 8 śukre śrīmad aṇahillapurapattana-vāstavyaṃ śrī  ābhyaṃtaranāgara-

A Prosopography of the Śaunakīya Atharvaveda Families Introduction

Ù  347

jñātīya paṃcolī śrī 6 Devajī-suta-paṃcolī-nyāyā-suta-Lomajīkena likhitaṃ paṃcakalpī Br̥haspatī-paṭhanārthaṃ

The Ms. was written in the spring of 1641 by the Pañcolī Lomajī, son of Devajī, living at Patan; for Br̥haspati cf. 1660- (B), (2.B). 1645-(J) An AVŚ Padapāṭha Ms. (ŚPP); 1645/50 CE At the end of AV 19, we find: saṃvat 1702 varṣe jyeṣṭha vade 4 śanidine āgarā-madhye poṭhi lyaṣī che || lyaṣitam (=+likhitam) Vāṃmana || paṃcoli-nyāmnyā (+nāmnā)-putra Somajī Vāṃmana-samasta-bhrāta-paṭhaṇārthaṃ || maṅgalam Vāṃmane svara dīdhā che || Ānaṃtajī ne bhāgya che ||

This part was written in Jyeṣṭha (May) 1645 CE at the Moghul capital of Agra in Uttar Pradesh, by Vāmmana (= Somajī?) for all his brothers; he belongs to another branch of the Pañcolī than the Alavesara line, see the diagram, below. The accents (svara) have been added by Vāmmana. And, at the end of book 20 we read further: saṃvat 1702 varṣe aśvana- (+āśvina) māse śuklapakṣe 3 guruvārena likhitam idaṃ pustakam || lyakitam (+likhitam) Vaṃmaṇa || śrī || cha || kalyāṇam astu || śubhaṃ bhavatu || śrīr astu || paṃcoli-nyāmnya (+nāmnā) Somajī Vāmmana bhrātā-Ciṃtamaṇa (+Cintamaṇī) samasta-bhrātā-pāṭhaṇartham || ... || Ānaṃtajī ne bhāgya che ... ||

This part of AV, too, was written by Vāmmana in Āśvina (Sept.) 1645 CE, thus just a few months later. It is not clear where, but presumably also at Agra. But then, in 1650 CE, another part of the Saṃhitā, book 17, was written by Vāmmana, back at Patan in Gujarat. The colophon at the end of book 17 reads, in part again in Gujarati: svasti saṃvat 1706 varṣe jyeṣṭha śuda 3 guruvāsare śrī aṇahilapurapattane vāstavyaṃ vr̥ddha-nāgara-nyātīya (+jñātīya)-paṃcoli-nyāṃnyā(+nāmnā)suta Somajī bhrāta Vāmaṇa Ciṃtāṃmaṇa (+Cintamaṇī) Anaṃtajī paṭhanārthaṃ || liṣitaṃ (+likhitam) Vāṃmaṇa || || śrī || paṃcolī-nyānyātmaja(+nāmnātmaja)-Goviṃdajī-Vāṃmaṇakena padasaṃhitāyāṃ Sūradattāpathī-pāṭhaṇa || … Anaṃtajī ne bhāgya che ||

The Ms. was copied in Jyeṣṭha (May) 1650 CE at Anahilapura (Patan), and corrected (?) according to the recitation of Sūradatta(?), by Goviṃdajī Vāṃmana, for study use of his brothers Somajī, Cintāmaṇi and Ānantajī. It belonged to Ananta.

348  Ù

The Vedas in Indian Culture and History

1650-(A) Undated Ms of the AV-Prāyaścitta, c. 17th century CE śrī-viśveśvarāya namaḥ | śrī-sarvavidyā-nidhāna-KavīndrācāryaSarasvatīnām atharvavede vaitānasūtre prāyaścitti-prasaṅgapustakam.

The Viśveśvara location and the name Kavīndra point to 17th cent. Benares: see on the Kavīndrācārya Mss. collection, above, Ms. 1421-SPr, and the description by the famous Mss. collector R.A. Sastry, Kavīndrācārya Sūcīpatram. Gaekwad Oriental Series 17, 1921. 1651-(K) A Kauśikasūtra Ms. (ed. Bloomfield), 1651 CE svasti śrī saṃvat 1708 varṣe kārttika śuda 10 ravau adyeha śrī  maphalīpura-vastavyam ābhyaṃtaranāgara-jñātīya trivāḍī śrī 5 Śivajīsūt Jeyakr̥ṣṇena svayaṃ likhitam idaṃ | adyeha maphalīpuravāstavyaṃ ābhyaṃtaranāgara-jñātīya paṃcakalpī Gaṇapatya-bhrātr̥Gaṃgādhara-bhrātr̥- Rāhāgavajī-bhrātr̥- Gadādhara-bhrātr̥-Vidyādharapaṭhanārthaṃ.

The Ms. was copied by the Ābhyantaranāgara Brahmin Jayakr̥ṣṇa, son of the Trivedin Śiva, in the town of Maphalī  in the autumn of 1651 CE, for the use of his Pañcakalpin brothers Gaṇapati, Rāhāgava (+Rāghava?), Gadādhara and Vidyādhara. The location of Maphalī is unclear; perhaps it is the modern Māvlī / Māhōḷī, some 40 km NE of Udaipur in Mewār, or rather Mahuvā in SE Saurastra? At any rate, Maphalī does not correspond to the later location of Atharvavedins in Maharastra at Māhulī near Sātārā in Sangam Dst. (where this Ms. was discovered in c. 1880), as Gujarat Atharvavedins moved into Maharastra only about 1730 CE (see below). As mentioned, pañcakalpin54 refers to the study of the five kalpas of the AV, such as the Nakṣatrakalpa, etc. 1658-(NŚ) Naigeya-śākhānukramaṇī, 1658 CE saṃvat 1715 varṣe aśvina vadya amāvāsyā śanau ’dyeha muphalī-puravāstavyam ābhyaṃtara-nāgara-jñātīya trāvādī (+trivedī) Śevarāja-sunu Devakraṣṇena svayam likhitaṃ < … > | atra pautrāṇāṃ pāṭhanārthaṃ ||

This Ms. giving a list of contents of the Naigeya school of the Sāmaveda was written in the autumn of 1658 CE, by “himself,” the Ābhyantara-Nāgara Brahmin Devakr̥ṣna, son of Śivarāja, living at Muphalī (or Maphalī) see above 1651-(K). 54 See Bloomfield 1889: lvii; see below, Ms. 1651-(K), 1661-(3.J) mentioning pañcakalpin instead of the usual pañcolī.

A Prosopography of the Śaunakīya Atharvaveda Families Introduction

Ù  349

1660- (B), (2.B) An AV-Prātiśākhya MS (ed. Deshpande B, Surya Kanta 2.B), 1660 CE saṃvat 1717 varṣe bhādrapadamāse kr̥ṣṇapakṣa 11 ravivāsare adyeha śrī anahalapurapattana-madhye-vāstavyaṃ ābhyaṃtaranāgara-jñātīyapañcolī Somajī-suta Br̥haspatijī-paṭhanārtham.

Written in the summer of 1660 at Patan by Br̥haspati Pañcolī, son of Somajī, for study purposes. 1661-(E) An AV-Prātiśākhya MS (Caturādhyāyikā, ed. Deshpande); 1661 CE Pañcolī-Nāgajit-sununā kaṇvālaya-nivāsinā Bhavadevena idaṃ granthaṃ likhitam | saṃ. 1718 kārtika śu di 11 budha |

The Ms. was copied by Bhavadeva, son of Nāgajit Pañcolī, at Kāṇvālaya in the autumn of 1661. Kāṇvālaya is a forest area called Papadnod in Kathiavar (Saurastra), in the former Khachar state; it is situated between the Sarangashringa hills and the river Samburi in the east, the river Macchu in the west, the Chotila hills (Chotīlā, NE of Rajkot) in the south and the Taladhar mountains in the north. See now BORI cat. no. 403. 1661-(3.J) An AV-Prātiśākhya Ms. (ed. Sūrya Kanta) saṃvat 1718 vārṣe kārttika śudi 12 gurau kaṇvālayanivāsinā paṃcakalpī śrī 5 Rāmacandrātmaja Nāgajit-sununā Bhavadevākhyenedaṃ graṃthaṃ likhitaṃ

The Ms. was copied by Pañcolī (Pañcakalpī) Rāmacandra’s son Nāgajit, known as Bhavadeva, living at Kāṇvālaya, in the autumn of 1661. See now BORI catalog no. 396. This Ms. does not seem identical with Deshpande’s (E), though most details, even the year, month and pakṣa, though not the actual day and designation of the weekday, overlap. 1661-(or.fol. 621c) A Kauśika-Paddhati Ms. (Berlin Ms. or.fol. 621c); 1661 CE saṃvat 1674 varṣe mārgaśīrṣa śu di 13 ravau likhitam idaṃ | paṃcolī Bhavadeva Nāgajī nī pothi karmāṃtara nā patra 48

The Ms. was copied in the late autumn of 1661 by(?) the Pañcolī Bhavadeva Nāgajī. This apparently is the same person as in Ms. 1661-(3.J): Bhavadeva, son of Nāgajit.

The Vedas in Indian Culture and History

350  Ù

1663-(P) An AVŚ Padapāṭha Ms. (ŚPP), 1663 CE saṃvat 1720 varṣe jyeṣṭhamāse kr̥ṣṇapakṣe saptamyāṃ tithau ravivāsare || idaṃ pustakaṃ saṃhitāyāṃ pada saṃpūrṇaṃ samāptaṃ |< … > adya śrī  gūrjarakhaṃḍe uttaravibhāge aṇahilapurapattana-vāstavyaṃ ābhyaṃttara-nāgaraṃ-jñāti-paṃcolī  śrī Viśrupa-suta paṃcolī Vasīṣṭasuta paṃco˚ | Lahūā–suta paṃco˚ | Vāsaṇa-suta paṃcolī  śrī Nārāyaṇasuta paṃcolī Bacyāṃ-suta paṃcolī Vāchā-suta Bhavāṃnīdāsa paṭhanārthe saṃhitāyāṃ pada lakhite ||

The Ms. was copied in the summer of 1663 CE, four months before ŚPP’s Ms. 1673-B, for the same person, Bhavāni Dās. The family tree coincides with the second one of AV-Par Ms. 1584-R, and for the earlier part with ŚPP’s AVŚ Ms. 1594-R. 1663-(B) An Atharvaveda Ms. (ŚPP); 1663 CE saṃvat 1720 varṣe aśvanamāse śuklapakṣe 3 tr̥tīyāyāṃ tithau somavāsare adya śrī Aṇahilapurapaṭaṇa-vāstavyaṃ ābhyaṃtara-nāgaraṃ-jñāti || paṃcolī śrī-Viśrupa-suta paṃco˚ Vasīṣṭa-suta paṃco˚ Lahūā–suta paṃco˚ Vāsaṇā-suta paṃco˚ Nārāyaṇa-suta paṃco˚ Bacyāṃ-suta paṃcālī Vāchāsuta Bhavaṃnīdāsa likhitaṃ

This Ms. was written by Bhavānī Dās in the summer of 1663 CE. The family tree given above coincides with the second one of AV-Par Ms. R of 1584 CE, and for the earlier part with ŚPP’s AVŚ Ms. R of 1594 CE. 1669-(Ady 528) An Atharvaveda Saṃhitā Ms. (Adyār Library 528, 41 B 69) 1669 CE saṃvat 1816 varṣe bhadrapadamāse kr̥ṣṇapakṣe 8 ravivāsare || adya śrī anahalapurapattana-madhye vāstavya ābhyaṃtaranāgara-jñātīyapaṃcolī-nyānyā-suta-Vāmana Nīlakanta Sadāśiva paṭhanārtham Vaiṣṇava Nārāyaṇa Dāsena likhitaṃ

The Ms. was written by one Nārāyaṇa Dāsa for the study of Nīlakanta Sadāśiva Vāmana, a Pañcōlī, at Patan in the Summer of 1669. Cf. the Pañcolī Vāṃmana at 1645 who however does not seem to have a Śaiva link. 1670-(Jaṭha17). A Jaṭhapāṭha Ms. (ed. Deshpande, HOS 61) saṃvat 1727 varṣe śāke 1593 nā prathama-vaiśākha va di 9 ravau vāsare adyeha śrīmad-aṇahillapurapattana-vāstavyam ābhyaṃtaranāgarajñātīyā [paṃ]colī-nyāyā-suta-Goviṃda-sū-pāṭhena paropakārārthe likhitaṃ | śrī sa Goviṃda … svasti 55 55

Given according to the BORI catalogue no. 388; small deviations (anusvāra)

A Prosopography of the Śaunakīya Atharvaveda Families Introduction

Ù  351

The Ms. was copied in the spring of 1670 at Patan for study purposes of others by Govinda Pañcolī (and his sons?) (Deshpande points to Govinda in AV Prāt. (P), and AVŚ (ŚPP) Ms. (C) (J). (cf. Mss. 1650, 1645 CE). See now BORI cat. No. 388 (Deccan College 83/1880/81). 1675-(4.M) An AV-Prātiśākhya Ms. (ed. Surya Kanta), c. 1675 CE? paṃcolī śrī  6 Jayakr̥ṣṇa-suta Rāmajī-susurajī-bhrātr̥- Sāṃmajībhrātr̥- Mādhavajī-paṭhanārthaṃ Mādhavajīkena likhitam | śrī śrī  Lūṇanāthasa che

The Ms. was written by Mādhava, son of Pañcolī Jayakr̥ṣṇa, for his own study purposes and that of his brother Sāmma, and his susura (śvaśura, father in law) Rāma, apparently in the area of Lūṇāvaḍā (E. Gujarat) 1678- (K) An Atharvaveda Saṃhitā Ms.; 1678 CE Written saṃvat 1735, śake 1600, 1678/9 CE, by Eṃmavā Gaṇeśa under King Anūpasiṃha, at Pattananagara. 1678-(Bi) A Kauśikasūtra Ms. (ed. Bloomfield), 1678 CE saṃvat 1735 | varṣe mārgaśīrṣa śudi 4 gurudine śrī-pattananagare likhitaṃ | granthāgram 1800 | mahārāja Anuṃpasiṃhajī vijeṃrājye śubham bhavatu

The Ms. was copied in Patan in the autumn of 1678 CE, under the reign of King Anūpasiṃha. Unfortunately the scribe’s name is not given. The Ms. is from the Royal (Anūp) Library of Bikaner (# 91). 1681-(Ba) A Kauśikasūtra Ms. with Keśava-Paddhati (ed. Limaye et al.); 1681 CE svasti śrī  saṃvat 1736 varṣe phālgunamāse śuklapakṣe pratipattithau bhr̥guvāsare adya śrī  pattanavāstavya ābhyantara-nyāti-nāgara-pacoliGaṇapa[ti]-suta pacolī-Gaṃgādhara-suta pacoli Kahana (Kr̥ṣna)-suta pacoli-Madhachuda-suta paṃcoli-Devajī-suta paṃcoli-Udhava-suta Paṃcoli-Somajī-suta bhrasyatyajīkasya(?) keśikagrajya-padānī pustaka || sa yāta lakhi pacolī-Vāchā-suta Tragavatī (Bhavānī/Bhagāvatī?)Dāsakenya lakhitam |

in spelling from Deshpande. However, the last words read śrī ras[…] paṭhitaḥ in Deshpande’s edition.

352  Ù

The Vedas in Indian Culture and History

It appears that this ms. was copied in the spring of 1678 CE, in Patan, for the Pañcōli family starting with Gaṇapati (cf. 1481 CE) and ending with Somajī and Govindajī. The copyist was +BhavānīDāsa. (The Gujarat style Nāgarī  letters for tra and bha are very similar). 1681-(C) An AVŚ Saṃhitā Ms. (ŚPP); 1681/1691 CE A first colophon, in Gujaratī script has: pacasī Govindajī || cha || bhraspatmaja (+Br̥haspaty-ātmaja?) śrīr astu || cha || saṃ 1738 nā caitra śuda 10 sa Balarāṃma śrī saṃvat 1746 varakhe (+varṣe) je[ṣ]ṭha śuda 12 sa Dhaneśvara.

A second hand colophon, again in Gujarati script (p. 72): … pa[ṃ]cā Govaṃdajī nī vīme … paṃcā Govandajī nī pāthī che (the same repeated in Nāgarī : paṃcā Gāvaṃdajī nī pāthī che)

From this, ŚPP concludes that in VS 1738 and 1748 Balarām and Dhaneśvar must have used this Ms., while Govindjī Pañcolī  must have been the owner at some time. 1684-(P) A Kauśikasūtra Ms., (ed. Bloomfield), 1684 CE The Ms. has two colophons: (a) saṃvat 1740 varṣe śāke 1606 pravartamāne adyeha brugukṣetramāhāprabhustāne Dave śrī 5 Jayakr̥ṣṇa-suta-Devakr̥ṣṇena likhitam … saṃvat sattara-śyālīśa-varṣe (= 0471) śāke solaśe saṃpravartamāne phālguna vadi śukreṇa likhitam kauśikam pūrvārdhaṃ likhitam asti (b) svasti saṃvat 1740 | varṣe śāke 1606 | pravartamāne caitra śudi da budhe adyeha brugukṣetravāstavyaṃ Dāve śrī  5 Jayakr̥ṣṇa-suta Devakr̥ṣṇena likhitam

The Ms. was copied by Devakr̥ṣṇa, son of Jayakr̥ṣṇa Dāve, at Bhr̥gukṣetra (Broach), in the late winter of 1684 CE. The Dave/ Dave clan is not attested in Yajnik’s list (JBBRAS 10:93-119). Bhr̥gukṣetra is Bhr̥gukaccha, modern Bharuch (Broach), in S. Gujarat where they survive still. 1684-(D) An Atharvaveda–Padapāṭha Ms. saṃvat kurauvedasapteṃdu [1471 = saṃvat 1741] | śrāvaṇe śuklapakṣe tr̥tīyā maṃda-vāsare hi | lekhitā cattane-pure || 1 || saṃvat 1741 varṣe śāke 1606 prathame śrāvaṇa śudi 3 śanivāsare idam pustakaṃ likhitaṃ

A Prosopography of the Śaunakīya Atharvaveda Families Introduction

Ù  353

1684-(P2) An AVŚ Padapāṭha Ms., 1684 CE saṃvat kuvedasapteṃdu [1471 = VS 1741] | śrāvaṇe śuklapakṣake | tr̥tīyā maṃda (+maṅgala-)vāre hi | lekhitā pattane pure ||1|| saṃvat 1741 varṣe śāke 1606 prathamaśrāvaṇa śudi śanivāsare idaṃ pustakaṃ likhitaṃ …

The Ms. was copied in the summer of 1684, on Saturday, the third day of the bright half of the month in the first (intercalary) month of Śrāvaṇa of VS 1741 of Śāka 1606 in (Anahila)purapattana (Paṭaṇ). 1688-(ŚB) 1688 CE (Jammu, Cat. Stein) Śatapatha Brāhmaṇa, saṃvat 1745, written by an Ābhyaṃtaranāgara Brahmin, perhaps: Janīveja-nāma, at Giripura (Girnar?) under a King Gaṅgadāsa. 6th kāṇḍa, complete. 25 fol.; no extract in Stein’s catalog. This ŚB Ms. is dated in Stein’s Catalogue p. 8 : saṃvat 1145 = 1088 CE. However, on autopsy (on my film), the number rather looks like saṃvat 1745 and the type of Nāgarī used is not that of the 11th century either. 1691-(D) An AVŚ Saṃhitā Ms. (ŚPP); 1691/2 CE The colophon, probably by a second hand, reads: svasti saṃvat 1748 va[r]ṣe pauṣa śudi 5 dine li˚ Śivadeva || Paṃḍyā śrī Puṣīmajī Nr̥siṃhe paṃcolī Kr̥ṣṇadeva tathā bhrātr̥- Śivadeva ta˚ Śaṃkarena śivārpaṇabuddhe āpi ||…

An additional note by a third hand reads: pānāṃ 40 te Svarvadat (!) paṃcolī Kraṣṇarāma Rāvarāme

ŚPP understand this as “Kr̥ṣṇa Rāma Pancholī surnamed Sūryadatta having borrowed or read 40 leaves.” He describes the Ms. as “very old.” 1692-(N) A Padapāṭha Ms. in Kathmandu, 1692 CE Nepal National Archives Ms. pra- 801,56 No. 380 of the Section Dharma: Veda, listed under the misleading title Vedaprapāṭhaka, 56 That is, prathama, the first collection of the former Bir and Darbar Libraries. The accession date is: śrī 3 Candra Samser 1970 = 1913/14 CE.

354  Ù

The Vedas in Indian Culture and History

film A 597/31 of NGMPP. Its date corresponds to the spring of 1691 CE. The Ms. was written by order of the Malla king Bhūpatīndra57 of Bhaktapur, one of the three Newar kingdoms of the Kathmandu Valley. Written in Newari script with red accent marks. The colophon runs: samvat 812 vaiśaṣe vadi 7 thva kuhnu śrī 2 bhūpatīdra-malla-devayā ājñāna śrī hemarājena coyā dina “in the year 812 [1692], on the 7th day of the dark half of the month of Vaiśākha, on this day, by order of Śrī Śrī Bhūpatīndra Malla, it has been written by Śrī Hemarāja”58 (a typical Newar name).

The Ms. has some peculiar ways of marking the accents.59 As in other late AV Mss., the basic system underlying the accent marking is the usual, well-known one of the R̥V and TS: the Udātta is not marked at all, while the preceding Anudātta is marked by a dot below the syllable in question. However, differently from the well-known R̥V system, where the Svarita is marked by a vertical stroke on top of the syllable, it is represented here by a small dot after the syllable in question. The standard sequence therefore is seen in: rū.pāṇio = rūpā́ṇi. This is fairly close to the system of the Śaunaka and Maitrāyaṇi Saṃhitā that is found in several old Mss. in Gujarat. However, instead of the Padapāṭha usage of Avagraha to indicate internal word division there is a small circle (o). This is quite similar to the Gujarat AV Saṃhitā Mss. P (1663 CE) of ŚPP, but it differs as far as the independent (jātya) Svarita is concerned: it is not written by a “crooked line” but by a half-circle below the line, thus AV 1.1.1: śā 60 | yé | tri 61 o saptā́ḥ | pari o yáṃti | víśvā | rūpā́ṇi | bíbhrataḥ | In sum, it is very likely that this Ms. was copied into Newari characters from a Gujarat original written in Nāgarī. Due to the close cultural relations between Nepal and Benares it most likely was obtained from Pañcolī Brahmins at Benares, where they 57 This could be before he actually became king, cf. Regmi 1965: 2.235: his coin is of 1696 only. 58 This is partly in Newārī language (Nepāl Bhāṣā): thva = this, kuhnu = day, -yā = genitive ending, coyā = it was written. 59 In red ink after the Ms. had been copied, as is common practice (see already M. Müller 1890-92 passim, and Śaṅkar Pāṇḍuraṅ Pandit in 1895: 15, who reports that some Ms. was accented after the Vaidika “had seen the svaraśāstra”; see above, Ms. 1645-(J) of 1645/50 CE. 60 Perhaps a version of the siddham symbol or for the Paippalāda-inspired beginning of some AVŚ Mss. with śam no devīr … 61 For this sign, see Witzel 1974: § 1.4.5: in compounds, this part is treated as separate word without accent.

A Prosopography of the Śaunakīya Atharvaveda Families Introduction

Ù  355

had moved around 1550 CE.62 There are a few other Mss. in Nepal that stem from Gujarat Vaidika Brahmins, such as a Lilāvatī Jyotiṣa Ms. (see below). 1696-(D) An AV-Prātiśākhya MS (Caturādhyāyikā, ed. Deshpande) = (6.P) of Surya Kanta’s AV Prātiśākhya (Pañcapaṭalikā); 1696 CE The colophon is given according to Deshpande: saṃvat 1753 varṣe caitra śudi 2 ravau adya 63 śrī  anahīlapurapattane vāstavyaṃ ābhyaṃtara-jñātīya- tulāpuruśa-brahmāṃḍamahīmahādānādi atirudrakartāhitāgnī 64-paṃcakr̥tva-cāturmāsyayāji-tripāṭhi śrī Anaṃtajī-suta-Vrajabhūṣaṇena śrī sāṃbasadāśi[vā]rpaṇabuddhyā likhyāpitam 65 idaṃ paropakārāya.

This Ms. was copied by Vrajabhūṣana, son of Anantajī, for someone else’s use at Patan in the spring of 1696. Ananta was a reciter of three Vedas and a great ritualist who had performed the Cāturmāsya and various post-Vedic rituals. Note that the religious affiliation is not with Kr̥ṣṇa but with Śiva.

62 For more on AV Mss. in Nepal see Witzel 2001. There is another Atharvaveda Saṃhitā from Bhūpatīndra’s time, with the usual accent marks (Ms. 1-711), as well as an Atharvaṇavākya (NGMPP E 128/8 (śabdavyākhyānam nirṇayo nāma ekādaśamaḥ  siddhāntaḥ). Later Mss., of the past 250 years, all in Devanāgarī, include: Atharvaveda Saṃhitā, some Atharvaveda-Pariśiṣṭas, Nakṣatrakalpa, Śāntikalpa, Gāyatrīkalpa, Atharvavidhāna, etc. They are, in some detail: Atharvaveda Saṃhitā (Śaunaka): Ms. 4-204, Veda 7, NGMPP B 492/2, accented in modern AV style, dated VS 1942/1885 CE, copied at Kathmandu, Bhosiko Tol; Ms. 4-2118, Veda 6, AV accents, dated VS 1949/1892 CE, B 492/7-493/1; various excerpts from the AV: NGMPP A 588/11 (accented), 12, 15, 16, 17 (accented), 18, 20 (incl. RV 1.90), 25, B 503/26, B 504/39; Vaitāna Sūtra: NGMPP A 597/18; Atharveda Pariśiṣṭa: A 588/26; AV-Nakṣatrakalpa: B 358/5; AV-Śāntikalpa: B 358/6; Śaunakīya-Gāyatrīkalpa: B 380/23; Atharvavidhāna (“from Agnipurūṇa,” “by Kāśīnātha Bhaṭṭa”); Yajurvidhāna A 596/30; Śaunakīyasaṃgraha: Stotra coll. no. 1592; Atharvavedīya-viniyoga (a part of Viniyogamālā): A 588/13, B 402/15; Śaunakamūlavacana: Karmakāṇda (KK) section no. 2192; a part of Adbhūtasāgara, from the Paippalāda school: Karmakāṇḍa no.179; Kauśika GS, Madhuparkaprayoga: KK 2112; Caraṇavyūha (and Nakṣatrakalpaśrāddhavidhi, fols. 80 sqq.) A 490/2, A 609/7, B 500/18, 21, 26; Gopatha-Brāhmaṇa 1-2: 4-207, Veda 8, A 588/10; AV Upaniṣads: Atharvaśīrṣa Upaniṣad (includes Sūtra and Devī Up.): A 588/14; Praśna Up.: A 844/1, 7, A 860/1; Praśna Up. Ṭīkā: B 658/9 etc. See JNRC 12. 63 adyeha Surya Kanta. 64 atirudrakarmāhitāgni Surya Kanta. 65 likhāpitam Deshpande.

The Vedas in Indian Culture and History

356  Ù

1696-(BORI365) An Atharvaveda Saṃhitā Ms. (Deccan College 125/1879-80); 1696 CE saṃvat 1753 varṣe caitra śu di 2 ravau adye śrī anahīlapurapattane vāstavyaṃ ābhyaṃtanāgara-jñātīya tulāpuruśa-brahmāṃḍamahīmahādānādi atirudrakartāhitāgnī-paṃcakr̥tva-cāturmās[ya]yāji-tripāṭhi śrī Anaṃtajī-suta-Vrajabhūṣaṇena śrī sāṃbaśivārpaṇabuddhyā likhyāpitam idaṃ paropakārāya.

Cf. Colophon of 1696-(D). 1700-(C) An AV-Prātiśākhya MS (Caturādhyāyikā, ed. Deshpande); c. 1700 CE From Anahilapurapattana, written by a Pañculī-Īśvara-sutaNīlakaṇṭha-paṭhanārtham, c. 1700 CE (M. Deshpande). 1709-(Cp) An AVŚ Saṃhitā Ms., 1709 CE saṃvat 1766 varṣe jyeṣṭha vadi 7 buddhe samāptaṃ || staṃbhatīrthavāsī udīcya-ṭalakīā bhaṭṭa-Bakātmaja Ātmārāmeṇa likhitaṃ ||

The Ms. was copied in the summer of 1709 CE by one Ātmārāma, son of Bhaṭṭa Baka, an Udīcya Ṭoḷakīya Brahmin,66 living at Stambhātīrtha; cf. above Murāri/Nr̥siṃha at the same location (1619 CE). (The ms. includes this post-colophon benediction: śrī gurubhyo namaḥ | Rāüla śrī someśvara-caraṇa-kamalebhyo namaḥ, which points to the famous Someśvara (or Somanātha) Temple on the coast of southern Saurashtra). See now BORI cat. No. 381. 1717-(BORI 372) An AVŚ Saṃhitā Ms. (Deccan College 126/1879-80); 1715-1717 CE saṃvat 1774 varṣe śāke 1640 pravartamāne hemānta r̥tau … śrāvaṇamāse kr̥ṣṇapakṣe 4 somavāsare lakhitaṃ Dave Vīreśvara-sūtala Dave Pītaṃbara lakhitaṃ Dave Vīreśvara-suta Dave Maheśvara paṭhanīyaṃ

The Ms. was copied in the summer (month of Śrāvaṇa) of saṃvat 1772-1774 (1717 CE) by Pītaṃbara, son of Vīreśvara Dave for the use of his brother Maheśvara. The scribe calls the summer month of Śrāvaṇa one “at the end of winter.”

66

See Yājnik 1871-74: 109, no. 9.

A Prosopography of the Śaunakīya Atharvaveda Families Introduction

Ù  357

1720-(BORI 382) An Atharvaveda Padapāṭha Ms. (Deccan College 1/1891-95); 1720 CE svasti śrī saṃvat 1777 varṣe āṣāḍhamāse śuklapakṣe ekādaśāṃ puṣpatithau ekādaśī śanīvāsare śrī || rājanagrā[ma]-madhye || likhitaṃ gaṃdholakīya || nāgara-jñātīya vr̥ddhanagrā || paṃco. Devarāja-suta paṃco. Deneśvara tathā || kaniṣṭha bhrā. Bhavaṃnīśaṃkare likhitam idam ||

The Ms. was copied in the summer of 1720 at Rājanagrāma for(?) the Pañcolī Brahmin Deneśvara and his younger brother Bhavaṃnīśaṃkara, son of Devarāja who may in addition to Pañcolī also be called a Vr̥ddhanagarā Brahmin, known as Gaṃdholakīya Nāgara. 1756-(Mill80) An Atharvaveda Saṃhitā Ms. (Bodleian 938); 1756 CE adhika jyeṣṭḥa va dī budhavāre saṃvat 1812

This is the original colophon, however the current Ms. was copied from the 1756 Ms. only in c. 1840 CE. Additional note: likhitaṃ tr̥˚ travāḍī-śrī Kr̥ṣṇa-suta-Bālakr̥ṣṇa bhavaṇa-māṭhe lakhī che svārthaṃ ca paramārthaṃ ca vārāṇaśī-madhye, — that is, copied by the Trivedin (Tiwāri) Bālakr̥ṣṇa, son of Kr̥ṣṇa, at the Bhavaṇ Maṭh in Benares, “for someone else.”

1788-(A) A Gopatha Brāhṃaṇa Ms. (Munich: Haug 22); 1788/9 CE (a) idaṃ pustakaṃ Tryambaka-daivajñātmaja-Bāpu-daivajñasya khali iti || (b) Tryambaka-daivajñātmaja-Vāpu-bhaṭṭena likhāpayitvā dīyate Govindamalharasya… (c) idam  Tryambakasya sununā Bāpu-bhaṭṭena likhāpayitvā saṃvat 1710 (d) Tryambaka-daivajñātmaja-suta-Govindena likhyate

This Ms. obviously was written by the astrologer Bāpu, son of Tryambaka and then given to his brother Govinda. 1798-(Dc) An AVŚ Saṃhitā Ms., 1798 CE saṃvat 1855 varṣe kārtika śukla 14 guruvāsare

The Vedas in Indian Culture and History

358  Ù

A Ms. apparently from the Māhulī school of Atharvavedins in Maharashtra, copied in the autumn of 1798 CE, with G.D. Bhaṭ’s “revised” AV text. 1810-(Daś) A Daśakarma Ms. (ed. Bloomfield), 1810 CE saṃvat 1867 kārtikamāse śuklapakṣe 6 bhr̥guvāsare adyeha Siddhapurvāstavyaṃ udīcyā-jñātīya Śukla-Sukharāma-ātmaja Hr̥dayarāma likhitaṃ | pathaṇārthaṃ Paṃcolī-Rāmakr̥ṣṇa | idaṃ pustakaṃ vārāṇasī-kṣetre kāśyāṃ lipikrtaṃ | ādiviśveśvara-samīpe

The Ms. was copied in the autumn of 1810 by Hr̥dayarāma, son of Sukharāma belonging to the Udīcya and Śukla group of Gujarat Brahmins, living in Siddhapur (some 20 km east of Patan in N. Gujarat). The Śuklas were reciters of the Maitrāyaṇi Saṃhitā in the mid 1800s.67 The Ms. was copied for study purposes, apparently for that of Pañcolī Rāmakr̥ṣṇa, who got it copied and who lived near the Ādiviśveśvara temple in Kāśī (Benares) in Varāṇāsikṣetra. This reflects the move of Gujarat Brahmins to Benares several centuries earlier, but also the continuing relationships with their homeland. 1824-(R) An Atharvaveda Ms. 1824 CE Many of these books have their own brief colophons (all dated śāke 1746), the most extensive one is at the end of book 14: śāke 1746 tāraṇanāma-saṃvatsarai | māde pauṣa śukla saptami saha aṣṭami vyatīpāta-parvaṇi sāmavāsare prātaḥkāle śataghaṭiye kayāma 7 tasmin samaye atharvaṇasaṃhityāḥ saṃpūkāḍe yathāśaktyā Paṭhuvardhana Biṭhala Rāmakena saṃpūrṇalekhanaṃ kr̥taṃ

At the end of book 20, there is: saṃvat 1856 = 1799 CE; in Roth’s handcopy of his AV edition, Roth writes “A.D. 1770”(!). Apparently a copy of an older colophon. Whitney, transl. p. CXV, reports that this is saṃvat 1926 = 1870 CE(!). —Multiple confusions. 1827-(Bh) An AVŚ Saṃhitā Ms. (ed. ŚPP), 1827 CE saṃvat 1884-nā śrāvaṇamāse śuklapakṣe caturddasyāyāṃ 14 ravivāsare likhitaṃ gaḥ saṃpūrṇaḥ śrīḥ ||

This Ms. is dated in the summer of 1827 CE; a second colophon reads: 67

See Schroeder 1881-86: Introduction, xxii f.

A Prosopography of the Śaunakīya Atharvaveda Families Introduction

Ù  359

marusthaly-aṃtagata-pradeśa-viśeṣa-yodhapurāṃtargata-surapura – vāstavyaṃ nāṃdīmukhājñātīya Ojhākelā-suta-Vohorāgiridhareṇa brāhmaṇena revāṃtargata-rājanagaravāstavyaṃ rā || Brāhmanāgaravisallagarā atharvaṇa āppu

The Ms. was copied(?) by the Brahmin Bohora Giridhara, living at Rājanagara (Rājpur?) that is part of Revā (Narmadā River area); he was the son of an Kelā Ojhā known as Nāndimukha, living in Surapura, a part of Jodhpur (Rājasthān) in Marusthalī (Thar Desert). This Ms. was “revised” by Gaṇeś Bhaṭ Dādā (as per ŚPP). 1828-(ga) A Kauśikasūtra Ms. with Keśava-Paddhati (ed. Limaye et al.), 1828 CE rāma saṃvat 1885 || śrīkṣetra kā || śyām nā || gara-brāhma || ṇena Vedamūrtiguru Ganeśa Bhaṭajī Dādā  Gore atharvavedi tasya caraṇasevā me a˚ || arpaṇaṃ aśvina śu[klapakṣe] 5 gu[ruvāre] saṃ 1875 (sic)

This Ms. belongs to the Gore family of Gwalior, but it was written at Śrīkṣetra in Kāśī (Benares) by the Nāgara Brahmin Vedamūrti Gaṇeśa Bhaṭṭa Dādā Gore in 1828 CE, for the support of his Atharvaveda–caraṇa. Note that the current title vedamūrti occurs here for the first time in an AV colophon. This Gaṇeśa Bhaṭṭa apparently is identical with the Gaṇeśa Bhaṭṭa Dādā who visited Śrṅgeri in the early 1800s, and found there a commentary on the Atharvaveda composed by “Sāyaṇa.” 1837-(Vā) A Kauśikasūtra Ms. with Keśava-Paddhati (ed. Limaye et al.), 1837 CE chā saṃvat 1893 miti māgha śudi 2 bhaumavāsare …

Copied in the winter of 1837; (photocopy from the Ms. of V.R. Ratate of Varanasi). 1879-(Bü) An undated modern ms. of the Kauśikasūtra (ed. Bloomfield); c. 1879 CE likhāvime tātrikaṃ ajīttanaya kr̥pra Rāma likhāvitaṃ Kāśipuri-madhye idaṃ pustakaṃ paṃcolī-Rāmakr̥ṣṇasya dattam

It seems that this modern Ms., acquired by the Bombay Govt., in 1879/80, # 150, was copied by one Kr̥ṣna(?) Rāma in Benares and was given to a Pañcolī Rāmakr̥ṣṇa, confirming the new home of the Pañcolīs there.

