1. Early-Upanisads Patrick Olivelle OUP ( See Sanyasa Upanishads in this folder, Dharma folder, Ascetism Folder)

Page 1

The Early Upanisads

Annotated Text and Translation

Translated and Edited by: Patrick Olivelle

Page 2

THE EARLY UPANIṢADS

Page 3

SOUTH ASIA RESEARCH

Series Editor

Richard Lariviere

A Publication Series of

The University of Texas Center for Asian Studies

and Oxford University Press

THE EARLY UPANIṢADS

Annotated Text and Translation

Patrick Olivelle

INDIAN EPIGRAPHY

A Guide to the Study of Inscriptions in Sanskrit, Prakrit,

and the Other Indo-Aryan Languages

Richard Saloman

A DICTIONARY OF OLD MARATHI

Anne Feldhaus

DONORS, DEVOTEES, AND DAUGHTERS OF GOD

Temple Women in Medieval Tamilnadu

Leslie C. Orr

Page 4

THE EARLY UPANIṢADS

ANNOTATED TEXT AND TRANSLATION

Patrick Olivelle

New York Oxford

Oxford University Press

1998

Page 5

Oxford University Press

Oxford New York

Athens Auckland Bangkok Bogotá Buenos Aires Calcutta

Cape Town Chennai Dar es Salaam Delhi Florence Hong Kong Istanbul

Karachi Kuala Lumpur Madrid Melbourne Mexico City Mumbai

Nairobi Paris São Paulo Singapore Taipei Tokyo Toronto Warsaw

and associated companies in

Berlin Ibadan

The Introduction, Translation, and Notes to the Translation

first appeared in slightly different form as an Oxford World's Classics paperback

Published by Oxford University Press, Inc.

198 Madison Avenue, New York, New York 10016

Oxford is a registered trademark of Oxford University Press

stored in a retrieval system, or transmitted, in any form or by any means,

electronic, mechanical, photocopying, recording, or otherwise,

without the prior permission of Oxford University Press.

Library of Congress Cataloging-in-Publication Data

Upaniṣads. English

The early Upaniṣads : annotated text and

translation / Patrick Olivelle.

p. cm. — (South Asia research)

Includes bibliographical references and index.

ISBN 0-19-512435-9

I. Olivelle, Patrick. II. Title.

III. Series: South Asia research (New York, N.Y.)

BL1124.52.E5 1998

294.5′9218—dc21 98-17677

3 5 7 9 8 6 4 2

Printed in the United States of America

on acid-free paper

Page 6

For

Ludo Rocher

and

Richard Gombrich

And in memory of

Thomas Burrow

and

R. C. Zaehner

॥ एतेन मे निबन्धेनार्पिता वै गुरुदक्षिणा ॥

Page 8

Preface

In the summer of 1996, after the publication of my translation of the Upaniṣads in the Oxford University Press series World’s Classics, the thought occurred to me to produce an edition of the Upaniṣads containing both the text and the translation, as well as variant readings and scholarly conjectures. I consulted with my two colleagues, Richard Lariviere and Gregory Schopen, who enthusiastically endorsed the project and encouraged me to publish the volume in the new series South Asia Research published jointly by the Center for Asian Studies at the University of Texas at Austin and by Oxford University Press, New York. I want to thank Richard (the editor of the series), Gregory, and Cynthia Read (executive editor at OUP in charge of this series) for their support and encouragement. Insightful comments and criticisms of the two outside readers were very useful and deeply appreciated. Anna Shtutina read the entire Sanskrit text and the variant readings given in the notes; I thank her for her patience and diligence. I want to thank the editors of World’s Classics for permission to reproduce here much of the material published in that edition. My gratitude to the many individuals, especially Joel Breton, who helped me in the preparation of the translation has already been acknowledged in the preface to that edition. Madhav Deshpande created the beautiful fonts in which this book is set.

It may be useful to say a few words about some of the decisions and compromises that have gone into the preparation of this volume. The translation and the introduction as originally published in World’s Classics were intended for a nonspecialist audience. I had the choice of completely reworking them for the more specialized readership of the current edition. Given the constraints of time imposed on me by other research projects and academic responsibilities, I decided against it. Such a reworking would have postponed this volume almost indefinitely. I hope that even the readers of this volume will find the present translation and introduction of some use in their exploration of the Upaniṣads. Another desideratum would have been a Sanskrit word index. I started working on one, but once again constraints of time made me decide against it. Readers can find an exhaustive index in Vishva Bandhu (1945) and useful lists in Limaye and Vadekar (1958) and Jacob (1891). I go into greater detail regarding specific decisions involving the edition and the translation in my notes on the edition and on the translation. My hope, however, is that even though I have fallen short of my own ideal of what a scholarly edition

Page 9

Preface

of the Upaniṣads should be, this volume will spur further scholarly interest in these important documents of India's religious and cultural history.

Preface

The publication of this volume was aided by a University Cooperative Society Subvention Grant awarded by The University of Texas at Austin and a subvention from the Marlene and Morton Meyerson Endowment for South Asian Studies of the University of Texas at Austin.

Preface

It is always the family that bears the brunt of a project such as this. My wife, Suman, not only provided comfort and encouragement but also proofread the entire manuscript several times. She has an eye for the detail and caught many errors, especially in the Sanskrit text. My daughter, Meera, was the model of patience as I stared at a computer screen for hours on end. They provided an environment of love and peace.

Preface

Austin, Texas

Preface

April 1998

Preface

P. O.

viii

Page 10

Contents

Abbreviations, xi

Note on the Edition, xv

Note on the Translation, xx

Introduction

Introduction, 3

The Social Background of the Upaniṣads, 4

The Literary History of the Upaniṣads, 7

Vedic Rituals, 16

Vedic Cosmologies, 19

Human Physiology and Psychology, 22

Cosmic Connections, 24

Bṛhadāraṇyaka Upaniṣad, 29
Chāndogya Upaniṣad, 166
Taittirīya Upaniṣad, 288
Aitareya Upaniṣad, 315
Kauṣītaki Upaniṣad, 324
Kena Upaniṣad, 363
Kaṭha Upaniṣad, 372
Īśā Upaniṣad, 405
Śvetāśvatara Upaniṣad, 413
Muṇḍaka Upaniṣad, 434
Praśna Upaniṣad, 457
Māṇḍūkya Upaniṣad, 473

Page 11

Contents

Appendix: Names of Gods, People, and Places, 478

Contents

Notes

Contents

Bṛhadāraṇyaka Upaniṣad, 487

Contents

Chāndogya Upaniṣad, 532

Contents

Taittirīya Upaniṣad, 571

Contents

Aitareya Upaniṣad, 578

Contents

Kauṣītaki Upaniṣad, 581

Contents

Kena Upaniṣad, 596

Contents

Kaṭha Upaniṣad, 599

Contents

Īśā Upaniṣad, 611

Contents

Śvetāśvatara Upaniṣad, 614

Contents

Muṇḍaka Upaniṣad, 629

Contents

Praśna Upaniṣad, 636

Contents

Māṇḍūkya Upaniṣad, 641

Contents

Bibliography, 643

Contents

Index, 653

Page 12

Abbreviations

Aitareya Āraṇyaka. Ed. and tr. Keith (1909). Ed. with Sāyaṇa’s commentary (AnSS 38, 1959).

