Evidence page · measured, not asserted
The honest answer is: much less than the argument usually assumes — and that makes the real finding stronger, not weaker.
red: hostile-outsider terms · green: praise and request terms · grey: everything else
It is liturgy. Overwhelmingly, it asks named gods for cattle, victory, sons, rain, long life and wealth, and praises them in order to ask. Deity names alone account for more of the corpus than every hostile term combined, by more than an order of magnitude.
| Category | Tokens | % of corpus | Hymns containing it |
|---|---|---|---|
| Named deities | 8,388 | 5.1% | 981 of 1,028 — 95% |
| Praise, request, wealth, fame | 5,421 | 3.3% | 956 — 93% |
| Hostile-outsider vocabulary | ~500 | ~0.3% | 264 — 26% |
An argument that the Rigveda is a document of hatred dies the moment anyone opens it. It is a book of hymns, and an inflated number hands away the easiest possible refutation.
The measured finding is harder to dismiss and worse in its implications — and the reason is in the next section.
Where the hostility is concentrated, and what happened to it
The hostile vocabulary is rare, unevenly distributed, and load-bearing. Two hymns carry a large share of it; most of the corpus has none.
| Term | Occurrences | Hymns |
|---|---|---|
| dāsa (both lemmas) | ~49 | across the family books, Archaic-weighted |
| dásyu | 67 | 58 |
| rákṣas | 80 | 52 |
| paṇí | 53 | 44 |
| ádeva | 31 | 28 |
| brahmadvíṣ | 13 | 12 |
| áyajvan | 10 | 6 |
| mṛdhrávāc | 6 | 6 |
| śiśnadevá | 2 | 2 |
| kīkaṭa · śígru · vaikarṇá · piśā́ci | 1 each | 1 each |
Four of the most cited "ethnic" terms in the Rigveda occur once. Once each, in the whole book.
A handful of words, in a fraction of the hymns, in a corpus that is otherwise a request for cattle — and from them a legal order was built that graded human beings by birth for two thousand years and still does.
The disproportion between the textual base and the social consequence is the finding. Not volume of hatred. Leverage. The Puruṣa Sūkta is one hymn, late, in Book 10. It became the charter of a hierarchy governing hundreds of millions.
Everything that made caste operative — the pratiloma taxonomy, the toponym-to-caste lists, the degradation-through-non-observance clause, the exclusion of Śūdras from the Veda — is later, in the Dharmasūtras, in Manu, in the commentaries. The Rigveda supplied the vocabulary. The law code supplied the machinery. Blaming the hymns overstates their contents and understates what was done with them.
The comparison that matters is between humans, not against the whole corpus
A book of hymns praises its gods; counting that proves nothing. The question is how humans are spoken about — who is the "we," who is the "they," and where the affection goes.
| In-group and affection | n | Out-group and contempt | n |
|---|---|---|---|
| sákhi / sakhyá — friend, friendship | 425 | dásyu | 68 |
| priyá — dear | 237 | dā́sa / dāsá and compounds | 96 |
| jána — people, folk | 327 | paṇí | 50 |
| átithi — GUEST | 45 | ádeva / ádevayu | 40 |
| jāmí / sajātá — kinsman | 45 | árāvan — ungenerous | 14 |
| sumánas — well-disposed | 32 | brahmadvíṣ — hater of the sacred word | 13 |
| bándhu — kin | 9 | áyajvan — non-sacrificer | 10 |
| ā́rya — the self-designation | 38 | anyávrata — of other observances | 3 |
They name the enemy four and a half times more often than they name themselves.
ā́rya — the word an entire racial theory was built on — appears 38 times in 1,028 hymns. jána, the ordinary word for "people," appears 327 times. These are not people preoccupied with an ethnic label for themselves.
They are preoccupied with an enemy. The identity is not asserted; it is constructed negatively, out of who is being fought, robbed of cattle, or excluded from the sacrifice. Every out-group term is either an ethnonym-turned-slur (dāsa, dásyu, paṇí) or a privative — a-deva, a-yajvan, anya-vrata: godless, non-sacrificing, of other observances. The other is defined by lack.
The direction matters more than the volume
Affection vocabulary outweighs hostility roughly three to one. That is real and should be stated. But the two run on different axes.
| Direction | Vocabulary | Directed at |
|---|---|---|
| Vertical, upward | sákhi, priyá, sumánas | Gods — Indra is repeatedly the friend — and human patrons who pay (sūrí, maghávan) |
| Horizontal, inward | jána, víś, jāmí, bándhu | Own clan and settlement |
| Horizontal, outward | dásyu, dāsa, paṇí, ádeva | Other humans |
The love is vertical and internal. The hostility is horizontal and external. There is abundant warmth in the Rigveda, and almost none of it crosses the boundary of the group.
The guest. 45 occurrences — more than ā́rya — and it is an inherited Indo-European institution of obligatory hospitality, with cognates across the family.
