The lamp that will not cross over

To translate is, at root, to carry across: trans-latio, the Latin says it plainly, a bearing of something from one bank to the other. The image is honest about the risk. Something is picked up here, ferried, and set down there — and anyone who has carried water across a room knows that the carrying is where the loss happens. A Kabir doha is, in this respect, the worst possible cargo. It is small, it is dense, and almost everything that makes it itself lives in the very things a translator cannot lift: the sound of the words, the grammar of a language that thinks differently about being, and a name for God that is not quite a name.

Consider one couplet. It is among the most quoted lines attributed to Kabir, the kind of verse Indians who can recite nothing else of his can still recite:

जब मैं था तब हरि नहीं, अब हरि हैं मैं नाहिं।

सब अँधियारा मिट गया, जब दीपक देख्या माहिं॥

jab maĩ thā tab hari nahī̃, ab hari haĩ maĩ nāhī̃.
sab am̐dhiyārā miṭ gayā, jab dīpak dekhyā māhī̃.

A flat, almost word-for-word English might run: When I was, Hari was not; now Hari is, and I am not. All the darkness was wiped away when I saw the lamp within. That is the sense. It is also, already, a small betrayal — and watching where the betrayals fall is the most exact lesson in what a Kabir verse actually is.

(A first flag, since this essay is partly about honesty in attribution: the doha is transmitted in dozens of slightly different forms across the popular sakhi tradition, and its precise recensional home is itself contested. I give here the version that circulates most widely in print and recitation, and I make no claim that one manuscript line is the "original." That instability is not a footnote to the argument; it is the argument.)

Why eight words will not move

Look at the first line. In the Devanagari it is eight short words, and it holds an entire theology of the dissolved self: as long as the "I" took up room, the divine had none; the moment the "I" emptied out, only the divine remained. The economy is the meaning. Kabir is not describing ego-death at leisure; the line enacts it, compressing a whole spiritual biography into a breath you could speak between two steps of a loom.

English cannot keep that economy. Our grammar makes us spend words on the scaffolding Hindi leaves implicit — articles, the verb "to be" dragged out into "was" and "is," the pronoun "I" repeated where the original lets a single maĩ do double duty. By the time an English sentence has said what the line says, it has used twice the words and lost the snap.

It loses the sound, too. Hear the braid of negation and rhyme: nahī̃ … nāhī̃ … māhī̃ — three line-ends chiming on the same nasal vowel, the word for no and the word for within made to rhyme so that absence and interiority answer each other in the ear. And hear the doubled maĩ, the "I" that the verse names twice precisely in order to abolish it. None of this is decoration. In a doha — a form built for memory, meant to be carried in the body and not on the page — sound is structure. Strip it and you have the proposition without the proof.

Then there is the hardest word: Hari. We reach, in English, for "God" or "the Lord." But Hari is a name — one of the old names of Vishnu, "the one who takes away" — and Kabir, the nirguna poet who spent a lifetime refusing every fixed image of the divine, uses it anyway, drained of its sectarian body, as a sound that points past itself. (His suspicion of received forms is the spine of his Teachings & Quotes essay.) "God" imports a whole Christian theology the verse never asked for. "The Lord" kneels in a register Kabir distrusted. "Hari," left untranslated, asks the reader to do work most readers won't. Every choice here is a small doctrinal decision disguised as a vocabulary problem.

Tagore and Vaudeville: the smoother and the literalist

Set two temperaments against each other and the stakes get visible.

Rabindranath Tagore's Songs of Kabir (1915), shaped from Kshitimohan Sen's Bengali anthology and prefaced by the Christian mystic Evelyn Underhill, is the version through which the English-reading world first met Kabir. Its register is devotional, hushed, faintly Edwardian — "O servant, where dost thou seek Me?" — and it smooths Kabir's burrs into something a Theosophist could love. The smoothing is not a failure of skill; it is a reading. Tagore heard in Kabir a universal mystic, a singer of the One, and he translated that Kabir faithfully. What goes missing is the grain: the abrasiveness, the marketplace bluntness, the deliberate roughness of a weaver who insulted his listeners awake.