Aitareya Brāhmaṇa. Ed. with Sāyaṇ’s commentary. 2 vols. (AnSS 32, 1896). Tr. Keith (1920).

ALB

Adyar Library Bulletin.

AnSS

Ānandāśrama Sanskrit Series, Poona.

Aitareya Upaniṣad. Ed. with the commentaries of Śaṃkara and Ānandagiri, and Vidyāraṇya (AnSS 11, 1931).

Atharva Veda. Ed. with Sāyaṇ’s commentary by Vishva Bandhu. Vishveshvaranand Indological Series 13–17 (Hoshiarpur, 1960–64). Tr. W. D. Whitney (HOS 7–8, 1905).

BhG

Bhagavad Gītā. Ed. and tr. W. G. P. Hill (London: Oxford University Press, 1928).

BKSGW

Berichte über die Verhandlungen der königlich sächsischen Gesellschaft der Wissenschaften zu Leipzig, Philologisch-historische Classe.

O. Böhtlingk and R. Roth. Sanskrit-Wörterbuch. 7 vols. Reprint of 1855–75. Delhi: Motilal Banarsidass, 1990.

Brahma Sūtras of Bādarāyaṇa. Ed. with Śaṃkara’s commentary (Bombay: Nirnaya Sagar Press, 1948). Tr. with Śaṃkara’s commentary by G. Thibaut. 2 pts. (SBE 34, 1890; SBE 38, 1896).

BSS

Śaṃkara’s commentary on the BS.

Bṛhadāraṇyaka Upaniṣad. Ed. with Śaṃkara’s commentary (AnSS 15, 1939).

BU(K)

Bṛhadāraṇyaka Upaniṣad, Kāṇva Recension.

BU(M)

Bṛhadāraṇyaka Upaniṣad, Mādhyandina Recension. Ed. in SB by Weber.

Chāndogya Upaniṣad. Ed. with Śaṃkara’s commentary in Śrīśaṃkaragranthāvalī, 5 (Śrīraṅgam: n. d.).

EVP

Études védiques et pāṇiniennes. Renou 1955–69.

Frenz

Frenz (1967–68).

HOS

Harvard Oriental Series, Cambridge, Mass.

IHQ

Indian Historical Quarterly.

IIJ

Indo-Iranian Journal.

IndT

Indologica Taurinensia.

Page 13

Int.

Introduction to this volume.

Īśā Upaniṣad. Ed. with the commentaries of Śaṃkara, Ānandagiri, Śaṃkarānanda, etc. (AnSS 5, 1888).

IU(K)

Īśā Upaniṣad, Kāṇva Recension. Variants given in VS.

IU(M)

Īśā Upaniṣad, Mādhyandina Recension in VS.

JAOS

Journal of the American Oriental Society.

Jaiminīya Brāhmaṇa. Ed. Raghu Vira and Lokesh Chandra (Nagpur, 1954). Partly tr. Bodewitz (1973).

JIP

Journal of Indian Philosophy.

JOIB

Journal of the Oriental Institute, Baroda.

JRAS

Journal of the Royal Asiatic Society.

Jaiminīya Upaniṣad-Brāhmaṇa. Ed. Limaye and Vadekar (1958). Ed. and tr. Ortel (1896).

KaSU

Kaṭha-Śikṣā-Upaniṣad. Ed. and tr. Witzel (1979).

KaU

Kaṭha Upaniṣad. Ed. with the commentaries of Saṃkara, Ānandagiri, and Gopālayatīndra (AnSS 7, 1935).

KeU

Kena Upaniṣad. Ed. with the commentaries of Śaṃkara, Ānandagiri, Śaṃkarānanda, and Nārāyaṇa (AnSS 6, 1926).

Kāṭhaka Saṃhitā. Ed. L. von Schroeder. 3 vols. (Leipzig: F. A. Brockhaus, 1900–10).

KsB

Kauṣītaki Brāhmaṇa. Ed. Shreekrishna Sarma. Verzeichnis der Orientalischen Handschriften in Deutschland. Ed. W. Voigt. Supplementband 9,1 (Wiesbaden: Franz Steiner, 1968). Tr. A. B. Keith (HOS 25, 1920).

KS–H

A. Hillebrandt, Kleiner Schriften. Ed. R. P. Das (Stuttgart: Franz Steiner, 1987).

KS–O

H. Oertel, Kleiner Schriften, 2 vols. Ed. H. Hettrich and T. Oberlies (Stuttgart: Franz Steiner, 1994).

KS–T

P. Thieme, Kleiner Schriften (Wiesbaden: Franz Steiner, 1984).

KsU

Kauṣītaki Upaniṣad. Ed. with Śaṃkarānanda’s commentary by Cowell (1861).

Les Upanishad. Ed. L. Renou (Paris: Adrien-Maisonneuve, 1943–76).

Limaye, V. P. and Vadekar, R. D. (1958).

MaU

Māṇḍūkya Upaniṣad. Ed. with Gauḍapāda-Kārikā by R. D. Karmarkar. Reprint (Poona: Bhandarkar Oriental Research Institue, 1973).

Mbh

Mahābhārata. Ed. V. Sukthankar et al. 19 vols. (Poona: Bhandarkar Oriental Research Institute, 1927–59). Tr. J. A. B. van Buitenen. 3 vols. [containing Books 1–5] ( Chicago: Chicago University Press, 1973–78).

metri causa, for the sake of meter.

Maitrāyaṇīya Saṃhitā. Ed. L. von Schroeder. 4 vols. (Leipzig: F. A. Brockhaus, 1881–86).

ms(s)

manuscript(s).

MtU

Maitrāyaṇīya (Maitrī) Upaniṣad. Ed. and tr. J. A. B. van Buitenen (The Hague: Mouton, 1962).

MuU

Muṇḍaka Upaniṣad. Ed. with the commentaries of Śaṃkara, Ānandagiri, and Nārāyaṇa (AnSS 9, 1935).

Page 14

Praśna Upaniṣad. Ed. with the commentaries of Śaṃkara and Ānandagiri (AnSS 8, 1922).

Rgveda. Ed. with Sāyaṇas commentary by F. Max Müller. 6 vols. (London: Wm. H. Allen & co., 1849–74). Tr. K. F. Geldner (HOS 33–6, 1951–57).

Śāṅkhāyana Āraṇyaka. Ed. V. G. Apte (AnSS 90, 1922). Tr. Keith (1908).

Śatapatha Brāhmaṇa. Ed. with extracts from the commentaries of Sāyaṇa and Dvivedagaṅga by A. Weber, 1855. Reprint. Chowkhamba Sanskrit Series 96 (Varanasi: 1964). Tr. J. Eggeling (SBE 12, 26, 41, 43, 44, 1882–1900).