The Vedas in Indian Culture and History

360  Ù

1884-(bhā1) A modern Ms. of the Vaitāna Sūtra (ed. Garbe = bhā1 ed. Hoshiarpur); 1884 CE svasti śrī  vikramārka-samayasamaye saṃvat 1941 || śrāvaṇamāse śuklapakṣe tithau 14 bhaumavāsare samapto (’)yam || śrī  śubham || lipyakr̥tau brāhmaṇa parīka Haragovinda vāsa Savāī Jayapura.

The Ms. was transcribed by one Brahmin Haragovinda living at Savāī in Jodhpur, in the summer of 1884. § 2. Data from the 18th-20th centuries There is piecemeal information from various Mss. of the past two or three centuries. As pointed out earlier, there is some evidence for AV tradition in Nepal around 1700 CE. The AV is needed at great para-Vedic fire rituals such as the Lakṣahoma,68 when all four Vedas must be recited from beginning to end. However, nowadays manuscripts or printed books are used that are merely read out by four Brahmins, as I have observed at the 14-day Lakṣahoma performed at Bhaktapur in Nepal in the Summer of 1976 (cf. Witzel 1976). The Mss. of the AV Saṃhitā commentary, attested to be Sāyaṇa’s in the colophons and the introductory verses,69 are not of interest here as they do not represent the direct oral tradition of Gujarat but a written one derived from Gujarat and Maharastrian Mss.70 These commentary Mss. come from the Śr̥ṅgeri Maṭha in Karṇataka and from places nearby (such as Kūḍle in the former Mysore state). In c. 1730 CE, a number of Gujarat AV reciters moved into the Maratha territories, especially into the Peshwa lands in Maharashtra proper. Śaṅkar Pāṇduraṅ Paṇḍit describes this vividly in the very much neglected introduction to his AV edition of 1895.71 See AV Par no. XXX, and on Koṭihoma no. XXXI. Which is unlike what W. Slaje (2010), has recently shown for the author of the R̥V commentary that has been wrongly attributed, since Max Müller, to Sāyaṇa. The AV commentary in his name is much worse than that of the SV (which is a “graduate student” rehash of “Sāyaṇa’s” R̥V comm.), and note already Whitney’s scathing discussion of the quality of the AV commentary (see in transl. p. lxvii f.). 70 See ŚPP 15 f., 24. 71 However his introduction has hardly been read and even less used. I did so in the late Sixties (see my still unpublished AV Prolegomena, read by a few students and colleagues since it was written in 1982/4); recently, Madhav Deshpande and Shrikant Bahulkar have made use of ŚPP’s accounts. For information on ŚPP, see the encyclopedia Mahārāṣṭra Jñāna Kośa. 68 69

A Prosopography of the Śaunakīya Atharvaveda Families Introduction

Ù  361

Subsequently, one Gaṇeś Bhaṭṭa Dāda (Gore) of Maharastra and Benares (see above: Ms. 1828-ga) visited Śr̥ṅgeri in the early 19th century and found the commentary on the Atharvaveda, allegedly composed by Sāyaṇa.72 As this was written in Nandināgarī character he misread quite few passages but incorporated these misreadings into his Ms. and his recitation which he then taught to his students in Maharastra (see ŚPP, AV ed., p. 15). This was then perpetuated by Maharastrian Veda reciters, whom ŚPP recorded around 1890.73 These Atharvaveda reciters and priests survived into the 20th century at Sānglī and Sātārā, in the small independent Marathi kingdom left untouched by the British. This is the Māhulī School, attested in Mss. since at least 1798 CE (Ms. Dc).74 From the notes inserted in ŚPP’s and other editors’ discussion of their Mss., we can glean the following data for the late 19th century. The locations are concentrated, first and foremost, in Central and Eastern Saurastra, but locations also include Southern Gujarat, Southern Maharastra, Gwalior in N. Madhya Pradesh and, as expected, Benares. Saurastra 1540?-(Sm) An AVŚ Saṃhitā Ms. (ŚPP) that belonged to Sadāśaṅkar of Junāgadh (Saurastra). 1594-(R) An AVŚ Saṃhitā Ms. (ŚPP) that belonged to Sundarjī Durgaśaṅkar of Junāgadh. 1645-(J) An AVŚ Padapāṭha Ms. (ŚPP) that belonged to Vaidika Sundarjī Durgāśaṅkar of Junāgaḍh. 1827-(Bh) An AVŚ Saṃhitā Ms. (ed. ŚPP), returned to its owner Gaṇpatrām Veṇilāl Ojhā at Bhāvnagar. (Ant) An undated Antyeṣṭi Ms. from Ahmedabad (ed. Bloomfield), belonging to a Paṃcolī Kr̥ṣṇa. S. Gujarat 1681-(Bh) A Kauśikasūtra Ms. (ed. Bloomfield). The Ms. belonged to Dave Narabherāma at Bharūch (Broach). 1638-(Bü) Atharvaveda Saṃhitā Ms. (Bühler, Gujarat Cat.). The Ms. belonged to Dave Narabherāma at Bharūch (Broach).

ŚPP could make extensive use of it, see his ed. p. 15, 24. See his AV ed., p. 8, 14. 74 S. Bahulkar tells me (Cambridge, April 2010) that most of these 50-60 families have left the area by now, including a Gore who is now at Vaidika Samshodhana Mandala, Pune. For details see his forthcoming discussion. 72 73

362  Ù

The Vedas in Indian Culture and History

Southern Maharashtra 1651-(K) A Kauśikasūtra Ms. (ed. Bloomfield). The Ms. was discovered at Māhulī in S. Maharastra in c. 1880. 1798-(Dc) An AVŚ Saṃhitā Ms., apparently from the Māhulī school of Atharvavedins in Maharashtra. 1819-(F) An Atharvaveda Prātiśākhya Ms. (ed. Deshpande), originally from Māhulī near Sātarā. 1828-(11) A Kauśikasūtra Ms. with Keśava-Paddhati (ed. Bloomfield), belonged to Venkaṇ Bhatji Gore = Venku Dāji, at Sānglī (Mah.), see 1895-(K), apparently of the same Gore family as in MS. 1828-(ga). See also Ms. sā in Kauśikasūtra Ms. with Keśava-Paddhati, ed. Limaye et al. (no colophon). 1895-(K) AV reciter (ŚPP, before 1895) Keśav Bhaṭ bin Dāji Bhaṭ at Māhulī, who knew all of AV, except the inauspicious book 18; his text included the “corrections” made by Gaṇeś Bhat Dādā (see ŚPP, ed. p. 15). He also knew the Padapāṭha. 1895-(V) AV reciter (ŚPP, before 1895) Veṅkan Bhaṭjī or Veṅkan Dājī, at Sānglī. He was born around 1820. He used the AV in his daily ritual. His text included the “corrections” made by Gaṇeś Bhat Dādā. He also knew the Padapāṭha. (Note ŚPP, AV ed., p. 15 on accent marks added from the svaraśāstra). Northernmost Madhya Pradesh 1700-(I) An AV-Prātiśākhya MS (Deshpande) belonged to Bāḷaśāstri Garde from Gwalior. 1828-(ga) A Kauśikasūtra Ms. with Keśava-Paddhati. This Ms. belongs to the Gore family of Gwalior, but it was written by the Nāgara Brahmin Vedamūrti Gaṇeśa Bhaṭṭa Dādā Gore at Śrīkṣetra in Kāśī (Benares) in 1828 CE, for the support of his Atharvaveda–caraṇa. Benares 1879-(Bü) An undated modern Ms. of the Kauśikasūtra (ed. Bloomfield), was given to a Pañcolī Rāmakr̥ṣṇa, at Benares. 1837-(Vā) A Kauśikasūtra Ms. with Keśava-Paddhati (ed. Limaye et al.), (photocopy) belonged to V.R. Ratate, Benares. 1700-(J) An AV-Prātiśākhya Ms. (Deshpande). The Ms. (photocopy) belonged to the late Pt. Narayan Shastri Ratate, of Durga Ghat, Benares. 1810-(Daś) A Daśakarma Ms. (ed. Bloomfield), given to Pañcōli Rāmakr̥ṣna.

A Prosopography of the Śaunakīya Atharvaveda Families Introduction

Ù  363

20th century In the mid-Fifties S. Raghavan toured most parts of India for the Sanskrit Commission.75 Based on what he saw and heard then, he published a short report about then current Vedic traditions.76 Regarding AV tradition in Gujarat he says that “Atharvaveda Sva-śakhins still exist in Karnali, Lunavada and Bhavanagar,”77 and at Sinor he “met Pandit Vasudewa Ramaniklal Pancoli of Karnali or Kaṇvālaya,” who could recite the AV as his svaśākhā. There were Vedapāṭhaśālās then at Bhavanagar, Dvārakā and Baroda. He also heard of AV tradition in Benares: there were two Atharvedins then, Bhavani Sankar Dvivedin and his son Mohanlal Dvivedin. The AV was also recited “on ceremonial occasions by a non-Atharvavedīya Mahārāṣṭrian Pandit, Rama Bhatta Atata and his pupils.” The recordings made by John Levy between 1957 and 1966 indicate that AV tradition survived at Jamnagar in northwest Saurastra and at several other places in northern Saurastra, though J.F. Staal says “until recently.” Samples (AV 19.9.1-7 and GB 1.5.7) have been published, as soundtracks, by Levy and Staal on two LP records, The Four Vedas.78 However, Shrikant Bahulkar tells me (April 2010) that this tradition has by now disappeared: there was a Pāṭhaśālā run by Vāsu Pañcolī in 1976, but already by 1983 students only recited with the help of Mss. Likewise, the AV that is included in the (expensive) Veda recordings now stored at the Royal Library at Copenhagen, recited by Vāsudev Pañcolī at Pune in 1987, depended on Mss. In March 1974, I met with a rather conservative representative of the Benares AV tradition at Banaras Hindu University, who taught the AV in a Pāṭhaśālā situated on the campus; this also had the three Vedic sacrificial altars (which he did not allow to be photographed). Again, S. Bahulkar informs me that this tradition has by now disappeared.79 75 See Report of the Sanskrit Commission 1956-1957. Manager, Govt. of India Press/Manager of Publications, Delhi 1958. 76 Raghavan 1962. On Gujarat see p. 16 sq; on Benares p. 19. This agrees with what the Skt. Comm. says on p. 210 of its report. It adds however that there were a few Nagara Vaidika families who preserve the recitation of the AV as their sva-śākhā. About Benares it merely says that all Vedas and their Śākhās are maintained. 77 Of which All India Radio then made tape recordings. When I wrote to AIR in the late Sixties they could not find them. 78 The Four Vedas. Recordings and notes by John Levy and J.F. Staal. Asch Mankind AHM 4126 (2 LPs) [Smithsonian Folkways F 04126] See the accompanying booklet p. 11. The AV recordings are on record one, side II, bands 8-9. 79 See his forthcoming discussion.

364  Ù

The Vedas in Indian Culture and History

In the Summer of 1992, I could extensively observe Veda teaching at the Veda school of Kapileshwarapuram, situated a few miles away from the ocean on the Gautamī  branch of the Godāvarī  river. All Vedas were taught, upon the instigation of the then Śaṅkarācārya of Kanchi; this expressively included the AV.80 The students of this school would go on, after graduation, to recite the Vedas at various South Indian temples. I observed one such case (Taittirīya YV) at the famous Tirupati temple in S. Andhra already in October 1973, where the recitation was broadcast outside by loudspeaker.81 In sum, over last few decades, the recitation of the AV has deteriorated and disappeared almost everywhere82 where it still was surviving the Sixties. But recently, there is evidence of some revival, as I could observe at the Pāṭhaśālā at Baripada in Orissa in 1983, and at the Kapileshwaram Pāṭhaśālā in 1992. The same is true elsewhere in the South, I was informed then, where the AV had been absent for about a millennium. This small but very important oral tradition thus may survive well into this century–and we must hope, beyond that. § 3. Gujarat Brahmin clans As is obvious from the colophons reported above, the Śaunakīya Atharvaveda was mainly transmitted by the Pañcolī families of Gujarat, especially so in Saurastra. They belong to the ĀbhyantaraNāgara group, while a second line is that of the Pañcolī Vr̥ddhaNāgaras. The Pañcolī Ābhyantara-Nāgara are represented from 1431 CE onward, though in this early case still without naming individuals. The Nāgara Brahmins are one of the major Brahmin groups of Gujarat, that were, however, in the late 19th century divided into: Vaḍnagarā, Visalnagarā, Sāṭhodarā, Praśnorā, Kr̥ṣṇorā, Citroḍā (Yajnik, JBBRAS 10, 1871-74: 93-119). The Ābhyantaranāgara and Vr̥ddhanāgara are not mentioned there. (1) Pañcolī (Ābhyantara-Nāgara) The known Pañcōli lineage begins with the Ms. 1431-R of the Gujarat Minister Mohan Singh (Muhaṇasiṃha), perhaps preceded by AV Par U and V (dated 1414, 1474 CE), and soon followed 80 See my Kapileshwar report: “English commentary on Vedic chanting, video cassettes 1 & 2.” Rahway NJ: International Foundation for Vedic Education (and Cambridge, MA) Febr. 1993 [available at the Library of Congress]. 81 I was of course not allowed to enter the temple, but my Buddhist Newar photographer, Balaram Chitrakar, did so and positioned himself next to the reciter. 82 For details see S. Bahulkar’s forthcoming discussion.

A Prosopography of the Śaunakīya Atharvaveda Families Introduction

Ù  365

by the first surviving AV Mss. (of 1465, 1477, 1484 CE). Several Mss. give the same or very similar names, as listed below from Mss. dated 1584-1673 CE. The oldest known ancestor, Viśrupa, would have lived close to the date of the minister Mohan Singh. Towards the end of the known linage, Nārāyaṇa apparently had 3 sons, Alavesara, Śiva and Bācyāṃ, as well as his son Vāchā and grandson Bhavānīdāsa. (Muhanasiṃha 1431) (Ms. 1584-R)

(Ms.1594-R)

(Ms.1663-P)

(Ms.1673-B)

Viśrupa (+Viśvarūpa?)

Viśrupa

Viśrupa

Viśrupa

Vasīṣṭa

Vasīṣṭa

Vasīṣṭa

Lahūā

Lahūā

Lahūā

Vāsaṇa

Vāsaṇa

Vāsaṇḁ̄

Nārāyaṇa

Nārāyaṇa

Nārāyaṇa

Bacyāṃ 1663/73

Bācyāṃ

Bācyāṃ

Vāchā 1594

Vāchā 1663/73

Vāchā

| Vasīṣṭha | Lahuā | Vāsaṇa | Nārāyaṇa | Alavesara 1584, Śiva 1594

Bhavāṃnīdāsa Bhavaṃnīdāsa 1663, 1673, 1681 (cf.1678)

The Vedas in Indian Culture and History

366  Ù

(2) Another prominent Pañcolī lineage at Patan (and temporarily at Agra) starts around 1475 CE with Gaṇapati and ends with Vrajabhūṣaṇa (1696 CE). Gaṇapati (1678, 1681) Gaṅgādhara (1678, 1681) (cf. 1613) Kahana (1678, 1681) Madhachūda (1678, 1681) Devajī (1678, 1681) Udhava (1678, 1681) Anantajī (1645)

Somajī (1645, 1678, 1681)

Vāmmana (= Govinda 1650)

|

1681, 1878

“trivedin” 1696

|

|

|

Br̥haspati 1660 ?

?

Vrajabhūṣaṇa 1696

Cintāmaṇī (1645)

| Balarāma 1681 | Dhaneśvara 1691 The mentioning of Bhavanīdāsa and his father Vāchā is due only to the fact that Bhavanīdāsa (“Tragavatī”) was the scribe for a manuscript written for this lineage, see Ms. 1681 and the postcolophon note in Ms. 1678, 1681. Both Pañcolī lineages were living at Patan at the time.

A Prosopography of the Śaunakīya Atharvaveda Families Introduction

Ù  367

(3) Pañcolī Vr̥ddhanāgara Another lineage, different from the Ābhyantara-Nāgara, is that of the Vr̥ddhanāgara. They, too, do not appear in Yajnik’s late 19th century list of Gujarat Brahmins. Perhaps their name indicates that they claim an older, more elevated lineage than the other Nāgara Brahmins. They appear in the AV Padapāṭha Mss. of 1645-(J) AVŚ Padapāṭha Mss. (ŚPP) of 1645/50 CE; however their lineage begins about the same time as that of the ĀbhyantaraNāgaras but then disappears, only to reappear briefly in 1720 CE with the Pañcolī Brahmins Deneśvara and Bhavaṃnīśaṃkara, sons of Devarāja who is called a Vr̥ddhanagarā Brahmin, known as Gaṃdholakīya Nāgara. (4) Śukla The Śuklas are Udīcya Brahmins. As the name indicates, the Udīcya, who are the most numerous among the Gujarati Brahmins,83 originally came to Gujarat from the north.84 However, they represent another group of Gujarat Brahmins in the AV colophons quoted. They are divided into, three branches, Siddhpurā and Sihorā (both known as Sahasra Udīcya, and the Ṭoḷakīya Udīcya.85 They are attested in three Mss.: 1593-(Ch8) An Atharvaveda Padapāṭha Ms. (Berlin, Chambers 8); 1593 CE 1709-(Cp) saṃvat 1766 varṣe jyeṣṭha vadi 7 buddhe samāptaṃ || staṃbhatīrthavāsī Udīcya-ṭalakīā bhaṭṭa-Bakātmaja Ātmārāmeṇa likhitaṃ 1810-(Daśakarma) Siddhapur-vāstavyaṃ udīcyā-jñātīya ŚuklaSukharāma-ātmaja Hr̥dayarāma likhitaṃ The Śuklas apparently were Caturvedins, just like the Moḍha and Pañcolī: some were reciters of the Maitrāyaṇī Saṃhitā, for example at c. 1880 at Ahmedabad, as Schroeder reported.86 Unfortunately, no recording of these Maitrāyaṇīya Śuklas, exists.87 83 See Yajnik 1871-74. They have their own traditional history, the Udīcya Prakāśa. Cf. further Dalpatrām Dayabhāi, Jñātinibandh. 84 See Yajnik 1871-74, cf. Witzel 1986a, n. 121. 85 See Yajñik 1871-74: 109. 86 See Schroeder 1881-86: introduction. 87 Existing ones at Vaidika Samshodhana Mandala, Pune, are those of N. Maharasthrian Maitrāyaṇīyas living around Nandurbar on the Gujarat border;

The Vedas in Indian Culture and History

368  Ù

The earliest Veda recordings were made on wax rolls in the first decade of the 20th century at Madras; however, they just represent the common Veda schools of the R̥V, SV and YV only. As expected, there was no one at Madras who could recite the AV then. (5) Ojhā The occurrence of Ojhā, a clan name that is frequent with Maithila/Bengal Brahmins, is remarkable. However, Yajnik’s list (JBBRAS 10: 110 sq.) has several Brahmin groups stemming from the north, such as the Śrī Gauḍ, Gayāwāḷ, Kanojiā, Nāpal, Jāmbu, etc. S. Bahulkar tells me (Harvard, April 2010) that an Ojhā from Kathiavar recorded the SV at Tirupati in the Seventies. There are two Mss. that are of interest here. As early as 1594 CE, one Rudra Ojha wrote an AV Saṃhitā manuscript at Anahalapurapattana (Patan) in 1594/5 CE, see Ms. 1594-(R). Second, 1827-(Bh), an AVŚ Saṃhitā Ms., was copied or owned by the Brahmin Bohora Giridhara, living at Rājanagara (Rājpur?), a part of the Revā (R. Narmadā) area. He was the son of an Kelā Ojhā known as Nāndimukha, who however lived at Surapura, a part of Jodhpur (Rājasthān) in Marusthalī (Thar Desert). It seems that Giridhara had moved south to the Narmada area. (6) Dāve The Dāve (Dave) clan, too, is not listed in Yajnik, JBBRAS 10. However, it continues to exist in Saurastra.88 It appears in two Mss.: at 1623/4 in the Ms. 1624-(15/bha2) of the Kauśikasūtra with Keśava’s Paddhati, written by Gaṇeśa Dāve, a student of the AV, at Yahyanagara-Saraṃgapura. A few decades later, Ms. 1684-(P), a Kauśikasūtra Ms., was copied by Devakr̥ṣṇa, son of Jayakr̥ṣṇa Dāve, at Bhr̥gukṣetra (Broach), in the late winter of 1684 CE.

they preserve some features of older pronunciation, such as -s s- instead of -ḥ s-; cf. the introduction to the edition of MS by Satavalekar (who used Mss. and recitation of Nasik Brahmins). Cf. also the presentation by Madhav Deshpande, “Predicament of the Maitrāyaṇi community in Maharashtra: Migration, Acculturation and Identity Crises,” at the 13th World Sanskrit Conference, Edinburgh, 10th-14th July, 2006. 88 See Witzel 1989: §6.4: letter no. 14 by Kanhaiyālāl Bhāīśaṅkar Dave of Patan (N. Gujarat) in the book by Y.K. Deshpande and R.G. Śāstrī on the Caraka Brahmins of Maharastra (on which see Witzel 1981 and 1982).

A Prosopography of the Śaunakīya Atharvaveda Families Introduction

Ù  369

§ 4. Historical table The combined data found in the early Gujarat Mss. tradition can be tabulated as follows, according to the lineages appearing in the colophons. Dates are approximate (such as ~1400), unless given in exact figures (such as 1481). A generation has been calculated at 25 years, due to the early marriages common then. Dates given are those of the Mss. that these persons appear in, inserted between parentheses, though their lifetime may be earlier. Thus, Viśrūpa did not live in 1584 or 1593 CE but nearly two hundred years earlier. Actual dates have been given without parentheses.

The Vedas in Indian Culture and History

370  Ù

TIME

GUJARAT ATHARVAVEDA FAMILIES

Ābhyan-tara(main Nāgara Pañcolī, branch) at Pāṭān

Ābhyantara-Nāgara Pañcolī at Patan  

Vr̥ddha-Nāgara Pañcolī, at Vīrama-grāma

~ 1400 CE

Lakṣmīdhara (1484)

~ 1425

Mā[—]na 1484 ========

(Mohan Singh, minister1431) ~ 1450 +Viśrūpa (1584, 1594,1663, 1673)

Ābhy.-Nāgara Pañcakalpin at Stambhatīrtha (Khambat)

~ 1475 Vasiṣṭha (1584, 1594, 1663, 1673)

—————— Gaṇapati (1678, 1681)

~ 1500 Lahūā (1584, 1594, 1663, 1673)

Gaṅgādhara (1678, 1681) (cf. 1613)

~ 1525 Vāsaṇā(1584, 1594,1663)

Kahana (1678, 1681)

~ 1550 Nārāyaṇa (1584, ← Madhachūda 1594,1593??, (Nārāyaṇa) (1678, 1681) 1663, 1673) (1663) ~ 1575

Murāri

Alavesara 1584; Bācyām — (1663, brother: 1673) Śiva 1594 ~ 1600

~ 1625

Devajī (1678, 1681); [hisson: Lomajī, 1641]

|

Udhava (1678, 1681)

============

[Lomajī 1641]; Somajī ←(Anantajī (1645, 1678, 1681); Trivedin)(1696) Vāmmana (1645; 1669? = Govinda 1650), Anantajī (1645) → At Agra 1645

Nr̥siṃha, at Stambhatīrtha (Khambat) 1619 =========

A Prosopography of the Śaunakīya Atharvaveda Families Introduction

Abhy.-Nāgara Pañcakalpin at Maphalī

Other Pancolī

Nāgara at Vācala

Udīcya Ṭoḷakīya —————— and others

Ù  371

Dāve & Gore Clans

Ūṃdākana, at Bālīṃba (1465) —————— (1421 SV Ms., Jammu) Lāpā, son of Ūṃdākana (1465) Vācīyā Duve (1477) —————— Kr̥ṣṇa, son of Lapā (1465)

1476 Yavarāja

1517 Padhāka, Nāgara

Rāmacandra (1661) ——————

Scribe Aṃkāka, a Kanya-kubja-vāstavyajñātīya (1517)

Rāala (1614) Kaṇodiyā, a SahasraUdīcya yajñika, at Śrīnara & Nāgara Govinda —————— Rudra Ojhā, copyist at Patan, 1594

Bhaṭṭa Gaṅgādhara Mādhava & at Benares 1613, cf. Viśvanātha, 1651 at Vācāla ————≤—1614 Trivedin Śiva (1651); Śivarāja

Nāgaji/˚t Bhavadeva, son of Rāmacandra (1661) —————— Īśvara; ←Govindajī (1627) 1677; cf. 1700 ——————-

Gaṇeśa Dāve at YahyanagaraSaraṅgapura 1624 (→1684, 1826/7)

Ganeśa Dāve 1624

Jayakr̥ṣṇa Dāve (1684)

372  Ù

The Vedas in Indian Culture and History Somajī (1645, 1650), ←(Br̥haspati → Cintāmaṇī (1645, 1660, at Patan; 1650), Soma, son of Somajī) → Ānantajī (1645, 1650) at Patan 1645/50

~ 1650

Vāchā Anantajī → 1663, 1673 Cintāmanī Govinda | (1670?), Udhava (= 1681, 1678)

~ 1675

Bhavānīdāsa 1663, 1673, 1681

← Ananta āhitāgni, cāturmāsyayājin, tripāṭḥī (1696) —————— Nīlakanta Sadāśiva Vāṃmana 1645; (cf. V. at 1645)

Somajī 1678 (son of Udhava); ← son: Tragavatī?/ ← = Bhavānī-dāsa> (1681) see 1663/1673>

~ 1700

Govindajī, Balarāma 1681

Vrajabhūṣaṇa, son of Anantajī 1696 —————— Pañcolī Deneśvara & Bhavaṃnīśaṃkara, son of Devarāja, a Gaṃdholakīya Vr̥ddhanāgara, at Rājanagrāma 1720

~ 1725

Dhaneśvara 1691 ~ 1750

(cf. Bhaṭṭa Baka 1709)

A Prosopography of the Śaunakīya Atharvaveda Families Introduction Devakr̥ṣna 1684, at Bhr̥gukṣetra ——————

Jayakr̥ṣṇa (son of Śiva) at Maphalī 1651; bro.s: Gaṇapati, Gaṅgādhara, Rāhāgavajī Gadhādhara, Vidyādhara

Devakr̥ṣṇa, son of Śiva, at Muphalīpura, 1658

Bhavaveda, son of Nāgajit at Kāṇvālaya 1661 = 1661 Bhavadeva (son of Nāgajī) —————— Mādhavajī, son of Jayakr̥ṣṇa; Rāmajī ‘Susura’; Sāṃmajī bro., at Lūṇāvaḍā c. 1675 —————— Nr̥siṃha, Kr̥ṣṇadeva, Śivadeva, Śaṃkara, (Sūryadatta?), Kr̥ṣṇarāma 1691; cf. 1658? ——————

Ābhy.-Nāgara Janīveja(?) at Giripura, King Gaṅgādāsa 1688 —————— 1670, 1678 for King Anūpasiṃha

Nīlakaṇtha, son of Īśvara ~ 1700

Bhaṭṭa Baka (1709)

Vīreśvara Dave (1717)

Ātmarāma at Sthambhtīrtha 1709

Pītaṃbara & Maheśvara Dave 1717

——————

—————-

Gore clan

Ù  373

374  Ù

~ 1775

~ 1800

~ 1825

~ 1850

~ 1875

~ 1900 ~ 1925 ~ 1950

~ 1975

~ 2000

The Vedas in Indian Culture and History

A Prosopography of the Śaunakīya Atharvaveda Families Introduction Sukharāma, Udīcya Śukla at Siddhapura Govinda, son Bāpu Bhaṭṭa, son of of Tryambaka Daivajña (1788) Vedamūrti Hr̥daya-rāma, & Rāmakr̥ṣṇa Pañcoḷi,. Ganeśa Bhaṭṭa Dādā Gore at at (sons of Sukharāma) at Benares, for Benares 1828; Pañcōli Rāmakr̥ṣna (and Śr̥ṅgeri!) 1810 later: ========= Gore family at Raghunātha, son: Gwalior Mārtāṇḍa Bhaṭṭa 1819 ========= Īśvara-candra Mukho-padhyāya, Benares (1818) ========= Paṭhuvar-dhana Biṭhala Rāmaka 1824 —————— Bohora Girishara, at Rājana-nagara; son of Kela Ojhā Nāndimukha, at at Jodhpur 1827 ========= Viṣṇunā-rāyaṇa Devadhara 1863 Pañcolī Rāmakr̥ṣna, at Benares c. 1879; cf. 1756 Bālakr̥ṣṇa at Bhavaṇā-Maṭha

Atharva-vedins at Karnali, Lunavada, Bhavanagar; Bhavani Sankar Dvivedin & son Mohanlal Dvivedin at Benares (V.R. Ratate, Narayan Shastri Ratate, Benares) New AV tradition in the South

Venkaṇ Bhaṭjī Gore ŚPP)

Ù  375

The Vedas in Indian Culture and History

376  Ù

Excursus As a small excursus, it may be underlined that the information gleaned from colophons is not merely useful for a prosographic study as undertaken here. Even this small sample teaches us something about the religious history of Gujarat. One can notice that Kr̥ṣṇa-related names suddenly appear around 1600 CE and are then sustained, though the worship of this deity is many centuries older, as evidenced for example in Jayadeva’s famous Gītagovinda of the 12th century.89 In the AV Mss. quoted, evidence begins with Ms. 1619-(N.5) of 1619 CE mentioning Murāri, Rāmacandra, Govinda, Mādhava, followed by data in 1651-(K), 1675-(4.M), 1681-(Ba) Kahana (if this indeed stands for Kaṇha, Kr̥ṣṇa), 1684-(P), etc. Note however already 1584-(R) with the post-colophon kiṃcit kr̥ṣṇārpaṇam astu. Without going deeper into this kind of investigation, it may however be pointed out that the same development is also seen in contemporary Nepal, where the newly independent kingdom of Pātan (south of Kathmandu, 1621 CE) established a strong Krishnaite tradition, beginning with its first king, Siddhinarasiṃha Malla. He built an uncharacteristic stone temple for Kr̥ṣṇa at the new Patan Durbar square, and did not do so in Nepalese pagoda style but in purely North Indian style. § 5. Conclusions Summing up this detailed Ms. study, it is very important to note that the oral tradition of the AV between c. 1400-1700 is represented only by a two well-attested families in Saurastra, belonging to the Pañcolī clan.90 All other known AV reciters emigrated from there to Benares around 1550 as well as to the neighboring areas of Rajasthan, and from Benares91 maybe also to Nepal around 1700;92 finally some moved to Maharastra around 1730. This underlines the narrow, unstable AV transmission even durFor the centuries preceding, see for example, Dimock 1999: 5f. The early lineage of the Vr̥ddhanāgara (c. 1400) is not heard of again. 91 From where it has recently spread again, also to South India (see above, and my Kapileshwara report). 92 If the Newari Ms. indeed indicates transfer by a Veda reciter (see above), not just a transcript from an existing Ms. The unusual accentuation of this Padapāṭha speaks against a simple copy. —The Vedavr̥k ṣa (Journal of Vedic Studies, 1931) lists the existence of the Paippalāda school in Nepal, but no trace of it has been found, except for a copy Bloomfield-Garbe’s facsimile in a Kathmandu Brahmin’s house, see Witzel 2001. 89

90

A Prosopography of the Śaunakīya Atharvaveda Families Introduction

Ù  377

ing the late Middle Ages, seen also in other numerically small schools (śākhās).93 The Saurastra area definitely was a very small basis for the transmission of a Veda tradition. Actual loss of texts94 and even more so, mistakes could easily creep into the narrow written and weak oral transmission. Such mistakes were then perpetuated, like mutations in biology, as there were so few means to countercheck the text by other reciters and Mss.95 This kind of change has indeed happened on a major scale after the AV had been transferred to Maharastra, as reported by ŚPP about Gaṇeś Din Bhaṭṭ’s Sringeri expedition. By 1890, due to the mistaken changes made by him in Maharastrian AV recitation, there were two schools of reciters in Maharastra, the old (Gujarat) style one and the new (wrong) one started by Gaṇeś Din Bhaṭṭ. There is no wonder that in traditions that have a small basis of tradition easy proliferation of mistakes can occur. This is indeed noticed in the Maitrāyaṇī Saṃhitā, for example, in the medieval writing mistake cch > ñś (ucchiṣṭa > uñśiṣṭa) that has become normal for MS (see Lubotsky 1983). Small schools like the AV Śaunaka (Gujarat), AV Paippalāda (Orissa), Kaṭha (Kashmir), Vādhūla (Kerala), Āgniveśya (Tanjore area), or other geographically isolated ones like the Kāṇva who are widespread, but always in small and historically changing numbers, here and there (“sarvatra”) in the subcontinent.96 Regarding the AV, immediate conclusions, therefore, are as follows. – We cannot take very seriously the obvious mistakes in the Śaunaka AV tradition (mostly based on Mss. only). – We must pay close attention to writing and pronunciation peculiarities of medieval Gujarat in order to restore corrupt passages of AVŚ. (The situation of the Paippalāda AV is similar, but it involves both Orissa and Kashmir as well as their Gujarat archetype).97 93 See Witzel 1986a for small śākhās, including percentages for various Vedas and their schools. 94 Such as the lost Brāhmaṇa of the Śaunaka AV, which was substituted by the Gopatha Br. that originated in the Paippalāda school. For a discussion see now Griffiths 2007. 95 As carried out, for example, in the yearly Veda recitations held by the Nambudiri Brahmins of Kerala, where chanting occurs in chorus (ghoṣa). 96 For example, the settling of Kāṇvas (and Atharvavedins) in the Tanjore area under the Coḷa kings around 1000 CE; nothing is heard of them subsequently. An exception of the piecemeal settlement of the Kāṇvas is their strong representation in Orissa, see Witzel 1985a. 97 Witzel 1985b; see now Griffiths 2007 and Schmiedchen 2007.

378  Ù

The Vedas in Indian Culture and History

– We cannot not count much on the weak, narrowly local medieval AV tradition in establishing apparently deviating Vedic forms in the AV. This applies even more so for the AV’s use in linguistic studies (such as IE accent represented in the AV). Instead, individual cases must first be corrected according to manuscripts, palaeography, AV recitation, the general picture of Vedic grammar and that of related Vedic texts. In fact, we must first create a new edition of the Śaunaka text: the manuscripts in Roth-Whitney’s edition are badly reported, if at all,98 and they must be correlated with those extensively documented in ŚPP’s edition,99 as well as with the two AV Prātiśākhyas. Many other AV Mss. have not yet been used, such as the Bhaktapur one, and a stemma of AVŚ Mss. is not yet in sight.100 In sum, close study of Mss. and colophons can teach us a lot, not just about the Veda, its traditions and transmitters, but about medieval India in general. More attention is required. References Bendall, Cecil. 1883. Catalogue of the Buddhist Manuscripts in the University Library Cambridge. Cambridge: Cambridge University Press. Bhattacharyya, D.M. 1964 (Vol. 1); 1970 (Vol. 2). Paippalāda Saṃhitā of the Atharvaveda. First Kāṇḍa. Calcutta: Calcutta Sanskrit College. Bloomfield, Maurice. (ed.). 1889. The Kauśika Sūtra of the Atharva Veda. JAOS 14. Bolling, George Melville and Julius von Negelein (eds.). 1976. The Pariśiṣṭas of the Atharvaveda. Benares: Chaukhambha Orientalia (first edition Leipzig: Harrassowitz, 1909-1910). Brough, John. 1996. Collected Papers. Ed. by Minoru Hara and J.C. Wright. London: SOAS. Chaube, V.B. 2001. Vādhūla-Anavākhyāna. Hoshiyārpur: Kātyāyana Vaidika Sāhitya Prakāśana. ———. 2009. Āśvalāyanaśākhīyā R̥gvedasaṃhitā: padapāṭhaviśadabhūmikā-anukramaṇīdvaya saṃvalitā. New Delhi: Indirā Gandhī Rāṣṭrīya Kalā Kendra. 98 Only in Whitney’s word index to AVŚ published in JAOS 12, and in the translation, with such vague mentioning such as “some Mss.,” etc. 99 They are frequently misprinted in the Hoshiarpur edition of the AV. 100 I briefly tried in 1982/4, but could not progress much beyond some subdivisions in the two branches of transmission, the Gujarat and the Maharastrian ones. Now, newly noticed Mss. such as those from Rajasthan and Nepal will have to be taken into account.

A Prosopography of the Śaunakīya Atharvaveda Families Introduction

Ù  379

Deshpande, Madhav. 1997. Śaunakīya Caturādhyāyikā. Harvard Oriental Series, 52. Cambridge, MA: Harvard University Press. ———. 2002. Recitational Permutations of the Śaunakīya Atharvaveda. Harvard Oriental Series, 61. Cambridge, MA: Harvard University Press. Dimock, Edward C. 1999. Caitanya Caritāmr̥ta of Kr̥ṣṇadāsa. Harvard Oriental Series 56. Cambridge, MA–London: Harvard University Press. Diskalkar, D.B. 1959-60. “Atharvaveda and Epigraphi.” Journal of the Asiatic Society of Bombay, New Series 34/35, 75-83. Felber, Erwin. 1912. Die indische Musik der vedischen und der klassischen Zeit. Studie zur Geschichte der Rezitation. Sitzungsberichte der Kais. Akad. der Wissenschaften, Phil.-Hist. Kl. 170, 7. Abh. Garbe, Richard. 1899. Verzeichnis der indischen Handschriften der Kgl. Universitätsbibliothek (Zuwachs der Jahre 1865-1899). Tübingen. Gray, J.E.B. 1959a. “An Analysis of R̥ g vedic Recitation.” BSOAS 22, 86-94. ———. 1959b. “Analysis of Namdudiri R̥ g vedic Recitation and the Nature of the Vedic Accent.” BSOAS 22, 499-530. Griffiths, Arlo. 2007. “The Ancillary Literature of the Paippalāda School.” In: Griffiths and Schiedschen 2007, 141-194. Griffiths, Arlo and Annette Schmiedchen (eds.), The Atharvaveda and its Paippalādaśākhā. Aachen: Shaker. Haug, Martin. 1871. Brahma und die brahmanen. München: Verlag der Kgl. Akademie. ———. 1874. Über das Wesen und den Werth des vedischen Accents. München: Verlag der Kgl. Akademie. Ikari, Yasuke. 1996. “Vādhūla Śrautasūtra I.1-4 (Agnyādheya, Punarādheya). A New Critical Edition of the Vādhūla Śrautasūtra, I.” Zinbun (Journal of Institute for Research in Humanities) 30, 1-129. Kirste, Johann. 1908. Die altindischen Platten. Sitzungsberichte der Kais. Akad. der Wissenschaften in Wien, Phil.-Hist. Kl. 160, 1. Abh. Limaye, Viṣṇu Prabhākara et al. (eds.). 1982. Keśava’s Kauśikapaddhati on the Kauśikasūtra of the Atharvaveda. Shri Balmukund Sanskrit Mahavidyalaya Research, Series 1. Pune: Tilak Maharashtra Vidyapith. Lokesh, Chandra. (ed.). 1981. Pañcaviṃśa-Brāhmaṇa. Śatapiṭaka Series 270. New Delhi: Sharada Rani. Lubotsky, Alexander. 1983. “On the external sandhis of the Maitrāyaṇi Saṃhitā.” IIJ 25, No. 3, 167-179. Müller, F. Max. (ed.). 1890-92. R̥ gveda-saṃhitā. 2nd edition. 4 Vols.

