The Bhagavad Gita

  The author and publisher have provided this e-book to you for your personal use only. You may not make this e-book publicly available in any way. Copyright infringement is against the law. If you believe the copy of this e-book you are reading infringes on the author’s copyright, please notify the publisher at: us.macmillanusa.com/piracy.

  Contents

  Title Page

  Copyright Notice

  Dedication

  Epigraph

  A Note on This Translation

  A Note on the Pronunciation of Sanskrit

  Introduction

  Chapter One

  Chapter Two

  Chapter Three

  Chapter Four

  Chapter Five

  Chapter Six

  Chapter Seven

  Chapter Eight

  Chapter Nine

  Chapter Ten

  Chapter Eleven

  Chapter Twelve

  Chapter Thirteen

  Chapter Fourteen

  Chapter Fifteen

  Chapter Sixteen

  Chapter Seventeen

  Chapter Eighteen

  Notes

  Selected Bibliography

  Acknowledgments

  About the Author

  Copyright

  For

  Susan, Akira, Nik, and Ruth,

  and for

  my entire kulaparamparā

  có akáre paramé vyòman yásmin dev ádhi víśve niedú |

  yás tán ná véda kím c kariyati yá ít tád vidús tá imé sám āsate ||

  —RIGVEDA 1.164.39, THE RIDDLE HYMN ON “OM”

  That syllable of the Rigveda, that highest space where all the gods have taken their places. What will he do with the Rigveda who does not know that syllable? Only those who know that sit together here with me!

  A Note on This Translation

  Charles Wilkins published the first translation of the Bhagavad Gītā into English in 1785, and there has been a steady stream of new translations ever since then. So inevitably a question arises that a translator of a new one must face: Why yet another version? This is my best attempt at an answer.

  I have been teaching this text for many years, and I have used many of the available translations. There are quite a few reasonably reliable and accurate translations among them, but students have generally not responded well to the scholarly ones. I have no elaborate theory of translation, but because I am a philologist, I prefer an accurate translation to a loose one that may be more eloquent. Some eloquent translations of the Bhagavad Gītā have been made by people with little or no training in Sanskrit. Occasionally a student will ask me for permission to use one of them. I take such opportunities to compare translations, to show my students how different they can be. In this way, I emphasize the importance of choosing a translation that is accurate over one that may be easier to read or that appeals to a preconception that we might have of a given text, or religion, or culture. The popular literature that is readily available, not only in print but also increasingly online, concerning the traditions that I study and teach tends to conceal or ignore the little details that can make one tradition or text or passage distinctive, and therefore interesting.

  I think it is necessary to try to present a text like the Bhagavad Gītā not only as accurately as possible, but also with a commitment to conveying its importance, both to those within its social and historical context—which in the case of the Bhagavad Gītā is very great indeed!—and to those whom we invite (or, in the case of students, require) to read it.

  One of the reasons that I have taken up this translation is that I have long felt the need for one that is both accurate and engaging—and perhaps even eloquent. Over the years my students have complained that the verse translations of the Bhagavad Gītā were awkward, stiff, and sometimes hardly poetic at all. But my decision to make a vigorous, rhythmic prose translation is also rooted in my aim to make, first of all, an accurate translation. The Bhagavad Gītā is composed in a metrical form, mostly in a quatrain of eight-syllable lines known in Sanskrit as the śloka, or the anuubh. Occasionally the Bhagavad Gītā shifts into longer quatrains of four lines of eleven syllables, in interesting ways and at interesting times. But I think it has been a mistake to try to imitate or reproduce the quatrain stanzas in English.

