I Take Counsel With My Own Body: The Rigveda's Word for Self

A Man Sits Up at Night and Interrogates His Own Body

Something has gone wrong between a poet of the Vasiṣṭha family and the god Varuṇa, and the poet cannot work out what. The hymn he leaves behind, RV 7.86, is one of the few places in early Indo-European poetry where a person is shown thinking a problem over rather than announcing a conclusion. In the second verse he describes his method:

uta svayā tanvā saṃ vade tat kadā nv antar varuṇe bhuvāni

kiṃ me havyam ahṛṇāno juṣeta kadā mṛḷīkaṃ sumanā abhi khyam

“And I take counsel with my own body about this: when will I be inside Varuṇa? What offering of mine would he accept without anger? When will I see, well disposed, his mercy?”

Rigveda 7.86, verse 2. My rendering from the Saṃhitā text, checked against Griffith (1896). [1] [2]

Svayā tanvā saṃ vade: literally, I converse with my own tanū́. And tanū́ is the ordinary Vedic word for a body, the thing with hands and feet that can be burned, wounded, decorated or lost in battle.

Griffith could not bring himself to print it. His 1896 version reads “With mine own heart I commune,” which is graceful and is not what the Sanskrit says. [2] The evasion is understandable. English wants a soul to have the conversation with, and English has been told for two hundred years that Sanskrit supplies one, ātman, the Self of the Upaniṣads, the term that ends up on one side of the most famous equation in Indian philosophy. So a translator meets a Rigvedic poet talking things over with himself and reaches for heart, or soul, or self, because “I discussed it with my body” sounds like a joke.

It is not a joke. It is the normal Vedic way of saying it, and the next verse proves the poet meant it as a real procedure rather than a flourish: having asked his own tanū́ and got nowhere, he goes and asks other people. “I ask about this offence, wanting to see; I approach those who know, to ask around. The poets all told me the same thing: this Varuṇa is angry with you.” RV 7.86.3. Two consultations, in order. Himself first, then the experts. The self he consulted was his body.

Run the vocabulary of personhood across the whole Rigveda and this stops looking like one poet’s oddity and starts looking like the structure of the language.

Methods note. These counts come from my own scan of the romanised Saṃhitā text of all ten maṇḍalas, matched by stem and then sorted by hand. The tanū figure is the roughest of them, because Vedic sandhi fuses words together and because forms of the noun tanū́ sit next to forms of the verb tan, “to stretch”, which look identical for the first four letters: tanvate and tanvaḥ are different words. I have removed the verbs and the compound name Tanūnapāt, which leaves about 140 verses. Grassmann’s 1873 dictionary of the Rigveda remains the authority for an exact figure and should be preferred over mine. [3] Read every number here as “roughly this many verses”.

The Word Means Something Stretched

Tanū́ is a feminine noun in -ū. The standard etymology attaches it to the root tan, “to stretch, extend, span”, the same root that gives tántu, a thread on a loom, and tantra, the warp, and eventually a whole genre of texts named after the idea of a continuous thread. Its Indo-Iranian siblings are Avestan tanū- and Old Persian tanūš, both meaning body and both, like the Vedic word, doing reflexive duty. Further out the family includes Latin tenuis, thin, and English thin itself. [4]

So the Vedic word for a person is, on the usual account, the extended thing: what is spread out, drawn like a thread, occupying space.

Aside. The etymology is standard but not airtight. Mayrhofer reconstructs a Proto-Indo-Iranian tanúHs and traces it to a Proto-Indo-European form that may be a primary noun for “body” rather than a derivative of the verb “to stretch”. [4] If so, the resemblance between tanū́ “body” and tanoti* “he stretches” is inherited rather than transparent, and the poets who heard “the stretched one” in it were doing folk etymology. Vedic poets did that constantly and with enthusiasm, so the folk etymology is still evidence about how the word felt to them, even if the historical linguistics does not license it.

Whichever way that goes, the semantic starting point is spatial and external. The Rigvedic self is not a hidden kernel. It is an extent.

One Word, Six Jobs

Sort the roughly 140 verses by what the word is actually doing and six distinct jobs come out. They are not metaphorical extensions of a core meaning so much as six ordinary uses that a Vedic speaker moved between without noticing.

The fourth row is the one that repays staring at. Here is the verse in full, from a short hymn to the improbable pairing of Soma and Rudra:

somārudrā yuvam etāny asme viśvā tanūṣu bheṣajāni dhattam

ava syataṃ muñcataṃ yan no asti tanūṣu baddhaṃ kṛtam eno asmat

“Soma and Rudra, place all these remedies in our bodies. Untie, release from us the committed offence that is bound in our bodies.”