380  Ù

The Vedas in Indian Culture and History

Oxford: Oxford University Press. Oppert, Gustav. 1880. Lists of Sanskrit Manuscripts in Private Libraries of Southern India. Vol. 1. Madras: Government Press. Petech, Luciano. 1984. Mediaeval History of Nepal (c. 750-1482). 2nd, thoroughly rev. ed. Rome: Istituto italiano per il Medio ed Estremo Oriente. Peterson, Peter. 1887. A Third Report of Operations in Search of Manuscripts in the Bombay Circle. Royal Asiatic Society. Extra number (45) of the JBBRAS. Bombay: Society’s Library. Raghavan, Venkatarama. 1962. The present position of Vedic recitation and Vedic Sakhas. Kumbhakonam: The Veda Dharma Paripalana Sabha. Regmi, D.R. 1965. Medieval Nepal. 4 Vols. Calcutta: Mukhopadhyay. Rosenfield, Susan. 2004. Kaṭha Brāhmaṇa Fragments: A Critical Edition, Translation and Study. PhD Thesis, Harvard. Sarma, K.V. 1971. “The Manuscript Collection of the Jaḍe Family of Varanasi and the Literary Output of the Jaḍe Authors.” Vishveshvaranand Indological Journal 9, 347-356. Satavalekar, S.D. (ed.). n.d. Maitrāyaṇīyasaṃhitā. Pāraḍi: Svādhyāya-Maṇḍala. Schmiedchen, Annette. 2007. “Epigraphical evidence for the history of Atharvaveda Brahmins.” In: Griffiths and Schmiedchen 2007, 355-384. Schroeder, Leopold von. 1879. “Ueber die Mâtrâyani Samhitâ, ihr Alter, ihr Verhältniss zu den verwandten Çâkhâ’s, ihre sprachliche und historische Bedeutung.” ZDMG 33, 177207. ———. 1881-86. Mâitrâyaṇî Saṃhitâ. Die Saṃhitâ der MāitrāyaṇīyaÇākhâ. 4 Vols. Leipzig: Verl. der Dt. Morgenländischen Ges. Slaje, Walter. 2010. “Sāyaṇa oder Mādhava? Verfasserschaft und Reihenfolge der Saṃhitā-Kommentare aus Vijayanagara.” ZDMG 160, 383-414. SPP = Pandurang Pandit, Shankar. (ed.). 1894-98. AtharvavedaSaṃhitā with the Commentary of Sāyaṇācārya. 4 Vols. Bombay: Department of Public Instruction. Staal, J. Frits. 1961. Nambudiri Veda Recitation. The Hague: Mouton. ———. 1983. Agni. 2 Vols. Berkeley, CA: Asian Humanities Press. ———. 1986. “The Fidelity of Oral Tradition and the Origins of Science.” Mededelingen KNAW, Afd. Letterkunde d. 49, No. 8, 251-288. Surya, Kanta. (ed.). 1968. Atharvaprātiśākhya. Delhi: Mehar Chand Lachman Das. Turner, Ralph L. 1969. A Comparative Dictionary of the Indo-Aryan Languages. London, New York: Oxford University Press. Whitney, William Dwight. 1905. Atharva-Veda Saṁhitā. With a Critical and Exegetical Commentary. Rev. and ed. Charles Rockwell

A Prosopography of the Śaunakīya Atharvaveda Families Introduction

Ù  381

Lanman. 2 Vols. Harvard Oriental Series, 7-8. Cambridge, MA: Harvard University Press. Witzel, Michael. 1974. “On Some Unknown Systems of Marking the Vedic Accents.” Vishvabandhu Commemoration Volume. Vishveshvaranand Indological Journal 12, 468-502. ———. 1975. “Eine fünfte Mitteilung über das Vādhūla-Sūtra.” Studien zur Indologie und Iranistik 1, 75-108. ———. 1976. “On the history and the present state of Vedic tradition in Nepal.” Vasudha 15, No. 12, 17-24; 35-39. ———. 1980. “Die Kaṭha-Śikṣā-Upaniṣad.” WZKS 24, 21-82. ———. 1981. “Materialien zu den vedischen Schulen: I Über die Caraka-Schule.” StII 7, 109-132. ———. 1982. “Materialien zu den vedischen Schulen: I Über die Caraka-Schule.” [continuation of Witzel 1981] StII 8-9, 171240. ———. 1985a. “Die mündliche Tradition der Paippaladins von Orissa.” Festgabe für K. Hoffmann, I = MSS 44, 259-287. ———. 1985b. “Die Atharvavedatradition und die PaippaladaSamhita.” ZDMG, Supplementband 6 (XXII. Deutscher Orientalistentag, Tübingen, March 1983), 256-271. ———. 1986a. “Regionale und überregionale Faktoren in der Entwicklung vedischer Brahmanengruppen in Mittelalter.” In: H. Kulke and D. Rothermund (eds.), Regionale Tradition in Südasien. Beträge zur Südasienforschung, 104. Stuttgart: F. Steiner Verlag Wiesbaden, 37-76. ———. 1986b. “On the Archetype of Patañjali’s Mahābhāṣya.” Indo-Iranian Journal 29, 249-259. ———. 1989. “Tracing the Vedic Dialects.” In: Colette Caillat (ed.), Dialects dans les litteratures indo-aryennes. Actes du Colloque International organise par UA 1058 sous les auspices du C.N.R.S avec le soutien du College de France, de la Fondation Hugot du College de France, de l’Universite de Paris III, du Ministre des Affaires Etrangeres, Paris (Fondation Hugot) 16-18 Septembre 1986. Paris: College de France, Institut de Civilisation Indienne, 97-264. ———. 1997. “Introduction.” In: M. Witzel (ed.), Inside the texts, Beyond the Texts. New Approaches to the Study of the Vedas. Harvard Oriental Series, Opera Minora, Vol. 2. Cambridge, MA: Harvard University Press. ———. 2001. “Medieval Veda Tradition as Reflected in Nepalese Manuscripts.” Journal of the Nepal Research Centre 12, 255-299. ———. 2004. Katha Āraṇyaka. Critical edition with a translation into German and an introduction. textkritische Edition mit Übersetzung und Kommentar. (Also PhD Thesis, Erlangen 1972; Teildruck, Erlangen–Kathmandu 1974). ———. (forthcoming). Prolegomena to the Atharvaveda.

382  Ù

The Vedas in Indian Culture and History

Yajnik, Javerilal Umiashankar. 1871-74. “Notices of Hindu Tribes and Castes of Gujarat.” JBBRAS 10, 93-119. Select Abbreviations

NGMPP JNRC JBBRAS

Nepalese German Manuscript Preservation Project Journal of the Nepal Research Centre Journal of the Bombay Branch of the Royal Asiatic Society

Introduction

Ù  383

V. The Continuing Life of the Veda

384  Ù

The Vedas in Indian Culture and History

Introduction Brides of the Buddha

Ù  385

Brides of the Buddha, or How Vedic Marital Customs Served Buddhist Ends K aren Muldoon-Hules The University of California at Los Angeles This paper focuses on an unusual use of Brahmanical marital imagery in two closely-related stories from the Avadānaśataka, a collection of Buddhist tales from Northern India usually dated circa the 2nd C.E.1 The two stories in question are the first and the sixth avadānas of the eighth varga of the Avadānaśataka, the avadānas of Suprabhā and Kāśisundarī. This varga or section contains ten stories dealing with women who attain arhatship, the highest achievement of a Buddhist disciple. Surprisingly, these ten women’s stories have been largely overlooked by scholars of Indian Buddhism, apart from a single article published by Léon Feer in 1883. At a point partway into the sixth (or seventysixth) avadāna, the text reads:2 yadā kāśisundarī dārikā krameṇa mahatī saṃvr̥ttā, tadā prātisīmaiḥ ṣaḍbhi rājabhī rājñe brahmadattasya dūtasaṃpreṣaṇaṃ kr̥tam | śrutam asmābhir yathā tava duhitā jāteti | tad arhasy asmākaṃ putrāṇām anyatarasmai anupradātum iti || tato rājā śokāgāraṃ praviśya kare kapolaṃ 1 Speyer proposed a date of ±100 C.E., while Sharma in 1985 suggested a date of late 1st century C.E. or early 2nd century C.E. Both based their dating, in part, on the supposed date of 3rd century C.E. for the Chinese translation (T. 200). However, based on a close analysis of the differences between the Sanskrit text edited by Speyer and the Chinese translation, Mitsuyo Demoto argued in her 1998 doctoral dissertation for a revised date for the Chinese translation of late 5th or early 6th century C.E. The significance of this in relation to the date of the Sanskrit AvŚ is unclear. For a summary of current research on the Chinese translation and an analysis of Sanskrit fragments of the AvŚ in the Schøyen Collection, see Demoto 2006: 207-244. 2 Translations from these two avadānas herein are mine, based on Speyer’s Sanskrit text. The only complete translation of the AvŚ in a Western language to

386  Ù

The Vedas in Indian Culture and History datvā cintāparo vyavasthitaś cintayati | yady ekasmai dāsyāmi apareṇa me saha virodho bhaviṣyatīti || kāśisundarī dārikā sarvālaṅkāravibhūṣitā pituḥ sakāśam upasaṃkrāntā | tayā pitā śokārto dr̥ṣṭaḥ pr̥ṣṭaś ca: tāta kimarthaṃ śokaḥ kriyata iti | pitrā asyā yathābhūtaṃ samākhyātam || tataḥ kāśisundarī pitaram uvāca | kriyatāṃ tāta prātisīmānāṃ rājñāṃ dūtasaṃpreṣaṇaṃ | saptame divase kāśisundarī dārikā svayaṃvaram avatariṣyati yena vo yat karaṇīyaṃ sa tat karotv iti || yāvat saptame divase ṣaṃ prātisīmā rājānas saṃnipatitāḥ | kāśisundary api ratham abhiruhya kāṣāyaṃ dhvajam ucchrāpya buddhapaṭaṃ hastena gr̥hītvā rājasabhāṃ gatvovāca | śr̥ṇvantu bhavantaḥ prātisīmā rājānaḥ | nāhaṃ bhavatāṃ rūpayauvanakulabhogaiśvaryaṃ tulayāmi api tu nāhaṃ kāmair arthinī | ya eṣa eva me bhagavān buddhaḥ paṭe likhitas tasyāhaṃ śrāvikā asya śāsane pravrajiṣyāmīti || (Avadānaśataka [Speyer] 1906-09 (cited after 1958 reprint): ii.32-33) In due time, when the girl Kāśisundarī had grown up, six neighboring kings sent envoys to King Brahmadatta. “We have heard that a daughter was born whom it would be fitting for you to present to one of our sons.” Then the king went into a mourning hall, 3 put his head in his hands and sat there depressed, thinking: “If I give her to any of them, there will be conflict with the rest.” The girl Kāśisundarī, magnificent in all her ornaments, approached her father. She saw that her father was troubled and asked, “Papa, what is making you sad?” Her father explained what had happened. Then Kāśisundarī said to her father, “Let messengers be sent to the neighboring kings to say that in seven days the girl Kāśisundarī will hold a ceremony of self-choice; whoever has something to do with regard to this, he should do it.” Then, in seven days the six neighboring kings assembled. Kāśisundarī, having mounted a chariot, unfurled an ochre banner. Taking in her hand a painting of the Buddha on cloth, she went to the assembly-hall of the kings and said: “Listen, honorable neighboring kings, I am not one who is considering your honors’ beauty, youth, and family and might. Moreover, I am not one who is pursuing passions. That very one who is painted on this cloth is for me, the Blessed One, the Buddha. I will go forth as a disciple in the teaching of that one.”

The story of Suprabhā, the first avadāna in this section (or the date is the French translation entitled Avadāna-çataka, cent legendes (bouddhiques) by Léon Feer. See Feer 1891: 260-263, 284-287, for his translation of these two stories. 3 Śokāgāra, according to Edgerton, could be a “grief-house, hall of lamentation ... boudoir” or even the rather delightful “sulking room.” (Edgerton 1953: II, 533a) Monier-Williams suggests that it was a room for women to weep in, yet the situation here clearly has a man using it. (Monier-Williams 1899: 1091a) The Tibetan for Śokāgāra, unfortunately, does not shed much light on the problem, as it is a literal translation: mya ngan gyi khang bu, meaning “a small hut [or room] of sorrow.” (cf. G. Schopen, pers. comm., 5/2005)

Introduction Brides of the Buddha

Ù  387

seventy-first of the whole collection) mirrors Kāśisundarī’s story quite closely. yāvad asau dārikā krameṇa mahatī saṃvr̥ttā tadā tasyā bahavo yācanakā āgacchanti rājaputrā amātyaputrāḥ śreṣṭhiputrāś ca | tair upadrūyamānaḥ pitā cāsyāḥ kare kapolaṃ dattvā cintāparo vyavasthitaḥ | yady ekasmai dāsyāmi anye me amitrā bhavisyantīti || yāvad asau dārikā pitaraṃ cintāparam avekṣyovāca | tāta kim asi cintāpara iti | tena so ’rtho vistareṇa samākhyātaḥ || dārikā kathayati | tāta na te śokaḥ kartavyaḥ svayam evāhaṃ saptame divase svayaṃvaram avatariṣyāmīti || tataḥ śreṣṭhī rājñāḥ prasenajito nivedya śrāvastyāṃ ghaṇṭāvaghoṣaṇaṃ kārayām āsa | saptame divase suprabhā dārikā svayaṃvaram avatariṣyati yena vo yat karaṇīyaṃ sa tat karotv iti || tataḥ saptame divase suprabhā dārikā rathābhirūḍhā kāṣāyaṃ dhvajam ucchrāpya buddhaṃ bhagavantaṃ citrapaṭe lekhayitvā abhiṣṭuvatī vīthīm avatīrṇā || sā tatra rājaputrair amātyaputraiḥ śreṣṭhiputraiś ca sotkaṇṭhodvīkṣyamāṇā vicitrābhiḥ kathābhiḥ saṃjapyovāca | sarvathāhaṃ na kena cid aṃśena bhavatāṃ paribhavaṃ karomi kevalaṃ tu nāhaṃ kāmenārthinī buddhaṃ śaraṇaṃ gatāsmi tasya sakāśe pravrajiṣyāmīti | tatas te nirbhartsitāḥ pratinivr̥ttāḥ || (Avadānaśataka [Speyer] 1906-09: ii.2-3) In due time when that girl had grown up, many suitors arrived: kings’ sons, ministers’ sons, and bankers’ sons. Besieged by them, her father put his head in his hands and sat there depressed, saying, “If I give her to any one of them, the others will become my enemies.” When that girl observed her father depressed, she said, “Papa, why are you depressed?” He explained the reason in detail. The girl said, “Papa, don’t be sad; I myself will undertake a self-choice ceremony in seven days.” Thereupon the merchant, having made it known to King Prasenajit, had the bell struck, [proclaiming], “In seven days, the girl Suprabhā will hold a ceremony of self-choice; whoever has something to do with regard to this, he should do it.” Then on the seventh day, the girl Suprabhā, having mounted a chariot, with an ochre banner unfurled and the Buddha, the Blessed One, painted on a cloth, alighted in the street, praising him. There, while the king’s sons, ministers’ sons and bankers’ sons stared ardently at her, she got their agreement with charming words, saying, “I do not mean to insult you in any way, sirs, but I am by no means one who seeks worldly desire. I have gone for refuge in the Buddha. I will go forth in his presence.” Thereupon, having been refused, they went away.

As can be seen from this excerpt, Suprabhā also proposes to hold a svayaṃvara to ease her father’s worries and then arrives in a ratha (”chariot” or “wagon”) bearing a picture of the Buddha and announces her decision to become a Buddhist nun. It would seem that here we have two women who, in entering the religious life, both seem to be presenting themselves, in effect, as brides of the Buddha. Each one does so at what is explicitly labeled a

388  Ù

The Vedas in Indian Culture and History

svayaṃvara. To my knowledge, this specific use of Brahmanical bridal motifs as a metaphor for Buddhist female renunciation occurs only in the Avadānaśataka and in a later version of the Suprabhā story that was included as the eighth avadāna in the Kalpadrumāvadānamālā (folios. 39-54) (Feer 1891: 263-265).4 The only tale that I know of in the classical literature that is remotely comparable to these two Buddhist stories is the complicated story of Gālava and Mādhavī in the fifth book of the Mahābhārata. Mādhavī, gifted with conveniently renewable virginity, is used by the young brahman Gālava as a bargaining chip in his fantastic bid to obtain 800 white horses with one black ear as a gift for his guru. After four unions with kings that produced four sons, Mādhavī, having returned to her father a virgin yet again, is given a svayaṃvara by him. She passes over a remarkable assortment of highly eligible suitors, both human and non-human, and chooses the forest as her bridegroom where she lives the rest of her life in chastity, practicing austerities. (Mbh 5.104-21, especially 112-118; Jamison 1996: 208-210.) Beyond this single story, I have not found any other evidence so far for the motif of choosing a religious life in a svayaṃvara in the classical literature. In these two Buddhist avadānas of Suprabhā and Kāśisundarī, the monk-redactors make clever use of Indian marital customs, as part of what I believe is a larger agenda operating in the eighth varga of the Avadānaśataka. To understand exactly what is going on here, we will have to unpack the marital imagery of the Brahmanical “envelope” the redactors are using. First, however, a very quick review of the position of women vis à vis marriage in the Brahmanical tradition would be helpful. In the Indian world in which Buddhism arose, the married gr̥hapati functioned as the lynchpin of the Brahmanical system, as evidenced by a familiar quote from the Mānava-Dharmaśāstra: yathā vāyuṃ samāśritya sarve jīvanti jantavaḥ | tathā gr̥hastham āśritya vartanta itarāśramāḥ || 77 || yasmāt trayo ’py āśramiṇo jānenānnena cānvaham | gr̥hasthair eva dhāryante tasmāj jyeṣṭhāśramo gr̥hī || 78 || sa saṃdhāryaḥ prayatnena svargam akṣayam icchatā | sukhaṃ cehecchatātyantaṃ yo ’dhāryoḥ durbalendriyaiḥ || 79 || 4 There are other appearances of the svayaṃvara in Buddhist literature. See, for example, jātakas 31, 82 and 536 in Fausbøll’s collection [or Cowell’s translation], and the Mahāyāna Candrottarādārikāparipr̥cchāsūtra (Derge Mdo-sde (tsa) 224b-243b; Lancaster 1979: 144, gives other occurrences of this sūtra). In all of these, however, the svayaṃvara merely fulfills its normal function of enabling the choice of a marriage partner.

Introduction Brides of the Buddha

Ù  389

As all living beings exist dependent on air, so people in other orders of life exist dependent on the householder.78 Because it is householders who sustain people in all three orders of life every day by giving them knowledge and food, the householder represents the most senior order of life.79 This is the order that must be shouldered assiduously by anyone who desires undecaying heaven and absolute happiness, an order that cannot be shouldered by people with feeble faculties. (MDh 3.77-79 [Olivelle 2005: 148-149, 594-598])5 77

However, taking on that role was only one of several options open to men. For women, on the other hand, marriage was presented by Manu as the only possible order of life,6 and, through marriage, a woman became “half-sharer in the ritual” as is affirmed not just once but twice in a verse of the Mānava Śrauta Sūtra. MānŚS 8.23.10 sahāvad etau mithunau saṃbhavataḥ sahāgnīn ādhattaḥ saha prajāḥ prajanayataḥ ... 12 ... tasmād ardhabhāginī bhavati 13 yājyāḥ striyo ’rdhabhāk patnī yajñe yajamānasya This couple comes together mutually; together they establish the fires; together they produce offspring ... . Therefore she is half-sharer (in the ritual). Women are fit for ritual. The wife is half-sharer in the ritual of the Sacrificer.7

The catch was that not all forms of marriages turned a woman into a patnī, the kind of ritually empowered wife referred to in the Mānava Śrauta Sūtra. For example, of the eight forms of mar-

5 While a number of law codes were available at the period under discussion, in the interests of brevity —and not from any desire to oversimplify a rich and diverse cultural landscape— I will take Manu as representative of the Brahmanical position. For further information on the “orders” discussed in this particular quote, see Olivelle 1993: 131-177, especially 137-141 and 183-190. 6 Some scholars see a decline in the position of women in Indian society between the Vedic period and the classical era. For example, Findly argued that there was a considerable change in the social status of women and a significant narrowing of their options after the Vedic period drew to a close about 600 BCE. (Findly 1985: 39-40) Olivelle noted a similar overall trend in his study of the āśrama system, “The history of Brāhmaṇical theology is a constant movement in the direction of an ever increasing restrictive ideology regarding the status and role of women.” (Olivelle 1993: 184) 7 Quoted in Jamison 1996: 31. The English translation here provided is Jamison’s. See also Mānava Śrauta Sūtra, edited and translated by J.M. van Gelder 1921, cited here after the 1961-63 reprint (MānŚS 1961: 182 and transl. 1963: 247).

The Vedas in Indian Culture and History

390  Ù

riage that Manu lists in his Dharmaśāstra,8 which are summarized in the chart below, only the first four were held to fall fully within the law (Jamison 1996: 212) and therefore were able to produce a patnī with the ability and the responsibility to act in the sacrifice. Jamison attributes this to the nature of the exchange involved: “... the giving of a gift implies a conscious and willing recipient, who by accepting the gift acknowledges a relationship of obligation with the giver.” (Jamison 1996: 212) Notice that in the first four marriages, the father as the “giver” is the active agent in choosing the husband, and, in Manu’s eyes at least, the law is upheld. Although the last four marriages were clearly considered less desirable or condemned outright by Manu, the labeling of situations such as abduction and rape as forms of marriage may have been useful in protecting the community standing and rights of both the girl and her family, while also controlling or preventing reprisals and what are now termed “honor killings.” Table 1. Eight Forms of Marriage in the Mānava Dharmaśāstra (Source: MDh, 3.20-35 [Olivelle 2005: English 109-10; Sanskrit 450-53]). Name

Characteristic

Brāhma

Girl, adorned, bestowed by father on suitable man

Daiva [Divine]

Girl, adorned, given by her father to a priest officiating at a sacrifice

Ārṣa [Seer’s]

Father accepts 1-2 pairs of cattle from bridegroom in exchange for the girl in accordance with law

Prājāpatya

Father honors girl and gives her to a man, while instructing the couple, “May you jointly fulfill the Law.”

Āsura [Demonic]

Girl given to a man after a voluntary payment of money to her relatives and herself, insofar as the man is able to pay

Gāndharva

Voluntary sexual union, based on love/lust

Rākṣasa [Fiendish]

Violent abduction with resistance by the girl

Paiśāca [Ghoulish]

Rape of a woman who is asleep, drunk, or mentally deranged

Marriage was also intricately woven into the enforced depend8 The law codes of Āpastamba (2.11.17-12.4), Gautama (4.6-14), Baudhāyana (1.20.1-1.21.3.), and Vasiṣṭha (1.28-38) differ somewhat in the number of marriage forms they present, describing six or eight forms. Depending on the number of forms described, three or four of these were fully sanctioned. Interestingly, Gautama also notes some disagreement about which forms were fully acceptable, while Baudhāyana mentions that some praised the Gāndharva marriage. See Olivelle 1999: 54-55, 84-85, 161-63, 250.

Introduction Brides of the Buddha

Ù  391

ence of women on men within the Brahmanical system, as is underscored by the well-known statements about women’s dependence found throughout the dharma literature. Turning again to Manu as a prime example, note in particular the final verse of the following passage: bālayā vā yuvatyā vā vr̥ddhayā vāpi yoṣitā | na na svātantryeṇa kartavyaṃ kiṃ cit kāryaṃ gr̥heṣv api ||147 || bālye pitur vaśe tiṣṭhet pāṇigrahasya yauvane | putrāṇāṃ bhartari prete na bhajeta svatantratām || 148 || pitrā bhartrā sutair vāpi necched viraham ātmanaḥ | eṣāṃ hi viraheṇa strī garhye kuryād ubhe kule || 149 || Even in their own homes, a female —whether she is a child, a young woman, or an old lady— should never carry out any task independently. 148As 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’. 149She must never seek to live independently. She must never want to separate herself from her father, husband or sons; for by separating herself from them, a woman brings disgrace on both families. (MDh 5:147-149 [Olivelle 2005: 146, 588]) 147

It is useful to have this particular image of proper womanhood in front of us in considering the avadānas of Suprabhā and Kāśisundarī, for becoming a Buddhist nun seems to have entailed exactly the same kind of separation —and, therefore, the same kind of disgrace for the family— that Manu warned against. Hence, this intense pressure on women to marry would have been a major problem for Buddhists seeking to recruit women into the order of nuns. Given that marriage was the only dharmically correct option for women in the Brahmanical system, and the right kind of marriage conferred on a woman both social status and ritual significance as a patnī, as well as guaranteeing her entry into heaven,9 how could the redactors make a convincing 9 An instructive story from the Mahābhārata underscores the importance of marriage to a woman’s chances to attain heaven. In the ninth book of the Mahābhārata, we find the story of a female ascetic named Subhrā who, despite her great austerities, is told by the sage Nārada that unmarried women cannot go to heaven. (Mbh 9.51) Since heaven had been the object of Subhrā’s austerities, she promptly offers half her tapas (power gained from ascetic practice) to an assembly of sages, and even this incentive only wins her a reluctant husband for a single night. When the marriage is performed, Subhrā suddenly becomes quite beautiful. Because of this sudden improvement, her new husband wants to extend their one-night marriage, but the lady turns him down and resumes her ascetic life, now assured of her celestial goal. (Sur 1973: 50, Jamison 1996: 248-249).

392  Ù

The Vedas in Indian Culture and History

case for becoming a Buddhist nun? Moreover, why did they think that using the svayaṃvara in these two avadānas would make that case more convincing? Despite the enforced dependence of women on men just mentioned, Stephanie Jamison has observed that a svayaṃvara was one of the few occasions in ancient India when “a girl not only can but must assume legal responsibility for herself.” (Jamison 1996: 237) According to the dharmasūtras and the Mānava Dharmaśāstra, a father was responsible for marrying his daughter off before the end of a set period after the onset of her menses, or he was guilty of bhrūṇahatya, “embryo-killing,” for each menstrual period that passed.10 In theory, bhrūṇahatya was a sin that required penance in a strict Brahmanical household.11 (Schmidt 1987: 80) Could this be partially a cause of the depression of the two fathers in these two avadānas? In any case, the svayaṃvara provided a useful way to meet the father’s obligation and avoid sin. It also, significantly, afforded one of the few instances in which a woman could exert her own will in choosing a husband —at least, to some limited degree— and a time-honored forum in which to do so. Indeed, the svayaṃvara seems to have very deep roots. Schmidt suggested strongly that the svayaṃvara had probably operated in Vedic times and even earlier than that, given similar customs in Iran (Schmidt 1987: 81-82). More recently, Jamison has compiled proof for Vedic svayaṃvaras (Jamison 2001 and 2003), concluding that the R̥ gveda contains “a fair amount of indirect evidence for the svayaṃvara, and … this institution was not only known to the R̥ gveda, but … it already had fixed verbal expression there.” (Jamison 2001: 303) While Schmidt identified three forms of svayaṃvara, Jamison has condensed these into two types (Schmidt 1987: 76; 10 Technically, each r̥tu or “menstrual period” after the onset of menstruation which the girl passed unmarried amounts to bhrūṇahatya, “embryo-killing,” a sin which was charged to the father and which required penance in a strict Brahmanical household. MDh 9.90-93 (Olivelle 2005: 194, 763) stipulates a period of three years after the onset of puberty before a girl can choose her own husband. See, for comparison, the Dharmasūtra of Gautama, in which the period is three months. (GDhS 18.20-22 in Olivelle 1999: 110-111) 11 Detailed discussions of the father’s responsibilities for arranging a marriage for his daughter can be found in Schmidt 1987: 76-109 and Jamison 1996: 237-240. Schmidt notes that the actual punishments prescribed were not overly severe. (Schmidt 1987: 80) However, Sternbach’s research turned up a wide range of penalties for the father who failed to marry his daughter at the appropriate time ranging from virtually none to falling to the level of a śūdra (sometimes applied to both parents and daughter). (Sternbach 1967: 30-57, especially 48-49)

Introduction Brides of the Buddha

Ù  393

Jamison 1996: 246-247 and 1999: 244-247). The first type is a simple svayaṃvara, best exemplified by Damayantī’s svayaṃvara in Book 3 of the Mahābhārata (MBh 3.50-54, especially 3.54) and by Indumatī’s svayaṃvara in Sarga (Cantos) 5 and 6 of the Raghuvaṃśam (5.36-6.86).12 The second type is the vīryaśulka svayaṃvara, which required the performance of a heroic feat (the vīryaśulka) to win the girl, the example par excellence being Draupadī’s svayaṃvara in the Mahābhārata at which Arjuna must demonstrate his prowess in a contest to win her (MBh I.175-81). In these discussions, both Schmidt and Jamison see the vīryaśulka and the simple svayaṃvaras as closely related forms, with Jamison treating the vīryaśulka in effect as a subcategory of the simple svayaṃvara. Moreover, both forms of svayaṃvara co-exist through time; one does not replace or supercede the other, a point to which we will return later. (Jamison 1999: 246-47) The basic steps of these two types of svayaṃvaras have been summed up in a table based on Jamison’s work, given in Appendix A, which also presents the corresponding elements in these two avadānas. Briefly, the first three steps are (1) the invitation and assembly of the suitors; (2) the “display” of the bride, and (3) the proclamation of the suitors and their positive attributes. The fourth step, the vīryaśulka, or “heroic deed,” was required only for the vīryaśulka svayaṃvara, of course. In the final step, the bride designates her choice of bridegroom —or accepts the winner of the vīryaśulka contest, as the case may be— often by throwing a garland of flowers over his head. In the two Buddhist avadānas of Suprabhā and Kāśisundarī, we find several steps of the basic svayaṃvara, starting with the invitation and the assembly of the suitors. Next, both Suprabhā and Kāśisundarī come to meet the assembled suitors, so we do have a “display” of the bride. The third step, the proclamation of the suitors appears at first to be absent, but notice that Suprabhā does praise the Buddha who would be —for a Buddhist audience— the pre-eminent suitor, albeit a very unlikely one. So we might consider this a truncated proclamation, and strikingly this is not done by a relative or member of the girl’s household as is the norm in the classical literature,13 but by the girl herself. This proclamation step is, however, entirely absent in Kāśisundarī’s story. Finally, in both of our avadānas, the moment of the bride’s choosing actually happens offstage before the svayaṃvara, just 12 In Devadhar 1984: 85-118. Cited after 1993 reprint. This is a particularly detailed account of a royal svayaṃvara. 13 See, for example, the elaborate announcements of the suitors at Indumatī’s svayaṃvara by a female door-keeper in Raghuvaṃśa 6.20-80. (Devadhar 1984: 102-116)

394  Ù

The Vedas in Indian Culture and History

as it does in Damayantī’s case; instead, we have the formal announcement of that choice. However, both Suprabhā and Kāśisundarī have been communicating that choice from the moment they arrived at their svayaṃvaras. Suprabhā and Kāśisundarī do not arrive at the svayaṃvara either on foot or in their fathers’ vehicles, but rather they each appear mounted on a specially marked ratha. In both avadānas, the ratha is carefully decorated with two items: an ochre banner —the same color as the robes worn by Buddhist monks and nuns— and a portrait of the Buddha. This use of the ratha has important implications drawn from the visual vocabulary of Brahmanical wedding ceremonies, vocabulary that dates back to the R̥ gveda. First of all, rathas have strong overtones in Indian marriage ceremonies. They are the vehicle of choice in abducting a bride in the Rākṣasa form of marriage advocated by both Bhīṣma and Kr̥ṣṇa in the Mahābhārata,14 and they also feature prominently in other Brahmanical wedding ceremonies. The importance of the chariot is signaled by the special mantras used to prepare the bridegroom’s chariot for the wedding and the journey to the groom’s house; examples of these can be found in the Śāṅkhāyana Gr̥hya Sūtra. (ŚGS I.15.1-22 [Oldenberg 1886: 39-41]) In fact, the moment of mounting the ratha marks the climax of the wedding, as Stephanie Jamison points out: It is only when the bride mounts his own [the bridegroom’s] chariot (svaratha) that the groom takes actual charge of her, rather than the symbolic charge he has acquired during the previous parts of the ritual. She is now ensconced on a tiny island of his property in the midst of all of that of her family —the alien turf on which he has so far been acting. (Jamison 1996: 224)

Indeed, the phrasing found in our two avadānas, suprabhā dārikā rathābhirūḍhā and kāśisundary api ratham abhiruhya, is very close to the phrasing, ratham āropayat svakam used in Bhīṣma’s abduction of the King of Kāśi’s three daughters in the Mahābhārata (Mbh 1.96.13). It also echoes the phrasing, svaratham āropya “having made [her] mount [his] own chariot,” used in the directions for the climax of the wedding ceremony in the Baudhāyana Gr̥hya Sūtra. (BGS 1.5.4; Jamison 1996: 224) 14 For Bhīma’s use of a ratha to abduct three princesses from their joint svayaṃvara, see Mbh 1.96.4-24. This section also contains Bhiṣma’s own list of marriage forms, the only list of marriages to include the svayaṃvara that I have found so far. For Arjuna and Kr̥ṣṇa’s discussion about marriage by abduction, see Mbh 1.211.14-25.

Introduction Brides of the Buddha

Ù  395

We find similar evidence for the symbolic importance of the ratha in the Vedas. Jamison noted that “The Rigvedic wedding hymn (X.85) has more to say about the wedding vehicle than any other single element of the marriage (probably including the bride and groom).” (Jamison 2001: 306) She observed that the bridegroom’s ratha played such a integral part in both the marriage ceremony and the svayaṃvara that poets could play on the underlying svayaṃvara formula of svayáṃ (sā́) X [=name] pátiṃ vr̥ṇīte by substituting rátham for pátim, “husband,” a kind of “metonomy” or what Jamison calls “formulaic slippage” (Jamison 2001: 307). Hence, even as early as the R̥gveda, the ratha had become synonymous with the bridegroom who arrived in it. One implied the presence of the other. It seems that the same metonomy was applied in our two Buddhist avadānas. Moreover, to ensure the listeners did not mistake the meaning, the redactor(s) of our stories added unmistakable signage to the ratha to mark its “ownership.” Not only is it marked with the color of Buddhist monastic robes, but the woman carries a portrait of the Buddha, echoing the wellknown use of portraits to further romantic plotlines in classical Indian drama and kāvya.15 Here the portrait of the Buddha not only marks the “ownership” of the ratha as belonging to the absent “bridegroom,” if you will, it also literalizes the religious choice being made by the two young women in a striking manner. Hence, Suprabhā and Kāśisundarī are metaphorically saying that their marriage to the Buddha has already happened! The monk-redactor of these two avadānas has cleverly twisted the “brahmanical envelope” to finesse a Buddhist meaning into a familiar cultural scenario of marriage with its promise of fertility and reproduction and used that scenario, ironically, to justify its polar opposite, becoming a celibate Buddhist nun. Yet this is not quite all that is going on here. The two svayaṃvaras in these Buddhist stores leave —just as the fathers feared— disappointed, angry suitors, and these unruly “loose ends” have to be dealt with. Both Suprabhā and Kāśisundarī subsequently have to confront and subdue their menacing suitors, and the timing of this confrontation is the key difference between the two stories. In Suprabhā’s avadāna, this happens very quickly: After her svayaṃvara, Suprabhā goes to the Buddha who preaches a sermon 15 This link between portraits and love is fully exploited in Kālidāsa’s play, Mālavikāgnimitra, in which the king falls in love with Mālavikā when he sees her in a picture of the queen and her attendants and then compares the real woman to her image in the picture, while Mālavikā in turn only gets to look upon the king to her full satisfaction when she sees his portrait, thereby discovering the full extent of her love for him. (Kālidāsa 1964: 6-7, 37-38, 105-110)

396  Ù

The Vedas in Indian Culture and History

suited to her character. She attains the fruit of the stream-enterer (srotāpatti-phalam), a kind of preliminary realization. Suprabhā then asks to be ordained, but the omniscient Buddha thinks to himself, anayā asmac-chāsane mahad-vineyākarṣaṇaṃ kartavyam, “This one will make many converts.” Therefore, instead of having her ordained, he sends her back to the very assembly she just left to “take leave of” or “obtain permission from” them, depending on how you choose to construe the phrase, parṣadam avalokayeti. (AvŚ 2.3.4-2.4.1) Now this is a very curious thing for, in most of the stories in the Avadānaśataka, permission of the parents is all that is needed to join the Buddhist order. This is true, for example, in all the stories in the seventh varga, which deal with males who become monks and then arhats, the seventh varga being a male counterpart to the female-oriented eighth varga in which our stories of Suprabhā and Kāśisundarī occur. When Suprabhā returns to the assembly, the suitors decide to take her by force and she has to subdue them. She does this by prātihārya, “miracle-working,” and r̥ddhi, “supernatural power,” both staple features of Indian Buddhist saintly biographies as has been noted by Phyllis Granoff and Reginald Ray among others.16 The text tells us that having risen into the air like a gooseking on extended wing (vitata-pakṣa iva haṃsa-rājño gagaṇa-talam abhiyudgamya), she begins to work various wonders (vicitrāṇi prātihāryāṇi darśayitum ārabdhā). In a clichéd aside also found in the Mūlasarvāstivāda-vinaya, the monk-redactor comments āśu pr̥thagjanasya r̥ddhir āvarjanakarī, “Thus magic quickly makes converts of ordinary people.”17 And the suitors, quite in keeping with this sentiment, are bowled over “like trees with their roots 16 See Granoff 1996: 79-96; Ray 1994: 51, 90, 108-09, 111, 115, 121-22, etc. Ray identifies possession of supernatural powers as the 13th theme or characteristic of the saintly paradigm in Buddhism. (Ray 1994: 51) Granoff notes, “Buddhist texts frequently acknowledged the importance of the performance of miracles as a means to impress unbelievers with the greatness of the faith.” (Granoff 1996: 79). 17 This is the only time the monk redactor shows himself in this story, and this in and of itself is worthy of note. In his translation of the first part of the Śāyanāsanavastu, Schopen commented on the prevalence of such editorial asides in the MSV Vinaya: Such insertions are common in and even characteristic of the Mūlasarvāstivāda-vinaya. They are generalizing statements that are syntactically isolated from the passages that they are inserted into which explain to the reader —in effect tell him how to read —both what has preceded and, more commonly, what follows them. They are often directed towards points which might otherwise cause problems. (Schopen 2000: 157-158, n. V.1.) Here we can see the same pattern in action. The editorial aside comes right before a rather challenging event: How could a mere girl so effectively and quickly cow a group of menacing men?

