Ille Loquitur: Two Centuries of Getting the Rigveda Into European Languages

A Veda That Voltaire Could Hold

In 1760 a French officer named Louis Laurent de Féderbe, Comte de Maudave, handed Voltaire a manuscript. It was written in French and it presented itself as the translation of an ancient Indian scripture: a dialogue between two Vedic sages, one of whom patiently explains to the other that behind the crowd of Indian gods there stands a single creator. Voltaire, who had spent decades arguing that the moral core of religion predated and outranked the church, was delighted. Here, apparently, was the proof in writing, older than Moses and untouched by Rome.

The manuscript was a fake. It had been composed in French by Jesuit missionaries in India, almost certainly for translation into Sanskrit rather than from it, and it was engineered to make Hinduism look like a monotheism that had lost its way and could be led back. Voltaire cooled on it within a year, downgrading it in his own mind to a commentary rather than a Veda. It was printed in 1778, the year he died. Doubts about its authenticity were raised in 1782 and settled in 1822. Ludo Rocher’s 1984 critical edition finally laid out the whole affair, textual sources and all, under a title that says everything: Ezourvedam: A French Veda of the Eighteenth Century.

That is where European access to the Rigveda begins: not with a text, but with a forgery that told its reader what he wanted to hear. Everything that follows is a two-hundred-year effort to get from that starting point to something a person could actually check. The effort produced a dead man’s unfinished Latin edition, a six-volume Sanskrit printing paid for by a trading company, a bitter quarrel about whether a fourteenth-century Indian commentator knew what the hymns meant, a mathematician who abandoned mathematics for a Vedic dictionary, and finally a complete scholarly English version that appeared in 2014. It also produced the version almost everyone actually reads, which is none of these, and which arrived at its dominance by an accident of copyright law.

The Man Who Reached Hymn 121

Friedrich August Rosen was born in Hannover in 1805 and died in London on 12 September 1837, ten days after his thirty-second birthday. He had already published a specimen of Rigvedic text in 1830. His larger project was an edition of the whole collection, Sanskrit on one side and his own Latin translation facing it, so that any European scholar with a classical education could read the hymns without waiting for a vernacular version.

He did not get far. Rigveda-Sanhita, liber primus, sanskritè et latinè appeared posthumously in 1838, printed for the Oriental Translation Fund of Great Britain and Ireland. It contains the first 121 hymns of Mandala 1. There are 191. Rosen had covered slightly less than twelve per cent of the Rigveda, and the remainder of the text waited another half century for a complete printed edition.

It is worth pausing on what Rosen was actually doing, because it sets the terms for everyone after him. He was not producing a devotional translation or a literary one. He was producing a bilingual scholarly instrument, in which the Latin exists to license inspection of the Sanskrit. Every serious edition since has had the same shape underneath: the translation is a claim, and the claim is only useful if the reader can see what it is a claim about.

The first thing a reader wants from such an instrument is the opening line, and the opening line is a good measure of how hard this is. Here is the first verse of the collection in a modern rendering:

I call upon Agni, the one placed in front, god and priest of the sacrifice, the Hotar, best bestower of treasure.

(RV 1.1.1, after Jamison and Brereton)

Several of those words are technical terms of a ritual system nobody was practising in Rosen’s London, and one of them, ratnadhātamam, has been rendered as “treasure-giver,” “bestower of wealth,” “most richly endowing,” and “the one who best establishes what is precious,” depending on which century you consult. The Rigveda does not resist translation because it is obscure. It resists because its vocabulary is technical, its institution is gone, and its poets assumed you already knew.

Six Volumes and a Trading Company

The complete Sanskrit text arrived through Friedrich Max Müller, a German scholar in Oxford, and the money arrived through the East India Company.

Müller’s Rig-Veda-Sanhita: The Sacred Hymns of the Brahmans, together with the Commentary of Sayanacharya was published in London by W. H. Allen in six volumes between 1849 and 1874. That span is not a typo. The project took twenty-five years, and the political ground shifted underneath it: volumes one to three carry the phrase “published under the patronage of the Honourable the East-India-Company” on their title pages, while volumes four to six, printed after the Company was wound up following 1858, carry the patronage of Her Majesty’s Secretary of State for India in Council. The state that had absorbed the Company absorbed the Veda project along with it.

