Hard to Catch as the Wind: The Rigveda's Oldest Love Story

“I am hard to catch as the wind”

A king has cornered his wife on a riverbank, and she is already half gone. He pleads. He reasons. He reminds her of the nights they spent together, of the child she carries. He threatens, in the end, to throw himself off a cliff and let the wolves have him. She listens to all of it, and then, with the flat courtesy of someone who has already made up her mind, she tells him to go home:

What can we accomplish through such discourse? I have passed away like the first of the dawns. Return, Purūravā, to your dwelling: I am as hard to catch as the wind.

(RV 10.95.2, after Wilson)

That is the crux of RV 10.95, eighteen verses of pure dialogue with almost no stage directions, and it is one of the strangest and most affecting texts in the Rigveda. There is no narrator to tell you what is happening. There are two voices, a man’s and a woman’s, and the woman is a celestial nymph named Urvaśī (अप्सरा, apsarā, a spirit “moving in the waters”), and the man is a mortal king named Purūravas, and their marriage is over. What survives is the argument at the end of it.

Later India could not leave this alone. The Śatapatha Brāhmaṇa built a prose legend around the hymn; the Purāṇas made Purūravas the founder of a royal dynasty; and Kālidāsa, a thousand years and more downstream, turned it into one of the great Sanskrit plays and gave it the happy ending the Rigveda pointedly withholds. Comparative folklorists went the other way, outward, and found the same story of a fairy wife who must not be seen or named, and who vanishes when the taboo breaks, in Sweden, Siberia, and Japan. This piece is about the eighteen verses at the source: what they say, what they refuse to say, and why a lovers’ quarrel from the second millennium BCE has never stopped being retold.

A hymn shaped like a play

Most of the Rigveda is addressed upward. A poet praises Agni or Indra, asks for cattle and sons and safe passage, and the god does not answer inside the poem. A small handful of hymns break that mould and stage a conversation between two or more speakers with no divine addressee at all. Tradition calls them saṃvāda (संवाद, “dialogue”) hymns, and there are only about ten of them.

These are the closest thing the Rigveda has to theatre, and they cluster in the late tenth book, the youngest layer of the collection, alongside the creation hymn and the wedding hymn. They also carry, more than any other hymns, the voices of women who talk back. Read next to the dialogue hymns where women are given the last word, RV 10.95 belongs to a small, sharp tradition in which the immortal woman wins the argument and the mortal man is left holding the grief.

The trouble is that a dialogue with no narrator is a dialogue with no instructions. The Sanskrit text does not label its speakers. Editors assign the lines by grammar (a feminine verb here, a vocative “Purūravā” there) and by sense, and they do not always agree. Whole verses are genuinely obscure. Verse 9 seems to describe the nymphs bathing:

When a mortal, consorting with these immortal nymphs, has spoken with them in words and deeds, they, like waterbirds, do not show their bodies, like skittish horses champing the bit.

(RV 10.95.9, after Wilson)

What is that doing here? The image of the apsarās as waterbirds who hide their bodies is exactly the swan-maiden motif we will come back to, but in the hymn it arrives without introduction and leaves without resolution. This is the texture of the whole poem: a quarrel we are overhearing from the next room, catching every third sentence.

Aside. The traditional index of the Rigveda, the Anukramaṇī, names the ṛṣi (seer) of RV 10.95 as Urvaśī herself and its subject as Purūravas Aiḷa. Take that as a data point about how the tradition read the hymn, a woman’s poem in a man’s collection, not as authorship in any modern sense. On the vexed question of who spoke for the Rigveda, see the rishikas and goddesses of the text.

What the eighteen verses actually say

Strip the later legend away and read only the hymn, and the shape is this. Purūravas opens by begging Urvaśī to stay and talk. She cuts him off: it is finished, she has gone from him like the first light of dawn, he should go home. He protests. He recalls their union, invokes the arrow and the bow, reminds her that she once found him good. She is unmoved, and worse, she is generous about it, the way people are generous when they have stopped needing anything from you.

The most famous line in the hymn is the one Purūravas spits out when reasoning fails:

With women there can be no lasting friendship: hearts of hyenas are the hearts of women.

(RV 10.95.15, trans. Griffith 1896)

The word behind “hyenas” is sālāvṛka, some wild dog or jackal or hyena; the point is that it is an animal of the wilderness, not the settled village. It is a cruel line, and readers have argued for a century about whose cruelty it is. Is it Purūravas lashing out, the mortal turning on the divine because he cannot hold her? Or is it a chorus voice, a bit of proverbial folk misogyny inserted to warn the audience’s own wives against following Urvaśī’s example? Either way, the hymn does not endorse it: the very next verses give Urvaśī the dignity and the last word.