  The Bhagavad Gītā is a small section of that massive Sanskrit epic the Mahābhārata, the great encyclopedia of classical Indian culture. At its core, this encyclopedia is essentially an epic tradition; that is, it is fundamentally a narrative of war. The meter of Sanskrit epic poetry is not an elegant, finely tuned, finely crafted musical composition. In Sanskrit literature, such musical compositions were known as kāvya, a refined and complex genre of courtly poetry, intended for a highly educated elite audience. Sanskrit epic poetry, on the other hand, was governed by a loose and irregular meter, the primary function of which was simply to move the narrative along, quickly and easily. The Sanskrit of the epics, and of the Bhagavad Gītā, is much closer in fact to the colloquial everyday Sanskrit of the first century of the Common Era (C.E.). It is a vibrant and rhythmic kind of Sanskrit, much closer to what the spoken Sanskrit of the Brahmins and the warrior elite (katriyas) must have sounded like at the time than to the earlier fundamentally esoteric poetry of the Rigveda, or to the later poetic language of Kālidāsa and the other courtly poets of classical India. Like the language of other epic poetries known to us, it is formulaic and improvisational. It is fast-paced and was readily accessible to its intended audience. What such an epic poetry needs in translation is a similarly colloquial, direct, and fast-paced English. That is what I have tried to produce in this translation.

  To be sure, the Bhagavad Gītā does have highly poetic, even lyrical moments. It uses poetic figures more frequently than the Mahābhārata does as a whole, which is one reason why it is considered a later addition to the Mahābhārata. Nonetheless, the Bhagavad Gītā’s poetic context, or rather its performance context, is much closer to the contemporary vernacular—that is, to ordinary speech—than to a high-register poetic language. And while the text undoubtedly contains many passages of genuine poetic brilliance and intensity, one also has to acknowledge that many passages in the Bhagavad Gītā are poetically very ordinary, and that a few are unconvertible, even by a great translator, into genuine poetry. Many of them come in the last six chapters, but even in the earlier chapters one sometimes encounters strange, awkward passages. One well-known example is Chapter 6, stanzas 5–7, where, in the course of two and a half brief ślokas, we encounter the word ātman no fewer than fifteen times. In stanza 6.5 alone—a mere thirty-two syllables in length—the word occurs seven times! I have spent a lot of time trying to work out how to translate this passage, and this is the best I could come up with:

  5. One should lift oneself up by means of the self. Do not degrade the self, for the self is one’s only friend, and at the same time the self is one’s only foe.

  6. The self is one’s friend when one has conquered the self by means of the self. But when a man neglects the self, then, like an enemy at war, that very self will turn against him.

  7. A peaceful man who has mastered himself has a higher self that is deeply concentrated, whether in cold or in heat, whether in pleasure or in pain, whether in honor or in disgrace.

  A more literal and more accurate translation of stanza 5 might be something like:

  One should lift up oneself by means of the self. Do not degrade the self, for the self is oneself’s only friend, and at the same time the self is oneself’s only foe.

  In this version, the seven occurrences of Sanskrit ātman in the stanza are represented by seven occurrences of self in the translation. There are other ways to translate this passage, but I have not
found a way to make this passage genuinely “poetic” in English. Nor, as far as I know, has anyone else.*

  There is also in fact another good reason to avoid translating the Bhagavad Gītā into versified quatrains. Traditionally, the Sanskrit epic text, including the Bhagavad Gītā, is presented in Sanskrit print editions in couplets (or pairs: hemistichs) of sixteen syllables each, because the basic semantic and syntactic unit of Sanskrit epic meter is really the couplet, not the quatrain. Sanskrit is fond of long compound words, as is readily apparent in the Bhagavad Gītā, with the often very long names of its heroic figures. Frequently, a quarter-stanza of eight syllables will consist of only one or two significant words—for example, a name accompanied by an epithet that is no more than metrical filler, or perhaps a verb accompanied by a preposition or an adverb—and also various emphatic particles that serve as mere metrical fillers as well. A more accurate imitation of the metrical form of the Bhagavad Gītā would consist of couplets of sixteen syllables each. Such a couplet would be roughly comparable to two lines of Homeric dactylic hexameter.