The decisive editorial choice was in the subtitle. Müller printed the hymns together with the bhāṣya, the running commentary of Sāyaṇa, a Mīmāṃsā scholar and minister at the Vijayanagara court who died in 1387 and composed his Rigveda commentary at royal request. Sāyaṇa’s commentary is the single most influential piece of Rigvedic interpretation ever written, and printing it alongside the text made it, for a generation of European readers, the default meaning of the hymns.

Müller himself did not believe that. Having spent a quarter of a century editing Sāyaṇa, he later issued text-only editions in 1873 and 1877, stripping the commentary out. Ralph Griffith, quoting him, records the position bluntly:

Aside. Müller on the commentarial chain: “As the authors of the Brâhmaṇas were blinded by theology, the authors of the still later Niruktas were deceived by etymological fictions, and both conspired to mislead by their authority later and more sensible commentators, such as Sâyaṇa.” He is describing the tradition he had just spent twenty-five years putting into print.

There is a persistent claim online that the Company paid Müller a specific sum per printed sheet, usually quoted at four pounds. I could not verify any figure from a reliable source. Müller’s own account, and the scholarly literature on the edition, record only that the funds for printing were granted. The number should be treated as folklore until somebody produces the contract.

One more thing about Müller, because it is routinely got wrong: he never translated the Rigveda. His contribution to the Sacred Books of the East series was volume 32, Vedic Hymns, Part I, published in 1891, covering hymns to the Maruts, Rudra, Vāyu and Vāta. Part II, the hymns to Agni, appeared as volume 46 in 1897 and was translated by Hermann Oldenberg, not by Müller. The most famous name in nineteenth-century Vedic studies produced an edition, not a translation.

Whose Rigveda: The Commentator or the Poets?

The quarrel that organises the entire nineteenth century is easy to state. Sāyaṇa wrote roughly 2,500 years after the hymns were composed, in a different part of India, in a different religious world, with an entire scholastic tradition between him and the text. Does that make him an authority or a witness?

Horace Hayman Wilson said authority. Wilson, born 1786, was elected the first Boden Professor of Sanskrit at Oxford on 15 March 1832, beating William Hodge Mill by 207 votes to 200. His Ṛig-Veda Sanhitá: A Collection of Ancient Hindu Hymns ran to six volumes between 1850 and 1888, the first three under East India Company patronage. His reasoning was that Sāyaṇa possessed “a knowledge of his text far beyond the pretensions of any European scholar, and must have been in possession of all the interpretations which had been perpetuated by traditional teaching from the earliest times.” Wilson accordingly translated what Sāyaṇa said the hymns meant. Griffith, a generation later, described the result with some accuracy as “rather a version of Sâyaṇa’s paraphrase.”

Rudolf von Roth said witness. Roth, at Tübingen from 1845 until his death in 1895, published Zur Litteratur und Geschichte des Weda in Stuttgart in 1846 and spent the next three decades building, with Otto Böhtlingk, the great St Petersburg Sanskrit dictionary. His position, in the formulation that circulated in English through John Muir and was quoted approvingly by Griffith, was this:

“We do not esteem it our first task to arrive at that understanding of the Veda which was current in India some centuries ago, but to search out the sense which the poets themselves have put into their hymns and utterances. Hence we are of opinion that the writings of Sâyaṇa and the other commentators do not form a rule for the interpreter, but are merely one of those helps of which the latter will avail himself.”

Roth went further in private and in effect: he held that a careful European interpreter could understand the Veda better than Sāyaṇa, and that Sāyaṇa “knew actually no more of the real content than we do.”

It is tempting to narrate this as European arrogance versus Indian tradition, and that framing is not empty: Roth was writing in a period when philology carried heavy assumptions about who was entitled to interpret what, and Michael Dodson’s work on the politics of Vedic translation in colonial India is the place to follow that thread properly. But the philological point survives the politics. Sāyaṇa is separated from the Rigveda by a longer stretch of time than separates us from Sāyaṇa. That is a fact about chronology, not about nationality, and it constrains any reader.