Rigveda 6.74, verse 3. My rendering, checked against Griffith (1896). [2]

One verse, two requests, one location. Medicine goes into the tanū́ and guilt comes out of it. The offence is not a stain on a soul that the body happens to carry around; the Sanskrit says it is tied up inside the tanū́ the way an illness is, and the cure is the same kind of operation, an untying. The Rigvedic vocabulary for moral trouble runs on knots and narrow places, which is why the request to Varuṇa is so often to loosen something. Readers of the pieces on Varuṇa and cosmic guilt and on áṃhas, the narrow place will recognise the machinery. What is new here is where the knot is tied.

Whose Fault Is It: A Vedic Distinction

Return to the Vasiṣṭha hymn, three verses on from the insomniac’s self-interrogation. The poet has established that Varuṇa is angry. Now he asks for release, and he sorts the possible sources of the damage:

ava drugdhāni pitryā sṛjā no ava yā vayaṃ cakṛmā tanūbhiḥ

“Release us from the deceptions of our fathers, and from those we ourselves have committed with our own bodies.”

Rigveda 7.86, verse 5. My rendering, checked against Griffith (1896). [2]

Inherited fault on one side, fault committed tanūbhiḥ, with our own selves, on the other. The distinction is drawn cleanly and without fuss, three thousand years before anyone needed a word for moral luck.

The same cut appears in a Soma hymn in the eighth maṇḍala, where the god is asked for protection against hostilities that are tanūkṛ́t and hostilities that are anyakṛ́t, self-made and other-made. RV 8.79.3. Griffith took tanūkṛ́t to mean “wasting the body” and produced “hatreds that waste and weaken us”, which is a defensible reading of a compound that can be parsed more than one way. [2] But the pairing with anyakṛ́t, “made by another”, pulls hard the other way, and if the natural reading is right then a Rigvedic poet had a technical term for the harm you do to yourself.

Aside. Compounds in -kṛt are where translations diverge most, because Vedic composition rarely marks which member is agent and which is patient. Tanūkṛ́t turns up again at RV 1.31.9 as an epithet of Agni, where it clearly means “body-maker”, the god who forms bodies. The same compound therefore means “maker of bodies” in one verse and, probably, “self-made” in another. This is not sloppiness in the text. It is a reminder that Vedic compounds are compressed to the point of ambiguity and that a confident translation is often a decision in disguise.

Grammar Catches Up With Poetry

None of this is a poetic conceit. It is the grammatical state of the language, and it has been studied precisely.

Vedic had no dedicated reflexive pronoun of the sort English has in “himself”. What it had was the middle voice, which marks an action as coming back on the subject, plus a noun pressed into service when the middle voice was not enough. In the Rigveda that noun is tanū́. By middle Vedic it has been ousted by ātmán, and in the Atharvaveda you can find both words in the same clause, doubled up into what linguists call a heavy reflexive, the way a language behaves when one word is taking over a job from another and speakers are briefly using both. Leonid Kulikov’s study of the Vedic reflexives traces the handover in detail and notes that the tenth maṇḍala of the Rigveda already contains cases of tanū́ functioning as a clear reflexive with an active verb, which is an innovation and the first attestation of the pattern that classical Sanskrit would settle into with ātmán. [5]

Set the two words side by side in the Rigveda itself and the asymmetry is stark.

The word that means Self in every later century of Indian thought is, in the Rigveda, a rare word that mostly means breath.

What Ātman Was Doing Instead

Twenty-nine verses is not many, and the company ātmán keeps in them is physical. It is the wind: the Aśvins travel “as the ātman, the wind, goes to its pastures” (RV 1.34.7). It is the sun, called the ātman of everything that moves and everything that stands still (RV 1.115.1), which is a statement about the world being animated from a single source of motion, not about an inner witness. It is Soma, “the ātman of the sacrifice” (RV 9.6.8), the liquid that makes the rite go. It is what a boat can have: the Aśvins rescue the shipwrecked Bhujyu in vessels that are ātmanvatī, possessed of breath, self-moving (RV 1.116.3). Ships with breath in them. A late healing charm tears the wasting disease out of “your whole ātman”, meaning the entire physical frame from hair to nails (RV 10.163.5). Surendranath Dasgupta put the position plainly a century ago: the Rigvedic words in this area are manas, ātman and asu, and ātman, the one that became famous, generally means vital breath. [6]

The cleanest evidence is a verse that uses both words a few syllables apart and keeps them apart in meaning. It comes from the horse sacrifice hymns, addressed to the animal being killed:

mā tvā tapat priya ātmāpiyantaṃ mā svadhitis tanva ā tiṣṭhipat te

“Let your dear breath not scorch you as you go. Let the axe not lodge in your body.”

Rigveda 1.162, verse 20. My rendering, checked against Griffith (1896). [2]

The ātman is what leaves, painfully, at the end. The tanū́ is what the blade goes into. They are not synonyms, they are not a pair of soul and shell, and the verse does not say the first survives the second. It says the horse should be spared distress on both counts.