Introduction Brides of the Buddha

Ù  397

cut” (mūla-nikr̥ttā iva drumāḥ), language that is more commonly associated with battles than with religion,18 and Suprabhā does indeed convert them. Her ordination follows promptly. (AvŚ ii.4.1-9). The strong marking of the fact that Suprabhā has to accomplish these conversions before she is even allowed to join the Buddhist order looks like a rather deliberate, conscious choice on the part of the monk-redactor to present this “trial” set up by the Buddha as a twist on the well-known vīryaśulka svayaṃvara. Like Arjuna in the Mahābhārata, our young woman has to prove herself before her “union”—in this case, her ordination— can be consummated. In contrast, things are much more straightforward for Kāśisundarī. She is immediately ordained upon request to the Buddha after her svayaṃvara, and only after she has become an arhat is she forced to subdue her disgruntled suitors. Moreover, that comes about without any involvement of the Buddha, either overt or implicit for as the text puts it: tatas te rājaputrās tasyā rūpayauvanaśobhāṃ samanusmr̥tya rāgamadamattāḥ pravrajitām api prārthayituṃ pravr̥ttāḥ “Then the king’s sons, remembering Kāśisundarī’s beauty, youth, and splendor, were overcome with the intoxication of passion [or, intoxication and passion] even though she had gone forth, [and] they pursued her.” (AvŚ 2.33.11) This rather different sequence of events would seem to make Kāśisundarī’s story the equivalent of the “simple” svayaṃvara, while Suprabhā’s story would represent the vīryaśulka svayaṃvara. Bearing in mind that both forms of svayaṃvara existed and persisted concurrently, this might explain why we have two very similar —and yet, in this one key respect, different— stories set almost side by side in the same varga of the AvŚ. In conclusion, then, the fact that marriage was the only option available for women within the dominant Brahmanical system around the beginning of the Common Era —according to Manu and the authors of the dharmaśāstras— seems to have posed a considerable challenge for the recuitment of Buddhist nuns in 18 For example, when Yudhiṣṭhira kills Trigarta in a battle to rescue the abducted Draupadī in Mbh III.255.14, we are told that Trigarta “fell before the Pārtha like a tree whose roots have been cut.” (papātābhimukhaḥ pārthaṃ chinnamūla iva drumaḥ) (transl. van Buitenen 1975: 719) The root-cut tree image also appears earlier in the same episode when Draupadī fights her abductor Jayadratha (transl. van Buitenen 1975: 719): jagrāha tām uttaravastradeśe; jayadrathas taṃ samavākṣipat sā tayā samākṣiptatanuḥ sa pāpaḥ; papāta śākhīva nikr̥ttamūlaḥ Jayadratha held her by her skirt, But with all her strength she pushed him away; And, his body repulsed by her, that miscreant Fell down like a tree whose roots have been cut. (III.252.23)

398  Ù

The Vedas in Indian Culture and History

Northern India around the 2nd century C.E., at least in the eyes of the redactors of the Avadānaśataka. Of the ten stories about female arhats given in the eighth varga of the Avadānaśataka, seven focus on the issue of renunciation versus marriage and specifically on the problem of getting out of the overwhelming obligation of women to marry.19 This preponderance of stories dealing with marriage in the eighth varga about female arhats stands in striking contrast to the seventh varga with its ten stories about male arhats, because not one of those stories about male arhats makes any mention of marriage as an issue in renunciation, nor, for that matter, do they make any mention of marriage whatsoever. For the males, all that is required is permission of the father and mother, and that is readily given. So, it would seem that choosing renunciation over marriage was a “hot-button” issue for women at the time of the redaction of the Avadānaśataka, whereas for men, it seems to have been virtually a non-issue. By exploiting the svayaṃvara, the only forum that guaranteed a woman the right to choose her own husband and an ancient forum at that, with roots deep in the R̥ gveda, the monk-redactors of the avadānas of Suprabhā and Kāśisundarī took that little bit of independence as the “inch” they needed to go for the “mile.” They cleverly twisted the Brahmanical marriage imagery to argue, ironically, that a young woman who could choose her own husband had, in fact, the right to a much larger choice —to opt out of the marriage-centered Brahmanical system altogether and become a celibate Buddhist nun. To choose, in effect, to become a “bride” of the Buddha.

19 Of the three stories in this varga that do not present ways to get out of the obligation to marry, avadāna 72 concerns a girl who shows great spiritual inclination from birth and who renounces and becomes a nun at age seven. Avadāna 78 deals with an old woman who had been the mother of the Buddha in five hundred previous lives, and who was therefore certainly not a candidate for marriage. Avadāna 80, and the third of this set, features an ugly daughter of King Prasenajit who, after her marriage, becomes beautiful through the power of the Buddha.

Introduction Brides of the Buddha

Ù  399

Appendix A. Steps in a Svayaṃvara Simple & Viryaśulka Svayaṃvara*

Suprabhā

Kāśikasundarī

Invitation of suitors & assembly

tataḥ śreṣṭhī rājñāḥ prasenajito nivedya śrāvastyāṃ ghaṇṭāvaghoṣaṇaṃ kārayām āsa | saptame divase suprabhā dārikā svayaṃvaram avatariṣyati yena vo yat karaṇīyaṃ sa tat karotv iti ||

tataḥ kāśisundarī pitaram uvāca | kriyatāṃ tāta prātisīmānāṃ rājñāṃ dūtasaṃpreṣaṇaṃ | saptame divase kāśisundarī dārikā svayaṃvaram avatariṣyati yena vo yat karaṇīyaṃ sa tat karotv iti | yāvat saptame divase ṣaṭ prātisīmā rājānas saṃnipatitāḥ |

“Display” of bride

tataḥ saptame divase suprabhā dārikā rathābhirūḍhā kāṣāyaṃ dhvajam ucchrāpya buddhaṃ bhagavantaṃ citrapaṭe lekhayitvā abhiṣṭuvatī vīthīm avatīrṇā ||

kāśisundary api ratham abhiruhya kāṣāyaṃ dhvajam ucchrāpya buddhapaṭaṃ hastena gr̥hītvā rājasabhāṃ gatvā

Proclamation of suitors

abhiṣṭuvatī (see above)

[none]

Trial by heroic feat or contest

Confrontation with suitors before ordination and arhatship (instigated by the Buddha)

[none]

Bride’s choice

sā tatra rājaputrair amātyaputraiḥ śreṣṭhiputraiś ca sotkaṇṭhodvīkṣyamāṇā vicitrābhiḥ kathābhiḥ saṃjñapyovāca | sarvathāhaṃ na kena cid aṃśena bhavatāṃ paribhavaṃ karomi kevalaṃ tu nāhaṃ kāmenārthinī buddhaṃ śaraṇaṃ gatāsmi tasya sakāśe pravrajiṣyāmīti |

kāśisundarī ... rājasabhāṃ gatvovāca | śr̥ṇvantu bhavantaḥ prātisīmā rājāno | nāhaṃ bhavatāṃ rūpayauvanakulabhogaiśvaryaṃ tulayāmi api tu nāhaṃ kāmair arthinī | ya eṣa eva me bhagavān buddhaḥ paṭe likhitas tasyāhaṃ śrāvikā asya śāsane pravrajiṣyāmīti ||

References Avadānaçataka: A Century of Edifying Tales Belonging to the Hīnayāna. 1906-1909. Ed. by J.S. Speyer. Bibliotheca Buddhica 3. St. Petersburg–The Hague: Imperial Academy of Sciences– Mouton ‘S-Gravenhage. Repr. 1958. Demoto, Mitsuyo. 出本充代. 1998. “Avadānaśataka no bon kan hikaku kenkyū.” Avadānaśataka の梵漢比較研究. PhD Diss., Tokyo University. * The steps of the svayaṃvara were adapted from Jamison 1999: 248-249, though my terminology differs somewhat from hers.

400  Ù

The Vedas in Indian Culture and History

———. 2006. “Fragments of the Avadānaśataka.” Buddhist Manuscripts in the Schøyen Collection 3, 207-244. Devadhar, C.R. (ed.). 1984. Works of Kālidāsa. Vol. II: Poetry. Delhi: Motilal Banarsidass. Edgerton, Franklin. 1953. Buddhist Hybrid Sanskrit Grammar and Dictionary. 2 Vols. Delhi: Motilal Banarsidass. Repr. 1998. Feer, Léon. 1883. “Comment on devient arhatī.” Journal Asiatique (Ser. 8) 1, 407-440. ——— (transl.). 1891. Avadāna-çataka, cent legendes (bouddhiques). Annales du Musée Guimet 18. Paris: E. Leroux. Findly, Ellison B. 1985. “Gārgī at the King’s Court: Women and Philosophical Innovation in Ancient India.” In: Y.Y. Haddad and E.B. Findly (eds.), Women, Religion, and Social Change. Albany, NY: State University of New York Press, 37-58. Granoff, Phyllis. 1996. “The Ambiguity of Miracles. Buddhist Understandings of Supernatural Power.” East and West 46, 79-96. Jamison, Stephanie W. 1996. Sacrificed Wife / Sacrificer’s Wife: Women, Ritual and Hospitality in Ancient India. New York–Oxford: Oxford University Press. ———. 1999. “Penelope and the Pigs.” Classical Antiquity 18, 227272. ———. 2001. “The Rigvedic svayaṃvara? Formulaic evidence.” In: K. Karttunen and P. Koskikallio (eds.), Vidyāravavandanam: Essays in Honour of Asko Parpola. Helsinki: Finnish Oriental Society, 303-315. ———. 2003. “Vedic Vrā:́ Evidence for the Svayaṃvara in the Rig Veda?” In: S. Adhami (ed.), Paitimāna: Essays in Iranian, Indo-European, and Indian Studies in Honor of Hanns-Peter Schmidt. 2 Vols. Costa Mesa, CA: Mazda Publishers, 39-56. ———. 2006. “Women ‘Between the Empires’ and ‘Between the Lines’.” In: P. Olivelle (ed.), Between the Empires: Society in India 300 BCE to 400 CE. Oxford–New York: Oxford University Press, 191-214. Jātaka or Stories of the Buddha’s Former Births. 1895-1913. 6 Vols. in 3. Ed. by E.B. Cowell and Transl. by H.T. Francis et al. Cambridge–London: University Press–Luzac. Repr. 1957.

Introduction Vedism and Brahmanism in Buddhist Literature

Ù  401

Vedism and Brahmanism in Buddhist Literature Shrikant Bahulkar Central University of Tibetan Studies, Sarnath

1. Names of Mañjuśrī: “Vedic” epithets Holding on to the great vow of austerity, (he) wears the cord of muñja-grass, observing chastity, (he is) best in holy practice, having great penance, intent upon penance —Gautama, the foremost snātaka, the knower of brahman, a brāhmaṇa, Brahmā himself, and one who has attained the brahmanirvāṇa. (He is) release, liberation, (has) for his body the true release, (he is) true release, peacefulness (and) Śiva.1

This is not a description of any Vedic sage named Gautama, but of Mañjuśrī, the Bodhisattva of the Tenth Stage (daśabhūmīśvara). The verses are from the Mañjuśrīnāmasaṅgīti (MNS), a Buddhist Tantric text. This text has for its background the early Buddhist tradition as well as the later Mahāyāna tradition; at the same time it evinces some element of Vedism and (later) Brahmanism. The text incorporates 812 names of Mañjuśrī in all, reminding us of the Viṣṇusahasranāma, “Thousand names of Viṣṇu,” a text popular in the Vaiṣṇava tradition. Some of the names in the MNS clearly show the influence of Vedic and post-Vedic religious traditions. This text was composed sometime between the sixth and the eighth century C.E. and was translated into Tibetan and Chinese. It is still popular amongst Tibetan and Nepalese Bud1

mahāvratadharo mauñjī brahmacārī vratottamaḥ | mahātapās taponiṣṭhas snātako gautamo ’graṇīḥ || brahmavid brāhmaṇo brahmā brahmanirvāṇam āptavān | muktir mokṣo vimokṣāṅgo vimuktiḥ śāntatā śivaḥ || MNS 8.18-19.

402  Ù

The Vedas in Indian Culture and History

dhists. Like this work, there are other Buddhist Tantric works, the Tantras and their commentaries, which have some element that can be connected with Vedism and Brahmanism. 2. Vedism and Brahmanism: Heesterman’s view As regards the terms “Vedism and Brahmanism,” one may refer to Jan C. Heesterman’s contribution “Vedism and Brahmanism” in the Encyclopedia of Religion (Vol. 15). He says at the outset: The somewhat imprecise terms Vedism and Brahmanism refer to those forms of Hinduism that revolve primarily around the mythic vision and ritual ideologies presented by the Vedas. These terms are classifications that have been used by historians to categorize in a typical manner a variety of religious beliefs and practices in ancient and contemporary South Asia. Vedic and Brahmanic religious sensibilities are thereby distinguished from Āgamic, Tantric, and sectarian forms of Hinduism, which look to a variety of non-Vedic texts as the source of religious authority. Vedism is older than Brahmanism, which developed from and remains true to the Vedic overview but accommodated and remoulded the religious ideas and practices of non-Vedic South Asian traditions. (Heesterman 1987: 217).

The case of Buddhist texts is somewhat different. While criticizing the Vedic tradition on one hand, they honour that tradition on the other and borrow some terms, concepts and practices from the Vedic and later Brahmanical traditions. Heesterman has therefore restricted himself to the texts in Hindu tradition. It was natural for the post-Vedic Hindu religious cults to show reverence toward the Vedas and their tradition and to seek the origin of their deities, religious practices, and philosophical concepts in the Vedic tradition, or even to call them “Vedic.” The Hindu Tantric cults have a similar position; however at times they show some anti-Vedic attitudes, despising the religious and social hierarchy based on the varṇāśramadharma. The Buddhist Tantric cults belong to the same Indian soil where various Tantric cults emerged and developed along with them about the same period of Indian history. 3. Veda and Vedism in the Pāli canon The period of the early Pāli canon overlaps with the late Vedic Sūtra period. Unfortunately, we do not have adequate knowledge of the background of the formation of Vedic schools and the institution of sacrifice that they represent (Witzel 1993: 265). The references to Vedic schools in the Pāli canon might be of some use

Introduction Vedism and Brahmanism in Buddhist Literature

Ù  403

for reconstructing the history of Vedic sacrificial institutions. The Vedic sacrifice, which was in ancient times the core of the Brahmanic life, is reflected in the Pāli canon. It displays the knowledge of Vedic ritual and Vedic texts. In his essay “Vedische Opfer im Pāli-Kanon,” Harry Falk examines the stereotyped mention of the sacrifices such as the Aśvamedha, the Puruṣamedha, the Vājapeya, the Niraggala, which probably means the Sarvamedha, and so on (Falk 1988: 225-254). i) Meaning of the term “Veda” When we try to find how the Buddhist looks at the Vedic and Brahmanical tradition, we may observe that there is a mixed feeling about Vedism and Brahmanism reflected throughout the Buddhist literature in the Pāli canon and in the later texts composed in Buddhist Hybrid Sanskrit, classical Sanskrit, and Buddhist Tantric Sanskrit. In the Pāli canon, there is, on the one hand, reverence towards the Vedic lore, the study of Vedas, and at times the Vedic sacrifices, and, on the other, there are views opposing the supremacy of Brahmins over the other Varṇas and their indulgence in sacrificial ritual. A well-known passage from the Ambaṭṭhasutta of the Dīghanikāya (D 3) may be cited in this connection. The passage describes the accomplishment of a Brahmin as: “a student of Veda who knows mantras, perfected in three Vedas, a skilled expounder of the rules and rituals, the lore of sounds and meanings, and fifthly, of the legends.”2 In the Pāli canon, the word veda has a two-fold meaning: 1) It is used in the compounds atthaveda and dhammaveda, in the sense of “enthusiasm for the truth (for the letter and the spirit) of Buddha’s teaching (Majjhimanikāya 1.37; Aṅguttaranikāya V.329ff., 333, 349, 352); and 2) (higher) knowledge (as “Buddhist” antithesis to the authority of the “Veda.”) (Rhys Davids and Stede 1993: 647B). ii) Vijjā and Tevijja The word vijjā is sometimes used in the sense of Veda, as we find in a verse from the Aṅguttaranikāya (A 3.6): “With these three vidyās a Brahmin becomes tevijja, the possessor of three vidyās. I call him tevijja, not any other so called in empty words.”3 The concept of three vidyās is common to the Vedic and the Buddhist tradition; however the implications are different. It is said 2 ajjhāyako mantadharo tiṇṇaṃ vedānaṃ pāragū sanighaṇḍu-keṭubhānaṃ sākkharappabhedānaṃ itihāsa-pañcamānam ... D 3.1.3 (Rhys Davids and Carpenter 1975: I. 884-5). 3 etāhi tîhi vijjāhi tevijjo hoti brāhmaṇo | tam ahaṃ vadāmi tevijjaṃ nāññaṃ lapitalāpanaṃ || A 3.6 (Warder 1961: 16531-32).

404  Ù

The Vedas in Indian Culture and History

therefore in the Aṅguttaranikāya (A 3.6): “Well, Brahmin, the Brahmins describe a Brahmin as the possessor of three vidyās in one sense and one is (described as) the possessor of three vidyās in the discipline of the Ārya in another sense.”4 The word tevijja- represents the three vidyās both in the Vedic and the Buddhist traditions: in Vedic, it represents the first three Vedas and thereby the dharma centred around the performance of solemn sacrifices (trayīdharma), while for the Buddhist, it denotes the three vidyās identified with the three out of six kinds of superhuman knowledge, termed as abhiññā (Skt. abhijñā) which the Buddha possessed. They are: 1) the knowledge of former births (pubbenivāsa), 2) of the (future) rebirths of beings (cutūpapāta), and 3) of the destruction of the depravities (āsava-khaya). It is said that during his enlightenment, the Buddha, having entered the fourth jhāna, attained the three vijjās (Bhayabheravasutta, M I.4; Trenckner 1964: 22-23). Though not as frequently as found in the Pāli canon, the word traividya occurs in the Mahāyāna texts, particularly, in the Saddharmapuṇḍarīkasūtra (3.90; 5.29; 7.59) (=Vaidya 1960: 117), the Lalitavistara (Tripathi 1987: 209, 291, 300, 352), and in the Mahāyānasūtrālaṅkāra (7.9; Bagchi 1970). The connotation of this term is however not the same in all these texts. iii) Yañña “a sacrifice”: Buddha’s view   The word yañña too has a two-fold meaning: in the context of the Vedic tradition it means “a sacrifice,” while in the Buddhist sense, it is often “almsgiving,” “charity,” “a gift to the Saṅgha,” or a “bhikkhu.” However, it is said in the Selasutta of the Suttanipāta (Sn 3.7): “The oblations in the sacred fire are the foremost of the sacrifices and the Sāvitrī, the foremost of the chandas.”5 There are numerous expressions of this kind, showing respect towards the Vedic lore. The Buddha himself was well-versed in the Vedas. In the Sundarikabhāradvājasutta of the Suttanipāta (Sn 3.4), the Buddha asks Bhāradvāja: “If you call yourself a Brahmin and myself a non-Brahmin, I ask you about the sāvitrī, having three pādas and twenty-four syllables.”6 Ultimately, Bhāradvāja realizes that the Buddha is a real Brahmin, expert in the Vedas, Brahmā himself, and one who deserves the oblation of the sacrificial bread

4 Aññathā bho brāhmaṇa brāhmaṇā brāhmaṇaṃ tevijjaṃ paññāpenti, aññathā ca pana ariyassa vinaye tevijjo hotî ti | A 3.6 (Warder 1961: 16623-24). 5 Aggihuttamukhā yaññā sāvittī chandaso mukhaṃ | Sn 568 (Andersen and Smith 1965: 111). 6 Brāhmaṇo ce tvaṃ brūsi mañ ca brūsi abrāhmaṇaṃ | Taṃ taṃ sāvittiṃ pucchāmi tipadaṃ catuvīsatakkharaṃ || Sn 457 (Andersen and Smith 1965: 81).

Introduction Vedism and Brahmanism in Buddhist Literature

Ù  405

(pūraḷāsa) (Sn 479).7 The metaphor of sāvitrī is again used in a different context in some Mahāyāna works, with which we shall deal later. References to the Vedic sacrifices are also frequently met with. While showing respect for the Vedic lore and even for the sacrifice, the Buddha however opposes the killing of animals and the waste of natural resources in the sacrifice. The Kūṭadantasutta (D I.5) describes in an elaborate manner a bloodless sacrifice. Kūṭadanta, a Brahmin living in the Brahmin village Khānumata, had planned a great animal sacrifice. Seven hundred bulls, seven hundred bullocks, seven hundred heifers, seven hundred he-goats and seven hundred rams had been brought to immolate in the sacrifice. At that time the Buddha was staying in that village. Kūṭadanta heard that the ascetic Gotama had known how to conduct successfully the triple sacrifice with sixteen requisites (D 5.4).8 He approaches Gotama, who narrates how in ancient times a king Mahāvijita performed a bloodless sacrifice. Gotama says: In this sacrifice, O Brahmin, no bulls were slain, no goats or ship, no cocks and pigs, nor were various living beings subjected to slaughter, nor were trees cut down for sacrificial posts, nor were grasses mown for the sacrificial grass. (Walshe 1987: 138).9

The Buddha further speaks of a sacrifice which is simpler, less difficult, more profitable than the threefold sacrifice with sixteen attributes; he mentions more and more profitable sacrifices consisting in giving gifts to ascetics, providing shelter for the Saṅgha, going for refuge to the Buddha, the Dhamma and the Saṅgha, undertaking precepts and finally, attaining the four jhānas, various insights and the cessation of the corruptions. Here it is interesting to see that the sacrifice involving immolation of “seven hundred bulls, seven hundred bullocks, seven hundred heifers, seven hundred he-goats and seven hundred rams” is found nowhere in the texts describing Vedic solemn sacrifices (Dharmadhikari 2006: 228). It appears that the story of that kind of sacrifice is used just to convince Kūṭadanta of the impropriety of a sacrifice involving animal killing. In the story, the Vedic concept 7

Hutañ ca mahyaṃ hutam atthu saccaṃ, yaṃ tādisaṃ vedagunaṃ alatthaṃ |

Brahmā hi sakkhi: patigaṇhātu me Bhagavā, bhuñjatu me Bhagavā pūraḷāsaṃ || Sn 479

(Andersen and Smith 1965: 85). 8 Sutaṃ kho pana me taṃ: Samaṇo Gotamo tividha-yañña-sampadaṃ soḷasaparikkhāraṃ jānātîti | D 5.4 (Rhys Davids and Carpenter 1975: I. 12829-30). 9 Tesu pi kho brāhmaṇā, yaññesu n’ eva gāvo haññiṃsu, na ajeḷakā haññiṃsu, na kukkuṭa-sūkarā haññim (sic ṃ) su | na vividhā pāṇā saṃghātaṃ āpajjiṃsu, na rukkhā chijjiṃsu yūpatthāya, na dabbhā lūyiṃsu barihisatthāya | D 5. 20. (Rhys Davids and Carpenter 1975: I. 14223-26).

406  Ù

The Vedas in Indian Culture and History

of sacrifice has been modified and is given a deeper meaning. iv) Indra and Buddha In the Pāli canon, we frequently come across the description of the Buddha with standard epithets: That Blessed One is an Arhat, a fully awakened one, abounding in wisdom and goodness, happy, with knowledge of the worlds, unsurpassed as a guide to mortals willing to be led, a teacher for gods and men, a Blessed One, a Buddha. (D 3.1.2; Rhys Davids 1969: 109).10

The expression “a teacher for gods and men” (satthā devamanussānaṃ), placing the Buddha above men and gods reminds us of the description of Indra at R̥ gveda 3.34.2: “O Indra, you are equally the leader of heavenly subjects and human generations.”11 Indra, when born, set in motion the wheel of the Sun. On the basis of this reference, P. Banerjee tries to show its similarity with the first discourse of the Buddha that is called “the setting of the Wheel of Law in motion” (dhammacakkappavattana). At the time of Enlightenment, Māra attacks the Buddha and attempts to distract his mind from his determination to attain Enlightenment (Banerjee 1989: 6-18). There is one more name for Māra: it is Namuci. According to Malalasekara, this name is “given because he does not allow either gods or men to escape from his clutches, but works them harm” (Malalasekara 1989: 31). The name Namuci, however, rather appears to have been borrowed from the R̥ gveda where he is said to be Indra’s enemy, whom the latter kills. These similarities are superficial and do not necessarily indicate a possibility that Indra was a prototype of the Buddha. v) Supremacy of Buddha’s Dharma over Brahmanical Dharma While showing reverence towards the Vedic lore and using the Vedic vocabulary, the Buddhist texts show the supremacy of their dharma over the age-old and well-established Vedic tradition. The story of Buddha’s Enlightenment says that he remained four weeks at the bodhi-tree after attaining Enlightenment. He was doubtful whether he would be able to teach dhamma to others. He finally decided to engage himself in preaching, only when Brahmā appeared and requested him to do so. This episode has been interpreted in different ways. It may suggest that the repre10 Iti pi so Bhagavā arahaṃ sammā-sambuddho vijjā-caraṇa-sampanno sugato lokavidū anuttaro purisa-damma-sārathi satthā deva-manussānaṃ buddho bhagavā | D 3.1.2 (Rhys Davids and Carpenter 1975: 1.8716-19). 11 indra kṣitīnām asi mānuṣīṇāṃ viśāṃ daivīnāṃ uta pūrvayāvā | R̥g veda 3.34.2.

Introduction Vedism and Brahmanism in Buddhist Literature

Ù  407

sentative of the Vedic tradition recognized the importance and the supremacy of dhamma over his own dharma. 4. Vedism and Brahmanism in Mahāyāna In the Mahāyāna texts, a similar attitude of reverence towards the Veda and the inclination for Vedism is found. Mātr̥ceṭa, a Buddhist poet who flourished around the first century C.E., has composed a beautiful hymn, the Varṇārhavarṇastotra, in praise of the Buddha. A section of that hymn is called brahmānuvādastava “The praise (of the Buddha) in Brahmanical terms.” Some of the verses from that section (VII. 1, 2 and 13) are relevant to the present discussion: O Sun (Savitr̥) of the Good Law, you have illuminated that sāvitrī of the Law, of twice twelve (i.e., twenty-four) members (pada), running both forward and backward, (in which consists) the nature of this world, (consisting) in the manifestation and the cessation (of the world), (and) which all worldly beings including the gods severally obey (1-2). You are Brahma, in you is the brahmanical path, you are the chief among brahmans, you are the guide and the preceptor, the priest and the chaplain. (Bailey 1951: 965, 967)12

How is the formula of the pratītyasamutpāda to be understood as the sāvitrī, i.e., the mantra of Savitr̥? The sāvitrī is an invocation of the deity Savitr̥, in the gāyatrī metre, and hence is also called the Gāyatrī mantra. The metre has three pādas having eight syllables each, thus having twenty-four syllables in all. The formula of the pratītyasamutpāda has twelve links (dvādaśa nidāna). The formula is to be recited both ways: the formula denoting the origination (pravr̥tti, anuloma) and that of the cessation (nivr̥tti, viloma). The anuloma order is: avijjāpaccayā saṅkhārā, saṅkhārapaccayaṃ viññāṇaṃ, and so on. Then the formula is to be recited in the viloma order where the cessation is mentioned: avijjānirodhā saṅkhāranirodho, saṅkhāranirodhā viññāṇanirodho, and so on. Thus the formula becomes one having twenty-four constituents. We find a similar metaphor in a philosophical work. Bhāviveka (or Bhavya), in his Madhyamakahr̥daya (Lindtner 2001), devotes a small chapter to “Taking recourse to the vow pravr̥ttau ca nivr̥ttau ca prakr̥tir jagato’sya yā | yāṃ sarve nātivartante pr̥thag lokāḥ sadevakāḥ || pratilomānulomāṅgā dvidvādaśapadā tvayā | saddharmasavitar dharmasāvitrī samprakāśitā || tvaṃ brahmā pratipad brāhmī tvayi tvaṃ brāhmaṇarṣabhaḥ | praṇetā copanetā ca tvam r̥tvik tvaṃ purohita || Varṇārhavarṇastotra VI. 1, 2 and 13 (Bailey 1951: 964, 966).

12

The Vedas in Indian Culture and History

408  Ù

of an ascetic” (Madhyamakahr̥daya, Munivratasamāśraya, Chapter II). He describes in a poetic manner how the ascetic, aspiring for Enlightenment, should engage himself in the practice of the vow of an ascetic: (One aspiring for Enlightenment), resides in a beautiful forest, meant for penance, in the form of Mahāyāna, takes for his meals the fruits of the joy of meditation. He takes the cows (in the form of his faculties) to graze (gocara) on the (four) foundations of mindfulness (smr̥tyupasthāna). He destroys all sin through the study (svādhyāya) of the vast and profound Sūtrāntas and relying on the two-fold truth, muttering (the mantra in praise) of Savitr̥ in the form of the formula of dependent origination (pratītyotpāda).13

Here the words svādhyāya and sāvitrī are of Vedic origin. In the Vedic tradition, svādhyāya means the study (literally, recitation) of one’s own śākhā (sva+adhyāya), while in the Buddhist tradition, it denotes the study of their religious scriptures. The word sūtrānta reminds us of the word vedānta, the study and practice of which are essential for the realization of the ultimate truth, according to the schools of Vedānta. This instance is just one that points to Vedism in the Mahāyāna. There could be many more instances of this kind in other texts. 5. Vedism and Brahmanism in Buddhist tantric literature i) Meaning of the term mantra/Mantrayāna Among the Mahāyāna texts, the Tantric works are particularly replete with the Vedic and Brahmanic vocabulary and are much more inclined towards Vedism and Brahmanism. The Buddhist Tantric way is known as Vajrayāna or Mantrayāna and is a part of Mahāyāna. The term mantra denotes the prominence of mantras to be recited in various Tantric practices. There are also the mantras associated with various Buddhas and Bodhisattvas, e.g., om maṇi padme hum (Tibetan pronunciation om mani peme hum), the mantra of Avalokiteśvara, occurring on the prayer-wheels in the Tibetan Buddhist temples, or namu amida butsu, the six-character mantra of Amitābha (popularly called nembutsu in Japanese), popular in the Pure Land sect of the Japanese Buddhism. The word mantra has however a peculiar meaning in the Tantric Buddhism. It is said in the Sekoddeśaṭīkā (SUT): 13

mahāyānamahāramyatapovanasamāśrayaḥ | dhyānaprītiphalāhāras smr̥tyupasthānagocaraḥ || gambhīrodārasūtrāntasvādhyāyahatakilbiṣaḥ | pratītyotpādasāvitrīṃ japan satyadvayāśrayaḥ || Madhyamakahr̥daya II.9-10 (Lindtner 2001: 6).

Introduction Vedism and Brahmanism in Buddhist Literature

Ù  409

Moreover, mantra is also called the supreme, imperishable knowledge, as it becomes the protection for the mind. The meaning of mantra, conveyed by the word mantra, is the imperishable knowledge of emptiness. Mantra is endowed with the sacred knowledge and is of the nature of (the union of) emptiness and compassion.14

Mantrayāna is therefore the way that aims at the realization of the union of emptiness and compassion, i.e., the great bliss (mahāsukha), the attainment of the state of the Completely Enlightened Buddha. ii) Vajra/Vajrayāna The word vajra in the name “Vajrayāna” has obviously been borrowed from the Vedic tradition and is associated in a unique manner with the Buddhist Tantric sect. Here again the association with Indra is to be noted. As Indra destroys his enemies with his mighty thunderbolt, the Buddha too subdues his foes (māra) in the form of defilements (kleśa) and so forth and hence is called arhat, ariha(n)t “foe-destroyer.” In the Buddhist Tantric works, the epithet bhagavān is interpreted in various ways. One of the interpretations offers the etymology of the word bhaga, from the root bhaj “to break, to destroy,” hence “destruction.” Bhagavān is so called, because he subdues defilements and passions. The Sekoddeśaṭīkā (SUT) quotes a verse from the Hevajra Tantra (HT): Bhaga is said to be “breaking,” since it destroys mental afflictions (kleśa-māra) and so forth. Those defilements are to be destroyed by means of wisdom; therefore, wisdom is called bhaga.15 As regards the word vajra, it has a special meaning in the Buddhist Tantric tradition. In a verse quoted in the Advayavajrasaṃgraha (AVS), vajra is equated with śūnyatā, “emptiness,” through which Enlightenment is attained: “Firm, substantial and solid, of uncuttable and unbreakable character, unburnable, indestructible, the void is said to be vajra.”16 tathā manastrāṇabhūtatvān mantro ’pi paramākṣarajñānam ucyate | Mūlatantre ca-kāyavākcittadhātūnāṃ prāṇabhūto yatas tataḥ | mantrārtho mantraśabdena śūnyatājñānam akṣaram | puṇyajñānamayo mantraḥ śūnyatākaruṇātmakaḥ || Sekoddeśaṭīkā (Sferra and Merzagora 2006: 193). 15 bhañjanaṃ bhagam ākhyātaṃ mārakleśādibhañjanāt | prajñāvadhyāś ca te kleśās tasmāt prajñā bhagocyate || Sekoddeśaṭīkā (Sferra and Merzagora 2006: 65). The verse is not found in the HT. However, it has a verse of similar import: bhagāni ṣaḍvidhāny āhur aiśvaryādiguṇākhilāḥ | athavā kleśādimārāṇāṃ bhañjanād bhagavān iti || HT I.v.15 (Snellgrove 1959b: 16). 16 dr̥ḍhaṃ sāram asauśīryam acchedyābhedyalakṣaṇam | 14

410  Ù

The Vedas in Indian Culture and History

iii) Oṁkāra/svāhā/vauṣaṭ/vaṣaṭ It is often said in the Buddhist Tantric texts that the mantra is to be preceded by om and followed by svāhā. This is just like any Vedic, post-Vedic Hindu or a Hindu Tantric tradition. The Kālacakra Tantra (KT III. 79ab) says: “In the beginning of the mantras of all gods and goddesses, there should be praṇava; while during the offerings, it should end with svāhā; it is the essence; also in (all acts of) worship, it should end with namas.”17 The text (KT III. 80) further mentions other utterances associated with the mantra to be employed in various rites: In a rite for prosperity, the mantra should end with svāhā; O lord of men, in a propitiatory rite, it should end with namas. In a rite for attracting (a person closer) (ākr̥ṣṭi), there is vauṣaṭ at the end; in (the rite for) subjecting (a person to one’s will) (vaśya), it should have vaṣaṭ at the end. In the sorcery rites, [i.e., creating enmity (vidveṣaṇa), eradication (uccāṭana) and killing (māraṇa)], it should end with hūm. Due to the intrinsic qualities, the mantra for the “nailing” (kīlana) rites [i.e., paralyzing (stambhana) and hypnotizing (mohana)], it should have phaṭ at the end. The mantra is white, red, black, or resembling the best gold in colour, depending upon the difference.18

The authors of these Tantric works were aware that the syllable om was called praṇava and was uttered at the beginning of the recitation of the Vedas. The Hevajra Tantra (HT) calls om “foremost of the Vedas” (vedānām ādima) and says that it should be placed at the beginning of the mantra. The commentary Yogaratnamālā on this Tantra calls it praṇava (Snellgrove 1959b: 82; 153). iv) Homa “sacrifice”: inner and outer All these mantras are to be used in the rites where the offerings in favour of deities are to be made in the fire. This sacrificial act is frequently termed homa, and is known by the same term in the Buddhist Tantric sects of other countries. For instance, in the Shingon sect of Japan, it occurs in its phonetic translation as goma. The Buddhist Tantric texts use the same Vedic and postVedic sacrificial terminology; but in particular contexts, the imadāhi avināśi ca śūnyatā vajram ucyate || AVJ (Shastri 1927: 23, 37). 17 sarveṣāṃ nāma pūrvaṃ praṇava iti bhaved devatādevatīnāṃ home svāhāntamantro hr̥dayam api tathaivārcane vai namo ’ntaḥ | KT III. 79ab (Dviwedi and Bahulkar 1994a: 78). 18 puṣṭau svāhāntamantro bhavati narapate śa(ā?)ntike ’sau namo ’nta ākr̥ṣṭau vauṣaḍanto bhavati ca vaṣaḍantaś ca vaśye tathaiva | hūṅkārānto ’bhicāre prakr̥tiguṇavaśa(ā?)t kīlanādye phaḍantaḥ śveto raktaś ca kr̥ṣṇo varakanakanibhaḥ karmabhedaiś ca mantraḥ || KT III. 80 (Dviwedi and Bahulkar 1994a: 79).

