The oldest cosmogony in any Indo-European language — the Rig Veda's Hymn of Creation, Nāsadīya Sūkta, 10.129 — does not end where a scripture is expected to end. It ends by saying that the gods came too late to have seen the beginning, and that the one who watches from the highest heaven só aṅgá veda, yádi vā ná véda — he only knows — or perhaps he knows not. This is the twenty-fifth entry of the Translation-Criticism Venue, and it is about the four words at the end, and who let them stand.
The Rig Veda is a body of some 1,028 hymns in an archaic Sanskrit, composed in north-west India in roughly the latter half of the second millennium BCE and transmitted, for most of that time, by memory alone. Most of the hymns praise or petition the gods — Agni the fire, Indra the storm, Uṣas the dawn. But the tenth and latest book holds a handful of hymns that turn away from the gods and ask, instead, where any of it came from. The most famous of these is the one traditionally called, from its first two words, the Nāsadīya Sūkta — the na-asat hymn, "the hymn of there-was-neither-non-being."
It was the first Vedic hymn to reach Europe: Henry Thomas Colebrooke translated it in Calcutta in 1805, and every generation of Sanskritists since has translated it again. It has been called the earliest philosophy in any Indo-European tongue, quoted by physicists and read aloud on television. And it is genuinely strange. It opens by refusing the most basic distinction language has — nā́sad āsīn nó sád āsīt, "there was neither the non-existent nor the existent" — and it closes, seven stanzas later, by refusing the thing a creation story exists to supply: an answer. A hymn bracketed by two refusals.
The closing refusal is the sharper one, and it is where the translations tear. The last line names an adhyakṣa, an overseer in the highest heaven, and then adds four words about him: só aṅgá veda, yádi vā ná véda. Read plainly, they say he indeed knows — or perhaps he does not know. A sacred text, ending in a shrug. Whether a translator lets that shrug stand turns out to depend on one thing: whether he read the Sanskrit directly, or read it through the eyes of Sāyaṇa, the fourteenth-century commentator for whom an all-knowing overseer cannot not know. This page lines up ten English renderings, from Colebrooke in 1805 to Macdonell in 1922 — every line transcribed verbatim from its named public-domain scan — and lets you watch the doubt open, close, and open again.
The whole hymn, in A. A. Macdonell's accented transliteration and his own literal prose (A Vedic Reader for Students, Oxford 1917), verified letter by letter against the printed page. Tap any lit word for its sense. The words in rose are the ones the scholars themselves mark as uncertain or contested — the places where even the grammar shrugs.
Here are all ten renderings of the final line — the four Sanskrit words só aṅgá veda, yádi vā ná véda and what leads into them. Two of these ten translators built their English on Sāyaṇa's medieval commentary; the other eight worked from the bare Sanskrit. Switch the filter and watch the split. Where a translator has added words that the six Sanskrit words do not contain — the words that close the doubt — they are marked in rose.
The pattern is not subtle, and it is not about the translators' skill — Colebrooke and Wilson were among the finest Sanskritists of their centuries. It is about what they were reading. The Sanskrit line ends with an open yádi vā ná véda, "or if not he knows." To close it, a translator has to add — Colebrooke adds "but not another can possess that knowledge"; Wilson adds "(no one else does)" — and both additions come straight from Sāyaṇa, who reads the whole hymn, in Wilson's own words, under "the Vedāntism of a later period."
Arthur Macdonell, Boden Professor of Sanskrit at Oxford, translated this hymn three times across twenty-two years. He never read it through Sāyaṇa; he let the doubt stand every time. But he could not decide how to say it — the same four Sanskrit words came out three different ways under his own hand. (In the 1900 version he also quietly dropped the fifth stanza entirely, the one W. D. Whitney called the hymn's most unintelligible.)
"Or e'n he does not know it" (1900), "or else he knows not" (1917), "or haply he may know not" (1922). Three settlings of a line that will not settle — from the one translator who was arguing only with himself.
The hymn's fame is nearly unanimous. Max Müller, introducing it to English readers in 1859, reached for Greece and Germany at once — the poet speaks, he wrote, "with a boldness matched only by the Eleatic thinkers of Greece, or by Hegel's philosophy" — and on the paradox of "That One breathed, breathless," he laid down one of the great sentences of nineteenth-century philology:
Macdonell, half a century on, was cooler but still admiring: the hymn shows, he wrote, "how philosophical speculation can be clothed in poetry of no mean order." And then there is William Dwight Whitney of Yale — who edited the Atharva Veda, wrote the Sanskrit grammar every student still uses, and read this same hymn in 1882 for the American Oriental Society, and was not moved at all. He worked through it stanza by stanza (the fifth, he said, "no one has ever succeeded in putting any sense into it, and it seems so unconnected with the rest of the hymn that its absence is heartily to be wished") and delivered a verdict that the record has never quite lived down:
We take no side. Whitney's dissent is here because it is real, and because a venue that shows only the praise would be doing to the record exactly what Colebrooke's translation did to the last line — quietly closing what the source leaves open. Even the closing doubt itself Müller would not resolve: the four words, he wrote, "may be interpreted in two ways" — either "a question of defiance addressed to all who might doubt," or "a confession of doubt on the part of the poet." He left both on the page. So do we.