SBE

Sacred Books of the East, Oxford.

SII

Studien zur Indologie und Iranistik.

SILP

Studies in Indian Literature and Philosophy: Collected Articles of J. A. B. van Buitenen, ed. L. Rocher (American Institute of Indian Studies; Delhi: Motilal Banarsidass, 1988).

Śvetāśvatara Upaniṣad. Ed. with the commentaries of Śaṃkara, Śaṃkarānanda, Nārāyaṇa, and Vijñānabhagavat (AnSS 17, 1966).

Taittirīya Āraṇyaka. Ed. with Sāyaṇa’s commentary. 2 vols. (AnSS 36, 1898).

Taittirīya Brāhmaṇa. Ed. with Sāyaṇa’s commentary. 3 vols. (AnSS 37, 1898).

Taittirīya Saṃhitā. Ed. with Sāyaṇa’s commentary. 9 vols. (AnSS 42, 1900–1908). Tr. Keith (1914).

Taittirīya Upaniṣad. Ed. with the commentaries of Śaṃkara, Ānandagiri, Śaṃkarānanda, and Nārāyaṇa (AnSS 12, 1929).

VaDh

Vāsiṣṭha Dharmasūtra. Ed. A. L. Führer. Bombay Sanskrit and Prakrit Series, 23 (Poona: 1930). Tr. G. Bühler (SBE 2, 1879).

VIJ

Vishveshvaranand Indological Journal.

variant reading(s).

Vājasaneyi Saṃhitā of the White Yajurveda. Ed. in the Mādhyandina and the Kāṇva recensions with the commentary of Mahīdhara by Weber (Berlin: Ferd. Dümmler’s Verlag, 1852).

Weber

Weber’s edition of BU(M).

WZKS

Wiener Zeitschrift für die Kunde Südasiens.

ZDMG

Zeitschrift der Deutschen Morgenländischen Gesellschaft.

Page 16

Note on the Edition

Let me state at the outset that this edition of the twelve early Upaniṣads does not in any way constitute a “critical edition.” A critical edition attempts first to reconstruct an archetype that is as close to the “original” text as evidence permits and second to reveal the textual history of its reception and transmission through the analysis of manuscript variants. Normally, a critical edition also involves the creation of a genealogical tree of manuscripts that permits an editor to select readings based not merely on his or her own preferences and biases but on objective criteria. A critical edition of this type has not been attempted with regard to any ancient Upaniṣad.1

It is not only unfortunate but also a disservice to the field that many scholars, going as far back as Otto Böhtlingk, have called their printed texts “critical edition,” or “kritische Ausgabe.”2 Few, if any, of these editors bothered to investigate the manuscript tradition. When manuscripts are cited, they are few and not scientifically collated. Even the two latest examples of “critical editions,” Maue (1976) and Pérez Coffie (1994), fall short of the ideal: much of their material comes from previous “editions,” which, as we will see, are often unreliable. Maue uses six mss; Pérez Coffie uses the same six and five others. Of these eleven mss., only three are from Indian collections. Although Maue and Pérez Coffie represent a move in the right direction and have in addition given us for the first time the accented editions of the Kāṇva recension, unless a systematic survey of the Indian materials is undertaken, all “critical editions” of the Upaniṣads will fall far short of the ideal.

The edition I present here is a “conservative” one. I have tried to be as faithful as possible to the traditionally transmitted text, and I explain below the few occasions when I depart from it. My decision to be a faithful link to the traditional transmission was sparked by the inexcusable and often scandalous liberties that previous scholars have taken with Upaniṣadic texts, liberties that often amount to textual butchery. This is not the place to examine this issue in full, but a few examples will illustrate the problem. The most egregious example of textual emendation

The only truly critical edition of a group of Upaniṣads is that of Schrader (1912). Salomon (1991, 48) also observes the lack of a critical edition of any Upaniṣad and wonders whether such an edition is even feasible.
So, for example, Böhtlingk 1889a, 1889b, 1890a; Hertel 1924; Hauschild 1927; Morgenroth 1958; Frenz 1968-69. Even Oberlies (1995) depends almost entirely on the evidence of printed editions; the only two mss. he takes into account are the ones already used a century ago by Max Müller.

Page 17

without any basis in manuscript evidence but based solely on "philological skill,"

where the philologist has the hubris to substitute his judgment of what is right and

wrong for manuscript evidence, is Böhtlingk's (1889a, 1889b, 1890a) editions of

five Upaniṣads, in particular BU and CU. Whitney (1890a) already took exception

to Böhtlingk's method, noting "Least of all to be approved, perhaps, is the tamper-

ing with the traditional text in pure prose passages" (409). What Böhtlingk did was

to put his conjectures in the text itself, relegating the traditional readings to the

notes, sometimes at the bottom of the page and sometimes in the endnotes, an inver-

sion of what should have been done.3 Later scholars often failed to read Böhtlingk's

notes. So, Senart (1930) follows Böhtlingk at CU 1.4.1: om ity etad akṣaram

upāsīta, taking this to be the traditional text and not realizing that Böhtlingk had

dropped udgītham of the traditional text, which he places in his endnotes. Incorpo-

rating conjectures into the text itself becomes even more problematic when the edi-

tor changes his mind later, as Böhtlingk did in his numerous articles in the

BKSGW. Later scholars, however, did not read these articles and continued to re-

peat the emendations of Böhtlingk's editions. Let me cite one example. CU 8.15.1

reads guroḥ karmātiśeṣeṇa. Böhtlingk (1889b) changes it to guroḥ karma kṛtvāvi-

śeṣeṇa. Senart (1930), somewhat reluctantly, follows Böhtlingk's emendation,

without realizing that Böhtlingk had changed his mind in a later article (1897a) and

had accepted the traditional reading.

Reasons of meter (metri causa) is a standard, and useful, principle of textual

criticism. When properly applied, however, it is used to decide between competing

readings given in the manuscripts. Many editors of the Upaniṣads, however, have

used this principle to restore the meter and to emend Upaniṣadic verses. Caution,

however, is warranted here because, especially in the later verse Upaniṣads, we are

not sure either about the metrical standards in use or about the actual pronunciation

of words; Alsdorf (1950), for example, has used Prakritic pronunciations (e.g., bhotī

for bhavati) to restore some meters without the need for textual emendations. Max

Müller's (1879, lxxii) wise observation has often been disregarded: "The metrical

emendations that suggest themselves are generally so easy and so obvious that, for

that very reason, we should hesitate before correcting what native scholars would

have corrected long ago, if they had thought there was any real necessity for correc-

tion."4

The spate of emendations and conjectures introduced into the texts has made it

difficult in some cases to recover the traditional text and the variant readings found

actually in the manuscripts. The edition of Limaye and Vadekar (LV) is one that is

used frequently by scholars (e.g., Salomon 1991, 48), but the variants they cite are

In the CU, for example, Böhtlingk makes the following global changes without any manuscript

evidence: ātadatmyā to etādatmaka, adhidaivam to adhidaivatam, and somya to saumya, besides adding

and deleting iti at will. For a detailed study of such questionable emendations, see Olivelle 1998.