What can actually be concluded is narrower than either camp claimed. Sāyaṇa is indispensable for ritual vocabulary, where a living tradition really did carry information forward, and unreliable for the archaic poetic vocabulary that had already gone dark by the time of the Brāhmaṇas. Jamison and Brereton state the modern consensus carefully in their 2014 introduction: Sāyaṇa is “temporally closer to our own age than to that of the Ṛgveda,” and it is “more useful to read Sāyaṇa not as a direct conduit of the ‘true meaning’ of the Ṛgveda but as a scholar grappling with the same problems as modern interpreters.” That is not a dismissal. It is a demotion from oracle to colleague.

The Mathematician Who Gave Up

While this argument ran, the practical tools were being built, and one of them came from an unexpected direction.

Hermann Grassmann, a schoolteacher in Stettin, had published in 1844 a work of mathematics that invented what we now call linear algebra and exterior algebra. Nobody read it. He revised it, republished it, and was ignored again. Disappointed by the reception, he lost contact with mathematicians and turned his attention to Sanskrit.

Aside. The output of that redirection: a roughly 2,000-page Wörterbuch zum Rig-Veda in 1873, indexing every word in the collection with its occurrences, and a complete German verse translation of the Rigveda in two volumes running past 1,000 pages, published 1876-1877. Grassmann also identified the sound law that still carries his name. Tübingen gave him an honorary doctorate in 1876, for philology. He died in September 1877. The mathematics he actually cared about was rediscovered after his death.

Grassmann’s dictionary is still on Sanskritists’ desks. It is the sort of object that only gets made by somebody with nothing else to do and an unusual tolerance for repetition, which is to say by a person whose first career had failed.

Alongside him, Theodor Aufrecht produced Die Hymnen des Rigveda, first in Albrecht Weber’s Indische Studien in 1861-1863 and then in a two-volume Bonn edition of 1877. Aufrecht’s edition was printed in roman transliteration rather than Devanagari, which sounds like a trivial typographic decision and was not: it made the Rigveda directly usable by Indo-Europeanists who worked on Greek, Latin, Gothic and Old Irish and had no interest in learning a new script. A large share of comparative Indo-European linguistics runs on Aufrecht’s text.

Alfred Ludwig, professor at Prague, published a German translation in six volumes between 1876 and 1888, with three of those volumes given over to notes. He also prepared a revised English version between 1886 and 1893 which was never printed in his lifetime and only appeared in an edited edition in recent years. That is the pattern of the field in miniature: the work gets done, and then it waits.

Ille Loquitur

Ralph Thomas Hotchkin Griffith was born in Wiltshire in 1826 and spent his working life in India. He joined Benares Government College as professor of English literature in 1853, became headmaster in 1854, succeeded Ballantyne as Principal of the Benares Sanskrit College in 1861, held that post until 1878, then served as Director of Public Instruction for the North-Western Provinces and Oudh until retiring in 1885. He died at Kotagiri in the Nilgiris in 1906.

His Hymns of the Rigveda, translated with a popular commentary appeared in four volumes at Benares from 1889 to 1892, published by E. J. Lazarus, with a two-volume second edition in 1896-1897 and a third in 1916. His method was metrical. He explains in the preface that he translated “each verse by a verse syllabically commensurate with the original and generally divided into corresponding hemistichs,” rendering octosyllabic Vedic verse in English iambic dimeter. It is unrhymed, archaic, and reads like a minor Victorian poet doing his best.

Except in two places, where it does not read like English at all.

Turn to RV 1.126. The hymn opens as a dānastuti, a praise of a patron’s gifts, and then swerves into a short erotic dialogue. Griffith translates verses 1 to 5 into English. Verses 6 and 7 he translates into Latin:

Ille loquitur. Adhaerens, arcte adhaerens, illa quae mustelae similis se abdidit, multum humorem effundens, dat mihi complexuum centum gaudia.

