The Upanishads

← Encyclopedia

A father tells his son to dissolve a lump of salt in water and bring it back the next morning. The boy cannot find it; the salt is gone. Taste the water at this end, the father says — salty. The middle — salty. The far end — salty. The salt was never lost; it had become the whole of the water, present everywhere and visible nowhere. That, the father concludes, is how the finest essence pervades all things, and that essence is the truth, and the self, and “thou, O Svetaketu, art it.” The lesson is recorded in the sixth chapter of the Chandogya, one of the oldest of a body of texts that the orthodox philosophy of India treats as its summit — the place where the Veda stops describing the cosmos and turns to ask what it is that knows the cosmos at all.

The name is a posture before it is a doctrine. Upanishad is built from the verb sad, “to sit,” with the prefixes upa, “near,” and ni, “down” — the session of a pupil seated below a teacher to receive instruction not meant for the crowd. Müller’s introduction parses it exactly this way, glossing the Sanskrit lexicographers who explain upanishad by samîpasadana, “sitting down near a person.” Later tradition heard a second meaning in the word — the secret teaching that loosens, that destroys ignorance — but the literal sense is the schoolroom: knowledge transmitted mouth to ear, between two people, in the open air.

A folio of the Chandogya Upanishad, Sama Veda, in an 1865 Devanagari manuscript. — Ms Sarah Welch, via Wikimedia Commons (CC BY-SA 4.0)

Genealogy: the end of the Veda

The Upanishads are not a freestanding scripture; they are a stratum. The Veda accreted in four layers, each answering a different need of an evolving priestly culture. First came the Samhitas, the collections of hymns — the Rig, Sama, Yajur, and Atharva — addressed to the gods of fire and storm and dawn. Around these grew the Brahmanas, dense prose manuals of sacrificial procedure, explaining which formula accompanies which oblation and why. Then the Aranyakas, the “forest books,” for those who had withdrawn from the village to interpret the rite inwardly, as meditation rather than performance. And at the far edge, the Upanishads, where the sacrifice has been almost wholly internalized and the questions have turned metaphysical: what underlies the fire and the worshipper alike. Because they close the corpus, the tradition calls the whole genre Vedanta — veda-anta, the end of the Veda, in the double sense of its final portion and its highest purpose. Müller’s note on the Chandogya states it plainly: together with the Brihadaranyaka, it “contributed the most important materials to what may be called the orthodox philosophy of India, the Vedânta, i.e. the end, the purpose, the highest object of the Veda.”

How many Upanishads there are depends on who counts. The late Muktika Upanishad enumerates a canon of one hundred and eight, and Müller records that figure while noting Burnell’s suspicion that it is artificial, 108 being a sacred number in the south. The Mughal prince who first carried the texts toward Europe translated fifty. Western scholars in the nineteenth century counted by title and reached far higher tallies — Weber proposed some 235, Müller’s own 1855 list reached 149 — because the same text travels under several names. But beneath the swarm of minor and sectarian Upanishads stands a stable inner canon: the mukhya or principal Upanishads, the ones the great philosophical schools all comment on. Robert Hume fixed the working number at thirteen in the title of his 1921 translation; ten are universally accepted, with the Shvetashvatara, Kaushitaki, and Maitri added to round out the set. These are the load-bearing texts, and two of them carry most of the weight.

The Brihadaranyaka and the Chandogya are the oldest and the longest. Patrick Olivelle, whose textual criticism is the current scholarly standard, places both before the rise of Buddhism — somewhere in the seventh to sixth centuries BCE, with the caution, which he repeats almost as a refrain, that any dating of these documents more precise than “a few centuries” is “as stable as a house of cards.” They are prose, sprawling, composed across generations during the urbanization of the Gangetic plain, and they preserve the texture of real instruction: kings who out-argue priests, women who corner sages with unanswerable questions, debates staged for prizes of a thousand cattle. The Brihadaranyaka belongs to the Yajur Veda, the Chandogya to the Sama Veda; later principal Upanishads attach to the Atharva Veda and to particular schools. What the academy treats as a layered, multi-authored anthology assembled over centuries, the tradition receives differently: as shruti, “that which was heard” — uncreated, authorless, eternal sound apprehended by the ancient seers rather than composed by them. On this internal view the Upanishads are not dated at all. They are heard.

The teaching: that thou art

Strip away the ritual scaffolding and the riddling dialogues, and the principal Upanishads pursue a single identity. There is an ultimate ground of all that is — Brahman, the absolute, infinite and unconditioned, from which the worlds proceed and into which they return. And there is the innermost self of a person — Atman, not the body, not the breath, not even the mind, but the bare witnessing awareness that none of those can be. The discovery on which everything turns is that these two are one. The ground of the universe and the ground of the self are not similar, not connected, but the same.