On the danger of correcting metrically incorrect verses, a good example is the three metrically

incorrect verses found with identical readings in both the Mādhyandina and Kāṇva versions of the IU,

which is part of the Saṃhitā of the VS. What would metrical corrections accomplish in such cases (see

Thieme 1965, 98)?

xvi

Page 18

Note on the Edition

sometimes suspect.5 For example, CU 4.9.2 reads kāme, a reading that caused problems for Śaṃkara himself (although unlike his modern counterparts, he tried to explain it rather than "correct" it). Deussen (1897) suggested kāmam in a note, and LV, suspiciously, gives kāmam as a variant reading; my guess is that Deussen's conjecture appears in the LV as a variant!6

The faithfulness of the native tradition of copyists and commentators stands in sharp contrast to the tampering of these texts by modern scholars. Space does not permit a lengthy defense of this statement, but a few examples will suffice. At CU 3.14.1 we have the famous expression tajjalān, which caused problems for Śaṃkara himself. He was, however, faithful to the text, although his explanation may sound fanciful to us. Böhtlingk (1889b) emends it to taj jānāni, a perfectly good and easy form, without explaining how such a simple word could have become so "corrupted" into an unintelligible jumble. Surely, the principle of lectio difficilior is the bedrock of textual criticism. Salomon (1991, 49) uses this principle to good effect in his excellent study of the linguistic peculiarities of the PU, observing that it is the existence of a large number of nonstandard Sanskrit forms that argues for the superiority of Śaṃkara's version. Indeed, it is the faithfulness of the traditional transmission that has enabled Salomon and other scholars to study the "dialectical" Sanskrit of ancient India. Surely, ancient commentators knew Pāṇini better than most modern scholars, yet they did not feel the need to correct forms and expressions that some philologists have characterized as "grammatical monstrosities." Let me cite a few examples. Hertel (1924) changes the "ungrammatical" kāmabhih of MuU 3.2 to karmabhih, but Salomon (1981, 94) sees it as a nonstandard form found also in Buddhist Sanskrit. Likewise, Hauschild (1927) changes the "ungrammatical" cetā of SU 6.11 to cettā, but the traditional reading permitted Rau (1964) to see it as an agent noun derived not from √cit but from √ci ("to punish") and to translate it as "avenger." And it was the "wrong" pronoun, the neuter tat rather than the masculine saḥ, that enabled Brereton (1986) to produce his brilliant reinterpretation of the famous "great saying" tat tvam asi.

These observations, I believe, justify a return to the traditionally transmitted text until truly critical editions of these documents become available. An editor's conjectures and insights cannot be substitutes for the received text, and they should not mediate between the reader and the text. Lest I be misunderstood, I am not suggesting that scholars should not seek to uncover the layers beneath the surface of the text. Indeed, much light has been thrown on these texts through the philological labors of scholars. I object only to conjectural philological emendations being incorporated into the edited texts of these documents. Scholarly insights and philological acumen are important for the understanding of these difficult documents, but they constitute "higher criticism" and are best relegated to notes or incorporated into translations and studies.

Among the many inaccuracies in the LV, some BU(M) variants noted in LV (e.g., BU 4.5.4) are not found in either Weber or in Böhtlingk 1889a. There are, on the other hand, many BU(M) variants that are not recorded in LV, e.g., BU 4.4.23–25; 5.5.1.
Other such suspicious variants are CU 4.9.3 prāpayatīti, CU 8.6.5 ūrdhva ākramate, both probably taken from Böhtlingk's (1889b) conjectures.

Page 19

Note on the Edition

I have given in the notes all the major recensional and manuscript variants that I could cull from the existing editions, as well as the emendations, conjectures, and suggestions offered by previous scholars.7 Given the enormity of this task, I do not claim that I have provided an exhaustive list of all such variants and emendations. But collecting at least the majority of them in one place, I believe, will be of some use to scholars and further Upaniṣadic scholarship. As Alsdorf (1950, 622) has said, “Too often emendations made by one editor or translator go unnoticed by his successor(s), so that the mere collection of all successful emendations made so far would be no useless task.” Variants recorded by an editor are noted as “vr in.” editions. It would have been preferable, as a reviewer has suggested, to identify the sources of all the variants given by different editors, but the sheer volume of those sources made that impractical. A reviewer also suggested that I explain the reasons for the emendations offered by editors or give their own explanations for the emendations. Given the vast number of such emendations, however, adding explanations to each would have made this volume unacceptably long for my publisher. So, unfortunately, readers who wish to pursue seriously the study of the variants and emendations recorded here will have to consult the editions and studies in which they originally occur. Since Böhtlingk (1889a) is based on BU(M), I have noted Böhtlingk’s variant readings only when they depart from the BU(M) readings as constituted by Weber.

I have departed from the traditional text in a few instances. In verses I have eliminated external vowel sandhi when it restored the meter; this is nonintrusive and does little damage to the text. Those of us who work with manuscripts know how inconsistent they are in noting external sandhi. I have, however, resisted reproducing in the printed text pronunciations of words required by the meter, for example, the pronunciation of semivowels as vowels (e.g., diauh for dyauh, dau for dvau); they appear strange to the eye, and even in Sanskrit it is counterproductive to attempt to duplicate in the written form all the subtleties of oral speech.

My most serious emendation of the traditional text is in the KsU. This Upaniṣad has been transmitted badly, probably because it lacked an old commentary. The rearrangement of KsU 1.3–4 by Frenz (1968–69), I think, restores a semblance of order into the narrative, and I have followed his rearrangement in my edition. Frenz also rearranges KsU 2.8–10, but his reasons for that are not as compelling. Frenz has put together two passages that deal with a rite to prevent the premature death of one’s children (KsU 8a and 8b in Frenz). But the traditional sequence has its own logic: KsU 8 = rite on the new-moon day; KsU 9 = rite on the full-moon day; KsU 10 = rite before sexual intercourse; KsU 11 = rite when one returns from a journey.

An exception is the minute emendations, especially of sandhi and of frequent words like somya (changed to saumya), of Böhtlingk. They are tiresome and insignificant. I have also ignored the variants in Little 1900 because he follows the text of Böhtlingk 1889b. I have not included obvious scribal or printing errors, but here I have erred on the side of inclusion, because one cannot always be certain whether a reading is an error and because sometimes even scribal errors can be helpful in determining the proper reading.

xviii

Page 20

Note on the Edition

The final issue relates to punctuation and the manner in which the Sanskrit text is reproduced on the printed page. This is where the hand and the mind of the editor intrude most sharply. Given the printed medium, an editor has to make choices; the alternative is to reproduce the manuscript format, which does not even divide words! It is an alternative few editors will follow. Manuscripts are notorious for their inconsistency of punctuation. I have inserted the danda punctuations in the most obvious places, taking into account both the meaning of the text and past editorial practices. Readers, however, should not take these punctuations too seriously in analyzing these texts. Having the Sanskrit text and the English translation on facing pages has its obvious advantages; it also creates serious difficulties because of the varying lengths of the text and the translation. I have had to divide the Sanskrit text into paragraphs with large spaces between them in order to accommodate the English. The paragraph divisions sometimes follow the traditional khaṇḍa division and sometimes the paragraph divisions I have made in the translation. Here, too, readers of the Sanskrit text should not pay too much heed to these divisions, knowing that the manuscripts themselves write the text in continuous and unbroken lines.