There is a deeper honesty problem here, and it bears directly on our doha. Tagore worked almost entirely from the pads — the longer sung lyrics — not from the sakhis, the doha-form couplets to which our verse belongs. So the clean fantasy behind this essay's working title — one doha laid side by side across the same six hands — cannot quite be honored, because two of the most famous "translators of Kabir" never translated this kind of verse at all. (I therefore do not reproduce a Tagore line for this doha; none reliably exists, and I will not manufacture one.) That gap is itself the recension problem wearing a different coat: not only which Kabir did a translator work from, but which form of Kabir his source even contained.

Against Tagore stands Charlotte Vaudeville, whose Kabir (1974) and A Weaver Named Kabir (1993) are the scholar's Kabir — literal, annotated, armored with apparatus. Vaudeville worked the western, Rajasthani sakhi tradition and rendered the dohas close to the bone, preserving paradox and idiom even when the English bruises. Where Tagore makes Kabir singable, Vaudeville makes him legible — you can see the joints of the original through her English glass. The cost is the reverse of Tagore's: clarity bought at the price of music; a verse you can study but would never spontaneously recite.

For our doha, a Vaudeville-temperament rendering would refuse to soften the bare grammar:

When I was, Hari was not; now Hari is, I am not.

All the dark was wiped out the moment I saw the lamp within. Bhakti Saints editorial · literalist register

I label that as ours, a demonstration of the approach rather than a quotation of Vaudeville's published line — the house rule of this site, and here the rule is also the subject. Notice what literalism keeps: the brutal symmetry of the first line, the refusal to explain. Notice what it still can't keep: the rhyme, the doubled maĩ, the breath.

Bly and Hess: the re-voicer and the performer

Robert Bly's The Kabir Book (1977) is the most-read Kabir in America, and it is not, strictly, a translation at all. Bly knew no Hindi; he reworked Tagore's English into spare modern free verse — "versions," he called them honestly. So Bly is a translation of a translation, twice removed from the Devanagari, and what he produces is genuinely new: a plainspoken, Midwestern, almost confessional Kabir who sounds like he could be talking to you across a kitchen table. Purists wince; readers swoon. Bly's gift is voice — he found an English in which Kabir feels urgent and alive — and his liability is fidelity, since the thing he is faithful to is Tagore's reading, not Kabir's words.

A Bly-temperament version of our doha would re-voice it into bare American speech:

As long as I was here, God wasn't.

Now God's here and I'm gone.

The dark just lifted —

turns out I'd been holding a lamp the whole time. Bhakti Saints editorial · free-verse register

Again: ours, a demonstration, not Bly's line. See how much it gains in immediacy and how much it spends to get there — "turns out," "the whole time," an American shrug Kabir never wore. The verse becomes reachable and slightly less strange, and Kabir's whole project was to keep you in the strangeness.

Linda Hess, with Shukdev Singh, gives us the fourth temperament in The Bijak of Kabir (1983) and later, more fully, in Bodies of Song (2015). Hess translated the Bijak — the eastern recension, the scripture of the Kabir Panth — and she translated it as something performed: heard, sung, argued with, alive in the mouths of singers she sat with for years. Her English is muscular, rhythmic, alert to the way a line lands in air rather than on a page. She is the translator most aware that a sakhi is not a text but an event. A Hess-temperament rendering would chase that orality, the stop-and-strike rhythm of speech:

When I was, He wasn't.

Now He is — I'm not.

Dark, all of it, gone

the second I saw the lamp inside. Bhakti Saints editorial · performance register

The short lines are the point: they are how the verse would actually be spoken, the caesura doing the work the Devanagari does with its sound. This is the version closest to what a doha is for — and still the rhyme is gone, still Hari has become "He," still eight words have become sixteen.