Because she answers, and her answer is the quietest devastation in the Rigveda:

When, changed in form, I wandered among mortals, and dwelt among them through four autumns, I ate, once a day, a small portion of butter; satisfied with that, I now depart.

(RV 10.95.16, after Wilson)

Four years, one mouthful of ghee a day, and she is done. There is no anger in it. Then, having refused every plea, she offers the one thing she can: not herself, but heaven. The immortals, she tells him, have decided his case. He is mortal and must die, but if his descendants keep the sacrificial fires burning, he will one day rejoice with her among the gods:

These gods have said to you, son of Iḷā: since you are indeed subject to death, let your progeny worship the gods with the oblation; and you yourself shall rejoice in heaven.

(RV 10.95.18, after Wilson)

The hymn ends there. Not with reunion, but with a postponement so long it might as well be a refusal, dressed as a gift. Purūravas gets his son, his fire cult, and a place in the sky measured out in generations. He does not get Urvaśī back. Here is the emotional map of the exchange, verse by rough verse:

Notice what is missing. There is no taboo in the hymn, not stated plainly. There is no lightning, no theft of lambs, no explanation of why she is leaving. The reasons that later India would supply so confidently are simply not in the eighteen verses. That gap is the whole story of this text’s afterlife.

The names

Two names carry a surprising amount of the argument.

Purūravas is transparent Sanskrit: puru (“much, many”) plus ravas (“cry, roar, din”), so “he who cries much” or “roaring loud.” It is a strange name for a hero, and it fits a man who spends a hymn wailing on a riverbank. Whether the etymology was invented to suit the story or the story grew to suit a name meaning “loud crier” is a genuine chicken-and-egg problem, and a good one.

Urvaśī is harder. The old and still respectable analysis is uru (“wide”) plus a root meaning “to pervade” or “to reach,” giving something like “widely extending,” an epithet that would suit the spreading light of dawn. Monier-Williams glossed it exactly that way and drew the obvious inference: that Urvaśī began as a personification of the dawn. Later Sanskrit, always fond of a vivid folk etymology, rederived the name from ūru (“thigh”), and told a birth story to match: she sprang from the thigh of the sage Nārāyaṇa. The philology and the mythology pull in different directions, which is a warning we will need in a moment.

Methods note. A Vedic name that “means” something is evidence about how the tradition understood a figure, not proof of that figure’s origin. Purūravas meaning “loud crier” tells us Vedic ears heard the name that way. It does not by itself establish that the character was invented as a weeping mortal. Etymology sets up hypotheses; it rarely closes them.

An apsarā, finally, is a water-spirit (ap, “water”), and her natural partner is the gandharva (गन्धर्व), a class of celestial male beings associated with the waters, the moon, plants, and the marriage rite. This matters for the plot: in the later legend Urvaśī does not simply leave, she returns to the gandharvas, her own kind, and Purūravas can only follow her by becoming one of them.

The taboo and the flash of lightning

Everything the popular story “knows” about Urvaśī and Purūravas, the secret condition, the pet lambs, the flash that reveals him naked, comes not from the Rigveda but from the Śatapatha Brāhmaṇa (XI.5.1), a ritual prose text several centuries younger. Its version runs like this.

Urvaśī agrees to live with Purūravas on conditions: she is to be shown her two pet ram-lambs, she will eat only ghee, and she must never see him naked. The gandharvas, wanting her back in heaven, steal the lambs by night. Urvaśī cries out that she is being robbed as if there were no man about; Purūravas, unwilling to be shamed, leaps from the bed to chase the thieves, and in that instant the gandharvas send a flash of lightning that lights up the room and shows him naked. The condition is broken. She vanishes. He wanders the land distraught until he finds her, with her companions, moving on a lake in the form of waterbirds. She tells him she is pregnant, sends him away for a year, and eventually the gandharvas grant him the boon of a sacred fire by which he can become one of them and rejoin her.