Page 21

Note on the Translation

Texts, especially ancient texts—whether they are the Upaniṣads or the Bible—composed in a different language, at a different time, and by people with social and cultural backgrounds and levels of scientific knowledge far different from our own, pose many and diverse problems of understanding and interpretation to their readers and, especially, to their translators.

A comparison with the methods we use to understand another type of data from ancient societies—archaeological discoveries—is instructive. Archaeological findings are only clues, and, just like clues in an investigation of a crime, archaeological clues require the interpretive efforts of the investigator, who alone can unravel the story behind the clues. A long process of interpretation, thus, precedes an adequate understanding, a process informed by our knowledge obtained from other sources regarding the culture and society of the people in question and by analogies with comparable cultures. A similar process of interpretation is also required for an adequate understanding of textual data and must precede any translation. Whether a translator is aware of this or not, a translation is always an interpretation.

In an important way, however, texts are different from archaeological data. Unlike archaeological remains, a text is a living reality; it is its very use and transmission by generations of readers, interpreters, and copyists that have preserved it for our examination. The interpretive history of a text is especially rich when it happens to be a sacred text, a text that is perceived by a community or a group of communities as religiously authoritative. Such, indeed, is the group of texts called “Upaniṣad” translated here.

Unlike archaeological data, therefore, texts, especially sacred texts, come to us already interpreted. Dealing now specifically with the Upaniṣads, their interpretive history consists both of formal commentaries and further commentaries on earlier commentaries, and of interpretations implicit in their use as scriptural texts within theological discourses and sectarian debates.

How, then, does a translator’s interpretation relate to this history of interpretation? The problem is further complicated by the fact that there is no one native interpretation of the Upaniṣads; across time and sectarian divides, we have a multiplicity of interpretations. If a translation has a theological purpose—if it is produced within a specific sectarian or theological context—a translator may choose one of these interpretations over the others. I have chosen not to do so, even though, like

Page 22

Note on the Translation

most translators, I have benefited by the insights of commentators. I want in my translation and notes to approximate, as far as our current knowledge permits, the understanding of these documents that their authors had and the meaning they desired to communicate to their contemporary audience.

Like any other historical work, mine is a reconstruction of the past. In this reconstruction, I want to distinguish the interpretive history of the documents, often separated from their composition by a millennium or more, from their original context. Even though there are significant differences in purpose and content between them, one may, nevertheless, profitably compare the Upaniṣads to the Constitution of the United States of America, which also has had an official, and often contradictory, history of interpretation by the Supreme Court. If I am translating the Constitution into Sanskrit, what interpretation should I follow? I would follow none of those interpretations and attempt to reconstruct for the Sanskrit reader the cultural and social context within which the Constitution was drafted. My translation would attempt to present the Constitution to the Sanskrit reader as a window into the world of America in the late eighteenth century, and not as a living document still guiding the destinies of the American people.

Both the Constitution and the Upaniṣads are living documents and play significant roles in the communities within which they are perceived as authoritative. Acts of interpretation of these documents are legitimate activities for lawyers and theologians of the respective communities, and the study of those interpretive histories is an important and legitimate part of historical scholarship; but these are not, I believe, the aim of a translation. My translation is not intended to be a vehicle for propagating religious truths (although, for some, it may perform this function) but for illuminating the distant past of India.

This translation was first published in the Oxford University Press series World’s Classics. For reasons spelled out in the Preface, I have not changed it substantially here, except for correcting a few errors and emending the translation of some passages. The major emendations are listed at the end of this note. This translation was not intended primarily for philologists but for ordinary readers who have little or no access to the original Sanskrit. I have not employed, therefore, the common defensive strategies of philological translations, such as placing within parentheses any English word added to draw out the sense of the Sanskrit. I expect my translation to be accurate without being literal, to be readable on its own without reference to the Sanskrit. I have used idiomatic and informal English, especially in translating dialogues and conversations, but avoided vulgarisms, keeping in mind that these are viewed by many as sacred writings. Although originally intended for a different audience, I hope this translation will be of use also to the more specialized and sophisticated readership of this edition in untangling the complex meanings of these documents.

One notable feature of Sanskrit is its frequent use of pronouns without clear antecedents; often the antecedent is not the noun that immediately precedes a pronoun. I have regularly repeated the noun when the use of an English pronoun would be confusing. In dialogues deictic pronouns—that is, pronouns used in conversations while pointing to something—are used frequently; in translating such oral

xxi

Page 23

usages into written prose I have often appended the item pointed to, for example, "this right eye." The use of these pronouns highlights not only the oral nature of the original dialogues but also the continuing oral transmission of the Upaniṣads.

Phonetic connections between words abound in these documents and play a significant role in Upaniṣadic thought. It is an impossible task to reproduce them in English. I have attempted to alert the reader to the connections hinted at by placing the Sanskrit terms within parentheses.

Three terms that cause special difficulty for the translator are prāṇa, ātman, and brahman; they have multiple meanings, but in the original Sanskrit the identity of the term recalls to the reader all the related meanings even when only one is primary within a given context. To alert the English reader to these connections, I have again placed the Sanskrit term within parentheses.

For the benefit of Sanskrit scholars who will be the primary readers of this edition, I note here some of the translation choices I have made. The Upaniṣads contain numerous nominal sentences in which two nouns in the nominative stand in apposition; these have been studied exhaustively by Gren-Eklund (1978). In translating these I have taken the first noun to be predicative (what Gren-Eklund calls the "comment") of the second (Gren-Eklund's "topic"). Thus vāg vai brahma (BU 4.1.2) I translate as "Brahman is speech." Likewise, when two accusatives stand in apposition as objects of the same verb, I have taken the second to be the direct and the first to be the indirect object. Thus, atha ha vācam udgītham upāsāṁ cakrire (CU 1.2.3) I translate as "Then they venerated the High Chant as speech." In translating the particles kila and khalu, I generally follow the insightful comments of Emeneau (1968-69) and Daalen (1988). And in translating iva I have followed Breton (1982); it indicates a clear affirmation, but in a general and undefined way, rather than an expression of a doubt or an attempt to hedge ("as it were").

Major Changes from the translation in World's Classics:

BU 1.4.17; 2.4.10; 3.7.1; 4.5.11; 6.2.8; CU 3.11.3; 5.1.2; 5.2.2; 5.3.2; 5.3.7; 5.4.2; 5.5.2; 7.8.1; 8.1.3; 8.4.2; TU 2.1, 3, 4, 5; 3.1; 3.10.4; AU 2.1, 3; KsU 1.4b; 2.14; KaU 2.10–11; SU 2.2; 4.5; MuU 1.2.3; 2.2.2, 6, 11; 3.2.2, 4.