This is compressed in a handful of utterances the later schools singled out as mahavakyas, “great sayings,” tradition pairing one with each of the four Vedas. From the Chandogya, in that dialogue between Uddalaka Aruni and his son Shvetaketu, comes tat tvam asi — “that thou art.” Müller’s text records the Sanskrit at Chandogya 6: Etadâtmyam idam sarvam, tat satyam, sa âtmâ, tat tvam asi Svetaketu — all this has the Self for its essence; it is the True; it is the Self; and thou art it. The father repeats the formula nine times, applying it after each illustration, the salt among them, until the son sees that the subtle essence underlying the world is the very thing he himself is. From the Brihadaranyaka, at 1.4.10, comes aham brahmasmi — “I am Brahman” — spoken of the primordial reality that, knowing itself, became the whole. The Aitareya gives prajnanam brahma, “consciousness is Brahman”; the short Mandukya gives ayam atma brahma, “this self is Brahman.” Four sayings, one claim, turned over from four directions.

The claim is approached as much by subtraction as by assertion. Because Brahman is the subject of all knowing, it can never be an object of knowledge; any predicate one fixes to it is already too small. The Brihadaranyaka answers this with the formula neti neti — “not this, not this.” Whatever can be pointed at, named, or conceived is set aside, and the negation moves inward through every layer of the apparent self until nothing remains to be discarded except the awareness doing the discarding, which cannot be negated because it is what every negation presupposes. This is the apophatic core of the Upanishads, a theology by refusal: the absolute is reached not by adding attributes but by exhausting them.

Adi Shankara with disciples, by Raja Ravi Varma (1904). — via Wikimedia Commons (public domain)

The Mandukya, the shortest of the principal Upanishads at twelve terse mantras, maps the same absolute onto the sounded syllable Om and onto consciousness itself. The syllable is heard in three components — A, U, M — and these are correlated with three states every person passes through. A is the waking state, turned outward to gross objects. U is dreaming, turned inward to subtle impressions. M is dreamless deep sleep, where the mind rests undivided and at peace. But the Upanishad insists on a fourth, turiya, “the fourth” — not a state alongside the others but the silent awareness that underlies and outlasts them all, corresponding to the silence after the syllable is sounded. That fourth is the Self; that Self is Brahman. Gaudapada, in the verse commentary known as the Karika composed around the seventh century, built this brief text into a full metaphysics of non-origination, and the tradition came to regard the Mandukya, properly understood, as sufficient by itself for liberation.

Woven through the metaphysics is a doctrine that would organize Indian religious life for millennia: the first systematic statement of karma and rebirth. The earlier Veda had imagined the dead passing to a world of the fathers; the Upanishads recast death as transition. In the Brihadaranyaka, a sage is asked privately what becomes of a person after death, and answers — out of earshot of the assembly — that one becomes good by good action and bad by bad. The Chandogya elaborates the two paths the dead may travel, the way of the gods that does not return and the way of the fathers that leads back to rebirth. Action binds; its fruit ripens across lives; the self transmigrates, fitted to a new condition by what it has done, until knowledge of its identity with Brahman cuts the chain. This is the seedbed of the later doctrines of karma and rebirth. Liberation, moksha, is not the soul’s journey to a better realm but its waking to what it always was. The bondage was only ignorance — the failure to taste the salt already dissolved through the whole.

Reception: schools, a prince, and a philosopher’s desk

The Upanishads do not present a system; they present a treasury of dialogues, parables, and lightning formulas, often pulling in different directions on the same page. Building a coherent philosophy out of them was the work of the Vedanta schools, and they did it through a shared canon. Each major Vedantic master commented on the same three foundations — the prasthanatrayi, the “three points of departure”: the principal Upanishads, the Brahma Sutras of Badarayana (which already aimed to systematize the Upanishads in terse aphorisms), and the Bhagavad Gita. To found a school of Vedanta was, in practice, to write three commentaries and let them disagree with everyone else’s.

The first and most influential reading was Shankara’s. Critical scholarship now places him around 700 CE — the older textbook dates of 788–820 have lost ground, and the tradition’s own ascription of him to a far earlier age is a separate, internal claim — and his commentaries on the prasthanatrayi are the cornerstone of Advaita, “non-dual” Vedanta. For Shankara the great sayings mean exactly what they say: reality belongs to Brahman alone, the self is identically that Brahman, and the plural world of objects and individuals is a superimposition (adhyasa) on the one reality, a misperception comparable to mistaking a rope for a snake. Liberation is therefore not an attainment but a correction of error, available the instant the identity is truly known. Thibaut’s translation of Shankara’s commentary on the Brahma Sutras — The Vedânta-Sûtras — remains the standard English window onto this architecture.