Ernst Arbman’s long study of Vedic soul-concepts, published across two volumes of Le Monde Oriental in 1926 and 1927 and still the reference point for the argument, sorted the Vedic evidence into what he read as two originally distinct lines: a breath-soul and a free-soul, the one bound to the living body and the other able to leave it. [7] Later scholarship has pushed back on the neatness of that scheme, and Bodewitz in particular has argued that the Vedic material resists being sorted into the tidy categories comparative religion brought to it. [8] [9] What nobody disputes is the raw lexical fact: early Vedic distributes the work of “person” over several words, and none of them is a soul in the sense that the word carries in English.

graph TD
    P["A Rigvedic person"] --> T["tanū: the extent, the self, what acts"]
    P --> M["manas: mind, intent, attention"]
    P --> H["hṛd: heart, seat of insight"]
    P --> A["atman: breath, what animates"]
    P --> S["asu: life force, what departs"]
    T --> V["can be harmed, healed, staked, adorned"]
    M --> W["can travel, wander, be yoked"]
    A --> X["shared with wind, sun, soma, ships"]

Gods Have Several

If the tanū́ were a soul, a god could have only one. Gods in the Rigveda have as many as the poets need.

agne trī te vājinā trī ṣadhasthā tisras te jihvā ṛtajāta pūrvīḥ

tisra u te tanvo devavātās tābhir naḥ pāhi giro aprayuchan

“Agni, three are your powers, three your stations, three your tongues, O born of Order, many of them. And three bodies of yours, savoured by the gods: with those protect our songs, unfailingly.”

Rigveda 3.20, verse 2. My rendering, checked against Griffith (1896). [2]

Three bodies, and the poet asks the god to use them all at once as instruments of protection. The Atris, in the fifth maṇḍala, ask “which of your tanū́s, Mitra or Varuṇa, is unpraised?” (RV 5.67.5), where Griffith sensibly gives “persons”. The Maruts deck their tanū́s with gold before a raid (RV 5.60.4), which is what a war band does to itself, not to its soul.

The most interesting case is a name. The Āprī litanies, the fixed sequence of invocations that opens an animal sacrifice, address in second position a figure called Tanūnápāt, “descendant of the body”, a title of Agni that makes him the offspring of himself. [10] One verse actually explains the naming:

tanūnapād ucyate garbha āsuro narāśaṃso bhavati yad vijāyate

“He is called Tanūnapāt while he is the embryo within; he becomes Narāśaṃsa when he is born apart.”

Rigveda 3.29, verse 11. My rendering, checked against Griffith (1896). [2]

A being is named one thing while unborn and another once it has come out. The name of the first stage is built on tanū́. Whatever else the word carried, it was the right word for the stuff a person is made from at the point where there is not yet a separate person. Readers who have followed the Āprī litanies through their fixed sequence will have met Tanūnapāt without being told that the name contains the Rigveda’s word for self.

The Other Body: What Śárīra Was For

Seven verses use śárīra, the word that classical Sanskrit will make the default term for a body. All seven are worth listing, because the distribution is not random.

Corpse, carcass, the thing an axe meets, the thing fire consumes, the thing spectators can see. Śárīra is the body from outside: the object, the remains, the visible surface. Tanū́ is the body from inside: the person, the agent, the thing you argue with at two in the morning.

The last entry makes the contrast explicit. The kéśin hymn, RV 10.136, describes wandering ecstatics who claim to have mounted the winds, and the boast they make to everyone else is precisely that the visible body is not the interesting part:

unmaditā mauneyena vātāṃ ā tasthimā vayam

śarīred asmākaṃ yūyaṃ martāso abhi paśyatha

“Maddened by our silent state, we have mounted upon the winds. You mortals see nothing of us but our bodies.”

Rigveda 10.136, verse 3. My rendering, checked against Griffith (1896). [2]

Set that next to the Vasiṣṭha hymn and you have the two poles of Vedic personhood inside one collection. The householder poet consults his tanū́ as the seat of his own judgment. The ecstatic dismisses his śárīra as the part the neighbours can see. The word chosen is doing the work in both cases, and the long-haired ascetic of RV 10.136 is exceptional in the Rigveda partly because he is one of the very few speakers who treats the visible body as a disguise.