Page 24

THE EARLY UPANIṢADS

Page 26

Introduction

The Upaniṣads translated here represent some of the most important literary products in the history of Indian culture and religion, both because they played a critical role in the development of religious ideas in India and because they are valuable as sources for our understanding of the religious, social, and intellectual history of ancient India. The Upaniṣads were composed at a time of great social, economic, and religious change; they document the transition from the archaic ritualism of the Veda into new religious ideas and institutions. It is in them that we note for the first time the emergence of central religious concepts of both Hinduism and of the new religious movements, such as Buddhism and Jainism, that emerged not long after the composition of the early Upaniṣads. Such concepts include the doctrine of rebirth, the law of karma that regulates the rebirth process, and the techniques of liberation from the cycle of rebirth, such as mental training associated with Yoga, ascetic self-denial and mortification, and the renunciation of sex, wealth, and family life. Even though theoretically the whole of the vedic corpus is accepted as revealed truth, in reality it is the Upaniṣads that have continued to influence the life and thought of the various religious traditions that we have come to call Hindu. Upaniṣads are the scriptures par excellence of Hinduism.

In translating and interpreting these ancient documents, I have drawn on the most current philological, historical, and anthropological research available to me, research that provides glimpses into the social and intellectual world of the ancient Upaniṣads. Some of that research is presented in the notes, and my debt to previous scholarship is thereby acknowledged. Much of it is implicit in the translation itself. In this introduction, I want to supplement and to organize the information scattered in the notes—information that will give the reader the social, religious, and cultural background of these documents, information without which much of what is said in them cannot be understood. Given the constraints imposed by the very nature of an introduction and the limits of space prescribed by the format of this series, it is impossible to explore fully that background. I have deliberately aimed these introductory remarks at nonspecialist readers, knowing that the specialists can fend for themselves. My hope is that the information given here, although inadequate, will provide a sufficient background to appreciate these documents within their historical contexts. My task is complicated by the fact that the Upaniṣads translated here were composed over a period of about six centuries, a period which saw many and far-reaching social, economic, and religious changes. My emphasis will be on the early period, which produced some of the major Upaniṣads, including the

Page 27

Bṛhadāraṇyaka and the Chāndogya, although I will deal with some of the doctrines

of the Sāṁkhya and Yoga traditions and with the monotheistic and devotional ten-

dencies that underlie several later Upaniṣads.

In this introduction I have avoided speaking of "the philosophy of the

Upaniṣads," a common feature of most introductions to their translations.1 These

documents were composed over several centuries and in various regions, and it is

futile to try to discover a single doctrine or philosophy in them. Different theologi-

ans, philosophers, and pious readers down the centuries both in India and abroad

have discovered different "truths" and "philosophies" in them. That has been, after

all, the common fate of scriptures in all religions. Even in the future, that is an en-

terprise best left to the readers themselves, and the prudent translator will try to step

aside and not get in their way.

The Social Background of the Upaniṣads

Modern scholarship has unveiled to some degree the distant and long-forgotten past

of northern India. Prior to the middle of the second millennium BCE, this region was

probably populated by numerous groups of diverse ethnic and linguistic back-

grounds. About most of them we know very little. About one group, however, we

have a wealth of information unearthed by archaeologists early in the twentieth

century. This group of people built an urban civilization along the Indus River

(hence the name Indus Valley Civilization); it lasted from around 2300 until the

middle of the second millennium BCE and centered on the two major cities of

Harappa and Mohenjo Daro.2 This urban civilization died without exerting any no-

ticeable influence on the civilizations that followed.

At about the time when the Indus Valley Civilization became extinct, according

to the commonly accepted theory, there took place a relatively large migration of

people from the west into the upper Indus Valley.3 They were a pastoral but mili-

tarily powerful people who called themselves ārya (literally, ‘noble’ or ‘honorable’;

whence the word Aryan). Modern linguistic studies have demonstrated that the lan-

guage of these people, the language in which the ancient vedic literature including

the Upaniṣads was written and which later came to be known as Sanskrit, belongs to

For an eminently readable yet accurate, unbiased, and brief presentation of the main doctrinal

themes of the Upaniṣads, I refer the reader to Brereton 1990.

For information about this civilization see B. and F. R. Allchin, The Rise of Civilisation in India

and Pakistan (Cambridge: Cambridge University Press, 1982); Gregory L. Possehl (ed.), Ancient Cities

of the Indus (Durham, NC: Carolina Academic Press, 1979); Gregory L. Possehl (ed.), Harappan Civi-

lization: A Contemporary Perspective (Delhi: Oxford University Press, 1982).

The Aryan migration theory has been challenged recently by several archaeologists: see Colin

Renfrew, Archaeology and Language: The Puzzle of Indo-European Origins (London: Jonathan Cape,

1987). Even though the migration theory possibly needs revisions and the Aryan migrations may have

occurred at different times and at a much earlier period, I think the preponderance of evidence still sup-

ports the view that there was an Aryan migration into India in the second millennium BCE and that the

language of the Veda reflects the language of those Aryans. For the most recent study of this question by

leading archeologists and textual scholars, see The Indo-Aryans of Ancient South Asia, ed. G. Erdosy

(Berlin: Walter de Gruyter, 1995).

Page 28

Introduction

a family of languages (generally referred to as Indo-European) that includes Greek, Latin, and most of the modern European and northern Indian languages, as well as the ancient and modern languages of Iran. This distribution of languages suggests that a related group of nomadic pastoral people migrated across Europe and toward the east, settling first in Iran and finally in northern India.

Introduction

The Aryans who came into India first settled the fertile land of the upper Indus tributaries (the area of present-day Punjab) but soon migrated farther east into the Ganges Valley. Although the material culture of the Aryans was much inferior to that of the Indus Valley Civilization--urban culture, for example, would not rise again for another thousand years--they left behind a vast corpus of literature, the Vedas, and for this reason the Aryan civilization from the time of the composition of the earliest Ṛgvedic hymns, probably in the last few centuries of the second millennium BCE, until about the fifth century BCE is known as the Vedic Civilization.4

Introduction

The Aryans dominated the native populations across much of northern India. Their military prowess may have had many causes, but the horse and the horse-drawn chariots that they possessed must have been an important factor. The horse remained a central symbol of royal power in ancient India, and the horse-sacrifice that figures so prominently in the opening chapter of the Bṛhadāraṇyaka Upaniṣad became the primary ritual expression of that power. But their domination was not necessarily numerical. "The evidence of widespread settlement in the Ganga Valley by the late 2nd millennium B.C.," Erdosy (1988, 101) points out, "would suggest, that numerically the native population would have been overwhelmingly dominant." It is this blend of indigenous and Aryan peoples that constituted vedic society.