(RV 1.126.6, exactly as printed in Griffith’s English translation of 1896)

And the woman’s reply, put back into the language Griffith declined to use:

Come close, come closer, and touch me softly. Do not think my body hair is scant: I am shaggy all over, like a ewe of the Gandhāris.

(RV 1.126.7, Englishing the Latin Griffith printed in its place)

The same thing happens at RV 10.61.5-8, where the passage describing a father’s violent act toward his daughter goes into Latin at the words Membrum suum virile. This was not eccentricity. Rendering indecent passages into Latin was standard Victorian editorial practice across classics and orientalism alike, on the theory that a reader with Latin was a reader who could be trusted. The Rigveda in English therefore has, embedded in it, a nineteenth-century judgement about which of its readers were adults.

The comparison that makes this vivid is with Wilson, forty years earlier. Wilson, working through Sāyaṇa, put the same verse into plain English: “Approach me, deem me not immature; I am covered with down like a ewe of the Gandhārins.” Sāyaṇa’s decorous gloss and Wilson’s willingness to follow it produced a translation that says what the verse says. Griffith, working closer to the Sanskrit, produced one that hides it. Fidelity to the original and fidelity to what the original says are not the same variable.

Methods note. I verified Griffith’s Latin directly against the text as published at sacred-texts.com and cross-checked the same verses on Wikisource; both preserve the Latin at 1.126.6-7 and 10.61.5-8. A related claim circulates that Griffith did the same at 10.86, the Vṛṣākapi hymn. That is not what the received text shows: verses 16 and 17 of that hymn are simply absent, with no Latin and no note, the sequence running 1 to 15 and then 18 to 23. Whether the printed Benares volumes carry a footnote at that gap I could not confirm, because the page scans I could reach are restricted. Treat the omission as documented and its explanation as open.

Why the Default Rigveda Is From 1896

Griffith’s translation is the one you will find if you search for the Rigveda in English. It is on sacred-texts.com, credited “Ralph T.H. Griffith, Translator [1896],” and it has been there since roughly the founding of the Internet Sacred Text Archive, which opened to the public on 9 March 1999 under John Bruno Hare. It is on Wikisource in full, preface included. It is quoted in blog posts, forum arguments, comparative religion syllabi, and a great many books that should know better.

It is there for one reason. Griffith died in 1906 and his translation was published in the 1890s, so it is out of copyright, and every complete alternative in English was either unfinished, unpublished, or protected. The web filled the slot with what the web could legally copy.

Jamison and Brereton, in the general introduction to their own translation, are unusually direct about the consequences: “the only readily available complete English translation, the nineteenth-century product of R. T. H. Griffith, conceals rather than reveals the wonders of the Ṛgveda.” It “really does not deserve as many rebirths as it has had. Its philology was already dated when it was published, and the English style of the translation is cloying and almost unreadable.”

Dated how? Griffith was translating before the great twentieth-century advances in Vedic accentuation, meter, and comparative Indo-Iranian syntax. He worked with a nineteenth-century understanding of the ritual system. He smoothed obscurity into decorous English rather than flagging it. And he made the choices a Victorian schoolmaster in Benares would make, including the Latin.

$$\text{what a reader encounters} = \text{what is legally free} \; \neq \; \text{what is currently correct}$$

That equation is not specific to the Rigveda. It governs Homer in English, the Zend Avesta, the Egyptian Book of the Dead, and most of the world’s classical literature on the open web. But it has an unusually sharp effect here, because the Rigveda’s difficulty means the gap between a good translation and a bad one is not a matter of style. It is the difference between reading a text and reading a guess about a text.

graph TD
  A["Oral Rigveda, c. 1400 BCE"] --> B["Padapatha and schools"]
  B --> C["Sayana's commentary, 14th c."]
  C --> D["Manuscripts reach London"]
  D --> E["Muller's edition 1849-1874"]
  D --> F["Aufrecht's roman text 1861-1877"]
  E --> G["Wilson: translate the commentary"]
  F --> H["Roth: translate the poets"]
  G --> I["Griffith 1896"]
  H --> J["Geldner 1951"]
  J --> K["Renou 1955-1969"]
  K --> L["Jamison and Brereton 2014"]
  I --> M["Public domain, so ubiquitous"]

The Long Wait

The good translations existed. They were just slow, foreign, or partial.