The Handover

What happens next is a lexical relay, and it runs in three directions at once over roughly five centuries.

graph LR
    subgraph RV["Rigveda"]
        A1["tanū: self, reflexive"]
        B1["atman: breath"]
        C1["sarira: corpse"]
    end
    subgraph MV["Middle Vedic and AV"]
        A2["tanū fading as reflexive"]
        B2["atman takes the reflexive"]
    end
    subgraph UP["Upanisads and after"]
        B3["atman: the Self"]
        C3["sarira: the body"]
        A3["tanu: slender, form"]
    end
    A1 --> A2 --> A3
    B1 --> B2 --> B3
    C1 --> C3

Ātmán rises from breath to reflexive pronoun to the innermost Self, and the Upaniṣads eventually set it equal to brahman, an equation whose Rigvedic prehistory is a separate and much-argued question taken up in the piece on the non-dual seed. Śárīra comes up from the corpse to become the ordinary living body, the śarīra of classical medicine and philosophy. And tanū́ does not die, but it is demoted: it keeps “body” and “form” as meanings, picks up “slender” as its most productive sense, and loses the grammatical job it held in the oldest text. By the classical period tanvī is a poet’s word for a slim woman. The word that once meant “myself” ends up meaning “willowy”.

Aside. It is tempting to read the promotion of ātmán as thought becoming more inward, and there is something to that. But a good deal of it is plain grammar. A language that lacks a dedicated reflexive pronoun tends to recruit a noun for the role, and the recruited noun is typically “body”, “head”, “bone”, “soul”, or “self”, a pattern well attested across unrelated language families. Vedic recruited “body”, then swapped it for “breath”. The philosophy came afterwards and had to work with whatever the grammar had already promoted. [5]

What It Was Like to Be a Person

Put the evidence together and a picture of personhood emerges that is not the one most readers bring to the Veda.

Three observations seem to me to follow.

First, the Rigvedic self is outward-facing. It is the extent of you, the part that occupies space, meets weapons, wears gold, gets sick and is visible to gods. When the poets want protection they do not ask for their souls to be saved; they ask for guards on the perimeter. One Aśvin verse lists the posts to be filled: chardiṣpā́, guardian of the shelter, paraspā́, guardian from beyond, jagatpā́, guardian of the livestock, and tanūpā́, guardian of the selves (RV 8.9.11). Shelter, perimeter, herd, person, in that order. The self appears in a list of assets that need watching, which is exactly where a herding people under raid would put it.

Second, the self is the moral unit, and it is bodily. Offences committed “with our own tanū́s” are distinguished from those inherited from fathers. Guilt is tied into the tanū́ and has to be untied out of it. Nothing in the Rigveda suggests that wrongdoing happens in a compartment separate from the arms and legs that carried it out. This is one reason the Rigvedic treatment of fault feels so physical: the poet of RV 7.86 blames liquor, anger, dice and inattention for his lapse, four causes that operate on a body, and then asks to be untied like a calf from a rope.

Third, the self is plural and gradable. Gods have three of them. A being has one name for the tanū́ in the womb and another after birth. Two armies can be described as contending “with their bodies bright”, a compound that treats the shining person as the unit of battle. Nothing here is committed to the idea that a person is one indivisible thing, and the later Indian appetite for analysing the person into sheaths, breaths and constituents has a natural starting point in a vocabulary that never insisted on unity in the first place.

None of this makes the Rigvedic poets proto-anything. They were not anticipating embodied cognition and they were not refuting Descartes in advance. They were speaking a language in which the obvious word for yourself was the word for your body, and they thought about themselves in the terms it gave them. That is what makes the vocabulary worth the attention: not that it agrees with us, but that it shows a coherent, fully worked out way of being a person that got built on different foundations and then, over a few centuries, got quietly rebuilt on ours.

Compounds Are Where a Word Shows Its Hand

A last piece of evidence, which is really a test. If tanū́ meant only “flesh”, its compounds would be about flesh. They are not.

Tanūtyáj is the one to keep. It appears twice, at RV 10.154.3 of warriors who fall in the front line and at RV 10.4.6 in a simile comparing something to “thieves who risk their bodies in the forest”. The same compound covers the heroic and the criminal, because the thing being described is neither heroism nor crime but a stake. To be tanūtyáj is to have put yourself up as the wager. That is a concept of self with skin in it, in the most literal available sense, and it is also, incidentally, the closest the Rigveda comes to saying that a person is what a person is willing to lose.

Where to Start Reading

Open RV 7.86 and read all eight verses in one sitting, late, when the hymn’s situation is closest to hand. It is short. A man cannot sleep, does not know what he has done, questions himself, questions the experts, gets the same answer from all of them, blames drink and dice and bad company, and asks to be untied. Nothing in it needs a footnote except the second verse, and now that one does not either.

Then read RV 10.136 immediately afterwards and watch a completely different kind of person dismiss his own visible frame as the least interesting thing about him. The two hymns sit in the same collection, use two different words for body, and mean opposite things by them. Between those two poles the Rigveda had already worked out most of what a culture needs in order to start arguing about the self, several centuries before it had the vocabulary to state the argument. What it did not yet have was a word for the part of you that is not your body. That word was coming. It just had not stopped meaning “breath” yet.