Read the Brāhmaṇa next to the hymn and something important becomes visible. The prose is not simply “explaining” the poem; it is bending the poem to fit a ritual it wants to establish, the fire cult that makes a mortal into a gandharva. Stephanie Jamison, whose work on this hymn is the sharpest we have, argues that the Brāhmaṇa in places misreads its source, straightening the hymn’s ambiguities into a tidy morality tale about a broken taboo. The waterbird verse, which in the hymn is a floating image of the nymphs’ elusive modesty, becomes in the Brāhmaṇa a literal plot point: she is a bird on a lake and he catches her there. The prose needs the poem to be a story. The poem was content to be an argument.

Aside. This is how a great deal of Indian mythology grows: a compact, obscure Vedic verse becomes the seed crystal, and each later text (Brāhmaṇa, epic, Purāṇa, drama) grows a fuller narrative around it, resolving what the original left open. Watching the same figures pass through the layers is one of the real pleasures of reading the tradition forward. The transmission machinery that kept the Rigveda’s exact words intact is precisely what let later authors treat those words as a fixed puzzle to be solved.

The swan maiden

Now turn the story outward instead of forward, and it opens onto one of the widest folktale patterns on earth. A supernatural woman, often a bird who can shed her feathers and become human, marries a mortal man under a condition: he must not see her in some state, or ask her origin, or do some forbidden thing. Domestic happiness follows, sometimes for years. The condition breaks. She resumes her true form and leaves, and the grieving husband must undertake a long quest to recover her, if he can recover her at all.

Folklorists call her the swan maiden, and she is catalogued in the standard indexes as a distinct international tale type. She turns up as the Valkyries who fly off when their swan-shirts are found, as the Japanese tennyo whose feather robe is stolen at Hagoromo, as Siberian goose-wives and Scottish selkies who slip back into their sealskins. In the nineteenth century Sabine Baring-Gould argued that the motif was old enough to predate the dispersal of the Indo-European peoples, and he reached for Urvaśī and Purūravas as very possibly its oldest attested form.

graph TD
    A[The fairy-bride motif] --> B[Condition on the union]
    A --> C[Mortal husband]
    A --> D[Supernatural wife, often a bird]
    B --> E[Condition broken]
    E --> F[Wife vanishes]
    F --> G[Husband's quest]
    D --> H[Urvasi RV 10.95]
    D --> I[Swan maidens, Europe]
    D --> J[Tennyo, Japan]
    D --> K[Selkies, Scotland]

The parallels are real and the temptation is to over-read them. Two cautions are worth keeping.

First, the swan-maiden pattern is so widespread and so psychologically basic (the beloved who cannot be kept, the condition that dooms the keeping) that its appearance in Vedic India need not mean genetic inheritance from a common ancestor. Independent invention is cheap for a story this human. A shared Indo-European origin is possible; it is not proven by resemblance alone.

Second, and more subtly, the swan-maiden shape is clearest in the Brāhmaṇa, not the hymn. It is the prose that gives us the bird on the lake, the stolen condition, the quest. The Rigvedic hymn has the departure and the elusiveness and a single glancing waterbird image, but not the full folktale machine. So when we say the Rigveda holds the “oldest swan-maiden tale,” we are partly crediting the hymn with a structure that its earliest interpreters, not its poet, made explicit.

How to read her leaving

Why does Urvaśī go? The hymn does not say, and into that silence three very different scholarly readings have poured. They are worth laying side by side, because they show how much the same eighteen verses can be made to mean.

The dawn reading. Friedrich Max Müller, the founder of comparative mythology as a Victorian science, read the whole thing as a solar allegory. Urvaśī is the dawn; Purūravas is the sun. “Urvaśī loves Purūravas” means the sun rises; “Urvaśī sees Purūravas naked” (or he is revealed) means the daylight comes and the dawn must fade; her flight is the vanishing of the morning light before the risen sun. The etymology, dawn as the “widely spreading” one, seemed to clinch it. It is a beautiful theory and almost no one holds it now. Later scholars found the solar key unlocked far too many doors: applied without restraint, every Vedic story dissolved into sunrise and sunset, which is another way of saying the method explained nothing in particular.

The ritual reading. The Marxist historian D. D. Kosambi turned the hymn on its head and read it as the fossil of a fertility rite. In his reconstruction, Urvaśī is not a fickle nymph but a mother-goddess, close kin to Uṣas, and Purūravas is the year-king, the consort chosen for a season, kept for the fathering of a child, and then discarded (in the older, harsher form of such rites, sacrificed). The hymn, on this view, is the pleading of the doomed consort, and its real subject is the clash between an old goddess-centred cult and the incoming, patriarchal, Indra-worshipping order. Kosambi’s reading is bold, materialist, and, like most single-key readings, more confident than the evidence strictly allows. But it did something valuable: it took the woman’s power in the hymn seriously as social fact, not decoration.