Introduction

The subjugated non-Aryans appear to have been by and large relegated to the lower class of an emerging quadripartite social structure: (1) the élite ruling cum military class generally referred to as Kṣatriya; (2) the hereditary priestly class of Brahmins; (3) the large group of peasants and artisans known as Vaiśya; and finally (4) the Śūdra group, which included a motley array of people, including subjugated non-Aryans, servants, and slaves. These four social groups were called varṇa (lit., 'color'), the first three consisting at least in theory of Aryans.

Introduction

These groups were not as watertight as the later caste divisions of Indian society, and there appears to have been some mobility across the groups. The varṇa division of society, however, has remained from the vedic period until modern times the primary theological and theoretical conception of society in India,5 and it underlies much of the discussion in the Upaniṣads. By the time of the latest hymns of the Ṛgveda, probably the early centuries of the first millennium BCE, the varṇa division of society had become theologically so central that a creation hymn (RV 10.90) depicts the emergence of the four classes from the mouth (Brahmin), arms (Kṣatriya), loins (Vaiśya), and feet (Śūdra) of a primeval man (puruṣa), whose sacrifice and dismemberment created the universe.

Introduction

The best and the most accurate description of vedic society drawn from literary sources is Rau 1957. Witzel (1987, 1997b, 1997c) has provided the first ever social history of the vedic texts. Accessible, though somewhat dated, studies on vedic religion and mythology are: Macdonell 1898; Keith 1925.

Introduction

For a discussion of the varṇa ideology and its use as a classificatory principle, see Smith 1994.

Page 29

The Early Upaniṣads

How much influence the non-Aryan religious and cultural traditions may have exerted on the dominant vedic culture has been a matter of much debate and controversy. That there must have been some influence is unquestionable, but to identify specific non-Aryan elements of vedic culture is, I believe, a nearly impossible and an altogether futile enterprise.6 As Erdosy (1988, 101) has argued, the very durability of the social order that the Aryan migrants created indicates that the Aryans may have utilized and exploited the social divisions of the non-Aryan groups themselves. It is likely that the élites of these groups were co-opted into the Kṣatriya ruling class and that the Brahmin priestly class incorporated both Aryan and non-Aryan ritual specialists. The Aryans themselves soon lost any collective memory they may have had of having migrated from the west; none of the early vedic texts preserves any such memory.

The Early Upaniṣads

The influence of the native peoples was probably most marked in the economic area. The Aryans, who had been nomadic cattle herders, adopted much of the economic system of the natives; the economy of the vedic society along the Ganges Valley during circa 1000–600 BCE was by and large an agricultural one, although animal husbandry, especially cattle, continued to play an important role, and cows were the symbol of wealth. Rice, a native cereal of the Ganges region, became their staple food. There is also evidence of crafts, especially pottery, textile, and metal work. The vedic texts themselves attest to the existence of trade and traders.7 The economy, however, was primarily agricultural and was based on villages; vedic society remained primarily village-based until about the sixth century BCE.

The Early Upaniṣads

Beyond the villages and their internal organization, there were broader political units run by chieftains who ruled over many villages. “Already in the 10th–7th centuries B.C.,” Erdosy (1988, 55) claims, “we can see the presence of a two-tier hierarchy of settlements, and suggest that the largest centre existed to control the movement and processing of raw materials, exploiting its position at the boundary of two zones with distinct resources.” These chiefdoms were probably the largest political units during this period, controlling crafts and commerce, collecting taxes and tributes, and exercising military and judicial powers.

The Early Upaniṣads

By the late vedic period these units had been sufficiently consolidated so that we can speak of the emergence of kingdoms and a monarchical form of government along the Ganges. By about the sixth century BCE—that is, during the period of the first Upaniṣads—many relatively large kingdoms had been created. Kings and royalty, as the Upaniṣadic evidence itself indicates, began to play an increasing role both in the political economy of northern India and in the area of religious thinking and institutions.

The Early Upaniṣads

A major issue regarding late vedic society is urbanization. After the demise of the urban civilization of Harappa and Mohenjo Daro, northern Indian society was essentially village-based. The available archaeological and literary data indicate the rise of cities along the Ganges Valley between the sixth and fourth centuries BCE.8

The Early Upaniṣads

For a discussion of this issue, see Olivelle 1993, 68-69.

The Early Upaniṣads

See Rau 1957, 28.

The Early Upaniṣads

The most recent and very readable study of the archaeological evidence for the emergence of cities and states in the Ganges Valley in the sixth to fourth centurics BCE is Erdosy 1988.

Page 30

Introduction

The cities functioned as administrative, commercial, and military centers of the kingdoms. "If the previous period's findings indicated a chiefdom," writes Erdosy (1988, 116), "here we may justifiably argue for the first emergence of a state level of political organization." These relatively large realms facilitated commerce and travel, permitting not only the distribution of goods but also the dissemination of religious ideas and institutions.

Introduction

It is, however, uncertain whether the urbanization of the Ganges Valley occurred before or after the composition of the early prose Upaniṣads and what influence, if any, it had on the development of Upaniṣadic thought. The society and culture reflected in these documents, nevertheless, are far different from those of the early vedic period. At least some of the new ideas and institutions, especially asceticism and celibacy,9 I believe, reflect, if not an urban environment, then at least one that is in the process of urbanization. The vast geography known to the Upaniṣads indicates the relative ease of travel and commerce across much of northern India. They refer to the Gandhāra region of the northwest and the Videha region of the southeast, two regions separated by over, 1,600 kilometers. People traveled long distances not only to trade but also to acquire knowledge (BU 3.3.1). It is also interesting in this context to note that there are very few agricultural metaphors and images in the Upaniṣads, while examples derived from crafts such as weaving, pottery, and metallurgy are numerous. These crafts, of course, could appear in village life, but the dominance of craft metaphors at least suggests a milieu somewhat removed from the agricultural routine of villages. A close reading of these texts suggests to me that, by and large, their social background consists of court and crafts, rather than village and agriculture. The later verse Upaniṣads were certainly composed after the rise of urbanization and possibly even after the creation of the Maurya empire in the late fourth century BCE.

The Literary History of the Upaniṣads

The Upaniṣads, on the one hand, are portions of a large body of sacred texts collectively known as the Veda and are thus an integral part of the fundamental scriptures of most people we have come to call "Hindu"; it is in this manner that the Upaniṣads have been transmitted through the centuries, and we should, therefore, consider their position within the broader vedic corpus. They are, on the other hand, documents composed and edited by individuals at given moments in history and in specific geographical locations; we, therefore, need to look at the history of their composition.

The Upaniṣads within the Vedic Corpus

In the brief introductions to individual Upaniṣads, I identify each as belonging to a particular branch of the Veda. The literature of the Veda was produced by and

The Upaniṣads within the Vedic Corpus

For a summary of how this new urbanization may have affected religious ideas and institutions, see Olivelle 1993, 55–58.

Page 31

largely intended for Brahmins. Different families of Brahmins became specialists in different aspects of the many and complicated sacrifices that dominated their lives and thought. These specialized family traditions developed into the vedic "branches" (śākhā, sometimes also translated as "school") within which specialized ritual texts were produced and transmitted. The literary structure of the vedic corpus, therefore, mirrors the division of the priestly community into branches specializing in different aspects of the complex sacrificial ritual.