But look at how it is used: Agni is the guest. The god of fire is the honoured visitor in every household, and the hymns address him as such. The hospitality ethic is overwhelmingly theologised rather than applied to actual strangers, and there is no sustained ethic in the corpus of obligation toward a foreign human being as such.
The nearest thing is 10.117: the man who eats alone eats his sin; wealth rolls like chariot wheels, coming now to one and now to another. An unconditional ethic of sharing, without ritual qualification — and it sits in the latest book.
And why the satire reading is probably wrong
Frogs lie silent through the dry year and then, when Parjanya sends the rains, they lift their voices. The hymn compares them, at length, to brahmins fulfilling a vow; to priests sitting round the brimming soma vessel at the Atirātra rite; and — the sharpest line — to a pupil repeating his teacher's words back to him. It names them by their calls: Cow-bellow, Goat-bleat, Green, Spotty.
Max Müller read it as satire, and that reading is still repeated. The hymn's own ending does not support it. It closes by asking the frogs for riches, cattle in hundreds and long life — which is exactly what every other hymn asks of a god.
The likeliest reading is that it is a rain-charm, and that frogs and brahmins are put side by side because both are producers of efficacious sound that brings the rain. The comparison is functional, not mocking.
The wry note, if there is one, is verse 5 — one frog repeating the other's language exactly as a pupil repeats a teacher's lesson. Rote transmission, described in a hymn that is itself an artefact of rote transmission. Whether that was meant as a joke cannot be recovered.
The debrahminising reading is not that the Rigveda mocks brahmins. It is that this hymn is a village rain spell, and it sits in a family book, in the old layer — evidence that the corpus contains a great deal of ordinary agricultural religion that the later priestly framing has flattened.
The Nāsadīya's "who really knows?" is general epistemic scepticism, and belongs alongside the śramaṇa milieu of the same period rather than with any self-examination by the tradition.
Across the entire normative corpus — Rigveda through the Brāhmaṇas, Dharmasūtras and Manu — there is no passage in which the tradition asks whether ranking human beings by birth is just. The doubt reached the origin of the universe and stopped at the varna order.
Is anyone both praised and reviled? Yes — and the switch is always the same
Counting hostile words against friendly words says little until you ask whether they ever fall on the same people. They do, in both directions, and the variable that flips them is not ethnicity.
| Who | What the text says |
|---|---|
| Balbūtha explicitly called a dāsa RV 8.46.31–33 |
The poet receives a hundred from him. A hundred camels are made to bleat for the singer. And then: a woman, adorned with gold ornaments, is led forth to the poet as part of the same gift. The hymn closes by calling Balbūtha's people Vāyu's own. |
| Bṛbu chief of the Paṇis RV 6.45.31–33 |
Set above the Paṇis, over their highest head, like a wide bush on the bank of the Gaṅgā. Thousandfold bounty, swift as the wind. "All our singers ever praise the pious Bṛbu's noble deed." In Book 6 — the oldest book. |
Dāsa and Paṇí are the two most consistently reviled categories in the corpus. The only named individuals from either who are praised are the two who gave the poet a large gift.
And note what is in the gift. Alongside cattle and camels, a woman is transferred to the singer as an itemised part of the payment, in the same list, in the same breath. She is not named.
The reverse works too. árāvan is usually translated "godless" — but the sense the contexts carry is one who does not give.
| RV 9.21.5 | A request that Soma grant his desires to the man who has as yet given us nothing — the árāvan here is a prospective donor, not an enemy |
| RV 8.60.10 | Preserve us from every fiend who brings the gods no gift |
| RV 1.124.10 · 4.51.3 | Rouse the liberal givers; let the niggardly traffickers sleep on unwakened. Dawn is asked to shine on those who pay and to leave the rest in the dark |
| RV 5.34.7 | Indra plunders the niggard's goods and gives excellent wealth to him who offers gifts |
| RV 6.53.6 | Pūṣan is asked to drive an awl into the miser's heart and bend him to the poet's will |
Across 24 stanzas where praise and contempt appear together, the axis is consistently the same: the liberal giver against the withholder. Not Ārya against Dāsa. Giver against non-giver.
A Dāsa who gives a hundred cattle becomes "Vāyu's own people." A member of the group who gives nothing becomes a fiend, and a god is asked to put an awl through his heart.
The ethnic terms are real. The operative variable is whether goods are flowing toward the poet.
This cuts two ways and both matter. It weakens a purely racial reading of the Rigveda — the boundary is permeable, and payment crosses it in both directions. And it exposes what the boundary was for. A hierarchy that can be bought across is not fundamentally about descent. It became about descent later, when the law codes closed the door that a hundred cattle could once open.
RV 10.117, and it is worth reading against everything above. The gods did not ordain hunger as our death; death comes to the well-fed too. The man with food who hardens his heart against a beggar in misery finds no one to comfort him. He who gives to the one who comes in want succeeds and makes a friend for the future. No friend is he who offers nothing to a comrade who comes imploring food — let him depart and seek elsewhere; that is no home. Riches roll from one to another like the wheels of a chariot. And, pointedly: the speaking brahman is worth more than the silent one.