Introduction Vedism and Brahmanism in Buddhist Literature

Ù  411

plications are different. The Cakrasaṃvara Tantra (CST) describes “the inner sacrifice (adhyātmahoma) and prescribes that one should consume caru. This substance of offering, caru, generally means “cooked rice”; but here, it has a typically Buddhist Tantric sense, i.e., the five sacred ambrosias (pavitrāṇi pañcāmr̥tāni, Pandey 2002: 499), namely, blood, semen, human flesh, urine, and excrement (Wayman 1961: 91; 1977: 291). Elsewhere this Tantra explains the meanings of other substances to be consumed by a practitioner within a group of Yogins and Yoginīs (gaṇacakra). Madhu is “semen,” karpūra “flesh,” rakta “urine” (Pandeya 2002: 25). It further says: “Having consumed these substances, like drinking Soma, one undoubtedly obtains siddhi” (CST I.13). The Saṃvarodaya Tantra (SaṁT) speaks of a two-fold homa, the mundane and the transcendent and says that one should perform the mundane homa in daytime and the transcendent one at the night: “At an assembly of Yoginīs and Yogins, he should offer food and drink; caru as a burnt offering” (CST 28.4; SaṁT 23.51; 26.28). v) Fruits of sacrifice The SaṁT mentions the fruits of all divisions of homa: Now I shall speak of the fruits which result from all the divisions of homa: the (sacrificial) ground increases landed property; the hearth-pit makes the house prosperous (54). Clarified butter brings about every (kind of) prosperity; fuel increases splendour; firewood increases heroism; kuśa-grass protects everything (55). White mustard pacifies (calamities); grains of rice are considered to be increasing welfare; sesame seed is known to destroy evil; corn brings grain and wealth (56). Beans produce great strength; barley gives wind-like speed; dūrvā-grass extends the life span; wheat removes sickness (57). Honey and milk bring about wisdom; boiled rice mixed with curd (dadhyanna) grants all kinds of happiness; (the sacrificial) fire makes the desired object fulfilled; one’s own guardian-deity grants liberation (58). Other things are known to effect the rites of pacifying and so on according to the (nature of) the rite; the pot (pātrī) is wisdom; the ladle (sruva) the means (upāya); their union is the practice of nonduality (59). Clarified butter poured from them (as offering the sacred fire) is considered to be the ambrosia of great wisdom; with it (the practitioner) should please the (sacred) fire, whose essence is the whole world (60). He who in this way carries out homa, brings about fulfilment and good luck (for the sacrificer) (61). (Tsuda 1974: 313).19 19

athānyatamaṃ vakṣye sarvahomāṅgajaṃ phalam | kṣetravr̥ddhikarī bhūmiḥ kuṇḍaṃ gehavivr̥ddhikr̥t ||

412  Ù

The Vedas in Indian Culture and History

vi) Sacrificial utensils: tantric interpretation The Vasantatilaka (VT) also mentions sruva, bhājana, pātrī, and kuṇḍa (VT 8.23-24) using a similar metaphor (VT 8.21-24). It speaks of the inner and the outer homa. The inner fire is the fire of wisdom (prajñāgni) and the outer is the one emerged from the friction of the two wooden blocks (araṇimanthana)(commentary on VT 8. 21-24; Rinpoche and Dviwedi 1990: 66-67). One of the MSS of the VT reproduces a passage from the Ḍākārṇava where we find a similar metaphor (Rinpoche and Dviwedi 1990: 69). vii) Sacred fire and four Varṇas The fire for the Tantric sacrifice is however not necessarily obtained by the churning of the wooden blocks. The Vimalaprabhā (VP) commentary on the KT gives a detailed description of the fire to be obtained from the houses of persons of various classes: Here, first, for all acts, the fire should be fetched from a temple and should be kindled by means of the logs of sandal tree; for propitiatory and prosperity rites (śānti-puṣṭi), from the house of a Brahmin; for killing (māraṇa) etc., from the house of a Śūdra; for the purpose of seduction (vaśīkaraṇa) etc., from the house of a Kṣatriya; for stupefaction (stambhana) etc., from the house of a Vaiśya; for killing (māraṇa) etc., it may also be fetched from the house of a Caṇḍāla and should be kindled with the logs of thorny trees.… Then one should invoke Vaiśvānara with these mantrawords —Om come, come; O the best of great elements, gods, sages and Brahmins.20 sarvasampattikr̥t sarpiḥ samit tejovivardhikā | śauryādhikakaraṃ kāṣṭhaṃ sarvarakṣākaraḥ kuśaḥ || śāntikr̥t sitasiddhārthāḥ puṣṭikr̥t taṇḍulo mataḥ | tilaṃ pāpaharaṃ vidyād dhānyaṃ dhānyārthakarṣakam || mahābalakaraṃ māṣaṃ vāyuvegapradaṃ yavam | āyurvr̥ddhikarī dūrvā godhūmo roganāśakaḥ || prajñāpade madhukṣīre dadhyannaṃ sarvasaukhyadam | iṣṭārthasiddhido vahnir muktiṃ dadyāt sveṣṭadevatā || śeṣaṃ karmānurūpeṇa jñeyaṃ śāntyādikarmakr̥t | pātrī prajñā sruvopāyas tacchleṣo ’dvayabhāvanā || tato vinirgataṃ sarpir mahājñānāmr̥taṃ matam | tena santarpayed agnim ātmanā sacarācaram || evaṃ karoti yo homaṃ siddhisaubhāgyasampradam || SaṃT 23. 54-61 (Tsuda

1974: 144). 20 iha prathamaṃ candanakāṣṭhair agniṃ prajvālya sarvakarmaṇi devagr̥hād ānayitvā śāntipuṣṭyor brāhmaṇagr̥hān, māraṇa(ā?)dau śūdragr̥hād, vaśyādau kṣatriyagr̥hāt, stambhanādau vaiśyagr̥hān, māraṇe punaś caṇḍālagr̥hād ānayitvā prajvālayet kaṇṭakakāṣṭhair iti | … tato vaiśvānaram āvāhayed ebhir mantrapadaiḥ— Om ehi ehi mahābhūta-deva-r̥ṣi-dvijasattama | VP on KT 3.74-75 (Dviwedi and Bahulkar 1994a: 74).

Introduction Vedism and Brahmanism in Buddhist Literature

Ù  413

It may be observed that, while the Buddhist Tantras and particularly the Kālacakra, vehemently denounce thue Vedic dharma for its hierarchy and discrimination in respect of the four varṇas, this ritual act of fetching the fire from the houses of different castes maintains the hierarchy of castes. In this connection, it is interesting to find how some Indian Buddhist masters attempt to show the relation of the Buddhist Tantras with the Brahmanical class system. The Buddhist Tantras are divided into four classes: Kriyā “Action (Tantras),” Caryā “Performance (Tantras),” Yoga “Yoga (Tantras),” and Anuttara “Highest (Yoga Tantras).” In his commentary on the Vajramālā, the Indian master Alaṅkakalaśa (?) says: Brahmins are taught Action Tantra and those of the royal caste (kṣatriya) are taught Yoga Tantra. Those in the merchant caste (vaiśya) whose desire and hatred are slight but whose ignorance is extremely great and who believe in Viṣṇu are taught Performance Tantra. Those in the merchant caste whose desire and hatred are great but whose ignorance is slight are taught Yoga Tantra such as the Guhyasamāja [which is actually a Highest Yoga Tantra]. Those who are related to the servant caste (śūdra) whose desire and hatred are the great of the great and whose ignorance is the slight of the slight are taught the Mother Tantras such as the Little Saṃvara Tantra (laghusaṃvara). (Hopkins 1977: 155).

Tsong-kha-pa (C.E. 1357-1419), in his Great Exposition of Secret Mantra refers to Alaṅkakalaśa’s opinion cited above, and pointing out the flaws in this argument, rejects the view relating the four Tantras with the four Varṇas (Hopkins 1977: 156). viii) Yajñopavīta In the Buddhist Tantras, there is a mention of yajñopavīta “the sacred thread” as one of the ornaments of the deities. In the Yoginīsañcāra Tantra (YST), Heruka, a manifestation of the Buddha, the main deity, his consort Vajravārāhī and other deities have been described and their attributes mentioned: All the deities, Śrīheruka and others are standing in the ālīḍha posture and are adorned with a necklace (kaṇṭhikā), a pair of bracelets (rucaka), ear-rings (kuṇḍala), a jewel(-like round-shaped ornament) on the head (śiromaṇi), the sacred thread (yajñopavīta) and ashes (bhasman). These ornaments are also called six mudrās.21 śrīherukādayas sarve ālīḍhāsanasaṃsthitāḥ | kaṇṭhikārucakakuṇḍalaśiromaṇivibhūṣitāḥ || yajñopavītaṃ bhasmeti mudrāṣaṭkaṃ prakīrtitam | YST 6. 12-13 (Pandeya 1998: 63-64). 21

414  Ù

The Vedas in Indian Culture and History

Tathāgatarakṣita, in his commentary Nibandha, explains the yajñopavīta as brahmasūtra “a Brahmanical thread” or mekhalā “a girdle” tied upon the waist (Pandey 1998: 63-64). In the context of Tantric sacrifice, the Vimalaprabhā commentary on the KT speaks of various types of Agni (vaiśvānara) and in that connection quotes a mantra from the R̥ gveda, (thinking that it is from Vedānta). It says: Here, Vaiśvānara is threefold: dakṣiṇāgni, gārhapatya, and āhavanīya. Here the dakṣiṇāgni is the lightning, the gārhapatya, the sun, and the āhavanīya, the kravyāda. The fourth is the eternal, gnosis-fire, possessing bliss as its quality. Therefore every offering is to be made below that (fire). It is also said in the Vedānta: “I send afar flesh-eating Agni, bearing off stains. May he depart to Yama’s subjects. But let this other Jātavedas carry oblation to the gods, for he is skilled.” 22

The verse quoted in the VP belongs to the hymn in the R̥ gveda (10.16.9) which, along with the next verse (R̥ V 10.16.10), is employed in the funeral rite to be performed in the event of the death of an Āhitāgni. According to Sāyaṇa, they are to be recited while the aupāsana fire is being extinguished. The relevance of this R̥ gvedic verse in the Buddhist Tantric ritual is difficult to understand. However, it denotes the element of the Brahmanical influence on the KT. ix) Buddhist and other tenets The Buddhist philosophical texts make a distinction between those belonging to their own tenets (bauddha) and the outsiders (bāhya), frequently termed tīrthikas, “heretics.” The Buddhist Tantric texts use the same terminology (Upadhyaya 1986: 195, 266, 270; Dviwedi and Bahulkar 1994b: 118). Sometimes this distinction is made with the terms svayūthya “belonging to their own group” and parayūthya “those from other groups.” The VT says that the svayūthyas have five great disciplines of learning (pañca mahāvidyā) and the parayūthyas fourteen branches of learning. The reference to the fourteen vidyās is obviously to those mentioned in an oft-quoted verse: “(The six) limbs (i.e. Vedāṅga texts), 22 iha vaiśvānaras trividhaḥ | dakṣiṇāgniḥ, gārhapatya, āhavanīya iti | dakṣiṇāgnir atra vidyut | gārhapatyaḥ sūryaḥ | āhavanīyaḥ kravyādaḥ | satyaś caturtho jñānāgnir ānandadharmā | atas asyādhaḥ sarva eva homaḥ kriyate | tathā vedānte cāha— kravyādam agnim prahiṇomi dūraṃ yamarājño gacchatu ripravāhaḥ | ihaivāyam itaro jātavedā devebhyo havyaṃ vahatu prajānan || (R̥ gveda 10.16.9) VP on KT 3.74-75 (Dviwedi and Bahulkar 1994a: 75).

Introduction Vedism and Brahmanism in Buddhist Literature

Ù  415

(four) Vedas, Mīmāṃsā, extensive Nyāya (system), Dharmaśāstra, and Purāṇa —these are the fourteen vidyās.”23 The enumeration of branches of knowledge as vidyā or kalā is an ancient Indian tradition, reflected in various Brahmanical texts, religious and secular as well. The Buddhist tradition too mentions them as five great or ten minor branches of learning. x) Buddha: origin of all knowledge In the Hindu tradition, the Tantras have been recognized as an authority and have been given the status of revelation (śruti). This is stated by Kullūka quoting Hārīta, in his gloss on the Manusmr̥ti (2. 1). He says: “Therefore Hārīta (says); ‘And therefore (we) explain Dharma. Dharma has the testimony of the Śruti. And Śruti is two-fold: Vedic and Tantric.’”24 The Hindu Tantras thus get recognition and the status equal to the Vedas. The Buddhist Tantras attempt to establish their supremacy over the Vedic as well as the other traditions, including the tenets of Śrāvakas. The HT (I. vii. 54) says: “With all the Vedas and the Siddhāntas, and also with proliferation of acts, there is no accomplishment, nor with (any act of) purification. …”25 The YST (12. 9-10) says: All the sciences, the Vedas and the Siddhāntas have arisen from this very doctrine (of tantra) and have survived. There is no accomplishment (of the goal) without this (tantra) either in this world or the yonder (world). This (tantra) is established (as) the only essence (of all knowledge).26

The commentary (nibandha) on this Tantra explains that the Vedas are the doctrines outside the Buddhist parlance and that the word siddhānta denotes the scriptures of the Śrāvakas and so on. The KT (5. 49ab) maintains that all Vedas and other doctrines have arisen from the Buddha, manifested in the form of aṅgāni vedāś catvāro mīmāṃsā nyāyavistaraḥ | dharmaśāstraṃ purāṇaṃ ca vidyās tv etāś caturdaśa || Cf. purāṇanyāyamīmāṃsādharmaśāstrāṅgamiśritāḥ | vedāḥ sthānāni vidyānāṃ dharmasya ca caturdaśa || Yajñavalkyasmr̥ti I.3 (Pāndey 1967: 3). 24 Ata eva Hārītaḥ- athāto dharmaṃ vyākhyāsyāmaḥ | śrutupramāṇako dharmaḥ | śrutiś ca dvividhā | vaidikī tāntrikī ca | Kulluka on Manusmr̥ti 2.1 (Nene 1970: 35). 25 samastavedasiddhāntaiḥ karmaprasaraṇādibhiḥ | siddhir na syād bhavec chuddhyā punarjanma bhavāntare || HT I. vii. 54 (Snellgrove 1959b: 30). 26 śāstrāṇi ca samastāni vedasiddhāntasaṃsthitāḥ | na ca tena vinā siddhir iha loke paratra ca || samastavedasiddhāntasāram ekaṃ pratiṣṭhitam || YST 12. 9-10 (Pandey 1998: 114). 23

The Vedas in Indian Culture and History

416  Ù

the deity with four mouths: “The Lord of Jinas proclaims the R̥gveda from his rear mouth; the Yajurveda from the right mouth; from the left, the Sāmaveda; and the Atharvaṇa, in the clan of Supreme Hari, from the mouth at the front…”27 The Tantra goes on enumerating various tenets other than the Vedas, namely, Kaula, Bhūtatantra associated with Garuḍa, Siddhānta, i.e., probably the Śaiva Siddhānta, and the Dharma (proclaimed by) Viṣṇu that have arisen from the respective mouths of the Buddha. (KT 5. 49cd; Dvivedi and Bahulkar 1994b: 34). In the same Tantra (1.156), it is said elsewhere that the gnosis-body of the Jina proclaims the past, the present, the future, and all the sciences beginning with the Vedas, logic and so on.’’28 xi) Buddha identified with Vedic and Brahmanical deities The Buddhist Tantras thus claim that the Buddha is the creator of all knowledge in this world. Moreover, they identify the Buddha with Brahmanical deities, namely, Brahmā, Viṣṇu, Śiva, and so on. The HT (I. v. 12-14) says:29 This unity is known as Vairocana, Akṣobhya, Amogha, Ratna, Ārolik, and Sāttvika, as Brahmā, Viṣṇu, Śiva, Sarva, Vibuddha, and Tattva.30 He is called Brahmā because he has gained nirvāṇa and Enlightenment; Viṣṇu because he is all-pervading; Śiva because he is propitious; Sarva because he abides in all things; Tattva because he experiences real bliss; and Vibuddha because he is aware of this happiness.31

R̥ gvedaṃ paścimāsyād api gadati yajur vāmavaktrāj jinendraḥ savyāsyāt sāmavedaṃ paramaharikule ’tharvaṇaṃ pūrvavaktrāt | KT 5. 49ab (Dvivedi and Bahulkar 1994b: 34) 28 etaiḥ sañcārarūpair avatarati vibhur jñānakāyo jinasya bhūtaṃ bhavyaṃ bhaviṣytat pravadati sakalaṃ vedatarkādiśāstram || KT 1.156 (Upadhyaya 1986: 153) 29 vairocanākṣobhyāmoghaś ca ratnārolik ca sāttvikaḥ | brahmā viṣṇuḥ śivaḥ sarvo vibuddhas tattvam ucyate || brahmā nirvr̥tito buddhaḥ viṣaṇād viṣṇur ucyate | śivaḥ sadā sukalyāṇāt sarvaḥ sarvātmani sthitaḥ || satsukhatvena tattvaṃ ca vibuddho bodhanād rateḥ | HT I. v. 12-14ab (Snellgrove 1959b: 16). 30 Snellgrove 1959b: 62, note 5: S (= Saroruha in Padmini, Padma can, Narthang Tenjur, xv. 142a-194b): equates the buddhas with their six Brahmanical counterparts, associating the five with the five wisdoms: Vairocana Brahmā ādarśajñāna, Amoghasiddhi Śiva kr̥tyānuṣṭhānajñāna, Akṣobhya Viṣṇu suviśuddhadharmadhātujñāna, Ratnasambhava Sarva samatājñāna, Amitābha Tattva pratyavekṣājñāna, Vajrasattva Vibuddha sahajānandamahāsukhapratīti, He defines Vibuddha as the special and noble Buddha. 27

Introduction Vedism and Brahmanism in Buddhist Literature

Ù  417

And also elsewhere (HT II. v. 37): “Their sets are Brahmā, Indra, Upendra, Viṣṇu, Śiva, Yama, Kubera, Nirr̥ti, and Vemacitrin.”32 Thus the HT identifies the Buddha with Brahmanical deities; however, it explains the meaning of the names Brahmā and others in a different way, so as to fit in the Buddhist context. The verses from the MNS using Brahmanical names have already been quoted at the beginning of this essay. The commentators use all their knowledge and talent to offer a new interpretation. For example, the Amr̥takaṇikā commentary on that text attempts to explain the epithets or the names of Mañjuśrī in a very complicated manner. It is difficult at this stage to understand the correct meaning of the glosses. However, it can be said that the author tries his best to show that the words do not carry the usual meanings known in the Brahmanical tradition. For instance, the word muñja is used in the sense of not allowing the semen (bodhicitta) to ejaculate. Thus, in the context of Vedic tradition, the word mauñjī would mean “one wearing a girdle made of the muñja-grass,” while the commentator of the Nāmasaṅgīti understands it in the sense of “one possessing the ability of controlling one’s bodhicitta.” The commentator tries to connect the word muñja with the root muñc “to release” and explains it on the basis of verbal, or rather phonetic similarity.33 The word brahman in the designation brahmavid has a peculiar meaning: it is the knowledge of the union of the emptiness (śūnyatā) and the compassion (karuṇā). Mañjuśrī is called Brahmā, as he has the four heads in the form of the four brahmavihāras and the four dhyānas. The word brahmanirvāṇa is obviously borrowed from the Bhagavadgītā (2. 72). The commentator simply derives the word as a genitive compound and explains it as “bliss” (ānanda). Further he quotes a passage from some text, “bliss is the form or manifestation of brahman” (ānando brahmaṇo rūpam iti vacanāt). All this explanation might seem farfetched to establish that the terms have a typically Buddhist Tantric meaning and not usual the Brahmanical one.34 31 Some errors in the transliteration in Snellgrove’s edition have been corrected. 32 brahmendropendrarudrāś ca Vivasvata Vināyakaḥ | nairr̥tir Vemacitrī ca Gauryādīnāṃ tu vistaram || HT II. v. 37 (Snellgrove 1959b: 80). 33 The verbal connection of the root muñc with the word muñja is also found in the Kauśika Sūtra (26.2), a major ritual text in the tradition of the Atharvaveda, where a mantra having the verbal form muñca is employed (muñca śīrṣktyā … Atharvaveda I.12.3). 34 The passage, the basis of the statement made above, is reproduced here: Mahāvrateti | skandha-dhātvindriyādiniḥsvabhāvīkaraṇaṃ mahāvrataṃ tad dharatīti tathoktaḥ | muñjaḥ, bodhicittācyavanaṃ tadyogān mauñjī | acyutabodhicittatvād eva brahmacārī | ata eva vratottamaḥ | … prakr̥tiprabhāsvaraś śūnyatākaruṇābhinnaṃ jñānaṃ

418  Ù

The Vedas in Indian Culture and History

All these references and passages point to the possibility of the influence of Vedic and Brahmanical tradition on Buddhist literature. This is not to say that the Buddhist had a direct connection with the Vedic and later Brahmanical tradition or that they borrowed this material directly from the Vedic and the post-Vedic texts. On the contrary, it rather appears that the Buddhist had some knowledge of that tradition in a general way and that they did not rely upon some specific Vedic or Brahmanical texts. In the case of the early Pāli canonical texts, the compilers of those texts were close to the late Vedic period or to the end of that period. It is not surprising that the Buddhist used the vocabulary, concepts, and practices that were commonly known or popular. As for the Mahāyāna, there appears growing influence of the then Brahmanical tradition. In the case of Buddhist Tantras, it appears that the authors lived in the surrounding of the Hindu Tantras, particularly the Śaiva Tantras. The rituals such as the maṇḍala visualization, the concept of worshiping the deity in the maṇḍala and visualizing oneself as the deity (śivo bhūtvā śivaṃ yajet “one should worship Śiva, having become Śiva”), the vocabulary rich with ritualistic terms and so on, all this appears very close to the Śaiva tradition. It is possible to say therefore that the Buddhist Tantras had come under the influence of the Śaiva Tantric tradition, probably the Kashmir Śaivism, and not the Vedic or Brahmanical tradition. These concepts, common to Brahmanism and Śaivism, might have been used equally by the Buddhists. There was probably a common stock for various religious traditions. In this connection, it may not be out of place to make a brief reference to differing views on the interaction between the Buddhist and the Śaiva Tantricists. Alexis Sanderson has pointed out that the Buddhist Tantras, particularly the Yoginī Tantras, have borrowed considerable material from the Śaiva tradition. In his essay “Vajrayāna: Origin and Function,” he argues that “there is even direct borrowing of passages from the Śaiva texts” (Sanderson 1994: 94-95). In another essay, “History through Textual Criticism in the study of Śaivism, the Pāñcarātra and the Buddhist Yoginītantras,” he shows that … in the Laghusaṃvara (Tib. Bde mchog ñuṅ ṅu), which the Buddhist tradition identifies as the root-text (mūlatantram) of this corpus, long passages, amounting to some 200 verses, nearly a third of the whole, can be seen to have been redacted from brahma tattādātmyena vetty anubhavatī ti bramavit | pañcākṣararūpatvād haṃkāraḥ pañcatathāgatātmako brāhmaṇas tannādam āgamya yāvad uṣṇīṣalayena sarvavikalpatvaṃ vāhayatī ti brāhmaṇāḥ | ākāśāsaktacittatayāpratyāhārādiṣaḍaṅgasaṅkṣepa-caturaṅgabrahmavihāracaturdhyānacaturmukhasvabhāvatvād brahmā | brahmaṇo nirvāṇam

brahmanirvāṇam ānandaḥ | ānandaṃ brahmaṇo rūpam iti vacanāt | Amr̥takaṇikā on MNS 8. 18-19 (Banarsilal 1994: 67).

Introduction Vedism and Brahmanism in Buddhist Literature

Ù  419

Śaiva originals found in texts of the Vidyāpīṭha division, namely the Picumata, the Siddhayogeśvarīmata, the Tantrasadbhāva, and the Yoginīsañcāraprakaraṇa of the 3rd Ṣaṭka of the Jayadrathayāmala, or perhaps, from closely related version of these text-passages that are no longer accessible. (Sanderson 2001: 41-42) Recently, some scholars have attempted to refute his arguments. In his book Indian Esoteric Buddhism: A Social History of the Tantric Movement, Ronald Davidson points out the flaws in Sanderson’s assumptions and states that “decisions about textual borrowing are best made case by case, rather than corpus by corpus,” and that “a reciprocal appropriation model … will prove the most fruitful” (Davidson 2002: 386 n. 105). In the introduction to his study of the CST, David B. Gray accepts that the Buddhist Yoginītantras were significantly influenced by Śaiva Kāpālika practices, but says that the complex relationship between Śaiva and Buddhist Tantric traditions will only be determined conclusively once all of the surviving texts have been critically edited and published” (Gray 2007: 8-9). In this connection, it is important to refer to a recent monograph of Ruegg, The Symbiosis of Buddhism with Brahmanism/ Hinduism in South Asia and of Buddhism with “local cults” in Tibet and the Himalayan Region, where he discusses at length the question of “borrowing and substratum models for religious syncretism and/or symbiosis” (Ruegg 2008: 105-114) and refutes Sanderson’s view. Having analyzed Sanderson’s evidence, he says: It is, however, anything but evident how such a generalized conclusion is to be drawn from the kind of evidence adduced by Sanderson in his article of 1990/1994, or indeed from an accumulation of similar pieces of such evidence. Mere quantity of this kind can scarcely cancel out the view that Buddhism, having arisen in an Indian matrix and milieu, already shared common ground with Brahmanism/Hinduism. (Ruegg 2008: 107)

Sanderson has reiterated his stand in his latest article, “The Śaiva Age: The Rise and Dominance of Śaivism During the early Medieval Period” (Sanderson 2009). Having furnished more evidence, he clarifies his position. According to him, his intention “is not that the Śaiva parallels are the direct sources of the Buddhist versions but only that the Śaiva parallels are close enough to the Buddhist versions to reveal the direction of dependence” (Sanderson 2009: 191-192). 6. Conclusion We may put aside for the time being the yet inconclusive debate over the textual dependence between the Śaiva and the

420  Ù

The Vedas in Indian Culture and History

Buddhist Tantric tradition. What is important for the theme of the present essay is that there is seen a tendency toward Vedism and Brahmanism throughout the Buddhist literature, right from the early Pāli canon through the Mahāyāna to the late Buddhist Tantric texts. While borrowing the Vedic and Brahmanical vocabulary, concepts and ritual practices, the Buddhists did not necessarily adhere directly to particular traditions or texts. The proportion of the usage of such vocabulary and ritualistic practices has increased in the Mahāyāna and, more prominently, in late Buddhist Tantric tradition that involved the muttering of various mantras, offerings into fire and other practices, resembling the Vedic and Brahmanical sacrificial ritual. Abbreviations A AVJ BSOAS CST D HR HT JAOS KT M MCB MNS SaṁT Sn SUT VP YST

Aṅguttaranikāya. See Warder 1964. Advayavajrasaṃgraha. See Shastri 1927. Bulletin of the School of Oriental and African Studies. Cakrasaṃvara Tantra. See Pandey 2002. Dīghanikāya. See Rhys Davids and Carpenter 1975. History of Religions. Hevajra Tantra. See Snellgrove 1959a & 1959b. Journal of American Oriental Society. Kalacakra Tantra. See Upadhyaya 1986; Dviwedi and Bahulkar 1994a & 1994b. Majjhimanikāya. See Trenckner 1964. Mélanges Chinois et Bouddhiques. Mañjuśrīnāmasaṅgīti. See Davidson 1981. Saṃvarodaya Tantra. See Tsuda 1974. Suttanipāta. See Andersen and Smith 1965. Sekoddeśaṭīkā. See Sferra and Merzagora 2006. Vimalaprabhā. See Upadhyaya 1986; Dviwedi and Bahulkar 1994a & 1994b. Yoginīsañcāra Tantra. See Pandey 1998.

References Andersen, Dines and Helmer Smith (eds.). 1965. Sutta-Nipata. 1st edition 1913. London: Luzac & Co. (Published for the Pali Text Society). Bagchi, Sitansusekhar (ed.). 1970. Mahāyānasūtrālaṅkāra of Asaṅga. Buddhist Sanskrit Texts No. 13. Darbhanga: The Mithila Institute of Postgraduate Studies and Research in Sanskrit Learning.

Introduction Vedism and Brahmanism in Buddhist Literature

Ù  421

Bahulkar, Shrikant S. 1985. “Bauddha Stotra Vāṅmaya (Marathi).” In: Puruṣārtha, Special Issue on Stotras (Oct.-Dec.), 494-499. Banarsilal .(ed.). 1994. Ārya-Mañjuśrī-nāmasaṅgīti with Amr̥takaṇikā Ṭippaṇī by Bhikṣu Raviśrījñāna and Amr̥takaṇikoddyotanibandha of Vibhūticandrapāda. Sarnath: Central Institute of Higher Tibetan Studies. Banerjee, P. 1989. “The Story of the Birth of Gautama Buddha and its Vedic Parallel.” In: Buddhist Iconography 6-18. Delhi: Tibet House. Bloomfield, Maurice. 1889. The Kauśika Sūtra of Atharvaveda with Extracts from the Commentaries of Dārila and Keśava. JAOS 14. Davidson, Ronald M. (ed. and transl.) 1981. “The Litany of Names of Mañjuśrī. Text and Translation of the Mañjuśrīnāmasaṅgīti.” In: Tantric and Taoist Studies in Honour of R.A. Stein. Mélanges chinois et bouddhiques 20. Bruxelles: Institut Belge des hautes études chinoises, 1-69. ———. 2002. Indian Esoteric Buddhism: A Social History of the Tantric Movement. New York: Columbia University Press. Dharmadhikari, Trivikram Narayan. 2006. “Vedic Sacrifice and Buddhism”. In: Madhavi Kolhatkar and Pradnya Deshpande (eds.). Saṁvijñānam: Select Articles of Dr. T.N. Dharmadhikari. Pune. Dwivedi, Vraj Vallabha and Shrikant S. Bahulkar (eds.). 1994a. Vimalaprabhāṭīkā of Kalkin Śrī Puṇḍarīka on Śrī Laghukālacakratantrarāja by Śrī Mañjuśrīyaśas. Vol. II. Rare Buddhist Texts Series No. 12. Sarnath: Central Institute of Higher Tibetan Studies. ———. (eds.). 1994b. Vimalaprabhāṭīkā of Kalkin Śrī Puṇḍarīka on Śrī Laghukālacakratantrarāja by Śrī Mañjuśrīyaśas. Vol. III. Rare Buddhist Texts Series No. 13. Sarnath: Central Institute of Higher Tibetan Studies. Edgerton, Franklin. (ed.). 1953. Buddhist Hybrid Sanskrit Grammar and Dictionary. Vol. II: Dictionary. New Haven: Yale University Press. Falk, Harry. 1988. “Vedische Opfer im Pāli-Kanon.” Bulletin des études indiennes 6, 225-254. Gray, David B. (ed. and transl.). 2007. The Cakrasamvara Tantra (The Discourse of Śrī Heruka), Śrīherukābhidhāna: A Study and Annotated Translation. Treasury of the Buddhist Sciences Series. New York: American Institute of Buddhist Studies at Columbia University. Heesterman, Jan C. 1987. “Vedism and Brahmanism.” In: M. Eliade (ed.), Encyclopedia of Religion. Vol. 15. NewYork: Macmillan, 217-242. Hopkins, Jeffery. 1977. Tantra in Tibet: The Great Exposition of Secret Mantra by Tsong-ka-pa. London: George Allen & Unwin.

422  Ù

The Vedas in Indian Culture and History

Kashyap, Bhikkhu J. (ed.). 1958. Dīghanikāya. Vol. I. Nālandā Devanāgarī Pāli Series. Patna: Pāli Publication Board. ———. (ed.). 1960. Aṅguttaranikāya. Vol. I. Nālandā Devanāgarī Pāli Series. Patna: Pāli Publication Board. Lindtner, Christian. (ed.) 2001. Madhyamakahr̥dayam of Bhavya. Adyar Library series V. 123. Adyar, Chennai: Adyar Library and Research Centre. Malalasekera, George Peiris (ed.). 1974. Dictionary of Pāli Proper Names. Vols. I and II, Indian Texts Series. 1st edition, 193738. Repr. London: Pali Text Society. Nene, Gopāl Śāstrī. (ed.). 1970. The Manusmr̥ti with the Manvarthamuktāvalī Commentary of Kullūka Bhaṭṭa with the Maṇiprabhā Hindi Commentary by Pt. Haragovinda Śāstrī. Varanasi: The Chowkhamba Sanskrit Series Office. Pande, Govind Chandra. 1983. Studies in the Origins of Buddhism. 1st edition 1957. Repr. Allahabad: Department of Ancient History, Culture and Archaeology, University of Allahabad. Pandey, J. (ed.) 1998. Yoginīsañcāratantram with Nibandha of Tathāgatarakṣita and Upadeśānusāri ṇīvyākhyā of Alakakalaśa. Rare Buddhist Texts Series No. 21. Sarnath: Central Institute of Higher Tibetan Studies. ———. (ed.). 2002. Śrīherukābhidhānaṃ Cakrasaṁvaratantram with the Vivr̥ti of Bhavabhaṭṭa, Vols. I and II. Rare Buddhist Texts Series No. 26. Sarnath: Central Institute of Higher Tibetan Studies. Pandey, Umesh Chandra. (ed.). 1967. Yājñavalkyasmr̥ti of Yogīśvara Yājñavalkya with the Mitākṣarā Commentary of Vijñāneshwar, edited with the “Prakash” Hindi Commentary. Varanasi: The Chowkhamba Sanskrit Series Office. Rhys Davids, Thomas William. 1969. Dialogues of the Buddha. Part I: Sacred Books of the Buddhists. Vol. II: 1st Edition 1899, 2nd Repr. London: Luzac & Co. Rhys Davids, Thomas William and J. Estlin Carpenter. (eds.). 1975. The Dīgha Nikāya. Vol. I. 1st edition, 1890. 4th Repr. London: Pali text Society. Rhys Davids, Thomas William and William Stede. (eds.). 1993. Pāli-English Dictionary. 1st edition, 1921-25. London: Pāli Text Society. 1st Indian Edition, Delhi: Motilal Banarsidass Rinpoche, Samdhong and Vraj Vallabh Dwivedi. (eds.). 1990. Vasantatilaka of Caryāvratī Śrīkr̥ṣṇācārya, with the Commentary Rahasyadīpikā by Vanaratna. Rare Buddhist Texts Series No. 7. Sarnath: Central Institute of Higher Tibetan Studies. Ruegg, David Seyfort. 2008. The Symbiosis of Buddhism with Brahmanism / Hinduism in South Asia and of Buddhism with “local cults” in Tibet and the Himalayan region. Österreichische Akademie der Wissenschaften, Philosophisch-Historische Klasse Sit-

Introduction Vedism and Brahmanism in Buddhist Literature

Ù  423

zungsberichte, 774, Band. Wien: Verlag der Österreichischen Akademie der Wissenschaften (OAW). Sanderson, Alexis. 1994. “Vajrayāna: Origin and Function.” In: Buddhism into the Year 2000: Proceedings of an International Conference held in Bangkok in 1990, 87-102. Bangkok–Los Angeles: Dhammakaya Foundation. (Also available at: ). ———. 2001. “History through Textual Criticism in the Study of Śaivism, the Pāñcarātra and the Buddhist Yoginītantras.” In: François Grimal (ed.), Les sources et le temps. Pondichery: École française d’Extrême Orient, 1-47. ———. 2009. “The Śaiva Age: The Rise and Dominance of Śaivism during the Early Medieval Period.” In: Shingo Einoo (ed.), Genesis and Development of Tantrism. Institute of Oriental Culture Special Series, 2. Tokyo: Institute of Oriental Culture, University of Tokyo, 341-350. Schlingloff, Dieter. 1955. Buddhistische Stotras aus Ostturkistanischen Sanskrittexten. Berlin: Akademie Verlag. Sferra, Franscesco and Stefania Merzagora (eds.) 2006. The Sekoddeśaṭīkā by Nāropā (Paramārthasaṃgraha). Roma: Istituto Italiano per l’Africa e l’Oriente. Shackleton Bailey, David Roy (ed. and transl.). 1950. “The Varṇārhavarṇa Stotra of Mātr̥ceṭa (I).” BSOAS 13, No. 3, 671-701. ———. (ed.). 1951. “The Varṇārhavarṇa Stotra of Mātr̥ceṭa (II).” BSOAS 13, No. 4, 947-1003. Shastri, Hara Prasad. (ed.) 1927. Advayavajrasaṃgraha. Gaekwad’s Oriental Series 41. Baroda: Oriental Institute. Snellgrove, David Llewellyn (transl.). 1959a. Hevajratantra. Part I. London: Oxford University Press. ———. (ed.). 1959b. Hevajratantra. Part II. London: Oxford University Press. Thomas, Edward J. 1949. The Life of Buddha as Legend and History. London: Routledge & K. Paul, Ltd. Trenckner, Vilhelm. 1964. The Majjhima-Nikāya. Vol. I. 1st edition, 1888. Repr. London: Luzac & Co. (Published for the Pali Text Society). Tripathi, Shridhar. (ed.). 1987. Lalitavistara. Buddhist Sanskrit Texts No. 1. 1st edition by P.L. Vaidya. Darbhanga: The Mithila Institute of Postgraduate Studies and Research in Sanskrit Learning. Tsuda, Shinichi. (ed. and transl.). 1974. Saṃvarodaya-tantra (= Śrīmahāsaṁvarodayatantra), Selected Chapters. Tokyo: The Hokuseido Press. Upadhyaya, Jagannatha. (ed.). 1986. Vimalaprabhāṭīkā of Kalki Śrī Puṇḍarīka on Śrī Laghukālacakratantrarāja by Śrī Mañjuśrīyaśa