The social reading. The reading that wears best is the least reductive. Jamison places the hymn inside the anxieties of a patrilineal, patrilocal society about the wife who might leave, and reads the whole quarrel as doing cultural work: it lets the audience feel the full pull of Urvaśī’s freedom and the husband’s loss, and then contains it, channelling Purūravas’s grief into the socially useful outcome of a son and a fire cult. The nasty “hearts of hyenas” line and the consoling promise of heaven are two halves of the same management: the departing woman is both slandered and sublimated. On this account the hymn is not an allegory of anything. It is a real argument about a real fracture, marriage across an unbridgeable difference, staged with enough art that it still lands.

My own view, for what a reader’s view is worth, is that the third reading is right about the hymn and the first two are right about its resonances. The poem itself is a marital argument of unusual honesty. But it would not have haunted three thousand years of retelling if it were only that. Urvaśī carries the afterglow of the dawn goddess and the authority of something older than the men writing her down, and Purūravas carries every mortal’s certainty that what he loves most cannot be held. Compare Urvaśī with Uṣas, the dawn who comes and goes and never stays, and the family resemblance is unmistakable even if the genealogy is not.

From nymph to dynasty

The Rigveda gives Purūravas one consolation prize inside the hymn: his descendants. Later India cashed it in spectacularly. Purūravas Aiḷa, “son of Iḷā,” became the founder of the Lunar Dynasty, the candravaṃśa, the royal line that runs through his son Āyus to Nahuṣa, Yayāti, Pūru, and eventually to the Bharatas and the warring cousins of the Mahābhārata. The weeping king of a single obscure hymn became the fountainhead of the epic’s entire royal genealogy.

He also became a culture hero of the sacrifice. Several texts credit Purūravas with bringing the three sacred fires from the world of the gandharvas down to earth, so that the fire-worship at the heart of Vedic ritual traces back to a mortal’s grief-driven journey into the other world. It is a neat inversion: the man could not keep his celestial wife, but he brought back her people’s fire, and the fire is what the hymn told him to tend. The whole machinery of the Vedic fire cult gets, in this legend, an origin story rooted in loss.

graph LR
    A[RV 10.95, dialogue] --> B[Satapatha Brahmana XI.5.1]
    B --> C[Epic and Puranic genealogy]
    C --> D[Pururavas, Lunar Dynasty]
    B --> E[Kalidasa, Vikramorvasiyam ~400 CE]
    E --> F[Modern retellings and painting]

The final transformation is Kālidāsa’s. In his play Vikramorvaśīyam (“Urvaśī Won by Valour”), written in the Gupta era, the tragedy becomes a romance. There is a curse, a magical gem of reunion, a lost-and-found son named Āyus, a message from Indra, and, at the last, a reprieve: Urvaśī is allowed to stay with Purūravas for the whole of his life. Kālidāsa gives the king everything the Rigveda denied him. It is a measure of the hymn’s strangeness that the tradition’s greatest poet felt he had to reverse its ending to make it bearable. The Vedic poet did not blink; he let the nymph walk into the wind.

The argument that would not end

What makes RV 10.95 endure is not that it is a love story, but that it is a love story with the consolations deleted. The Brāhmaṇa gives you a taboo to blame, so the loss feels earned. Kālidāsa gives you a reunion, so the loss is undone. The Rigveda gives you neither. It gives you a man who has done nothing especially wrong, loving a woman who owes him nothing, and it lets her leave because leaving is simply what she is. She is “hard to catch as the wind” not as a punishment but as a fact of her nature, the way ṛta, the Vedic order of things, is simply the way the world is set. The immortal and the mortal touched, and the touch could not hold, and no one is quite to blame.

That refusal to assign blame or supply comfort is rare in any literature and astonishing in a text this old. It is also why the hymn keeps generating readings: dawn myth, fertility rite, social anxiety, folktale, each of them true to some facet and none exhausting it. The eighteen verses are a Rorschach with a broken heart at the centre.

Read it once straight through, ignoring the later stories, and let the obscurities stay obscure. Then read the Śatapatha Brāhmaṇa’s prose beside it and watch the tradition start, almost immediately, to explain away what it could not bear to leave unexplained. The distance between the two is the distance between a poem and its comfort. The Rigveda, here as in the creation hymn, is willing to end on a question. Very little written since has had the nerve.