  The standard Sanskrit śloka meter does not in general have the capacity to express complex abstract thought in a single stanza of thirty-two syllables. For this reason, the Bhagavad Gītā at times shifts to the longer triubh stanza of four lines of eleven syllables, thereby lengthening the individual stanza from thirty-two to forty-four syllables. We also frequently find single sentences extending across two or three or more stanzas, especially in those late chapters that present the popular Sākhya philosophy of the time as long lists of qualities or features or habits that are characteristic of various types of people.

  The Mahābhārata is a vast mosaic made up of huge and varied collections of smaller gems. While the size of these gems varies greatly, a significant number of them are roughly the same length as the Bhagavad Gītā, itself a “small classic” of seven hundred stanzas.* It is roughly as long as Greek and Roman dramas, and Sanskrit dramas as well. I would suggest that the Bhagavad Gītā, like other epic gems of similar size, was intended to be recited or sung in one session. It is likely that in such sessions the performer would recite or sing from memory, or perhaps from a text. The audience, consisting of devotees of Ka, might accompany the performer, reciting favorite stanzas from memory, or perhaps listening silently in rapt meditation. In such a context, the recitation would have been rapid and rhythmic, emotional and devotional, rather than slow and pensive. My translation is intended to capture such a performance context.

  On one matter I have decided to compromise literal accuracy for the sake of intelligibility, for a general non-Hindu audience. Throughout the Bhagavad Gītā, Arjuna and Ka address each other directly by name, but frequently their names are replaced by numerous popular epithets that are well known even today to an Indian audience. These epithets add nuance and texture to their personalities, and in the case of Ka also to the history of his assimilation to Viu (on which see “On the Organization of the Bhagavad Gītā”). However, to a contemporary American audience, and especially to an inexperienced student audience, these epithets create a good amount of confusion. I have preferred for the most part neither to retain them nor to translate them at all, but rather to substitute the names of Arjuna and Ka as appropriate. The alternatives would have required too much explanatory apparatus, which would have distracted readers from the more important issues that the Bhagavad Gītā presents to us for our edification.

  A Note on the Pronunciation of Sanskrit

  In this edition, the transliteration of Sanskrit words into a Romanized alphabet with a set of standard conventional diacritical marks has been adopted, after a good amount of consideration of the alternatives. It would have been easier to adopt a more casual and informal system of transliteration. The name Krishna is now nearly universally recognized in the English-speaking world as the name of that most prominent Hindu god and the central figure of the Bhagavad Gītā. But in this translation, his name is always represented in the form Ka, which is how it is represented in the correct transcription of Sanskrit. Likewise, Vishnu is here rendered more accurately as Viu. The reason for this technical accuracy is not mere pedantry. All informal transcriptions run the risk of obscuring important distinctions in Sanskrit pronunciation. The most important example of this, because the most common, is the failure to distinguish between short and long vowels. If we ignore the long vowels in important names like Mahābhārata and Rāmāyaa, we will inevitably mispronounce them. In both instances, the accent falls on the last long vowel: thus Mahābhārata and Rāmāyaa (where the syllable that receives the accent is underlined). We’ll get back to accent in a moment.

  The Sanskrit vowels should be pronounced as follows:

  a

  like u in cut or the definite article a; Sanskrit example mantra

  ā

  like a in father; Sanskrit example ātman or jñāna

  i

  like i in bit or sit; Sanskrit example Indra or nitya

  ī

  like ee in meet; Sanskrit example gītā or īśvara

  u

  like u in put; Sanskrit example buddhi or Buddha

  ū

  like oo in boot; Sanskrit example bhūta or kūastha

  like ri in risk or rig; Sanskrit example ta or i or Ka

  like ree in reel; Sanskrit example pitām (“of the fathers”)

