The Long Echo

In the year 1657, in Delhi, the eldest son of the emperor Shah Jahan finished a project that no one of his rank had attempted. Dara Shukoh, heir apparent to the Mughal throne, a Muslim prince of mystical temperament, had gathered Sanskrit scholars from Benares and set them to work translating the Upanishads into Persian. He believed the texts held something his own tradition pointed to and had not fully spoken, and he gave the collection a title that took the tradition’s own self-description and made it a banner. Sirr-i-Akbar. The Greatest Secret.

He did not have long to enjoy it. Within two years he was dead, defeated and executed in the war of succession by his younger brother Aurangzeb, his interest in the religion of the conquered counted against him. The prince was gone. The translation was not. It would travel further than he could have imagined, and the last chapter of this volume is the story of that travel, and of the older journey the texts made through the minds of the philosophers who could not agree on what they said.

The three sources

Long before the prince, the Upanishads had become the foundation of a philosophy, and the mechanism by which they did is worth naming precisely.

The tradition fixed three texts as the basis of Vedanta and called them the prasthanatrayi, the three sources, or the three points of departure. The first is the Upanishads themselves, taken as shruti, revelation, the foundational authority. The second is the Brahma Sutras, a work of terse, almost unreadable aphorisms attributed to the sage Badarayana, which sets out to systematize the scattered teachings of the Upanishads into a single ordered argument. The third is the Bhagavad Gita, the song embedded in the Mahabharata, which a later volume of this treatise will take up in its own right.

The rule that grew up around these three was simple and demanding. To found a school of Vedanta, to be taken seriously as a teacher of it, you wrote a commentary, a bhashya, on all three. The Upanishads supplied the raw vision. The Brahma Sutras supplied the skeleton of argument. The Gita supplied the bridge to practice. A complete philosophy had to answer to all of them. And three men did, and produced three philosophies that cannot all be true.

Three readings of three words

The test case is the sentence the second chapter of this volume turned on. Tat tvam asi. That thou art. Watch what three great readers do with three words.

Shankara comes first, in the eighth century, and his reading is Advaita, non-dualism, a-dvaita, not-two. He takes the sentence as an unqualified statement of identity. The self and brahman are one and the same reality, without remainder. The world of separate things, the very sense that you are a distinct person facing a distinct world, is maya, a misperception laid over the one reality by ignorance, real enough as experience and false as metaphysics. When the ignorance lifts, nothing is gained and nothing travels anywhere, because the self was brahman all along and only failed to know it. The salt was always dissolved.

Ramanuja comes some three centuries later, and he cannot accept this. His reading is Vishishtadvaita, qualified non-dualism. He agrees that there is one reality, brahman, but he refuses to dissolve the self and the world into an illusion. Selves and matter are real, he argues, and they are the body of brahman, as a body is to the soul that ensouls it, distinct from it and utterly inseparable from it and dependent on it. That thou art means, on his reading, that brahman is the inner self of the listener and of all things, not that the listener is simply identical with the absolute and his individuality a mistake. Love, devotion, the relation of a real worshipper to a real Lord, survive in Ramanuja in a way they cannot in Shankara, and that is much of why his school lived.

Madhva comes later still, in the thirteenth century, and he goes furthest in the other direction. His reading is Dvaita, dualism, and he holds the difference between the self and brahman to be real, permanent, and never overcome. The self is forever distinct from God, dependent on God, and the goal is right relation rather than dissolution into identity. So committed is he to this that he reads the famous phrase against its grain, parsing the Sanskrit so that it yields, by his analysis, something closer to thou art not that than to that thou art, the dependent self and the supreme reality held eternally apart.

Three men. The same verse. Non-dualism, qualified non-dualism, dualism. This treatise will not tell you which is right, because that is precisely the work the tradition assigns to a lifetime and not to a footnote. What the historian can say is what the disagreement reveals. The Upanishads are not a system. They are a spring, and the rivers that run from it run in different directions, and each of the three great philosophers found his whole vision genuinely present in the same words. A later volume of this work will give Vedanta the closer reading it deserves. Here it is enough to see that the contest was real and that the texts were large enough to sustain it.