424  Ù

The Vedas in Indian Culture and History

Vol. I. Bibliotheca Indo-Tibetica Series No. 11. Sarnath: Central Institute of Higher Tibetan Studies. Vaidya, Paraśurāma Lakṣmaṇa. 1960. (ed.). Saddharmapuṇḍarīkasūtra. Buddhist Sankrit Texts No. 16. Darbhanga: The Mithila Institute of Postgraduate Studies and Research in Sanskrit Learning. Walsche, Mourice. (transl.). 1987. Long Discourses of the Buddha: A Translation of the Dīghanikāya. Repr. New York: Wisdom Publications. Warder, Anthony K. 1961. The Aṅguttara-Nikāya. Part I. 1st edition by Richard Morris 1885. 2nd edition. London: Luzac & Co. (Published for the Pali Text Society). Wayman, Alex. 1961. “Totemic Beliefs in the Buddhist Tantras.” HR 1, No. 1, 81- 94. ———. 1977. Yoga of the Guhysamājatantra. 1st Indian edition, Delhi: Motilal Banarsidass. ———. (ed.). 1985. Chanting the Names of Mañjuśrī, The Mañjuśrīnāmasaṅgīti. Sanskrit and Tibetan texts. Boston– London: Shambhala Publications. Witzel, Michael. 1993. “Toward A History of the Brahmins.” JAOS 113, No. 2, 264–268. (Review article of Migrant Brahmins in Northern India: Their Settlement and General Impact c. A.D. 475-1030 by Swati Datta [née Sen Gupta]. Delhi: Motilal Banarsidass, 1989). Woodward, Frank Lee. (ed.). 1932. The Book of the Gradual Sayings. Vol. I. London: Pali Text Society.

ntroduction EIpic Aśvamedhas

Ù  425

Epic Aśvamedhas A lf Hiltebeitel George Washington University

My goal is to convince readers that the impregnations of the Kuru queens Ambikā and Ambālikā’s by the Mahābhārata’s author Vyāsa allude to the Aśvamedha scene where the chief queen or mahiṣī lies with the sacrificial horse. To put it bluntly, Vyāsa should come out smelling something like a horse. The case calls for four quick moves. The first is to recognize that the Aśvamedha scene most susceptible to veiled allusion is this very one. The second is to explore a suspicion that one way to allude knowingly to it would be to shift planes from the Aśvamedha to other rites where a woman is called upon to secure offspring outside marriage: especially via niyoga, with a live man rather than a dead horse. The main move is to note a symmetry where the two epics treat four episodes: three where Aśvamedhas occur in the main narratives, and the fourth, our target: ostensibly a niyoga or “levirate” story with an anti-niyoga sequel. In each case, we must keep track of the horse, the mahiṣī, and three or so other queens. The four episodes are quadrangulated in the following table: Table 1 Rām Book 1 Aśvamedha: Daśaratha’s Aśvamedha supplemented by a Putrīya Iṣṭi performed by R̥śyaśr̥ṅga, which enables the births of Rāma and his three brothers (Rām 1.8-17)

Book 7 Postwar Aśvamedha of Rāma (Rām 7.75-89)

426  Ù

The Vedas in Indian Culture and History

Mbh Book 1 Crypto-Aśvamedha scenes: Vyasa’s niyoga with the two Kāśi princesses that enables the births of Dhr̥tarāṣṭra and Pāṇḍu (Mbh 1.99100) (Sequel) Kuntī’s anti-niyoga tale of King Vyuṣitāśva told before she reveals her mantra, which enables the births of the five Pāṇḍavas (Mbh 1.112)

Book 14 Postwar Aśvamedha Of Yudhiṣṭhira (Mbh 4.87-91)

Finally, a fourth move is to ask, Who would these knowing allusions be known to? Here, my first and only necessary candidates are the epic poets, who could, moreover, be familiar with the way that R̥ gvedic poets handled similar things —if, as I do, we may believe Stephanie Jamison (1996: 74-88) that the potent monkey Vr̥ṣākapi is a veiled Aśvamedha horse in R̥ V 10.86, and Joel Brereton (2002) that the race of Mudgala and Mudgalānī is a veiled complement to a niyoga ritual in R̥ V 10.102. Moreover, since I believe the Mahābhārata is a little earlier than the Rāmāyaṇa, I think we may suspect that in squaring off our symmetry, Vālmīki would have understood the Aśvamedha overtones of our main episode and its sequel. He seems to make economy of a number of the vaster epic’s structural features to tell a far less ambiguous tale about a perfect man and woman.1 Granted, these four moves will be a little too quick. But considering the alternatives, better to be brief than labored. I will first treat the two postwar Aśvamedhas and then the two Book 1 scenes. A. Postwar Aśvamedha of Rām 7.75-89 Rāma chooses a postwar Aśvamedha because Bharata, one of his brothers, says it is less destructive than a Rājasūya, and Lakṣmaṇa, another of his brothers, says that it removes all sins (Rām 7.75.2).2 Lakṣmaṇa’s point recalls the rationale for Yudhiṣṭhira’s postwar Aśvamedha.3 But in Rāma’s case no explicit sins are mentioned, 1 See Hiltebeitel 2005a: 460-461 on the two epics’ similar structural features, and further discussion in Hiltebeitel 2011: Ch. 5. 2 It is curious that Rāma considers a postwar Rājasūya, having never, of course, performed one. Bharata’s point recalls the conversation in Mbh Book 2 where Yudhiṣṭhira chooses the Rājasūya over the Aśvamedha, despite Nārada’s warning of its greater danger. See Gehrts 1975, Hiltebeitel 1977. 3 Yudhiṣṭhira seeks benefit from the Aśvamedha’s sin-cleansing property because he suffers from massive war-guilt, Mbh 14.70.15-16; see Hiltebeitel [1976] 1990: 292, Jamison 1996: 76.

ntroduction EIpic Aśvamedhas

Ù  427

other than that Lakṣmaṇa tells how Indra’s Aśvamedha removed his Brahmanicide of killing Vr̥tra.4 Rāma orders Bharata to lead a procession trailed by all their mothers (who come from the inner quarters) and “my golden wife” (kāñcanīm mama patnīm) worthy of dīkṣā in sacrificial rites” (82.19ab). “All the mothers” must include Kausalyā, Kaikeyī, and Sumitrā. These three widows seem extraneous to Rāma’s Aśvamedha. But since Rāma is monogamous, they come in handy as shadow figures of the three subordinate Aśvamedha queens who would accompany the mahiṣī, who has been replaced by the golden statue. Even while Sītā is still alive in Vālmīki’s hermitage, Rāma has had this statue of her prepared, it seems, for this very rite. Rāma then assigns Lakṣmaṇa to follow the horse,5 and the last we hear of them is that Lakṣmaṇa protected it while it wandered (83.9). With the spotlight on Rāma’s hearing the Rāmāyaṇa,6 one never hears that Lakṣmaṇa or the horse returns!7 Further horse news comes only after Sītā vanishes into the earth.8 During Rāma’s ten thousand year reign, “in sacrifice after sacrifice there was a golden Jānakī for the sake of a wife” (89.4cd). He performed at least one more horse sacrifice, ten Vājapeyas, and other Śrauta rites while dedicating himself to dharma (89.1-7)— presumably using the same golden Sītā. Clearly, having replaced Sītā with a statue and forgotten the horse, Vālmīki has no interest in having Sītā lie down with a dead horse. Rāma’s continued use of a golden Sītā also seems to help him get over his despair at her loss. And from Sītā’s point of view, vanishing was no doubt better than this one more indignity that an Aśvamedha would have required. Her mind on Rāma, her last words ask the earth to engulf her (88.10) with no thought about a horse. 4 This could, however, like the threat of the Brahman carrying his dead son brought about by the tapas of Śambūka (7.64), allude to Rāma’s Brahmanicide in killing Rāvaṇa, a Brahman descendent of the sage Pulastya. Rāma now tells Lakṣmaṇa to gather Vasiṣṭha and other court Brahmans to officiate, to invite all their monkey and Rākṣasa friends, and the great R̥ṣ is together with their wives, and to prepare a vast sacrificial enclosure at Naimiṣa Forest (7.82.14). 5 He goes with his other two brothers to the sacrificial site (83.1-2). 6 And his eventual realization that the two boys reciting it are his sons. For now they are called “sons of the Muni.” The phrase munidārakau (dāraka meaning “boy, child, son”) is used thrice rapidly (7.85.9d, 17d, and 19b) just after Vālmīki has told the boys, “If Kakutstha [Rāma] should ask, ‘Whose two children (dārakau) are you?’ you may tell the lord of men so: ‘Just the disciples of Vālmīki (vālmīker atha śiṣyau hi brūtām evaṃ narādhipam).’” 7 It is Bharata whom Rāma addresses at 7.85.12-13, telling him to pay the twin singers, and not Lakṣmaṇa as the Shastri translation has it (1970: 3.612). 8 It does say here that “the sacrifice was concluded” (avasāne).

428  Ù

The Vedas in Indian Culture and History

B. Postwar Aśvamedha of Mbh 14.87-91 Unlike Lakṣmaṇa’s disappearance with the horse, Arjuna’s return with one is “marked”; unlike Sītā, Draupadī does seem to think about the horse;9 and unlike Rāma, Yudhiṣṭhira has explicit sins.10 As we enter the proceedings, Kr̥ṣṇa has just told Yudhiṣṭhira that Arjuna is returning, emaciated after many fights. Yudhiṣṭhira is glad to hear of Arjuna’s imminent return. But troubled by the report of his gaunt look, he asks Kr̥ṣṇa whether Arjuna bears some “unfavorable11 —aniṣṭa— mark by which he experiences such miseries (aniṣṭaṃ lakṣaṇaṃ kr̥ṣṇa yena duḥkhāny upāśnute)” (14.89.4ef). Ever the entertainer, Kr̥ṣṇa replies:12 O king, I surely do not detect this one’s having anything unfavorable except that this lion among men’s cheekbones13 are overly developed (na hy asya nr̥pate kiṃcid aniṣṭam14 upalakṣaye | r̥te puruṣasiṃhasya piṇḍike ’syātikāyataḥ). On account of these two, this tiger among men is always hitched to the roads (nityam adhvasu yujyate). I do not see anything else by which this Jaya has a share of misery. (14.89.7-8)

Yudhiṣṭhira is satisfied. 9 For an earlier treatment of his section, see Hiltebeitel 2007: 126-35. Most of the linguistic points are unchanged, but the comparative and intertextual points made here are new. 10 See Mbh 12.27.4-22, where he mentions his part in the slayings of his grandfather Bhīṣma, his guru Droṇa, his brother Karṇa, and the Pāṇḍavas’ princely children. See discussions in Hiltebeitel 1976 [1990]: 240-241, 2001: 6669, 2005b, 2011: Ch. 5. 11 For an-iṣṭa: “unwished, undesirable, disadvantageous, unfavorable; bad, wrong, evil, ominous” (Monier-Williams [1899] 1964: s.v.). 12 Kr̥ṣṇa speaks only after “reflecting for a very long interlude (dhyātvā sumahad antaram)” (6b). See 12.29.6, where Kr̥ṣṇa, speaking to Yudhiṣṭhira “disarmingly (abhivinodayan)” (Fitzgerald 2004: 228) or “entertainingly,” launches into amusing stories, among them that of Svarṇasthīvin, “Excretor of Gold,” in the Nārada-Pārvata-Upākhyāna (12.30), to begin to dissolve some of Yudhiṣṭhira’s postwar grief. See Hiltebeitel in 2005b: 254-255 on this passage, and on the generally gloomy and depressed character of Yudhiṣṭhira during the Aśvamedha, see Jamison 1996: 76, to which 277 n. 151 adds, with only slight but, as I am trying to suggest, significant exaggeration, “This, of course, is true for all of the postwar Mbh. The Pāṇḍavas seem to live in a state of clinical depression for parvan after parvan.” Biardeau 2002: 2.631-632, begins her discussion of this exchange noting that Kr̥ṣṇa is “le grand illusioniste,” and takes it as an ironic reference to the daiva (fate as divinely ordained) and to Kr̥ṣṇa’s relation to Arjuna as Nara. 13 Dumézil 1970: 164 n. 9: “Piṇḍikā, which designates ‘a globular swelling or protuberance,’ here certainly has the meaning ‘cheekbone.’” 14 Instead of aniṣṭam, the Vulgate (14.87.8b) reads saṃśliṣṭam, on which Nīlakaṇṭha suggests something confounded, bristling, fleshy, and extending broadly and from behind (Kinjawadekar 1929-33, Āśvamedhikaparvan: 6.119).

ntroduction EIpic Aśvamedhas

Ù  429

But Kr̥ṣṇā Draupadī indignantly glanced askance at Kr̥ṣṇa (kr̥ṣṇā tu draupadī kr̥ṣṇaṃ tiryak sāsūyam15 aikṣata). The Slayer of Keśin, Hr̥ṣīkeśa, approved that showing of her affection (pratijagrāha tasyās tam praṇayam) as if he were Dhanaṃjaya in person (sākṣād iva),16 a friend of a friend (sakhyuḥ sakhā). Having heard, Bhīma and the other Kurus and Yādavas there took pleasure in this story about Dhanaṃjaya whose manner was amusing17 (remuḥ śrutvā vicitrārthā dhanaṃjayakathā vibho). (14.89.10-11)

Note that the Vulgate reads yājakās, “sacrificial priests,”18 rather than yādavās, which would give us, “... the yājakas there took pleasure” in Kr̥ṣṇa’s story. What about this sakhyuḥ sakhā, “friend of a friend”? Who is whose friend? Georges Dumézil, discussing the facial contortions of Indo-European warriors, takes it as referring to Kr̥ṣṇa as the friend of Arjuna.19 This is plausible, since sakhyuḥ is masculine. But K.M. Ganguli takes it to refer to Kr̥ṣṇa (the sakhā) as the friend of Draupadī.20 Somewhat in favor of this reading, the preceding genitive tasyās, in “showing of her affection,” could point to Draupadī in the genitive masculine sakhyuḥ, which three Malāyalam manuscripts21 replace with sakhyāḥ to make it explicit that it means “of a female friend.” Actually, it is ambiguous, for as Whitney observes, “forms of sakhi are sometimes found used with feminine value” ([1889] 1960: 342).22 But Draupadī could have more to be indignant about than just this slight of Arjuna, whose return with the horse signals that she 15 For sāsūya, Ganguli [1884-96] 1970: 12, has “angrily,” which I would regard as the next best thing in this context. 16 Visibly, really, actually; with his own eyes. 17 Whose goal was varied; whose concern, manner, or meaning was charming, entertaining, amusing, colorful. 18 Mbh 14.87.12d; Kinjawadekar 1929-33: Āśvamedhikaparvan: 6.119. 19 Dumézil writes, “Draupadī, ... who has a preference for Arjuna, ... takes strong exception to a challenge of this kind to the hero’s perfect beauty; she throws an angry glance at Kr̥ṣṇa, who, in his own affection for Arjuna, enjoys her feminine reaction” (1970: 164 n. 9). He would probably have in mind, behind this, the Vedic precedent of Viṣṇu’s being the “intimate friend of Indra” (índrasya yújyaḥ sákhā; R̥V 1.22.19). 20 While reminding us in parentheses that Arjuna is also Kr̥ṣṇa’s friend. He translates, “The slayer of Kesi, viz., Hrishikesa, approved of that indication of love (for his friend) which the princess of Panchala, who also was his friend, displayed” —and adds in a note: “It is worthy of note that Draupadi was always styled by Krishna as his ‘sakhi’ or ‘friend.’” (Ganguli [1884–96] 1970: 12.165-166 and 166 n. 1). 21 Those called M1-3 in the apparatus of the Pune Critical Edition. 22 Moreover, since Kr̥ṣṇa sees Draupadī “as if he were Dhanaṃjaya in person (sākṣād iva),” it could even be saying that he is sympathetically seeing what Arjuna would be seeing as the friend in the nominative, leaving the genitive of whom Arjuna is friend to be either Draupadī or Kr̥ṣṇa.

430  Ù

The Vedas in Indian Culture and History

must now ramp herself up for this Aśvamedha highlight: When the bulls among yājakas had made the horse agree [i.e., killed it], they caused Drupada’s wise daughter to lie down beside it for three minutes (kalābhis tisr̥bhiḥ) according to rule. (14.91.2)23

Vyāsa is a supervising priest at both Yudhiṣṭhira’s Rājasūya and his Aśvamedha, and would thus be one of the chief “bulls among yājakas” present.24 It is also striking that Kr̥ṣṇa’s friendship with Draupadī comes into play explicitly only through two scenes of royal ritual: one, his response to her calling on him at her disrobing, itself an outcome of the dice match as a Rājasūya sequel; and this one where she is the mahiṣī exposing herself to the horse. I suspect that in each case Kr̥ṣṇa intervenes to lighten Draupadī’s sexual humiliation and attenuate her role as mediatrix and victim within the arena of the great Vedic royal sacrifices. For as Jamison (1996: 256) demonstrates, the “sacrificed” sacrificer’s wife plays the role of mediator between men and gods. This may be the Vedic ground of Draupadī and Kr̥ṣṇa’s friendship.25 As we have seen, Vālmīki spares Sītā such complications, for which Rāma is ill-equipped to intervene like Kr̥ṣṇa. But what about the horse? Kr̥ṣṇa seems to be referring not only to Arjuna but to the horse. If he is reading into his “friend” Draupadī’s mind, he would be hinting that Arjuna and the horse have a somewhat fused profile as they approach together. While piṇḍikā can designate ‘a globular swelling or protuberance’ on a man’s cheeks, I dare say it could also describe the same on a horse. For instance, although it is for piṇḍaka rather than piṇḍikā, Monier-Williams [1899] 1964 gives “a round protuberance (esp. on an elephant’s temples).”26 Kr̥ṣṇa’s bon mot would thus be a śleṣa: one appreciated by the Yādavas who, as Kr̥ṣṇa’s kinsmen, would know him better than most. Or alternatively, it would be especially amusing to the Yājakas, who are thinking not only of 23 This is mostly as it in Jamison’s translation, which is given with the comment that, “Three minutes sounds about right,” adding, “For whatever reason, at the Aśvamedha depicted in the Rāmāyaṇa a whole night is required” (Jamison 1996: 66), and offering a detailed description of what the traditional ritual would expect of Draupadī (66, 68-69). 24 See Sullivan 1990: 31-34; Hiltebeitel 2001: 50-51, 77-79. 25 She has a new friend to turn to in the text’s bhakti theology: God. But really, he is not a new friend but an old one, for as the earlier Vedic Brāhman. a texts are fond of repeating, yajño vai viṣṇuḥ, “Viṣṇu is the sacrifice” (see Hiltebeitel [1976] 1990: 105, 293-294, 333, 356). On Draupadī and Kr̥ṣṇa’s “unique friendship,” see Hiltebeitel 2007: 110-17. 26 Also, for piṇḍikā, f. a globular fleshy swelling (in the shoulders, arms, legs, &c., esp the calf of the leg [MW 625]). Similar meanings also apply to piṇḍa (idem).

ntroduction EIpic Aśvamedhas

Ù  431

Arjuna but more so of the horse he is bringing to be sacrificed. Indeed, the bull among Yājakas Vyāsa would be the first to understand such doubletalk. As “author,” Vyāsa not only stands behind everything Vaiśampāyana says, but is the chief of the priests overseeing the whole Aśvamedha. It is curious that this especially “knowing” variant is found only in Nīlakaṇṭha’s Vulgate. The narration and Kr̥ṣṇa’s own words offer further equine clues. Kr̥ṣṇa’s name Slayer of Keśin is curious here, since Keśin is a horse Kr̥ṣṇa slew in his childhood. The name Hr̥ṣīkeśa can mean “Master of the Senses” (Biardeau 2002: 1, 595) and, with that, convey the familiar Upaniṣadic image of “yoking” the senses like horses.27 But most tellingly, Kr̥ṣṇa says that Arjuna’s facial protuberances come from his being “always hitched to the roads (nityam adhvasu yujyate)” like a family workhorse. If so, no wonder Draupadī looks at him askance. Note that Draupadī goes it alone with no co-wives slapping their thighs, etc. She has some, but her main co-queens are Arjuna’s wives. Arjuna’s other three wives are important in Book 14: Subhadrā sees her grandson revived by Kr̥ṣṇa while the Pāṇḍavas are getting the wealth to perform this costly rite; Citrāṅgadā’s son nearly kills Arjuna while he is guarding the horse in her domain; and Ulūpī resuscitates Arjuna in that episode. These three even appear for the horse sacrifice (14.89.25-90.2) to welcome back the weary Arjuna, and can be related, as Madeleine Biardeau observes, to Arjuna’s having four wives as the “ideal king” and “real sacrificer” among the Pāṇḍavas (2002: 2, 615, 636). Indeed, their earlier Book 14 roles could be Aśvamedha-related: Subhadrā as Arjuna’s Vāvātā; the estranged Citraṅgadā his “discarded” Parivr̥ktī; and the nāgī Ulūpī his lower class Pālāgalī.28 But Yudhiṣṭhira is the real real king for whom Draupadī has ritual duties.29 Draupadī is thus as narratively exposed as Sītā, but at least she gets to stay alive. Let us also note how both epics connect their postwar Aśvamedhas with the continuity of their chief royal lines. Kr̥ṣṇa revives Parikṣit, and Rāma discovers Kuśa and Lava as his recovered heirs. When Sītā attests to her purity so that the earth engulfs her, she also ratifies the legitimacy of Rāma sons. 27 See Kaṭha Upaniṣad 3.3-9 and other Mbh ramifications discussed in Hiltebeitel 1984. On the principle that names of multiply named characters are often used selectively for their contextual fit, see Biardeau 2002, passim. 28 According to Bhatt, “the lowest wife of the king” and “daughter of a messenger or a courier” (1960: 445); “the fourth and least respected wife of a prince” (MW). 29 These others would not have Aśvamedha obligations where Yudhiṣṭhira is the yajamāna and may have been excused co-wife roles for that reason.

The Vedas in Indian Culture and History

432  Ù

C. Rām 1.8-17 Securing royal progeny is thus implied in these Aśvamedha scenes and may also be inferred from the ritual texts. Several texts say the king should remain chaste sleeping between the Vāvātā’s thighs during the horse’s year of wandering, while the horse itself is to be kept away from mares.30 Daśaratha makes this inference explicitly. After sixteen thousand years of sonlessness, he thinks, “Why don’t I offer worship by a horse sacrifice for the sake of a son (sutārtham vājimedhena kim arthaṃ na yajāmy aham)?” (1.8.2).31 Promised four sons if he brings the sage R̥ ś yaśr̥ṅga, he sets out with his wives (10.8-13). Returning, After showing R̥ś yaśr̥ṅga honor as per the śāstras, the king had him enter the women’s quarters. (10.10ab)

—where his queens would be (10.28-29).32 When spring comes, R̥ ś yaśr̥ṅga says it is time to release the horse,33 and Daśaratha 30 See Jamison 1996: 84: there is a penance if the horse impregnates a mare; 66, 76 on our Episode B and 242-243 on fertility parallels between thigh-slapping rites involving the four Aśvamedha queens and those done on the Mahāvrata day in the Gavāmayana sacrifice. Jamison interprets R̥g veda 10.86 about Indra’s monkey-companion Vr̥ṣākapi “as a veiled Horse Sacrifice” that describes among the benefits brought by “Indraṇī’s mating with Vr̥ṣākapi” the restoration of Indra’s worship with bulls and soma, the reaffirmation of his power, his recovery of good erections, and his attainment of sons (74-88 quoting 81, 82-83). As preeminent victim, the Aśvamedha horse, identified with both the royal yajamāna and Prajāpati, imparts through the queens, and the mahiṣī in particular, a mysterious embryo that takes the form of “progeny and cattle” for the king and the kingdom; see Dumont 1927: 17; Śatapatha Brāhmaṇa 13.2.8.5; Hiltebeitel 1991: 378. Goldman 1984: 74-75, 292, 298-299, makes a difficulty over the association of Daśaratha’s Aśvamedha with fertility and offspring, but his hesitation accompanies an argument that the whole episode is an interpolation. 31 Sumantra then tells him of a prophesy he has heard that Daśaratha would one day call upon R̥ ś yaśr̥ṅga to perform a sacrifice (unspecified) that would bring him four illustrious sons and heaven, and relates what Robert Goldman calls a “relatively colorless version” of the R̥ ś yaśr̥ṅga tale (1984: 75) that may be bowdlerized from a more colorful Mbh version. R̥ ś yaśr̥ṅga had stored up great chastity as a young innocent performing tapas until King Romapada sought him to end a twelve-year drought in Aṅga, where R̥ ś yaśr̥ṅga married Romapāda’s daughter after doing so. 32 The three are also there at Rām 1.15.23-25 when Daśaratha brings them the pregnancy potion. 33 Once Daśaratha sets his heart on beginning the rite “for the sake of the continuity of the dynasty (saṃtānārthaṃ kulasya),” R̥ ś yaśr̥ṅga says, “You may gather all the necessary articles and release the horse (sambhārāḥ sambhriyantāṃ te turagaś ca vimucyatām)” (11.3cd). Daśaratha summons Brahman experts in the Vedas —Kāśyapa (presumably R̥ ś yaśr̥ṅga, who is from that gotra), Vasiṣṭha, and others (4-6)— and directs them to perform the horse sacrifice he has re-

ntroduction EIpic Aśvamedhas

Ù  433

orders an escort.34 After about a year (12.1), all the earth’s lawabiding (dhārmikāḥ; 17) friendly (snigdhāḥ) kings arrive from the four directions.35 But nothing is said of the horse’s wandering, much less its having met any resistance.36 When the horse has been returned, the Veda-wise priests led by R̥ ś yaśr̥ṅga see correctly to the Vedic prescriptions in the king’s “great Aśvamedha sacrifice” (aśvamedhe mahāyajñe; 2-3).37 Great detail is then given to the preparation of the twenty-one sacrificial posts (yūpas) for the three-hundred animals to be sacrificed. Indeed, as Goldman points out, one reaches the total of twenty-one yūpas, the usual solved on for the sake of a son “according to the rite prescribed in the śāstras (śāstradr̥ṣṭena karmaṇā)” (8-9). Again he is promised he will get four sons. Putting R̥ ś yaśr̥ṅga in charge, the experts repeat, “You may gather all the necessary articles and release the horse (sambhārāḥ sambhriyantāṃ te turagaś ca vimucyatām)” (11.11cd), and tell Daśaratha that since he has made this “law-abiding resolution to obtain a son (dhārmikī buddhir iyam putrārtham)” (11.12cd), he is sure to have four of them, confirming the prophesy. 34 Daśaratha orders the horse set free to be “guarded by strong men and attended by our preceptor”—apparently R̥ ś yaśr̥ṅga. He also directs that the sacrificial terrain (yajñabhūmi) should be carefully set out on the northern bank of the Sarayū, and every caution be taken to perform the rite unerringly (1.11.14-18). 35 They all arrive from the four directions by the time craftsmen have prepared the site. This is one of the few places where Vālmīki mentions a cohort of kings contemporary with his story. Other than Janaka, Romapada, and the as yet unnamed Kaikeya king with his son (already Daśaratha’s father-inlaw, with his son Yudhājit [Goldman 1984: 301]), the rest are irrelevant to the Rām story. 36 Goldman 1984: 301 makes the interesting observation “that none of the famous kings of the lunar dynasty, the central royal house of the MBh, are mentioned. This tends to support the notion that even the later strata of the Rām are ignorant of the longer epic.” But Goldman also suggests that Vālmīki bowdlerized version of the Mbh’s R̥ ś yaśr̥ṅga story is part of the same “later strata.” I think that Vālmīki produces such effects intentionally, knowing that the Mbh itself mentions few other kings and makes the Rāma story more ancient. Moreover, only “friendly kings” are mentioned. The kings stay through the horse sacrifice and its sequel, the Putrīya Iṣṭi or putrakāmeṣṭi (1.17.3). 37 The same locative phrase occurs at 37c and at Mbh 1.113.12a. On the mahāyajña as a technical term describing both great Śrauta rites and the five daily offerings of a Brahman, and not only in Brahmanical but Pāli Buddhist texts, see Tsuchida 1991: 72, 88-89. The Rām indicates familiarity here with the kalpasūtras, Brāhman. as, and śāstras, and mentions that additional rites were included: a Pravargya and Upasad (4) and a Morning Pressing (prātaḥ-savana) (5) are mentioned at the beginning. And after the horse is sacrificed, with an account of the treatment of its remains (28-32), there is a summary indicating that this phase took three days of prescribed rites (Catuṣṭoma, Ukthya, and Atirātra), plus rites “prescribed in the view of some authors of ritual texts” (Jyotiṣṭoma, Āyus rites, two Atirātras, Abhijit, Viśvajit, and Aptoryāma [33-35]), followed by a conversation about the appropriate compensation (niṣkraya), i.e., dakṣiṇā (36-44). For days there is munificent giving, and in the intervals numerous philosophical debates or perhaps brahmodyas (hetuvādān bahūn; 14), as would better befit an Aśvamedha (Goldman 1984: 305).

434  Ù

The Vedas in Indian Culture and History

number prescribed in Vedic texts, only by counting the posts mentioned (1984: 305; Bhatt 1960: 444). As at the Mahābhārata’s postwar Aśvamedha, an Agnicayana-like fire altar is also built with eighteen layers of bricks in the shape of a golden-winged Garuḍa three times the usual height (17-24).38 Finally, the three queens go into action: Kausalyā walked reverently around the horse and then with the greatest joy cut it with three knives.39 Her mind unswerving, with her love of dharma (dharmakāmyayā) Kausalyā passed one night with the horse. The hotr̥, adhvaryu, and udgātr̥ saw to it that the mahiṣī as well as the parivr̥tti and vāvātā were united with the horse (hayena samayojayan). (13.26-28)

Since vāvātā fits Kaikeyī,40 Sumitrā would be the parivr̥tti. This term has no clear meaning, but it probably replaces parivr̥ktī, the “disliked or despised one” (from pari-√vr̥j; MW), and Vālmīki may use it to spare Sumitrā this probably undeserved Aśvamedha title.41 Kausalyā’s “love of dharma” seems to be Vālmīki’s way of giving her “Vedic” motivation and anticipates that her son will be the paragon of dharma. Yet her aggression toward the horse before uniting with it is surprising.42 Her wielding of three knives may condense a rite called asi-pantha or “paths of the knife,” yet this should come after the mahiṣī lies with the horse.43 Most strik38 Cf. Mbh 14.90.30-31: iṣṭakāḥ kāñcanīś cātra cayanārthaṃ kr̥tābhavan | śuśubhe cayanaṃ tatra dakṣasyaiva prajāpateḥ || catuś cityaḥ sa tasyāsīd aṣṭādaśa karātmakaḥ | sa rukmapakṣo nicitas triguṇo garuḍākr̥tiḥ. 39 Rām 1.13.26cd: kr̥pāṇair viśaśāsa enaṃ tribhiḥ paramayā mudā. A kr̥pāṇa is a sword or sacrificial knife. 40 Even if it is the only usage of this term for anyone in either epic. 41 As Goldman points out, there are difficulties with this solution (1984: 306-308). Cf. Bhatt 1960: 445 and Jamison 1996: 66, 87, 274 n. 104, favoring an amendment to parivr̥ktī. There is of course no fourth wife or Pālāgalī. 42 C.V. Vaidya took an interesting slant on this moment: “And what should we think when we are told that Kausalyā killed by her own hands the sacrificial horse with three sword strokes .... She must have been very strong and a true Rajput lady indeed” ([1906] 1972: 9; cf. Hiltebeitel 1998: 398-399). But I think some kind of sexual aggression would be more plausible. Or perhaps she was possessed. See Rāma’s last image of his mother as he departs for exile: “Kausalyā came weeping after the chariot, crying ‘Rama, Rāma! Oh Sītā, Lakṣmaṇa!’ He glanced back often at his mother, who seemed to be dancing (nr̥tyantīm iva mātaram)” (Rām 2.35.32). 43 This rite, which, according to the ritual texts makes the bridge to heaven, is, however, performed not by one queen with three knives but by three queens using needles: the mahiṣī uses gold needles, the vāvātā silver ones, and the parivr̥ktī copper or lead ones to trace the path that the carver (śamitar) should follow with his knife in cutting up the victim (see Bhatt 1960: 445, Malamoud 1996: 173-74, Hiltebeitel 1991: 377). If so, the sequencing is anomalous, since Kausalyā cuts the horse herself using the three knives before she lies with it.

ntroduction EIpic Aśvamedhas

Ù  435

ing is how the priests of the three older Vedas have the other two queens unite with the horse after the mahiṣī. This would have no Vedic precedent, and must be preliminary to the impregnation of all three by what turns out to be an Atharvanic rite, the Putrīya Iṣṭi that generates the celestial pāyasam that Viṣṇu infuses for the three queens to eat.44 As R̥ ś yaśr̥ṅga soon says, To procure sons for you, I must perform the son-producing sacrifice. It must be done in accord with the instructions of the ritual texts and rendered efficacious by potent verses proclaimed in the Atharvaśiras (atharvaśirasi proktaiḥ). (14.2)

Rounding off with the fourth Veda indicates that the two rites form a whole. But the best proof of an elision comes when R̥ ś yaśr̥ṅga says a single Aśvamedha dīkṣā has carried Daśaratha through both of them. Daśaratha also transitions to the Putrīya Iṣṭi by smelling the smoke from the horse’s fat (vapā) to free himself from sin (29-30). Goldman translates vapā, “omentum,” as “fat” since a horse does not have an omentum (Goldman 1984: 151, 308; Bhatt 1960: 445). It is expertly extracted from the horse by the chief r̥tvij. Daśaratha is now svakulavardhanaḥ, “an increaser of his lineage” (36b). Wrapping up by going through the motions of giving away the earth to the priests (r̥tvijas) and settling on the more appropriate compensation (niṣkraya) of “a million cows, a hundred million gold pieces and four times that amount in silver,” which the priests hand over to R̥ ś yaśr̥ṅga and Vasiṣṭha to apportion among them (41-43), he is “content at heart at having obtained that greatest of sacrifices (prāpya yajñam uttamam)” (44)—which I take to mean “having obtained” or “reached” the results of this rite in preparing for the impregnation of his wives. The poet now makes an elision between the Aśvamedha proper and the Putrīya Iṣṭi that will complete what has all along been the stated purpose of this Aśvamedha. When the Iṣṭi is finished R̥ ś yaśr̥ṅga says the Aśvamedha is now finished (nirvr̥tte tu kratau; 17.1ab), and indicates that Daśaratha’s dīkṣā has ended (samāptadīkṣāniyamaḥ; 17.2a); everyone, the kings included, can go home (17.2-5). Daśaratha also makes this elision himself after he has “obtained” the Aśvamedha: King Daśaratha then spoke to R̥ ś yaśr̥ṅga, “You are true to your vows. Please act so that my line may be extended (kulasya vardhanam)” (45). Moreover, since Vālmīki supplies no other account of the horse’s killing (which the same carver should do by suffocation), it would seem that Kausalyā actually kills the horse, perhaps right at the sacrificial stake where it was last mentioned. 44 1.14.18gh-19ab: viṣṇo putratvam āgaccha kr̥tvātmānaṃ catur vidham || tatra tvaṃ mānuṣo bhūtvā pravr̥ddhaṃ lokakaṇṭakam.