What survives, and what only the original holds

Lay the four temperaments together and something clarifying happens. The proposition survives every crossing: the self and the divine cannot occupy the same room; emptying of one is the arrival of the other; the light was always interior. You can carry the thought across in Tagore's hush or Bly's plain talk or Vaudeville's precision or Hess's struck rhythm, and it arrives intact. Kabir's mysticism, it turns out, is robust cargo — which is part of why he travels so well, why he is the project's most-read saint.

What does not survive is everything that made it a doha rather than a sentence. The rhyme that married no to within. The doubled maĩ that names the ego twice to kill it. The eight-word compression that performs the dissolution it describes. The single sound Hari, pointing past every name including its own. These are not ornaments laid over the meaning; in Kabir they are a kind of meaning — the formal argument that truth is sudden, compressed, and interior, made in the very shape of the line. That argument is exactly the part that will not cross the river. The lamp stays on the far bank.

Coda: translation as the test of a verse's spine

There is a useful severity in all this. Translation is the truest stress-test we have for a verse's spine — because it strips away everything inessential and shows you what was load-bearing. Run this doha through four temperaments and you learn precisely which of its features were structure and which were skin. A weaker verse dissolves entirely in translation; you carry across a damp idea and nothing else. Kabir's doha survives — bruised, robbed of its music, but standing — which is itself a verdict on its strength.

And the severity cuts toward Kabir's own preoccupation. He distrusted received forms, scriptures recited without understanding, the Pandit's book and the Mullah's call alike (the burden of his Teachings and the death-legend in the Bio & Impact essay, in which Hindu and Muslim quarrel over a body that has become flowers). A translation is a received form — someone else's reading, handed to you finished, asking to be trusted. To read Kabir across six hands is to be reminded, in the most concrete way, that there is no neutral, transparent version waiting behind the translations; there is only this reading and that one, each a decision, each a small theology. Whose Kabir is this? is not a question we can answer once and shelve. The Bijak's Kabir, the Granthavali's Kabir, the Adi Granth's Kabir, Tagore's, Bly's — these are different poets wearing the same name, and the differences are not noise. They are the most honest thing about the tradition: that it was always plural, always in motion, always being re-sung by the next mouth, exactly as a doha is built to be.

To carry the lamp across is impossible. But you learn the most about the lamp in the trying.

Sources & notes

• Rabindranath Tagore, Songs of Kabir (1915), from Kshitimohan Sen's anthology, intro. Evelyn Underhill — characterized here for its devotional smoothing; no line reproduced for this doha (Tagore worked the pads, not the sakhis).
• Charlotte Vaudeville, Kabir, vol. 1 (Oxford, 1974) and A Weaver Named Kabir (1993) — scholarly literalism; the western/Rajasthani sakhi recension.
• Robert Bly, The Kabir Book (1977) — "versions" reworked from Tagore's English, not from Hindi.
• Linda Hess & Shukdev Singh, The Bijak of Kabir (1983; OUP 2002); Linda Hess, Bodies of Song (2015) — the eastern Bijak recension, read as performance.
• David Lorenzen, Kabir Legends and Ananta-Das's Kabir Parachai, and the standard accounts of the Bijak / Granthavali / Adi Granth recensions, for the "whose Kabir" / transmission problem.

Attribution flags (house rule, and the subject of this essay): No published line from Tagore, Vaudeville, Bly, or Hess & Singh is reproduced here. Each translator's approach is characterized from the documented character of their work; the four bracketed renderings of the doha are labelled Bhakti Saints editorial renderings, written to dramatize each temperament, not to stand in for any translator's exact wording. The doha's text is given in its most widely transmitted popular form; its precise recensional source is contested and is not asserted.

Cross-references

Themes: Translation (primary) · The vernacular

Kabir essays: Bio & Impact · Teachings & Quotes · Saint page