Karl Friedrich Geldner, born in 1852, sent the manuscript of Der Rig-Veda aus dem Sanskrit ins Deutsche übersetzt und mit einem laufenden Kommentar versehen to his publisher in 1928. He died in Marburg in February 1929. German hyperinflation stopped the printing; the work was eventually produced in the United States under Charles Lanman’s supervision, then delayed again by the Second World War, and appeared as volumes 33 to 36 of the Harvard Oriental Series in 1951. Twenty-three years passed between a finished translation and a readable book, and the result was in German.

Louis Renou, who lived from 1896 to 1966, translated hymns across seventeen volumes of Études védiques et pāṇinéennes, published by E. de Boccard between 1955 and 1969, the last volumes posthumously. He covered 619 of the 1,028 hymns, roughly sixty per cent, with commentary of extraordinary density. The result is in French, arranged by deity rather than by mandala, and effectively unusable as a reading text.

H. D. Velankar produced partial English translations in the late 1940s through the early 1960s which Jamison and Brereton describe as a decided improvement on Griffith, but they were partial. Wendy Doniger O’Flaherty’s Penguin volume of 1981 gave English readers 108 hymns, about a tenth of the collection, in prose that finally sounded like something a person might have said. It remains the most widely read literary selection and it is, by design, a selection.

Meanwhile the infrastructure kept improving. Maurice Bloomfield’s Vedic Concordance of 1906 indexed every line of the Vedic corpus. Barend van Nooten and Gary Holland produced a metrically restored text in 1994 as volume 50 of the Harvard Oriental Series, undoing the sandhi and vowel contractions of the transmitted text to recover the syllable counts the poets actually used. The VedaWeb project at Cologne now serves a lemmatised, morphologically annotated Rigveda with parallel translations, searchable by grammatical category, which means a question that took Grassmann years of card-index labour now takes a query.

Then in 2014 Oxford University Press published Stephanie Jamison and Joel Brereton’s three-volume The Rigveda, 1,728 pages, described by the publisher as “the first complete scholarly translation into English in over a century.” The gap it closed was from Griffith’s second edition of 1896 to 2014: one hundred and eighteen years in which no complete scholarly English Rigveda existed.

What Survives Translation

There is a hymn in the tenth mandala about exactly this problem. RV 10.71 is addressed to speech itself, and it describes what happens when people who share a language do not share an understanding:

One who has looked has not seen speech; another who listens does not hear her. But to another she has laid bare her body, as a wife, finely dressed, does to her husband.

(RV 10.71.4, after Jamison and Brereton)

A few verses later the same poet sorts a priestly circle by what each member can actually do:

One sits making the blossoming of verses. One sings a melody in the sakvari measures. One, the expounder, states the knowledge of what is. One lays out the measure of the sacrifice.

(RV 10.71.11, after Jamison and Brereton)

The poet is talking about ritual competence, about which members of a circle actually grasp what they are all reciting. But the passage turns out to describe its own reception history with some precision. Voltaire looked and did not see. Wilson listened to Sāyaṇa and heard Sāyaṇa. Roth insisted on looking past him and sometimes looked past the ritual too. Griffith saw enough to know which verses to hide.

The useful lesson is not that the earlier translators were foolish. It is that a translation is a dated instrument, like a map or a table of astronomical constants, and that the honest ones say so. Rosen’s Latin faced the Sanskrit so you could check him. Grassmann’s dictionary gave you every occurrence so you could argue with him. Jamison and Brereton print the disagreements in their online commentary rather than resolving them silently. What all of these share is the assumption that the reader might want to look.

The version that dominates the internet does not share it. Griffith’s Latin verses are the perfect emblem: at the exact point where the text was most likely to surprise a reader, the translation stopped being a window and became a curtain, and it did so without saying that it had. Most people who meet the Rigveda in English meet it through that curtain, not because anyone chose it, but because when the web was being stocked it was the only complete English Rigveda a site could legally post.

The scholarship moved on in 1951, and in 1969, and in 2014. The default did not.