436  Ù

The Vedas in Indian Culture and History

Now as Goldman says, R̥ ś yaśr̥ṅga “is perhaps viewed as serving the purpose of Vyāsa or the other Mahābhārata practitioners of niyojana, or levirate, only through an act of sacrifice in place of direct sexual liaison with the king’s wives” (Goldman 1984: 77; see n. 42). I would extend this point to say that just as R̥ ś yaśr̥ṅga’s Aśvamedha has a hidden niyoga agenda, Vyāsa’s niyoga has a hidden Aśvamedha agenda. Moreover, if a niyoga agenda is hidden with regard to R̥ ś yaśr̥ṅga, it is less so with regard to Viṣṇu, who is twice said to have been “appointed” (ni-√yuj) to intervene: “O Viṣṇu, we shall appoint you (niyokṣyāmahe) out of a desire for the welfare of the worlds” (1.14.17cd); After Viṣṇu Nārāyaṇa was appointed (niyuktaḥ) by the best of gods, he asked them about the means ... . (1.15.11ab)

The Putrīya Iṣṭi effected by R̥ ś yaśr̥ṅga is thus a quasi-niyoga through Viṣṇu,45 and calls for no genetic intervention by a smelly author. D. Mbh 99-100 On to our main event, Ambikā and Ambālikā are, of course, the younger of three sisters, their elder being the accursed Ambā. When Bhīṣma abducts them, his intention and that of the queen dowager Satyavatī is to bring three “Mothers” into the lineage. As both Jamison and Biardeau have recognized, these three names in combination must be related to their use in the Aśvamedha, where they are invoked in a special formula that is uttered while the mahiṣī lies with the horse. A clean version is used when the mahiṣī is led up to the dead horse accompanied by her cowives: “O Ambā, Ambālī, Ambikā [var. Ambā, Ambikā,

45 If Vālmīki picks up on an Aśvamedha innuendo in Mbh Book 1, it could be on this point. If we allow that Vālmīki makes economy by having the four brothers born in one generation rather than two, we can see that R̥śyaśr̥ṅga and Viṣṇu supply much the same requisites as Vyāsa and Durvāsas, who supplies the mantra that empowers Kuntī and Mādrī to get pregnant by gods in another quasi-niyoga. R̥śyaśr̥ṅga has a reputation for tapas that seems not only to enhance his fertilizing powers to bring a drought-ending rain (1.9.28-29), but, at least in the Rām, to be able to extend this fertilizing power to the area of sexuality —even though, unlike Vyāsa, he deploys this power ritually rather than by sleeping with anyone. Moreover, if Vyāsa is about to remind us of a horse, R̥śyaśr̥ṅga’s name means “Having deer antlers.’” See Goldman 1984: 296, supplying a commentarial legend of his birth from a doe who imbibed some of his father, the sage Vibhāṇḍaka’s, semen, as told also in Mbh 3.110.11-116. With a bemused look, he bears a single antelope horn on his forehead in a nice old bust at the Mathura Museum.

ntroduction EIpic Aśvamedhas

Ù  437

Ambālikā]. No one is leading me. The horsikins is sleeping.”46 Then, before the Adhvaryu covers the mahiṣī and the horse with a linen blanket, she “lies down beside the horse and invites it to stretch out its forefeet along with hers,”47 and either the Adhvaryu or the king (her husband) utters the most erotic of all the rite’s mantras, focusing on her sexual pleasure. Covered by the blanket,48 the mahiṣī, having “manipulate[d] the dead horse into some sort of copulatory position,” in some texts modifies her mantra with yabhati (“is fucking”) instead of nayati (“is leading”). With or without these words, which “mock” or “scold” the horse, the rite clearly invites the dead horse to regain its sexual stamina (Jamison 1996: 66-69, 76; cf. 242). Meanwhile, the cowives and their attendants circulate back and forth around the horse and the mahiṣī slapping their thighs and fanning with their hems, and then exchange “slangy and crude” riddling mantras with the priests (65-66, 69-70). In importing these three names into a dynastic crises, the Mahābhārata thus draws on the Aśvamedha’s reputation for assuring lineage continuity. This overlaps with what Bhīṣma and Satyavatī continue to wish from the two younger sisters even after they have become widows. But as Jamison and Biardeau recognize, the link is obscure, in part because with Vicitravīrya dead there is no king to perform an Aśvamedha. Jamison proposes that the usage can be illumined by a “husband-finding” (pati-vedana) rite linked with the autumn Śākamedha, in which a Traiyambaka Homa to Rudra can remove a woman’s lack of marital success and secure her first unborn descendant.49 Biardeau proposes that the rite’s triple name would evoke one woman, the mahiṣī, “transformed into three” as the Kāśi sisters, and that the one would be Ambikā not only as the mahiṣī, but in anticipation 46 Jamison 1996: 67, transl. Taittirīya Saṃhitā 7.4.19.1ab; the usual opening line ambe ambāly ambike has, as Jamison observes (1996: 274 n. 107, 304 n. 87), a precise Mbh nomenclature and sequence in Vājasaneyi Saṃhitā 23.18: ámbe ámbike ’mbālike, with the same but for the initial vocative in Maitrāyaṇī Saṃhitā 3.12.20. According to Jamison (243), “the three vocatives are taken as variants on affectionate terms for ‘mother,’ but together they also add up to the ‘three Ambikās’ of Rudra Tryambaka and its vr̥ddhi derivative, the Traiyambakahoma.” “Leading” (nayati), according to Jamison (67, 274 n. 108), probably suggests the mahiṣī is being led, as it were, into matrimony. 47 See Jamison 1996: 67. How that would help achieve a position for sexual contact is not made clear. 48 Biardeau says, “It seems that the queen complains at being looked in the sexual act” (2002: 1.220). 49 See Jamison 1996: 242-246: it uses a mantra to Tryambaka found in R̥V 7.59.12 that the girl modifies to request a husband. As Jamison points out, both rites resemble the movements of fire-circling servant maids on the Mahāvrata day of the Gavāmayana. There is also more thigh-slapping.

438  Ù

The Vedas in Indian Culture and History

of her name’s becoming “one of the most frequent names of the Goddess.”50 I believe these are pieces to a puzzle that our epic poets have intentionally left incomplete, and that three more pieces crop up from the text.51 First, Vyāsa stipulates that the two widows undergo a year-long vow before he lies with them; second, he will then give his dead brother sons like Mitra and Varuṇa.52 The yearlong vow replicates the Aśvamedha requirement that the queens remain abstinent during the year the horse wanders.53 And the pair Mitra and Varuṇa certainly has a Vedic ring. Vedavyāsa54 could be alluding to ways that the Aśvamedha identifies the king with dharma,55 and also that the Rājasūya invokes Mitra as “lord of truth” and Varuṇa as “lord of dharma” in announcing the newly consecrated Bharata king (MS 2.6.6; TS 1.8.10.1-2).56 As with its postwar Aśvamedha, where the Mahābhārata also gives an enigmatic shape to the horse, it is in the domain of Vedic allusion that we must explain why Vyāsa would be encrypting himself as a sacrificial horse. Such occultation occurs where Vedic practices 50 Biardeau calls Ambikā here “the mother of Pāṇḍu,” but must mean the mother of Dhr̥tarāṣṭra (2002: 1.220). 51 Jamison 1996: 304 n. 94 “assumes that some version of the epic story of Ambā and her sisters already existed in early Vedic and that these girls were associated with the three Ambikās belonging to Rudra. Despite the difficulties ... , I think this is more likely than assuming that a later epic poet simply made up the story of the abduction and its aftermath and named the female protagonists by plucking some designations out of the onomastic repository of Vedic ritual.” 52 Vyāsa tells his mother, “Satyavatī, you know both the higher and lower dharma. And since your mind is set on dharma and is beneficent to the living, I shall indeed do what is needed with respect to dharma, and by your appointment (tvan niyogāt) do what you desire. For this is found to be the ancient practice. I shall give my brother sons the likes of Mitra and Varuṇa. Both ladies must punctiliously submit to a vow, which I shall describe, for the space of a year, so that they become sanctified. For no woman may lie with me without carrying out the vow (saṃvatsaraṃ yathā nyāyaṃ tataḥ śuddhe bhaviṣyataḥ | na hi mām avratopetā upeyāt kācid aṅganā)” (Mbh 1.99.36-39). 53 See Jamison 1996: 84: “During the year when the horse is journeying, the king lies nightly with his favorite wife (Vāvātā), but does not have sex.” In Śatapatha Brāhmaṇa 13.4.1.9, he thinks, “May I, by this austerity (anena tapasā), reach the end of the year successfully.” 54 It is important to remember who is speaking here and what he has been doing. The last we heard of Vyāsa before Satyavatī recalled him as needed was that he had gone off to the Himalayas to divide the Vedas and impart the Mahābhārata as the fifth Veda to his five disciples, who were to proclaim it as the Bhārata (1.57.73-75) 55 Bowles 2007: 93 n. 47 cites the Taittirīya Saṃhitā version of a formula that equates the Aśvamedha king with dharma: “with my two shins and my two feet I am dharma (dhármo ’smi), the king fixed firmly on his people.” 56 Such a momentary fusion of elements from of two great royal rituals would fit our scene, since Vyāsa would be engendering just such a king —or two.

ntroduction EIpic Aśvamedhas

Ù  439

are of dark and doubtful dharma from the standpoint of post-Vedic Brahmanical culture. Vyāsa also makes a singular appearance in the Nārāyaṇīya, which, I have argued (Hiltebeitel 2006: 249250), offers a bhakti encryption of Vedic and purāṇic allusions.57 The Nārāyaṇīya comes to its deepest disclosure58 when Vyāsa tells Janamejaya,59 just now told to perform a Horse Sacrifice on top of his Snake Sacrifice (12.334.8-9), about Nārāyaṇa’s manifestation, called Harimedhas, as the Horse’s Head (Hayaśiras). Janamejaya learns how Nārāyaṇa places the Horse’s Head as a “repository of the Vedas” and source of world-oriented dharma (pravr̥ttidharma) in the northeastern ocean.60 Vyāsa says he was61 originally born of Harimedhas, the “sacrificial sap (medhas) of Hari,”62 which seems to identify him with the cosmic Veda-chanting Aśvamedha Horse’s Head.63 All this deepens Vyāsa’s horseplay. Now Vyāsa only says he will sire sons the likes of Mitra and Varuṇa (Mbh 1.99.38b). We would love to know how this would have turned out, but he does not get to do this because Satyavatī says matters are too pressing to give the widows a year for such a vow. Our third puzzle piece crops up here where things have been willfully set askew.64 Before Satyavatī asks Vyāsa to do it, she first 57 Vyāsa’s Horse’s Head story (below) is called “purāṇa equal to the Veda (purāṇaṃ vedasammitam)” (Mbh 12.335.7b). 58 It does this by dipping to the outermost dialogue frame of Śaunaka and Ugraśravas so that Śaunaka can ultimately hear what Vyāsa said to Janamejaya about the very same matter —the Horse’s Head— that has made Śaunaka curious. As I show (Hiltebeitel 2006) with reference to these dips, and as Grünendahl confirms particularly with reference to this third dip (1997: 52-53), the Pune Critical Edition errs in substituting the Vaiśampāyana-Janamejaya frame for the Ugraśravas-Śaunaka one. Indeed, in coming to this disclosure in the fourteenth of its eighteen adhāyas (as reconstituted from the Pune Critical Edition’s misguided attempt to make it nineteen), the episode may occur at a specifically designed narrative core; see Brodbeck 2006: 29-30. 59 This is the only occasion where Vyāsa explicitly addresses Janamejaya at his snake sacrifice. Vyāsa is addressed by Janamejaya on two other occasions (1.54.18-22; 15.42-43), but he responds there mainly through actions rather than words. See Hiltebeitel 2006: 246 n. 59. 60 Mbh 12.335.10-72, quoting 54c, 69d. The Horse’s Head retrieves the Veda from two cosmogony-disrupting demons, Madhu and Kaiṭabha, who had sought to hurl it into the ocean. 61 Vyāsa reveals that he himself was first “born of that god Harimedhas’s grace” before he was born again “by Nārāyaṇa’s grace” as Vyāsa (12.337.54-55). 62 Perhaps the “tawny/yellow/reddish (hari) sacrificial essence” (see Hiltebeitel 2006: 249 n. 70). 63 A Horse’s Head with its neck formed by Kālarātrī, the Night of Time (Mbh 12.334.44-48). 64 Instead of being a case of lost Vedic symmetries. Dumézil seems to ignore this verse, perhaps because it does not help his case that an original set of incarnations has been effaced: that Pāṇḍu would originally have incarnated Varuṇa; Dhr̥tarāṣṭra and Vidura likewise the two “minor sovereigns” Bhaga (god of des-

440  Ù

The Vedas in Indian Culture and History

tries to appoint (vi-ni-√yuj) Bhīṣma to sire sons with his brother’s “two mahiṣīs” (mahiṣyau; 1.97.9a). Can there be two mahiṣīs? This is the only dual of mahiṣī in either epic. As we just observed, the Rāmāyaṇa’s first Aśvamedha incongruously has three queens lie with the horse, but on unequal terms: the two non-mahiṣī queens apparently do so only briefly on the same busy night.65 Just as the Rāmāyaṇa has reasons to give each one time with the horse, the Mahābhārata has reasons to have two mahiṣīs66 —curious ones, though, since neither Ambikā nor Ambālikā would be a mahiṣī if Ambā were still around. It is Satyavatī who mentions “two mahiṣīs,” not Vyāsa. Her usage in speaking to Bhīṣma may be a reminder that Ambā’s unavailability as a Mother complements Bhīṣma’s as a Father, and is an intimation that despite the legal incongruity, to speak of Ambikā and Ambālikā as two mahiṣīs gives them an equal chance to become the mother of the one desired heir. Here Biardeau and Jamison’s puzzle pieces may also reenter the picture. Ambikā should now be the single mahiṣī, empowered like Durgā in the Devī-Māhātmyā. And hers is the main name in Rudra’s epithet Traiyambaka, the one “possessing three Ambikās.”67 Our main event may allude to this Homa,68 whose tinies) and Aryaman (god of Ārya clans); and Yudhiṣṭhira would have been sired by Mitra were it not that a “clumsy retouch” replaced Mitra by Dharma (1968: 146-148, 152, 159-160, 170-174). Dumézil was convinced that he could recover a Mbh whose “primary form [was] contemporary with the oldest Vedic times, or anterior” (172). 65 Vyāsa also comes to Ambikā in the dead of night (Mbh 1.100.2-6). 66 Both epics use the term mahiṣī almost exclusively for single chief queens, and where they do otherwise it is with similar notes of discordance that spell misfortune for the real mahiṣī. Thus the gallivanting Arjuna, already among those wed to Draupadī, exclaims to Kr̥ṣṇa what good fortune he has that Subhadrā will be his mahiṣī (Mbh 1.211.19) —a discordance on top of an anomaly, since Yudhiṣṭhira, speaking for all five Pāṇḍavas, had said that Draupadī “will be the mahiṣī of us all” (1.187.22ab). Similarly, when the troublemaking Mantharā tells Kaikeyī she (rather than Kausalyā) is Daśaratha’s mahiṣī (Rām 2.7.19), Kaikeyī is so imperious that Sumantra as her messenger speaks of her as the mahiṣī when he tells Rāma that she and Daśaratha want to see him (14.11); and when Rāvaṇa invites Sītā to be his mahiṣī, it is with the odd qualifier that she will be his agramahiṣī, “primary chief queen” (3.4.24; 5.18.16)— a term used nowhere else in the epics (Mandodarī would seem to have that position although the term is not used for her). On the other hand, in two plural usages it means mainly “woman” (Mbh 1.187.26; Rām 2.36.7; see 36.1). 67 Jamison 1996: 241, 303 n. 76. Although “Tryambaka” may —as usually translated— refer in the epic to Śiva’s having three eyes (see Scheuer 1979: 23736, 255-256), the matter is uncertain (Hopkins [1915] 1969: 220), and in the one case where the Mbh gives an etymology (Vyāsa is telling Arjuna how Śiva preceded his chariot in battle), it refers to his having three goddesses: “And since the Lord of the universe possesses three goddesses —Sky, Waters, and Earth— he is remembered as Tryambaka” (Mbh 7.173.89). 68 See Jamison 1996: 243. The Mbh never mentions the pativedana or the

ntroduction EIpic Aśvamedhas

Ù  441

“‘husband-finding’ spell” is performed at an inauspicious crossroads for a spinster running out of time in finding a mate. Our Ambikā is in some such predicament. And it is the result of her larger failure to meet the reduced vow demanded by Vyāsa of bearing “my smell, my looks, my garb, and my body” (99.43ab) that her son will be born blind and unfit to rule. Jamison also relates the Traiyambaka Homa to Bhīṣma’s abduction of the three Kāśi sisters on two interesting points. But she assumes that an older version of the epic story complemented this ritual (cf. above n. 53). In favor of positing an older version, she suggests that without one, “the epic maidens would provide bad role models for the husband-seeking girl of the Śākamedha” (245). But if we take things as we find them and ask what the epic poets might have made of the model of the older Vedic rite, the “husband-finder” of the Traiyambaka Homa would be a good role model, given her bad situation, for something untoward to go further haywire in an epic series of turnabouts, which would include Bhīṣma’s turning of a husbandfinding ritual into a wife-finding ritual —something even more basic than his turning a “woman’s-choice” svayaṃvara into a man’s-choice svayaṃvara, which, as Jamison notes, he does with verbal precision (1996, 299 n. 38). Jamison then says that while Ambā became a murderous avenger after the abduction caused her to lose her husband, “[e]ven Ambikā and Ambālikā, though they settled happily enough into their married life after their unexpected abduction ... , might not have chosen this particular method of pati-vedana [husband-finding] if they had their druthers” (245). But we know they did have their druthers. They too were going to have a svayaṃvara. It would seem that as personifications of “the tryambaka,” the three Kāśi sisters would be embodiments of a “husband-finding” by svayaṃvara that goes awry, at least for them, because Bhīṣma carts them off Rākṣasa style to find their rather limited un-chosen husband Vicitravīrya, whom Ambā in fact rejects. Of course here we have a way to suggest that if the epic poets recall both Vedic rites, it is separately: the Traiyambaka Homa. Generally, the epics seem to overlook the Caturmāsyāni rites. The Mbh does mention a Traiyambaka Bali (7.56.1-4, esp. 3d): according to Scheuer 1982: 255-63, it is probably offered nightly throughout the war on Arjuna’s behalf by his servants, after which Arjuna sleeps on the ground surrounded by weapons. Cf. also Scheuer 258 n. 23 on the Pāṇḍavas’ offering (upahāra, bali) to Rudra Tryambaka (14.8.23-24; 64.1-8) after a night’s fasting to get hold of the wealth needed to perform their postwar Aśvamedha. Ambikā as Rudra’s sister is linked with autumn, which in some texts is Rudra’s season of “special murderousness” due to her influence (cf. 241, 245, 304 n. 94). The Mbh does know Śiva as [Pārvatī-] “Ambikā’s husband” (ambikabhartre), yet a brahmacārin (7.57.53).

442  Ù

The Vedas in Indian Culture and History

“husband-finding” ritual would underlie the beginning of the three sisters’ adventure into the Kuru dynasty; the “Aśvamedha” would underlie the two younger sisters’ nights with Vyāsa. In any case, the allusion to the Aśvamedha’s “three mothers invocation” would underline that, rather than Ambā, it is Ambikā and Ambālikā and Ambikā’s Śūdra servant who lie with the smelly author. Three mothers once again, and again something askew with the Śūdra servant replacing Ambā, setting the stage for the next set of three queens, Gāndhārī, Kuntī, and Mādrī. E. Kuntī’s anti-niyoga Aśvamedha tale (Mbh 1.112) Gāndhārī and Kuntī are again rivals to bear a single legal heir. But before Kuntī will tell Pāṇḍu that the R̥ṣi Durvāsas gave her a secret mantra that will enable him to win this race, she must divert his demand that she perform niyoga, even if her solution smacks of another quasi-niyoga in that she will be calling in the intervention of gods. When Pāṇḍu tells her about a woman who goes at night to a crossroads (like a pati-vedana spinstress) with a flower in her hand until she finds a niyoga partner (1.111.33-36), Kuntī answers him with an equally outrageous story of her own, called the Vyuṣitāśva-Upākhyāna (1.112). Like Vicitravīrya, King Vyuṣitāśva died sonless of “consumption” (yakṣmāṇam) after a bout of lust with his wife, Bhadrā Kākṣīvatī (112.16-17). The term yakṣmāṇam is used in the epics only for these two kings.69 Famed as a yajamāna (8) in numerous Soma sacrifices (9, 14), the glory of King Vyuṣitāśva’s reign was an Aśvamedha at which he “became Indra among kings, endowed with the strength of ten elephants” (12). Lamenting and wanting to join him in the next world, Bhadrā says, “Faithful as a shadow, I shall ever do your will, always loving to please you (nityam priyahite ratā)” (23).70 His name with -aśva means “the Daybreak Horse,” and perhaps also “One Who is Inhabited

69 See 1.96.57d and 5.145.23f on Vicitravīrya, always with the same verb samapadyata. 70 A mahiṣī is formulaically “dear” (priya) to her royal husband. Draupadī is the “dear” (priya)mahiṣī of Yudhiṣṭhira (Mbh 4.15.31, 16.12; 10.11.17) and of all the Pāṇḍavas (4.20.19, 5.80.22); so is Sītā to Rāma (Rām 4.48.18, 5.12.43-44, 13.46; Mbh 4.20.10 according to Draupadī), Śakuntalā at their happy ending to Duḥṣanta (1.69.43); Indrāṇī to Indra (5.11.13; 12.22, while coveted by Nahuṣa); and Tārā was dear to Vālin according to their son Aṅgada (Rām 4.54.3). Both Draupadī (Mbh 4.19.10; 20.10, 19; 5.80.9) and Sītā (Rām 6.38.3) speak of the title with pride, and Vālmīkī uses it even when he welcomes the banished Sītā to his ashram (7.48.8). When Vyuṣitāśva’s corpse impregnates Bhadrā she is also a pativratā (32b).

ntroduction EIpic Aśvamedhas

Ù  443

or Possessed by the Horse.”71 The first meaning resonates when Kuntī calls him a rising sun before his Aśvamedha (10) and a setting one after it (17b), dying after his amours with Bhadrā during which they would have been making up for a year’s lost time after she (or another queen) lay with the horse. The name “Daybreak Horse” is enough for an Aśvamedha innuendo. But “possession” may also be in the air since Vyuṣitāśva can speak from his corpse with an “inner voice” (vāk ... antarhitā) and impregnate Bhadrā.72 Recall that the king may utter the most erotic mantras to the mahiṣī while she is lying with the Aśvamedha horse. Possession is also hinted on Bhadrā’s side: uttering a long lament in which she vows to lie on kuśa grass “possessed (āviṣṭā) by sorrow and intent on seeing you” (27), she “embraces the corpse” (taṃ śavaṃ sampariṣvajya) and gets his instructions on her bath and timing for intercourse (29-31). “By that corpse the lady gave birth to her sons, the three Śālvas and the four Madras, O best of Bhāratas” (33)! In brief, Kuntī answers Pāṇḍu not in the name a niyoga like that incurred after the death of Pāṇḍu’s father Vicitravīrya, but with reference to an Aśvamedha king who is like Pāṇḍu’s father in one way but unlike him in another. Like Vicitravīrya, Vyuṣitāśva died of “consumption.” Unlike him, he remained potent in death like an Aśvamedha horse. Kuntī thus fills out our quadrangle with this Aśvamedha gloss on Vyāsā’s nights with the “two mahiṣīs.” The means she will use to get pregnant is also Atharvanic, like the Putrīya Iṣṭi through which R̥ ś yaśr̥ṅga gets Daśaratha’s three queens pregnant by the “appointment” of Viṣṇu. The mantra that Kuntī gets from Durvāsas is something “heard in the Atharvaśiras” (atharvaśirasi śrutam; Mbh 3.289.20), and is, in particular, said to be “sorcerous” when Pāṇḍu calls on her to use it to sire their first son through the god Dharma (1.113.34, 39, 42cd). As one would now expect, the Rāmāyaṇa scene is again cleaner, omitting any suggestion of sorcery, not to mention possession, in the birth of Rāma. F. Knowing Vedic allusions If it can be shown that such epic usages are matters of knowing allusion, one possible conclusion worth exploring would be 71 See MW 1040-41. As a neuter noun, vyuṣita, “daybreak,” derives from vi2.√vas, “to shine forth” (from vi-uṣ, “to dawn”). But as an adjective from vi-5.√vas, “to abide, dwell, live,” it can mean “inhabited by” in compounds. 72 On the prominence of possession in the Mbh, see Smith 2006: 250-55, 259, 265-68, 272-75, with discussion of āveśa and derivatives of ā-√viś, an additional example of which is cited below.

444  Ù

The Vedas in Indian Culture and History

that, rather than the standard view that the epics’ references to Vedic ritual, and particularly Vedic royal rituals, are distanced by desuetude and confusion about them, we should look at epic depictions of Vedic rituals, at least where they are narrated, as deft and cunning. Van Buitenen’s (1972) treatment of the Mahābhārata’s Book 2 Rājasūya-and-Dice Match sequence has been a supple illustration of this point for over thirty years now. The Mahābhārata poets do indeed devalue the Aśvamedha, Rājasūya, and Vājapeya relative to pilgrimage in Book 3, and debunk Yudhiṣṭhira’s Book 14 Aśvamedha relative to the practice of gleaning. But these asides come by way of exalting pilgrimage and gleaning in the name of the ordinary man who cannot afford Vedic rituals much less sponsor royal ones. It may also be that when the Mahābhārata poets speak of morally superior substitutes or equivalents for the Aśvamedha, Rājasūya, etc., that these rites were decreasingly performed during the time that the Mahābhārata was composed. But whether or not doing other things like pilgrimage or gleaning were deemed morally superior, and whether or not these royal rituals were becoming less common if not necessarily less prestigious during the period of epic composition, this would not seem to have diminished the epic poets’ interest in referring to them in ways that deepen their stories. Moreover, the Aśvamedha seems to have been revived under the Śuṅgas by Puṣyamitra, and remained on tap for later Brahmanically inclined kings such as Samudragupta (Thapar 2002, 284, 296), and the early Pallavas (329, 335). No doubt these kings and their priests would also have made knowing ritual adaptations. One also finds what seem to be traces of the Aśvamedha, and in particular some likely evocations of the ritual scenes we have been describing, in popular rites and folklores that are traceable, at last hypothetically, to regional kingdoms.73 Here we come to a different problem: that of accounting for continuity in circumstances where Vedic allusions seem to be unknown.

73 See Biardeau 2004: 174-76, 183-85, 217-22, 223 n. 169. Generally, Biardeau posits that local goddess traditions, often linked with the Mbh, offer some explanation.

ntroduction EIpic Aśvamedhas

Ù  445

References Bhatt, G.H. (ed.). 1960. The Bālakāṇḍa: The First Book of the Vālmīki Rāmāyaṇa. Baroda Critical Edition. Baroda: Oriental Institute. Biardeau, Madelaine. 2002. Le Mahābhārata: Un récit fondateur du brahmanisme et son interprétation. 2 Vols. Paris: Éditions du Seuil. ———. 2004. Stories about Posts: Vedic Variations around the Hindu Goddess. Eds. Alf Hiltebeitel, Marie-Louise Reiniche, and James Walker. Chicago, MI: University of Chicago Press. Bowles, Adam. 2007. Dharma, Disorder and the Political in Ancient India: The Āpaddharmaparvan of the Mahābhārata. Leiden: Brill. Brereton, Joel P. 2002. “The Race of Mudgala and Mudgalānī.” Journal of the American Oriental Society 122, 224-34. Brodbeck, Simon. 2006. “Ekalavya and Mahābhārata 1.121-28.” International Journal of Hindu Studies 10, 1-34. Dhand, Arti. 2004. “The Subversive Nature of Virtue in the Mahābhārata: A Tale about Women, Smelly Ascetics and God.” Journal of the American Academy of Religion 72, No. 1, 33-58. Doniger, Wendy. 1995. “Begetting on the Margin: Adultery and Surrogate Pseudomarriage in Hinduism.” In: Lindsey Harlan and Paul B. Courtright (eds.), From the Margins of Hindu Marriage: Essays on Gender, Religion, and Culture. Oxford: Oxford University Press, 160-183. Dumézil, Georges. 1968. Mythe et épopée. Vol. 1: L’Idéologie des trois fonctions dans les épopées des peuples indoeuropéens. Paris: Gallimard. ———. 1970. The Destiny of the Warrior. Transl. by Alf Hiltebeitel. Chicago, MI: University of Chicago Press. Dumont, Paul.-Emile. 1927. L’Aśvamedha: Description du sacrifice solonnel du cheval dans le culte védique d’aprés les textes du Yajurveda blanc. Paris: Paul Guethner. Ganguli, Kisari Mohan (transl.) 1970 [1884–96]. The Mahabharata. New Delhi: Munshiram Manoharlal. Gehrts, Heino. 1975. Mahābhārata: Das Geschehen und seine Bedeutung. Bonn: Bouvier Vlg. Herbert Grundmann. Goldman, Robert P. (transl.). 1984. The Rāmāyaṇa of Vālmīki. Vol. 1: Bālakāṇḍa. Princeton: Princeton University Press. Grünendahl, Reinhold. 1997. “Zur Textkritik des Nārāyaṇīya.” In: Peter Schreiner (ed.), Nārāyaṇīya-Studien. Purāṇa Research Publications–Tubingen, Vol. 6. Weiesbaden: Harrassowitz Verlag, 30-73. Hiltebeitel, Alf. 1977. “Review of Gehrts 1975.” Erasmus 29, Nos. 3-4, 86-92.

446  Ù

The Vedas in Indian Culture and History

———. 1984. “The Two Kr̥ṣṇas on One Chariot: Upaniadic Imagery and Epic Mythology.” History of Religions Journal 24, 1-26. ———. [1976] 1990. The Ritual of Battle: Krishna in the Mahābhārata. Albany: State University of New York Press. ———. 1991. The Cult of Draupadī. Vol. 2: On Hindu Ritual and the Goddess. Chicago, MI: University of Chicago Press. ———. 1998. “Empire, Invasion, and India’s National Epics.” International Journal of Hindu Studies 2, No. 3, 387-421. ———. 2001. Rethinking the Mahābhārata: A Reader’s Guide to the Education of the Dharma King. Chicago, MI: University of Chicago Press. ———. 2005a. “Not Without Subtales: Telling Laws and Truths in the Sanskrit Epics.” Journal of Indian Philosophy 33, 455-511. ———. 2005b. “On Reading Fitzgerald’s Vyāsa. Review of James L. Fitzgerald’s translation of the Mahābhārata, Vol 7. (University of Chicago Press).” Journal of the American Oriental Society 125, No. 2, 241-261. ———. 2006. “The Nārāyaṇīya and the Early Reading Communities of the Mahābhārata.” In: Patrick Olivelle (ed.), Between the Empires: Society in India 300 BCE to 400 CE. New York: Oxford, 227-255. ———. 2007. “Among Friends: Marriage, Women, and Some Little Birds.” In: Simon Brodbeck and Brian Black (eds.), Gender and Narrative in the Mahābhārata. London: Routledge, 110-143. ———. 2011. Dharma: Its Early History in Law, Religion, and Narrative. New York: Oxford University Press. Hopkins, Edward Washburn. 1969 [1915]. Epic Mythology. New York: Biblo and Tannen. Jamison, Stephanie W. 1996. Sacrificed Wife Sacrificer’s Wife: Women, Ritual, and Hospitality in Ancient India. Oxford: Oxford University Press. Kinjawadekar, Ramachandra. (ed.). 1929-33. Mahābhāratam with the commentary of Nīlakaṇṭha. 6 Vols. Poona: Chitrashala Press. Malamoud, Charles. 1996. Cooking the World: Ritual and Thought in Ancient India. Transl. by David White. Delhi: Oxford University Press (= Cuire le monde. Rite et pensée dans l’Inde ancienne. Paris: Éditions la découverte, 1989). Monier-Williams, Monier (= MW). 1964 [1899]. A Sanskrit-English Dictionary. Oxford: Clarendon Press. Scheuer, Jacques. 1982. Śiva dans le Mahābhārata. Bibliothèque de l’Ecole des Haute Études, Sciences Religieuses, Vol. 56. Paris: Presses Universitaires de France. Smith, Frederick M. 2006. The Self-Possessed: Deity and Spirit Pos-

ntroduction EIpic Aśvamedhas

Ù  447

session in South Asian Literature and Civilization. New York. Columbia University Press. Sullivan, Bruce M. 1990. Kr̥ṣṇa Dvaipāyana Vyāsa and the Mahābhārata: A New Interpretation. Leiden: E.J. Brill. Tsuchida, Ryutaro. 1991. “Two Categories of Brahmins in the Early Buddhist Period.” The Memoirs of the Toyo Bunko 49, 51-95. Vaidya, Chintaman Vinayak. 1972 [1906]. The Riddle of the Rāmāyaṇa. Delhi: Meherchand Lacchmandas. Van Buitenen, Johannes A.B. 1972. “On the Structure of the Sabhāparvan of the Mahābhārata.” In: J. Ensink and P. Gaeffke (eds.), India Maior: Congratulatory Volume Presented to J. Gonda. Leiden: E.J. Brill, 68-84. Whitney, William Dwight. 1960 [1889]. Sanskrit Grammar. Cambridge, MA: Harvard University Press.

To Be Good is To Be vaidika. On the Genesis of a Normative Criterion in the Mānavadharmaśāstra Federico Squarcini University Ca’ Foscari of Venice In the Mānavadharmaśāstra, the practice of calling upon “vedic” (vaidika) ideal types takes the form of a specific and rather innovative use of adjectives. Because of that work, the term vaidika —virtually unknown to the authors of the dharmasūtras— has gained wide use among brahmanical intellectuals, becoming a distinctive semantic taxon. By using the term vaidika as a means of value judgment, the author of Mānavadharmaśāstra decreed the positive and normative character of a wide number of practices, customs, beliefs, and behaviours. In this contribution I will provide examples that demonstrate that this use of the term vaidika is an invention of the Mānavadharmaśāstra, and further, an invention that served to classify texts, practices, and ideas as possessing authority and legitimacy. 1. The discourse on “what is vedic” As it has been recently said regarding the term dharma (which has to be treated as a signifier of a negotiable semantic field, rather than as a positively defined notion),1 it is important to reflect upon the cultural and political presuppositions of the brahmanic discourse on what is “vedic.” Within brahmanical intellectual contexts, starting from the centuries that preceded Aśoka, saying that something is “vedic” —or that “it is stated in the Vedas”— 1 This can be stated considering the results of a wide research recently published. See Olivelle 2004.

The Vedas in Indian Culture and History

450  Ù

was tantamount to saying that it was “old,” “valuable,” “legitimate,” “authorized,” “appropriate,” or “good.” To indicate that something was related to the Vedas was a way of classifying and qualifying it positively.2 Such an attempt to profit from the “vedic” semantic taxon is not new to modern scholars, who are surely familiar with Louis Renou’s statement on the matter from the 1960s: A y regarder d’un peu près, la révérence au Veda comporte plus d’une nuance. Sous les allusions génériques, imprécises (type “ainsi est-il dit dans la śruti …”), comme en présentent d’ailleurs toutes les adorations, il se dissimule certains malentendus. Du fait meme que le Veda est censé contenir toutes choses en son sein (yad ihāsti tad anyatra, yan nehāsti na tat kva cid “ce qui s’y trouve existe ailleurs, ce qui n’y est pas n’est nulle part”), à la facon d’un avyaktaṃ brahma, les auteurs en viennent à le citer pour legitimer des données qui sont visiblement anachroniques ou, ce qui est pire, subrepticement réinterprétées. On croit ètre toujours dans le sillage du Veda, alors qu’on lui tourne le dos. Le terme tend à servir de symbole et la notion qu’il couvre a été l’objet d’une véritable aliénation. […] Meme dans les domaines les plus orthodoxes, il arrive que la révérence au Veda soit un simple “coup de chapeau,” donné en passant à une idole dont on entend ne plus s’encombrer par la suite.3

In saying this, Renou appears to be aware that the resort to the symbolic incrementum —achieved by indicating that something is “Vedic”— was a central topos of the classical brahmanical intellectual discourse. Nevertheless, Sanskrit scholars have not devoted much attention to this issue.4 The question, then, remains: how, within later brahmanical intellectual production, have the content and the forms of symbolic legitimation been conceptualized? Among these later conceptualizations, the invention and circulation of the term vaidika is one of the more significant and long-lasting accomplishments: a lexical novelty that can be fully understood only through the analysis of the social and pragmatic factors closely related to its genesis. Seen from this angle, the word vaidika appears as a novelty produced by a specific collective agency that was struggling to justify the primacy of its culSee, on this ancient tendency, Smith 1994. Renou 1960: 1-2. This metaphor has been largely employed in Indological studies. See Halbfass 1991: 1-3, Gonda 1997: 7-8, Smith 1998: 20, Patton 1994: 1. Nevertheless, the awareness of the symbolic power derived from the association with the Veda can be seen in the usage of the term saṃvaraṇa (“which cover,” “secret,” “hidden”) in Kauṭilya, Arthaśāstra 1,2,4-5. 4 The term vaidika is not present in Mayrhofer 1986-96. 2 3

To Be GIntroduction ood is To Be vaidika

Ù  451

tural and symbolic capital over against that of others producers of symbolic good. Terminological novelties, neologisms, semantic coinages, and lexical innovations are generally the product of strategic intentions, strongly related to various forms of dialectical dispute. The lexical form vaidika is no exception. To the contrary: it exemplifies the idea that “something, to be good, has to be perceived and represented as having a thorough relationship with the Vedas.” 2. The semantic field of vaidika: lexical and terminological notes The word vaidika derives from the word veda, indicating that something is “related to the Veda,” “derived from or in conformity with the Veda,” “prescribed in the Veda,” “Vedic,” or “knowing the Veda.” It can therefore be used in various ways: when masculine, it indicates “a Brāhman versed in the Veda”; when neuter, it can indicate “a vedic passage”5 or “a vedic precept.” It also appears in declensional forms (i.e. vaidikāḥ, vaidikaṃ, vaidikāni, vaidikyaḥ, etc.) and morphological variants (i.e., the many compounds like vaidikalaukikānāṃ; see infra). The compounds (samāsa) and derivates (pratyaya) that contain the word vaidika exemplify the intentions behind the usage of this semantic indicator: here it is used to connote the quality of an action (vaidikakarman), to declare the level of conformity to the Veda (vaidikatva), to assert the status of a teaching (vaidikaśikṣā), to indicate a convention that has been scrutinized (vaidikācāranirṇaya), to decree the poverty of ones “Vedic learning” (vaidikapāśa), and to connote the character of an intellectual ouvre (i.e., Vaidikacchandaḥprakāśa, Vaidikadharmanirūpaṇa, Vaidikasarvasva, Vaidikasubodhinī, Vaidikārcanamīmāṃsā).6 3. A short history of the usage of vaidika The term vaidika is absent from texts such as the Śatapathabrāhmaṇa, the Aṣṭādhyāyī of Pāṇini and the Nirukta of Yāska.7 It then gradually begins to appear —albeit still rarely— in later ritual brahmanical literature and then in juridical texts. These early occurrences are rather scattered and rarely more than

As in Mānavadharmaśāstra 11,96. For other examples, refer to specific entries into technical and specialised dictionaries. See Kashikar 1994, Joshi 1937-2000, Vishva Bandhu 1935-65. 7 But then present in Amarakośa 3.5[1010] (ardharcādau ghr̥tādīnāṃ puṃstvādyaṃ vaidikaṃ dhruvam). 5 6

452  Ù

The Vedas in Indian Culture and History

a singular case within a single text, as in Baudhāyanagr̥hyasūtra,8 from Paraskaragr̥hyasūtra.9 An example of the “dialectical antagonism” that constitutes the semantic ground for the genesis of our term appears in one of the ancient dharmasūtras, where the procedure for begging is prescribed: [the parivrājaka, a wandering ascetic, has to go] 26. [...] claiming, “Rejecting vedic rites and cutting ourselves off from both sides, we embrace the middle course.” apavidhya vaidikāni karmāṇy ubhayataḥ paricchinnā madhyamaṃ padaṃ saṃśliṣyāmaha iti vadantaḥ ||10

Later on, apart from few other occurrences in more recent dharmasūtras,11 the term vaidika is used to foster a famous distinction: the binary opposition proposed by Patañjali to identify two different linguistic domains, the laukika and the vaidika. Right at the opening of his Mahābhāṣya,12 Patañjali deals with the question of which words grammar is supposed to account for, stating that those are laukikānāṃ vaidikānāṃ ca (śabdānām).13 The later commentator Kaiyaṭa develops these binary criteria applying them to the well-known couple śruti/smr̥ti. He identifies 8 See Baudhāyanagr̥hyasūtra 4,3,3 (3. na jātu śyenakākādīn pakṣiṇaḥ pratiṣedhayet tadrūpās tasya pitaras samāyāntīti vaidikāḥ iti vijñāyate || iti bodhāyanīyaṃ gr̥hyaśeṣasūtre caturthapraśne tr̥tīyo ādhyāyaḥ ||); 4,7,4 (4. prathamaṃ yat pibati tena r̥g vedaṃ prīṇāti yad dvitīyaṃ tena yajurvedaṃ prīṇāti yat tr̥tīyaṃ tena sāmavedaṃ prīṇāti prathamaṃ yat parimr̥jati tenātharvavedaṃ prīṇāti yad dvitīyaṃ tenetihāsapurāṇāni yan mukhaṃ tenāgniṃ yat savyaṃ pāṇim abhyukṣati tena nakṣatrāṇi yat pādam abhyukṣati tena viṣṇuṃ yac cakṣuṣī tena candrādityau yan nāsike tena prāṇāpānau yac chrotraṃ tena diśo yad bāhū tenendraṃ yad dhr̥dayaṃ tena rudraṃ yan nābhiṃ tena pr̥thivīṃ yad aṅguṣṭhayoḥ sravanty āpaḥ kuberādayaḥ sarvā devatāḥ prīṇanty agnir vāyuḥ prajāpatir arkacandrau maghavāniti vaidikāḥ ||). 9 See Paraskaragr̥hyasūtra 2,17,9 (9. […] saṃpattir bhūtir bhūmir vr̥ṣṭir jyaiṣṭhyaṃ śraiṣṭhyaṃ śrīḥ prajām ihāvatu svāhā | yasyā bhāve vaidikalaukikānāṃ bhūtir bhavati karmaṇām | indrapatnīm upahvaye sītāṃ sā me vannapāyinī bhūyāt karmaṇi karmaṇi svāhā | […]). 10 Baudhāyanadharmasūtra 2,11,26. Transl. from Olivelle 2000. 11 See Vaiṣṇavadharmasūtra 1,30,43 (43. laukikaṃ vaidikaṃ vā api tathā adhyātmikam eva vā ādadīta yato jñānaṃ na taṃ druhyet kadā cana || 44. utpādakabrahmadātror garīyān brahmadaḥ pitā |); 2,55,18 (18. kṣaranti | sarvā vaidikyo juhotiyajatikriyāḥ | akṣaraṃ tv akṣaraṃ jñeyaṃ brahmā ca eva prajāpatiḥ ||). 12 See Patañjali, Mahābhāṣya 1,1,1 (ed. Kielhorn v. 1, pp. 8-9). Furthermore, Deshpande 1993: 19-22. This same distinction between laukika and vaidika is employed also in other domains. See, for example, Mānavadharmaśāstra 2,117; Viṣṇusmr̥ti 30,43; Śabara, Bhāṣya ad Jaimini, Pūrvamīmāṃsāsūtra 1,1,1-32. Furthermore, on mīmāṃsā and Śabara’s treatment of the distinction, Clooney 1990: 131-137; D’Sa 1980: 34-40. 13 Candotti 2005: 391-395.