  e

  like ay in say or pay; Sanskrit example tejas or namas te

  ai

  like ai in aisle; Sanskrit example daiva or maitra

  o

  like o in open; Sanskrit example loka or moha or OM

  au

  like ou in sound; Sanskrit example Draupadī or Paura

  The Sanskrit consonants are pronounced more or less as they are pronounced in English, except in the following instances:

  c

  like ch in church (never like c in cat); Sanskrit example cakra

  Sanskrit has a set of aspirated consonants: kh, gh, ch, jh, h, h, th, dh, ph, bh. These are pronounced like the corresponding nonaspirated consonants, k, g, c, j, , , t, d, p, b, with the addition of a heavy breathing. Thus, for example,

  ph

  like ph in shepherd (never as in photograph); Sanskrit example phala

  bh

  like bh in clubhouse; Sanskrit example bhakti or Bhārata

  th

  like th in boathouse (never as in think or bath); Sanskrit example atharva or ratha

  In the same way all of the other aspirated sounds of Sanskrit end with a breathy h.

  Sanskrit also has a set of retroflex consonants (, h, , h, , ) that are distinguished from the corresponding dental consonants (t, th, d, dh, n, s). The dentals are pronounced almost as in English, but with the tongue pressed against the teeth (as in Italian and other languages). The retroflex consonants are pronounced with the tongue curled back in the mouth. Some Sanskrit examples: Ka, prāa, triubh, anuubh. These retroflex sounds continue to be prominent in many Indian languages today, and can be easily recognized in contemporary Indian English as well.

  Sanskrit has a set of six nasals (, ñ, , n, m, ) and a set of three sibilants (ś, , s). Of these, the dental consonants n, m, and s are pronounced more or less as in English. The remaining nasals are pronounced:

  like ng in angle; Sanskrit example saga

  ñ

  like ny in canyon; Sanskrit yajña or jñāna (note that jña is frequently pronounced in India today with a hard g; thus yajña will often sound like yagya and jñāna like gyāna)

  this is the retroflex nasal, as in Ka, prāa, maala, etc.

  this is a sign of secondary nasalization before a consonant: thus śakara, sākhya, sasāra, sanyāsin

  Of the three sibilants, s is pronounced as in English, ś is pronounced like sh in shoe, and is the retroflex sibilant, pronounced with the tongue curled back, as in Ka.

  Finally, there is another secondary sign, , which is pronounced like h in English, but with a faint echo of the preceding vowel. Thus the famous mantra om nama
śivāya, where nama will be heard as namaha.

  Back to the matter of long and short vowels, and how to pronounce the often very long names that we encounter in the Bhagavad Gītā. The rules for word accentuation in Sanskrit are very complicated, especially when it comes to the older language of the Vedas. Early Vedic Sanskrit had a variable pitch accent like Greek, but by the epic and classical period this pitch accent was lost and replaced by a stress accent that resembles the stress accent of Latin. The accentuation of Vedic words was variable also insofar as accent was sometimes placed on different syllables, or was even absent, depending on inflectional ending, sentence context, and so forth. Fortunately, we can ignore this extraordinary complexity when it comes to the Bhagavad Gītā and classical Sanskrit in general.

  The basic rule for accent in classical and contemporary Sanskrit centers on the distinction between light and heavy syllables. Light syllables consist of a short vowel that is either followed by a single consonant or is itself isolated or final: thus the verb form bhavati has three light syllables and the accent is on the first. Heavy syllables may consist of a long vowel (ā, ī, ū) or a diphthong (e, o, ai, au) or a short vowel followed by more than one consonant.* Thus, in the names Arjuna and Ka, the syllables Arj and K are both heavy, even though their vowels, the first a in Arjuna and in Ka, are both short. Since the syllable is heavy in both cases, it takes the accent. In general, the accent of a Sanskrit word falls on the next-to-last syllable if it is heavy (e.g., ahakāra). If this syllable is light, then the accent falls on the third from the last syllable, again if it is heavy (e.g., Mahābhārata or Rāmāyaa). This is just as we find it in the pronunciation of Latin.