The Veda is broadly divided into three sections: Rgveda, Sāmaveda, and Yajurveda, but in time a fourth, the Atharvaveda, came to be added to these three. The first identity of a Brahmin is determined by his affiliation to one of these vedic traditions; I will explain below the different priestly functions associated with each. Further "branching" of the vedic traditions took place as a result of numerous factors, including geographical location, ritual specialization, and doctrinal and ritual disputes.

Each of these vedic branches has as its foundational text a "collection" (Samhitā) of verses or liturgical formulas and a prose text (Brāhmaṇa) explaining the meaning of the liturgy (see Fig. 1). The Samhitā of each Veda is by and large common to all its branches, even though some may have their own recension of it, while each branch has its own Brāhmaṇa. The Brāhmaṇas were not the work of single authors, and periodic additions were made to them. These additions included esoteric material explaining the hidden meanings of ritual actions and words. Some of these esoteric sections of the Brāhmaṇas came to be called Āraṇyakas (texts that were to be recited in the wilderness outside the village), while others came to be called Upaniṣads. The distinction between these two groups of texts is not altogether clear, since both deal with similar material; some of the Upaniṣads, such as the Aitareya, are embedded within their respective Āraṇyakas, while others, such as the Bṛhadāraṇyaka (lit. ‘Great Āraṇyaka’), are viewed as both Āraṇyaka and Upaniṣad. Cosmological and metaphysical topics generally occupy a more central position in the Upaniṣads, however, than in the Āraṇyakas, and the Upaniṣads are, by and large, later than the Āraṇyakas. Figure 1 illustrates the position of the early Upaniṣads within the framework of the vedic corpus.

The vedic texts, including the Upaniṣads, were composed and at first transmitted from generation to generation orally and within their respective vedic branches. The reader will observe, and I have frequently pointed out in the notes, the many instances where the oral nature of the Upaniṣads is manifest, as when the author uses deictic pronouns ‘this’ and ‘this’ to refer to his two eyes, or ‘this (here)’ and ‘that (over there)’ to refer to the earth (or something on earth) and the sun.

For a survey of this literature, see M. Winternitz, A History of Indian Literature, tr. S. Ketkar, Vol. 1 (New Delhi: Oriental Books Reprint Corp., 1972); J. Gonda, Vedic Literature (Wiesbaden: Harrassowitz, 1975). For the social history of these texts, see Witzel 1987, 1989, 1997b, 1997c.
The Atharvaveda is not directly connected with the vedic liturgical tradition and incorporates beliefs and rituals from traditions of healing and magic. Upaniṣads ascribed to the Atharvaveda are generally late and were probably composed as independent documents outside the vedic branches.
The Samhitā of the Rgveda is in verse and contains 1,028 hymns divided into ten books; the Samhitā of the Sāmaveda consists mostly of Rgvedic verses set to music; and the Samhitā of the Yajurveda is in prose and contains formulas that are recited during a sacrifice.

Page 32

ṚGVEDA YAJURVEDA SĀMAVEDA ATHARVAVEDA

Black (Kṛṣṇa) White (Śukla)

Ṛgveda Saṃhitā

Ṛgveda Saṃhitā Taittirīya Saṃhitā Kāṭhaka Saṃhitā Vājasaneyi Saṃhitā Sāmaveda Saṃhitā Atharvaveda Saṃhitā (Śaunaka)

Aitareya Brāhmaṇa Kauṣītaki/Śāṅkhāyana Brāhmaṇa

Aitareya Brāhmaṇa Kauṣītaki/Śāṅkhāyana Brāhmaṇa Taittirīya Brāhmaṇa Śatapatha Brāhmaṇa Talavakāra (Jaiminīya) Brāhmaṇa Gopatha Brāhmaṇa Chāndogya Brāhmaṇa

Aitareya Āraṇyaka Śāṅkhānaya Āraṇyaka

Aitareya Āraṇyaka Śāṅkhānaya Āraṇyaka Taittirīya Āraṇyaka

Aitareya Upaniṣad Kausītaki Upaniṣad

Aitareya Upaniṣad Kausītaki Upaniṣad Taittirīya Upaniṣad Śvetāśvatara Up. Katha Upaniṣad Bṛhadāraṇyaka Upaniṣad Chāndogya Upaniṣad Muṇḍaka Upaniṣad Praśna Upaniṣad Māṇḍukya Upaniṣad

Īśa Upaniṣad Jaiminīya-Upaniṣad-Brāhmaṇa Kena Upaniṣad

NB This chart is not comprehensive, and there are other vedic branches and texts, which are not included because they do not have extant Upaniṣads. The chart is intended only to indicate the position within the broader vedic corpus of the Upaniṣads translated or referred to in this book.

Fig. 1. The Upaniṣads within the Vedic Corpus

Page 33

writing down of these documents did not take place for perhaps a thousand years,

but the transmission was, nonetheless, very faithful—more accurate, I might add,

than most manuscript traditions—and the priestly tradition developed many, often

artificial, devices to ensure a faithful transmission.

Toward the last centuries BCE and certainly by the first centuries of the common

era, the role of the sacrifice within religion and of the vedic branches within

Brahmanical learning became less significant. Specialized traditions of learning

(śāstra) came into existence, first Sanskrit grammar and then others such as relig-

ious law (Dharmaśāstra), political science (Arthaśāstra), and medicine, traditions

that cut across vedic branches. Even within ritual and religious practice, Brahmani-

cal thought came to consider the literature of all vedic branches—that is, the totality

of the Veda—as authoritative over individuals in every vedic branch. The Upa-

niṣads themselves became somewhat detached from their respective vedic branches

and became the common property of all Brahmins under the generic title “Vedānta,”

meaning the end or conclusion and, in an extended sense, the essence of the Veda.

An early effort, probably within the first five hundred years of the common era,

to make a systematic presentation of Vedāntic doctrines is the Vedāntasūtra (or

Brahmasūtra) ascribed to Bādarāyaṇa.13 The Upaniṣads came to be considered the

section of the Veda containing salvific knowledge (jñānakāṇḍa), whereas the other

sections contained information about rites (karmakāṇḍa). As the revealed source of

knowledge, therefore, the Upaniṣads became the basic scriptural authority for most

later Indian theological traditions.

Given the importance of these basic texts, numerous documents, often espous-

ing sectarian viewpoints, were composed with the title “Upaniṣad,” and, at least

among some segments of the population, they enjoyed the authority and sanctity

attached to the Vedas; most of these late texts are ascribed to the Atharvaveda. Such

Upaniṣads continued to be produced possibly as late as the sixteenth century CE and

number in the hundreds. In the first half of the second millennium CE, furthermore,

the early Upaniṣads were detached from the Brāhmaṇas of which they were a part

and, together with these later Upaniṣads, gathered into collections; the number of

Upaniṣads included in these collections varied according to the region, 52 being a

common number in the north and 108 in the south.14