To Be GIntroduction ood is To Be vaidika

Ù  453

instances of laukika with smr̥ti, and vaidika ones with śruti.14 The term vaidika is also used in such way within the epic, where it serves to qualify the status of textual materials.15 To say that something is vaidika is to say that it has a specific value. This brings us to the usage of the taxon vaidika in latest stratum of the Maitrāyaṇīyopaniṣad. 8. Now the obstacles of knowledge. O King, this net of delusion has its origin in that the godly associate with the ungodly [asvargyaiḥ saha svargyā]. And the others who, always jolly, always ajourney, always abegging, always living off skills — and the others, begging in towns, sacrificing improper substances [ayājyayājakā], accepting śūdras as pupils [śūdraśiṣyāḥ], and śūdras that know the scriptures [śūdrāś ca śāstra vidvāṃsaḥ] — and the others, rogues, wearing braided hair, dancers, fighters, homeless, wanderers, who give shows, degraded to royal service etc. — and the others who, placing ahead the cause of yakṣas, rākṣasas, spirits, ghouls, demons, serpents, planets etc., are saying: “We must appease them,” — and the others who hypocritically [vr̥thā] wear saffron [kaṣāya] robes, (glass) earrings, skulls, — and the others who wish to erect themselves as judges concerning Vedic matters [vaidikeṣu] by weaving illusions with logic, illustrations and sophisms [tarkadr̥ṣṭānta] — with all those one should not have intercourse [na saṃvaset]. Indeed they are conspicuous thieves and ungodly [asvargyā]. Thus, the text says: erring because of the sophisms, false illustrations and grounds [mithyā dr̥ṣṭānta hetubhiḥ] of the doctrine that holds there is no ātman [nairātmavāda], the world does not know what the conclusion of Vedic wisdom is [loko na jānāti vedavidyāntaraṃ]. 9. Br̥ haspati, having become Śukra, created this false knowledge [avidyām] for the security of Indra, and the ruin of the Asuras. Through it they point to what is auspicious as being inauspicious [tayā śivam aśivam ity uddiśanti], and say that one must ponder the injurious character of the scripture like the Veda etc. [vedādiśāstrahiṃsakadharma]. Hence one must not learn that knowledge, else it is like a barren woman: its fruit is mere concupiscence; even one who has fallen away from his proper conduct [vr̥ttacyutasyeva] must not embrace it. Thus the text says: “Widely opposed and differently directed are what are known as knowledge and ignorance [avidyā yā ca vidyeti jñātā]. I believe that Naciketas is desired by knowledge; the many objects of desire do not hanker after thee. He who knows these two, knowledge and ignorance [vidyāṃ cāvidyāṃ ca], will, having crossed to death by ignorance [avidyayā mr̥tyuṃ], reach non-death by knowledge [vidyayāmr̥tam aśnute]. Enveloped within ignorance, 14 See Patañjali, Mahābhāṣya 1,1,1 (ed. Kielhorn v. 1, p. 34, ad vt. 1 [laukikaḥ smr̥tyupanibaddhaḥ | vaidikaḥ śrutyupanibaddhaḥ ||]). 15 See Mahābhārata 1,117,26; 12,11,13; 12,67,5; 12,70,9; 12,70,21; 12,77,10; 12,78,2; 12,80,9 (eṣā vaidikī śrutiḥ); 12,80,13 (eṣā vaidikī śrutiḥ); 12,260,15; 12,262,23; 12,290,12; 12,324,4 (vai vaidikī śrutiḥ); 12,339,18; 13,107,1; 13,114,1; 14,13,9; 14,35,38 (eṣā vaidikī śrutiḥ); 14,36,17; 14,36,28 (eṣā vaidikī śrutiḥ).

454  Ù

The Vedas in Indian Culture and History the self-styled sages [svayaṃdhīrāḥ] who deem themselves learned run around in a rush, confused, like blind men led by blind man [andheneva nīyamānā yathāndhāḥ].” 10. The gods and asuras, being desirous of the ātman, betook themselves to Brahman. Having bowed to him they said: “Reverend, we are desirous of the ātman: teach us.” Thereupon, having pondered awhile, he thought: “The asuras are after a different ātman.” Therefore something different was taught them. Those who are confused live according to that [tad ime mūḍhā upajīvanti], being attracted to it, assaulting the Veda [taryābhighātinaḥ], they look upon untruth as truth — it is like an illusion. Hence that which is stated in the Vedas is the truth [ato yad vedeṣv abhihitaṃ 5 tat satyam]. On that which is declared in the Vedas the wise live [yad vedeṣūktaṃ tad vidvāṃsa upajīvanti]. Therefore the Brahmin should not learn non-Vedic doctrines [tasmād brāhmaṇo nāvaidikam adhīyīta], that is the meaning.16

If in Patañjali the distinction between vaidika and laukika was related to the grammarians’ urge to establish clearly defined linguistic fields, for other authors the same distinction fulfils the hermeneutic need to fix the difference between ordinary language and verbal “vedic” precepts, as in the case of Jaimini.17 This prescriptive hermeneutic partition was enthusiastically received and broadly employed, as shown in the works of Kumārila Bhaṭṭa.18 Medieval Sanskrit sources employ the notion of vaidika to confer value, purity, antiquity, and legitimacy to their points of view, to enforce and justify their ideological and exegetical innovations. Pertinent cases are those of the supplementary text Atharvavedapariśiṣṭa19 of Jayanta Bhaṭṭa’s Nyāyamañjarī,20 or of the Bhāgavatapurāṇa.21 16 Maitrāyaṇīyopaniṣad 7,8-10 (transl. from van Buitenen 1962 [the insertion of the relevant Sanskrit terms is mine]). A similar narration, although more ancient, can be found in Chāndogyopaniṣad 8,7-9. Furthermore, Br̥hadāraṇyakopaniṣad 1,3,1-10; Chāndogyopaniṣad 1,2,1-7. 17 See Jaimini, Pūrvamīṃāmsāsūtra 1,1,4-31; 1,3,30-35. Further, D’Sa 1980: 1933, Gachter 1990: 70-84, Bilimoria 1988: 84-162. 18 See Kumārila, Tantravārtika 1,3,6. 19 Atharvavedapariśiṣṭa 21,1,8 (śuddhātmāno japair homair vaidikair vītamatsarāḥ ||); 23,14,5 (niṣkāmeṇa tu yat kiṃ cit kartavyam iti vaidikam | tat sarvaṃ muktidaṃ jñeyaṃ parāparaparaṃ sukham ||); 70,2,3 (kr̥cchraṃ cāpi hitaṃ kr̥tvā kuryuḥ karma samāhitāḥ | śuddhātmāno japair homair vaidikair vītamatsarāḥ ||). 20 See, for detailed references, Freschi and Graheli 2005: 287-323. 21 See Bhāgavatapurāṇa 1,4,19-20 (19. cāturhotraṃ karma śuddhaṃ prajānāṃ vīkṣya vaidikam | vyadadhād yajñasantatyai vedam ekaṃ caturvidham || 20. r̥g yajuḥsāmātharvākhyā vedāś catvāra uddhr̥tāḥ | itihāsapurāṇaṃ ca pañcamo veda ucyate ||); 7,15,47 (47. pravr̥ttaṃ ca nivr̥ttaṃ ca dvividhaṃ karma vaidikam | āvartate pravr̥ttena nivr̥ttenāśnute ’mr̥tam ||); 8,6,9 (rūpaṃ tavaitat puruṣarṣabhejyaṃ śreyo ’rthibhirvaidikatāntrikeṇa | yogena dhātaḥ saha nastrilokān paśyāmy amuṣminnu ha viśvamūrtau ||). See, furthermore, Halbfass 1988: 359-367.

To Be GIntroduction ood is To Be vaidika

Ù  455

Interesting to note here that, later on, this use met with the disapproval of some intellectuals and religious reformers, who found it arbitrary. In 1818, for instance, Rāmmohan Rāy accused gauḍīyavaiṣṇavas of falsifying evidence by promoting the verses they composed as vaidika or paurāṇika.22 Nevertheless, due to its semiotic strength, the notion of vaidika eventually succeeded, as shown by its many appearances in later neo-Hindū publications.23 Finally, the English version of vaidika, “vedic,” gained a widespread following amongst the neo-Hindū movements and missionary agencies in Europe and North and Latin America. Having recounted the career of the term vaidika, let me come back to the text of the Mānavadharmaśāstra, where it first gained its reputation. 4. From anonymity to celebrity: the usage of the notion of vaidika in the Mānavadharmaśāstra The introduction of the term vaidika is in accordance with the aims of a “universal grammar” underlying the Mānavadharmaśāstra, a text eager to re-establish and defend specific interests. As Olivelle observed, [...] 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 Aśokan imperial polity to the foreign invasions toward the turn of the millennium.24

What now makes the defence of brahmanical interests a burning priority is the sociological and political situation, which radically changed after the spread of Buddhism and the rule of Aśoka, as a growing body of recent scholarship is convincingly demonstrating.25 The author of the Mānavadharmaśāstra wants to defend such interests by reconfiguring the ideological field. In reforming the normative and juridical classifications, he is trying to prompt a new phase of the old “vedification” and “brahmanization” processes.26 The notion of vaidika meets his needs, since it is referring to something well known by many and highly considered by See Rāmmohan Rāy 1982: 35-38. See, for example, Bhagavan Das 1917, Dayananda Sarasvati 1968, Harideva Arya 1996. 24 Olivelle 2002: 547. 25 See Bronkhorst 2007, Olivelle 2006, Bailey and Mabbett 2003. 26 See Wezler 2004: 643-646. 22 23

The Vedas in Indian Culture and History

456  Ù

the new patrons of the brahmanic elites. Those are, apparently, the main two targets of his ideological campaign, which had, in fact, a double agenda, as again stated by Olivelle: Manu’s agenda is two-fold: he wants to tell Brahmins how to behave as true Brahmins devoted to vedic learning and virtue, and he wants to tell kings how to behave as true kings, devoted to Brahmins and ruling the people justly. For this agenda he brings the authority of no less a person than the Creator himself, who is presented as the absent author of the text.27

It is in this context that the term vaidika came to be used, rapidly rising from anonymity and absence to celebrity and fame. Within a short time, the word vaidika went from having the status of a mere adjective to having that of a powerful symbolic taxon, one essential for all those whose intention was to seal a new pact between traditional lore and new political and cultural needs. What follows is the history of the textual career of the term vaidika within the Mānavadharmaśāstra. Omitting the passages that refer directly to the Veda in order to highlight the relationship between practices or objects and the Vedic lore,28 there are fourteen specific ślokas in which the term vaidika is used. All these occurrences share the following characteristics: a) Except for 2,2, none occurs in one of the interpolated sections of the Mānavadharmaśāstra;29 b) none appears to be derived or quoted from previous dharmasūtras (a fact indicated, hereafter, with the acronym NPPDhS = “Not present in preceding dharmasūtras”), and thus can reasonably be considered lexical innovations of the Mānavadharmaśāstra;30 Olivelle 2005: 41. See the index in Olivelle 2005: 1129-1130. 29 It is important here to consider the rationale of textual interpolations —attributable to later redactors— that changed the original quadripartite structure of the text, and added excurses to it. The following is a synoptic synthesis of the strata of the Mānavadharmaśāstra as indicated by the author of the critical edition: 1 (see Olivelle 2005: 52-54); 2,2-5 interpolation; 2,88-100 suspect of interpolation; 3,171-175 interpolation; 8,20-22; 8,27-40; 8,386-420 interpolation; 9,229-249; 9,294-311; 9,313-323 interpolation; 10,1-73 interpolation; 11,1-43 interpolation; 11,127-179 interpolation; 11,191-247 suspect of interpolation; 12,117126 suspect of interpolation (although the entire chapter is awkward. See Olivelle 2005: 60-62). Consequently, of the 2680 verses of the Mānavadharmaśāstra 329 —which constitute the 12% of the entire text— are indicated by Olivelle (2005: 62) as resulting from interpolations 30 The innovative character of the Mānavadharmaśāstra is clearly perceived by looking at the rate of “new” stanzas in it. While some chapters are full of quotes 27

28

To Be GIntroduction ood is To Be vaidika

Ù  457

c) except for 2,2, all have been quoted by later authors of dharma texts, suggesting the relevance of the new taxon vaidika within brahmanic discourse. The following table summarizes these points. First of all, the verses that I have classified as NPPDhS: vaidika° in Mānavadharmaśāstra 2,2 (interpolation) [NPPDhS, not re-quoted] vaidika° in Mānavadharmaśāstra 2,15 [NPPDhS, then quoted in Devaṇṇabhaṭṭa, Smr̥ticandrikā 2,426; Mādhava, Pārāśaramādhavīya 1,288] vaidika° in Mānavadharmaśāstra 2,26 [NPPDhS, then quoted in Yājñavalkyadharmaśāstra 1,10; Devaṇṇabhaṭṭa, Smr̥ticandrikā 1,36] vaidika° in Mānavadharmaśāstra 2,67 [NPPDhS, then quoted in Viṣṇudharmaśāstra 22,32; 27,14; Yājñavalkyadharmaśāstra 1,13 [Viśveśvara, Bālakrīḍā 1,15]; Devaṇṇabhaṭṭa, Smr̥ticandrikā 1,61; Aparāditya, Aparārka 908] vaidika° in Mānavadharmaśāstra 2,84 [NPPDhS, then quoted in Viṣṇudharmaśāstra Lakṣmīdhara, Kr̥tyakalpataru 3,99]

55,18;

vaidika° in Mānavadharmaśāstra 4,19 [NPPDhS, then quoted in Viṣṇudharmaśāstra 71,8; Yājñavalkyadharmaśāstra 1,99; Devaṇṇabhaṭṭa, Smr̥ticandrikā 1,132, 2,448] vaidika° in Mānavadharmaśāstra 8,190 [NPPDhS, then quoted in Devaṇṇabhaṭṭa, Smr̥ticandrikā 3,423; Mādhava, Pārāśaramādhavīya 3,208] vaidika° in Mānavadharmaśāstra 11,97 [NPPDhS, then quoted in Viśveśvara, Bālakrīḍā 3,250; Lakṣmīdhara, Kr̥tyakalpataru 3,331] vaidika° in Mānavadharmaśāstra 12,86-88 [NPPDhS, śloka 88 then quoted in Vijñāneśvara, Mitākṣarā 3,58; and paraphrases from previous sources (up to 80%), others are rather original, quoting precedent texts only for the 20% of their length.

458  Ù

The Vedas in Indian Culture and History

Aparāditya, Aparārka 1033; Lakṣmīdhara, Kr̥tyakalpataru 14,146147] Then, the remaining verses: vaidika° in Mānavadharmaśāstra 2,117 [the author elaborates on preexistent dictum (Āpastambadharmasūtra 1,5,19-20; 1,14,7-9; Gautamadharmasūtra 6,1-5; Baudhāyanadharmasūtra 1,3,25-28; Vasiṣṭhadharmasūtra 13,4143), while introducing the vaidika semantic incrementum; then quoted in Viṣṇudharmaśāstra 30,43; Aparāditya, Aparārka 54; Devaṇṇabhaṭṭa, Smr̥ticandrikā 1,97; Mādhava, Pārāśaramādhavīya 1,296, 1,301; Bhaviṣyapurāṇa 4,44-45] vaidika° in Mānavadharmaśāstra 6,75 [the author elaborates on preexistent dictum (Āpastambadharmasūtra 2,21,14-16), while introducing the vaidika semantic incrementum; then quoted in Aparāditya, Aparārka 960] vaidika° in Mānavadharmaśāstra 7,97 [the author elaborates on preexistent dictum (Gautamadharmasūtra 10,20-23), while introducing the vaidika semantic incrementum; not re-quoted] Let us now closely examine the ślokas that contain the word vaidika, starting from those from the second adhyāya.31 The first occurrence is in a section of the treatise devoted to the theme of desire: 2. To be motivated by desire is not commended, but it is impossible here to be free from desire; for it is desire that prompts vedic study and the performance of vedic rites. 3. Intention is the root of desire; intention is the wellspring of sacrifices; and intention triggers every religious observance and every rule of restraint — so the tradition declares. kāmātmatā na praśastā na caivehāsty akāmatā | kāmyo hi vedādhigamaḥ karmayogaś ca vaidikaḥ ||2.2||

The second occurrence is in a section concerning textual study and how to deal with contradictions in Law. A “vedic” text is invoked as an authority to solve exegetical conflicts: 31 All the following English quotes of the Mānavadharmaśāstra are from the translation presented in Olivelle 2005. I have inserted the Sanskrit text only for the more relevant passages

To Be GIntroduction ood is To Be vaidika

Ù  459

14. When there are two contradictory scriptural provisions on some issue, however, tradition takes them both to be the Law with respect to it; for wise men have correctly pronounced them both to be the Law. 15. After sunrise, before sunrise, and at daybreak —the sacrifice takes place at any of these times; so states a vedic scripture. udite ’nudite caiva samayādhyuṣite tathā | sarvathā vartate yajñaitīyaṃ vaidikī śrutiḥ ||2.15||32

A section devoted to consecratory rites invokes the sacral power of the vedic: 26. The consecration of the body, beginning with the ceremony of impregnation, should be performed for twice-born men by means of the sacred vedic rites, a consecration that cleanses a man both here and in the hereafter. vaidikaiḥ karmabhiḥ puṇyair niṣekādir dvijanmanām | kāryaḥ śarīrasaṃskāraḥ pāvanaḥ pretya ceha ca ||2.26||

Then another stanza from a section devoted to consecratory rites for women: 66. For females, on the other hand, this entire series should be performed at the proper time and in the proper sequence, but without reciting any vedic formula, for the purpose of consecrating their bodies. 67. For females, tradition tells us, the marriage ceremony equals the rite of vedic consecration; serving the husband equals living with the teacher; and care of the house equals the tending of the sacred fires. vaivāhiko vidhiḥ strīṇāṃ saṃskāro vaidikaḥ smr̥taḥ | patisevā gurau vāso gr̥hārtho’gniparikriyā ||2.67||

What follows is at the end of an important portion devoted to the “vedic” recitation of oṃ: 81. The three inexhaustible Great Calls preceded by OṂ and the three-footed Sāvitrī verse should be recognized as the mouth of the Veda. 82. When a man recites this verse tirelessly for three years, becoming wind and assuming an ethereal form, he reaches the highest Brahman. 83. The highest Brahman is the monosyllable OṂ; the highest ascetic toil is the control of breath; nothing is higher than the Sāvitrī; and truth is better than ascetic silence. 84. Offering ghee while seated, offering oblations while standing 32 See note 15 of this essay about the reference to vaidikī śrutiḥ in the Mahābhārata.

460  Ù

The Vedas in Indian Culture and History —all such vedic rites perish. The syllable (akṣara) OṂ should be recognized as imperishable (akṣara); it is Brahman, it is Prajāpati. kṣaranti sarvā vaidikyo juhotiyajatikriyāḥ | akṣaraṃ duṣkaraṃ jñeyaṃ brahma caiva prajāpatiḥ ||2.84||

This stanza states the ways of salutation, showing the regulative role attributed to the observance of public and clearly visible habitus: 117. He should greet first the person from whom he received knowledge —whether it is the knowledge of worldly matters, of the Veda, or of the inner self. 118. A well-disciplined Brahmin, although he knows just the Sāvitrī verse, is far better than an undisciplined one who eats all types of food and deals in all types of merchandise, though he may know all three Vedas. laukikaṃ vaidikaṃ vāpi tathādhyātmikam eva vā | ādadīta yato jñānaṃ taṃ pūrvam abhivādayet ||2.117||

Now the author of the text explains how to study śāstras, another practical way to control the level of observance and adherence to a norm: 19. Every day, he should explore the treatises —those that aid in the quick development of one’s mind, those that facilitate the acquisition of wealth, and those that promote well-being— as well as ancillary texts of the Veda; 20. for, the more a man studies treatises, the more he comes to understand and the more brightly shines his understanding. buddhivr̥ddhikarāṇy āśu dhanyāni ca hitāni ca | nityaṃ śāstrāṇy avekṣeta nigamāṃś caiva vaidikān ||4.19||

The following stanza deals with ascetic meditation, here delimited and controlled by the vaidika taxon: 74. When a man possesses right understanding, he is not fettered by actions; but when he lacks understanding, he enters the transmigratory cycle. 75. By ceasing to harm living creatures, by withdrawing the organs from their attachments, by performing vedic rites, and by practicing fierce austerities, individuals do attain that state here on earth. ahiṃsayendriyāsaṅgair vaidikaiś caiva karmabhiḥ | tapasaś caraṇaiś cograiḥ sādhayantīha tat padam ||6.75||

Now the author talks about war and the warrior ethic (with specific reference to “war booty”), topics that are also to be monitored and approved as bona fides:

To Be GIntroduction ood is To Be vaidika

Ù  461

96. Whatever a man wins —chariot, horse, elephant, parasol, money, grain, livestock, women, all goods, and base metal— all that belongs to him. 97. A preemptive share, however, should be given to the king —so states the Vedic scripture; and the king should distribute among the soldiers anything that has not been won in single combat. rājñe ca dadyur uddhāram ity eṣā vaidikī śrutiḥ | rājñā ca sarvayodhebhyo dātavyam apr̥thagjitam ||7.97||

In these stanzas, which comes from a section on deposits —under the heading of grounds for litigation— our taxon appears again: 190. Using all the investigative methods as well as vedic oaths, the judge should examine anyone accused of appropriating a deposit or of demanding a deposit that has not been made. 191. A man who does not hand over a deposit and a man who requests the return of a deposit he has not made —both these should be punished like thieves and fined an amount equal to the deposit. 192. A man who has appropriated an open deposit, as well as a man who has appropriated a sealed deposit —the king should compel both without distinction to pay a fine equal to its value. 193. If a man appropriates the property of others by fraudulent means, however, he and his accomplices should be put to death publicly using diverse modes of execution. nikṣepasyāpahartāram anikṣeptāram eva ca | sarvair upāyair anvicchec chapathaiś caiva vaidikaiḥ ||8.190||

The following quotation appears in a section devoted to drinks that can be consumed by dvija, a practice that must be regulated since it can easily cause unwanted effects: 94. Liquor is clearly the filth of various grains; sin is also called filth. Therefore, Brahmins, Kṣatriyas, and Vaiśyas must not drink liquor. 95. It should be understood that there are three kinds of liquor: one made from molasses, another from ground grain, and a third from honey. Just as drinking one of them is forbidden to Brahmins, so are all. 96. Intoxicants, meat, liquor, and spirits are the food of demons and fiends; they must not be consumed by a Brahmin, who eats the oblations to the gods. 97. When a Brahmin is intoxicated, he may tumble into filth, blabber vedic texts, or do other improper things. 98. If the brahmans resident in a man’s body is drenched with liquor even once, his Brahmin nature departs from him and he sinks to the level of a Śūdra. amedhye vā paten matto vaidikaṃ vāpy udāharet | akāryam anyat kuryād vā brāhmaṇo madamohitaḥ ||11.97||

Finally, while describing actions that lead to supreme good,

462  Ù

The Vedas in Indian Culture and History

the author applies the notion of vaidika to distinguish “good” from “bad” acts: 84. Among all these splendid activities, a particular activity has been declared as the best means for a man here to secure the supreme good. 85. Among all these, tradition holds the knowledge of the self to be the highest; it is, indeed, the foremost of all sciences, for by it one attains immortality. 86. One should understand that acts prescribed by the Veda are always a more effective means of securing the highest good both here and in the hereafter than the above six activities. 87. All these activities without exception are included within the scheme of the acts prescribed by the Veda, each in proper order within the rules of a corresponding act. 88. Acts prescribed by the Veda are of two kinds: advancing, which procures the enhancement of happiness; and arresting, which procures the supreme good. 89. An action performed to obtain a desire here or in the hereafter is called an “advancing act,” whereas an action performed without desire and prompted by knowledge is said to be an “arresting act.” 90. By engaging in advancing acts, a man attains equality with the gods; by engaging in arresting acts, on the other hand, he transcends the five elements. ṣaṇṇāmeṣāṃ tu sarveṣāṃ karmaṇāṃ pretya ceha ca | śreyaskarataraṃ jñeyaṃ sarvadā karma vaidikam ||12.86|| vaidike karmayoge tu sarvāṇyetānyaśeṣataḥ | antarbhavanti kramaśas tasmiṃs tasmin kriyāvidhau ||12.87|| sukhābhyudayikaṃ caiva naiḥśreyasikam eva ca | pravr̥ttaṃ ca nivr̥ttaṃ ca dvividhaṃ karma vaidikam ||12.88||

In the Mānavadharmaśāstra, then, the idea of the vaidika describes three kinds of behaviour: to believe “vedically” (in a specific soteriology, in a specific vision of ritual functioning [i.e., śraddhā], in a specific retributive and practical logic); to behave “vedically” (according to a specific set of norms); and to belong “vedically” (not only to a generally defined Weltanschaaung but also to a specific clan or school [śākhā]). Due to the normative goals that rule the logic of the Mānavadharmaśāstra, the term vaidika has gradually gained a broader and more inclusive significance. The old claim of “vedicity” enlarged its domain in order to exercise influence and suasion on a larger number of subjects and contexts.33 This demonstrates, then, that the semantic history of the term vaidika follows the rule that any normative discourse has to obey: 33 All the occurrences quoted indicates that the term vaidika has a larger symbolic meaning, as can be clearly seen in the case of Mānavadharmaśāstra 11,96, where a drunk brāhmaṇa may dangerously “blabber vedic texts.” In this case, vaidika is not a simple adjective indicating “of the Vedas,” but stands out as the exact ethical opposite to a context in which dvijas drink liquors and “sink to the level of a śūdra.”

To Be GIntroduction ood is To Be vaidika

Ù  463

the larger the audience, the broader significance certain notions have to assume. 5. To conclude In summary, a few pivotal issues show how innovative this use of the term vaidika is. First, it is evident that when all was “vedic” —when the cultural system based on the Veda was majoritarian and stable— there was no need to speak of “vedicity.” A reasonable majority of the intellectual field shared a world view that came from the Vedas and as a quality it remained unspoken. During the period of the Mānavadharmaśāstra, by contrast, the role, content, and customs of the Vedas were threatened and jeopardized by competing religious actors. The author of the Mānavadharmaśāstra faced social and religious rivals that threatened the very status of the brāhmaṇas. For this reason, he repeatedly points out the reciprocal bond between Veda and brāhmaṇas, stating that without the Veda there would be no brāhmaṇas and vice versa. While strategically embracing the criticisms presented by antagonistic religious leaders (in particular, the Buddha), he encourages brāhmaṇas to regain their symbolic and social primacy through strict adherence to the worldview that grants their strength, survival, and supremacy. In this way he reestablishes the old “vedic” dictum according to which the brahmanical scholar becomes the “preserver of the treasure of the Veda for men.” This is why the author of the Mānavadharmaśāstra rehabilitates this world and this culture in the eyes of brahmanical śākhās and political leaders. When everything was “vedic,” no one felt the need explicitly to qualify practices, behaviours, rituals, and textual materials as vaidika. When the socio-political context changed, the rhetoric asserting that “it has to be vaidika in order to be good” became the main tool for bolstering the identity of a specific group by calling upon their past. References Bailey, Greg and Ian W. Mabbett. 2003. The Sociology of Early Buddhism. Cambridge: Cambridge University Press. Bhagavan, Das. 1917. The science of Religion or Sanatana Vaidika Dharma. Varanasi: s.n. Bilimoria, Puruṣottama. 1988. Śabdapramāṇa: Word and Knowledge. Dordrecht: Reidel Publishing Company. Bronkhorst, Johannes. 2007. Greater Maghada. Studies in the Culture of Early India. Leiden: Brill.

464  Ù

The Vedas in Indian Culture and History

Candotti, Maria Paola. 2005. “Loke, vede, śāstre: Grammarians’ Partition of Tradition and Related Linguistic Domains.” In: F. Squarcini (ed.), Boundaries, Dynamics and Construction of Traditions in South Asia. Firenze–Delhi: Firenze University Press–Munshiram Manoharlal, 389-409. Clooney, Francis X. 1990. Thinking Ritually. Rediscovering the Pūrva Mīmāṃsā of Jaimini. Wien: De Nobili Research Library. D’Sa, Francis X. 1980. Śabdaprāmāṇyam in Śabara and Kumārila: Towards a Study of the Mīmāṃsā Experience of Language. Wien: De Nobili Research Library. Dayananda Sarasvati, S. 1968. Vaidika Manusmr̥ti. Hindī ṭīkāsahita. Ed. by S. Siddhanta Sastri. Repr. Dillī: Dehātī Pustaka Bhaṇḍāra. Deshpande, Madhav M. 1993. Sanskrit and Prakrit: Sociolinguistic Issues. Delhi: Motilal Banarsidass. Freschi, Elisa and Alessandro Graheli. 2005, Bhāṭṭamīmāṃsā and Nyāya on Veda and Tradition. In: F. Squarcini (ed.), Boundaries, Dynamics and Construction of Traditions in South Asia. Firenze–Delhi: Firenze University Press–Munshiram Manoharlal, 287-323. Gachter, Othmar. 1990. Hermeneutics and Language in Pūrvamīṃāmsā. A study in Śābara Bhāṣya. Repr. Delhi: Motilal Banarsidass. Gonda, Jan. 1997. Change and Continuity in Indian Religion. Repr. Delhi: Munshiram Manoharlal. Halbfass, Wilhelm. 1988. India and Europe. Albany, NY: State University of New York Press. Halbfass, Wilhelm. 1991. Tradition and Reflection. Exploration in Indian Thought. Albany, NY: State University of New York Press. Harideva, Arya. (ed.). 1996. Vaidika manusmrti: Manu Maharaja dvara racita prakshipta slokom se rahita grantha manusmrti. Dilli: Madhura-Prakasana. Joshi, Laxman Shastri. (ed.). 1937-2000. Dharmakośa. 17 vols. Wai: Prājña Pāṭhaśāḷā Maṇḍaḷa. Kashikar, Chintaman Ganesh. (ed.). 1994. Śrautakośa. Encyclopaedia of Vedic Sacrificial Rituals Comprising the Two Complementary Sections, Namely, the Sanskrit Section and the English Section. Pune: Vaidika Saṁśodhana Maṇḍala. Mayrhofer, Mandref. 1986-1996. Etymologisches Wörterbuch des Altindoarischen. Heidelberg: Carl Winter Universitätsverlag. Olivelle, Patrick. (ed.). 2000. Dharmasūtras. The Law Codes of Āpastamba, Gautama, Baudhāyana and Vasiṣṭha. Delhi: Motilal Banarsidass. ———. 2002. “Structure and Composition of the Mānava Dharmaśāstra.” Journal of Indian Philosophy 30, No. 6, 535-574.

To Be GIntroduction ood is To Be vaidika

Ù  465

———.(ed.). 2004. “Dharma: Studies in Its Semantic, Cultural, and Religious History. Special issue of Journal of Indian Philosophy.” Journal of Indian Philosophy 32, 421-873. ———. (ed.). 2005. Manu’s Code of Law. A Critical Edition and Translation of the Mānava-Dharmaśāstra. New York: Oxford University Press. ———. (ed.). 2006. Between the Empires. Society in India between 300 BCE and 400 CE. Oxford: Oxford University Press. Patton, Laurie L. 1994. Introduction. In: L. Patton (ed.), Authority, Anxiety, and Canon. Essays in Vedic Interpretation. Albany, NY: State University of New York. Rāmmohan, Rāy 1982. “Reply to a Gosvāmin, 1818.” In: D.H. Killingley (ed.), The Only True God, works on religion by Rammohun Roy selected and translated from Bengali and Sanskrit, with an introduction and notes. Newcastle upon Tyne: Grevatt & Grevatt, 30-45. Renou, Louis. 1960. “Le Destin du Véda dans L’Inde.” In: L. Renou, Études védiques et paninéennes 6. Paris: E. De Boccard. Smith, Brian K. 1994. Classifying the Universe. The Ancient Indian Varṇa System and the Origins of Caste. New York: Oxford University Press. ———. 1998. Reflections on resemblance, ritual and religion. Repr. Delhi: Motilal Banarsidass. van Buitenen, Johannes A.B. 1962. The Maitrāyaṇīya Upaniṣad. Hague: Mouton. Vishva, Bandhu. (ed.). 1935-1965. Vaidika Padānukramakośa (A Vedic Word-Concordance). 16 Vols. Hoshiarpur: Vishveshvaranand Vedic Research Institute. Wezler, Albrecht. 2004. “Dharma in the Veda and the Dharmaśāstras.” Journal of Indian Philosophy 32, Nos. 5-6, 629-654.

INTRODUCTION THE YĀJUṢA HAUTRA DISPUTE IN EARLY MODERN MAHARASHTRA Introduction The Yājuṣa Hautra Dispute in Early Modern Maharashtra

Ù  467 Ù  467

The The Yājuṣa Yājuṣa Hautra Hautra Dispute Dispute in in Early Early Modern Modern Maharashtra Maharashtra M M. DESHPANDE MADHAV adhav M. Deshpande University University of of Michigan Michigan While editing and interpreting an 18/19thth century text titled While editing and interpreting an 18/19 century text titled Vedavicāra ascribed to Śāmaśāstri Draviḍa Dvivedi, I came across Vedavicāra ascribed to Śāmaśāstri Draviḍa Dvivedi, I came across the following discussion: the following discussion:  rhCB!%g%I
 r k%Dx9 %*
+
r
% #B&c
J~%'%c
%
lg:'%Dx%
lg:'%c
%
(Dxj&
+
~(48
+
 r
lg:'%c
#(% % d& %Dx%
+
8"c
\ !%%%I
# n})\%nC~%% %*
+
r
(

!%(n r k9 %@!ntn%#%

#
k .
K%%

+
q\(&
!%(\lg:'%Dxj&
+

l rg*
+
&
&M/%
!ntn%#%(
,|I%C%\(

!%(%K%%\%&
&=u

\ n#~x~#~%(
,CL

$

%K%

! (
aw

%KLX{
%K%%

e

Eu&t%
&
#%& ~%g-%g{(
,&
+
gM/%rY%&#?o
(&
!%(\lg:&%Dx%
+
e
%%&
g!X%
&$%1(
% #B&(
J~%'%{
lg:&!%(/p
\ n#~%&r&YgC
%
\%n&
#CA% C%&€
sW%7 $((!%C9%&€
# k
,Fr&YC9L
+
e
^%
# mDx'c
lg:&!%%%v9
G%b9 % 45 1%+
S(