The greatest secret travels west

Return now to the prince’s translation, lying in manuscript after his death, and follow it.

More than a century later, a French scholar named Abraham Hyacinthe Anquetil-Duperron, who had already spent years in India recovering the scriptures of the Parsis, obtained a copy of Dara Shukoh’s Persian Sirr-i-Akbar. He made his own translation of it, not into French but into Latin, the scholarly language of Europe, rendering the Persian word by word with a literalness that made the result almost as difficult as the original. He published it in two volumes in 1801 and 1802, under a Latinized form of the word itself. Oupnek’hat. It was the first time a substantial body of the Upanishads had appeared in any European language, and it had reached Europe by a strange relay: Sanskrit to Persian under a Mughal prince, Persian to Latin under a French Orientalist, two languages and two religions and a hundred and fifty years between the forest and the printing press.

The book found its most famous reader in Germany. Arthur Schopenhauer encountered the Oupnek’hat around 1814, and it marked him for the rest of his life. He kept it by him. He read in it, he said, the way one returns to a well. His praise of it, recorded across his works and gathered in the scholarship of Urs App, is among the most extravagant any philosopher has paid to a foreign scripture:

In the whole world there is no study, except that of the originals, so beneficial and so elevating as that of the Oupnek’hat. It has been the solace of my life, it will be the solace of my death.

He marked the distance himself. The qualifier in that famous sentence, except that of the originals, is Schopenhauer conceding that the Sanskrit he could not read stood higher than the Latin he loved. There is something to weigh carefully here, and it should be said plainly rather than smoothed into a triumph. What Schopenhauer read was a Latin translation of a Persian translation of the Sanskrit, at two removes, filtered through a Mughal prince’s frame and a French scholar’s literalism. He did not read the Upanishads as a Sanskritist reads them. And yet the thing carried. Even doubly translated, even bent through two foreign tongues, the central vision came through clearly enough to reorganize the thought of one of the nineteenth century’s most serious minds. The texts that a civilization had whispered from teacher to student, under a rule of secrecy, in a forest, had crossed the world and kept their charge.

What we do not know, and what the tradition holds

One question remains, the one the texts themselves will not answer. When were they made, and by whom?

The honest answer is the one the fourth chapter already gave in outline, and it is worth ending on because it is the proper posture of this whole work. We do not know who wrote the Upanishads. They are anonymous. The names inside them, Yajnavalkya, Uddalaka, Gargi, are characters in the teaching, not signatures on the text. By their language scholars can order them, and the oldest, the Brihadaranyaka and the Chandogya, are placed by Patrick Olivelle and Michael Witzel before the Buddha, somewhere in the centuries around the seventh and sixth before the common era, with the verse texts later and the short prose texts later still, some of them after the Buddha. But the dates are estimates with centuries of give in them, and Olivelle’s own warning, that any precision closer than a few centuries is as stable as a house of cards, is the truest sentence anyone has written about the question.

That is the documentary record, and it runs out where it runs out.

The tradition holds something else, and it is owed its own clear statement, kept separate from the record rather than mixed into it. In the tradition’s account the Upanishads are not dated because they are not composed. They are shruti, that which was heard, and apaurusheya, not of human authorship, the eternal knowledge perceived by the seers and not invented by them, the final and highest portion of a Veda that was never made and has no beginning. On this view to ask who wrote the Upanishads is to misunderstand what they are, as one might ask who composed the laws of number. The first volume of this treatise set out the same claim for the Vedas as a whole. It returns here, at the end, in its sharpest form, because the Upanishads are where the claim is felt most acutely: that what is being transmitted is a seeing rather than a text, and that the seeing was always there to be seen.

The historian and the tradition do not have to be reconciled. They have to be told apart, and both told truly. What can be documented has been documented here, dated and named and flagged where it is thin. What the tradition holds has been set beside it, marked as what the tradition holds. Between the two, the reader is left where the Upanishads always left the one who came to sit: with the question still open, and the responsibility for it returned to the one who asked.

The thread of this volume ends here.

The sitting-near does not.