Shankara’s non-dualism did not go unanswered. Ramanuja, in the eleventh and twelfth centuries, read the same texts as teaching Vishishtadvaita, “qualified non-dualism”: the self and the world are real and form, as it were, the body of a personal God, distinct from Brahman yet utterly dependent on it, so that devotion rather than mere knowledge becomes the path. Madhva, in the thirteenth century, pressed further to Dvaita, outright dualism: the difference between the soul and God is permanent and irreducible, and the great sayings must be construed to preserve it. Three philosophers, one set of source-texts, three incompatible worlds — and the disagreement is itself a measure of how much the Upanishads will bear. Their oracular brevity is exactly what made them generative.

Arthur Schopenhauer in 1815, by Ludwig Sigismund Ruhl. — via Wikimedia Commons (public domain)

The Upanishads reached Europe by a route so circuitous it nearly defies belief. In 1657 the Mughal prince Dara Shukoh, Shah Jahan’s eldest son and a Sufi-minded seeker who sensed kinship between Vedanta and the mystical Islam he loved, had some fifty Upanishads rendered from Sanskrit into Persian under the title Sirr-i-Akbar, “the Great Secret.” A century and a half later the French Orientalist Abraham Hyacinthe Anquetil-Duperron — who had failed to master Sanskrit but knew Persian — translated that Persian version into Latin and published it in two volumes in 1801–1802 as the Oupnek’hat. Müller, who could read all the languages involved, gives a specimen of Anquetil’s word-for-word Latin and concedes it took a real philosopher to find a thread through “such a labyrinth.” The text Europe first read was thus triply mediated: Sanskrit to Persian to Latin, the Persian syntax retained against Latin grammar, the Sanskrit terms left untranslated and glossed.

That philosopher was Arthur Schopenhauer, who kept the Oupnek’hat on his desk for decades and folded its non-dualism into his own metaphysics of will and representation. His praise, which Müller quotes at length, is famous for its intensity: every line of the book displays “firm, definite, and throughout harmonious meaning”; the reader’s mind is “washed clean of all early engrafted Jewish superstitions”; and the verdict — “It has been the solace of my life, it will be the solace of my death.” Halbfass reads this rapture coolly, observing that the first major European absorption of Vedanta was an absorption of a translation of a translation, valorized as much for the uses Schopenhauer’s anti-Christian polemic could make of it as for any fidelity to the Sanskrit. The pattern he identifies would recur: Indian thought entering European discourse as a screen onto which Western critiques of the West could be cast.

Sober philology arrived with the next generation. Friedrich Max Müller, the German-born Oxford Sanskritist, translated the principal Upanishads directly from the Sanskrit for his great series the Sacred Books of the East, in volumes I (1879) and XV (1884) — the translation and introduction that became the standard English source for the texts and their European history. Müller’s edition placed the texts before the English-reading public in a form that was scholarly rather than mystical, complete with the etymology of the word, the manuscript history, and the chain from Dara Shukoh through Anquetil to Schopenhauer set down as record. From there the Upanishads passed into the Bengal Renaissance and the reform movements that followed Rammohan Roy, and then — through Vivekananda and the global circulation of Vedanta — into the vocabulary of a worldwide spiritual modernity, sometimes faithfully, often modernized at the point of departure. Twentieth-century criticism, above all Olivelle’s, turned back to the texts themselves, resisting the Advaita gloss that centuries of commentary had laid over words like maya and brahman and reading the early Upanishads as documents of a specific, datable world rather than a timeless philosophy. The salt is still dissolved through all of it. Whether one can taste it, the texts say, depends on sitting down near someone who already has.

→ In the library: The Upanishads (Müller, Sacred Books of the East I & XV) · Śankara — The Vedânta-Sûtras (Thibaut, SBE 34 & 38) · The Bhagavadgītā (Telang, SBE 8)

→ Related: Brahman · Vedanta Advaita Visistadvaita Dvaita · Hindu Vedanta Tantra · Hinduism · Brahmanism · Historical Vedic Religion · Neo Vedanta · Neo Advaita · Sankara · Gaudapada · Ramanuja · Madhva · Bhagavad Gita · Karma · Reincarnation · Yoga · Yoga Sutras Of Patanjali · Swami Vivekananda · Rigveda · Samaveda · Yajurveda · Atharvaveda