No ritual condition. No group boundary. No mention of who the hungry man is. The obligation is to whoever arrives hungry, and the reasoning is that fortune rotates and you will need it too.
It is one hymn, in Book 10, in the latest stratum. Set it beside the awl in the miser's heart in Book 6 and the distance is the whole argument: the corpus contains an unconditional ethic of sharing, arriving late, occupying a single hymn, and never becoming the basis of anything. What became the basis of things was the varna hymn, seventeen hymns earlier in the same book.
The dānastuti, and a hymn addressed to the fee itself
Neither Balbūtha nor Bṛbu is called a king. The text calls Balbūtha a dāsa and names Tarukṣa alongside him; Bṛbu is described as set above the Paṇis. Patron and pre-eminent figure, yes. Rājan, not stated. If an earlier page here implied otherwise, that is the correction.
And "local" is not established either. Dāsa has Iranian cognates — Avestan daŋhu-, Old Persian dahyu-, both meaning land or people, and the ethnonym of the Dahae. A dāsa may well be a rival Indo-Iranian group rather than a South Asian one already present. Balbūtha and Tarukṣa are both phonologically odd names. We do not know who they were.
Kakṣīvān records what he was paid by Svanaya Bhāvya, on the bank of the Sindhu: a hundred necklets. A hundred horses. A thousand cows. Ten chariots with mares to draw them. Sixty thousand cattle following after. Forty bays leading a procession of a thousand. And the reason is stated in the first verse: the patron desired glory.
śrávas — fame — occurs 145 times. vā́ja (prize, booty) 376. maghávan (bountiful) 370. gó (cattle) 543.
The poet's product is reputation. The patron's payment is livestock. The transaction is not hidden, implied or reconstructed by modern scholars — it is itemised, by the poet, in the hymn.
Not to a god. To dákṣiṇā, the sacrificial payment, personified and addressed. What it promises:
| 10.107.2 | Those who give horses dwell with the Sun forever. Those who give gold gain life eternal. Those who give robes prolong their lives |
| 10.107.3 | From the niggards no reward comes, and no satisfaction of the gods |
| 10.107.5 | The fee-bringer comes first among the invited, and is chief of the settlement. "Him I account the ruler of the people who was first to introduce the fee" |
| 10.107.8 | The liberal do not die. They are never ruined. They suffer neither harm nor trouble |
| 10.107.9–10 | The liberal gain a fragrant dwelling, a bride in fair apparel, drink, and victory over the unpaying. "The maid adorns herself and waits to meet him" |
A hymn promising immortality to whoever pays the priest, and describing the woman who will be waiting.
No, and the accurate answer is worse.
A con requires a deceived party. Nobody was deceived. The patrons knew exactly what they were buying, the poets stated the price, and the product — śravas, a name that outlives you — was real by the standards of a preliterate society where the only archive was a professional's memory. Kakṣīvān delivered: we are still saying Svanaya's name three thousand years later.
And it is an inherited Indo-European institution, not a South Asian invention. Greek poets sold kleos; Pindar was paid for victory odes and complains about it in them; Norse skalds were paid for praise-poems; Irish filid could ruin a king with a satire, and were paid partly not to. Professional praise-poetry with an explicit fee is a family trait.
The critique that lands is not about the payment. It is about what happened to it. A profession that in the Rigveda is openly transactional — with itemised fees, complaints about non-payment, and a hymn promising heaven to donors — later claimed hereditary purity, divine sanction, and the authority to bar others from the texts entirely.
The dānastuti and the apaśūdrādhikaraṇa are the same institution fifteen hundred years apart. One is honest about its fee schedule. The other calls access a birthright and argues that a Śūdra may not hear the Veda. The dishonesty is not in Book 8. It is in what Book 8 was later said to have been.
Griffith's 1896 translation renders two verses of RV 1.126 into Latin rather than English, because their content is sexually explicit. The standard English Rigveda that shaped a century of Anglophone understanding silently withheld material its translator found indecent — a reminder that even the transmission into English has a filter on it, and that the filter had a nationality and a century.
RV 10.61.5–7 — the passage Griffith put into Latin
RV 10.61.7, in the Sanskrit: pitā yat svāṁ duhitaram adhiṣkan kṣmayā retaḥ saṁjagmāno ni ṣiñcat — when the father mounted his own daughter, he poured out the seed upon the earth.
And the same verse continues: svādhyo 'janayan brahma devā vāstoṣ patiṁ vratapāṁ nir atakṣan — the gods, well-disposed, generated the sacred word (brahma) and fashioned Vāstoṣpati, guardian of the vow.
There is no condemnation of the father. The gods do not punish him. They respond by creating religious language and a deity to guard obligation.
The origin of sacred speech is the gods' answer to an incestuous rape. Not a verdict on it. A liturgy generated by it.
And Prajāpati remains in living ritual. He is invoked in the Vedic marriage rite and in garbhādhāna, the conception ceremony. RV 10.121, with its prajāpataye svāhā refrain, is in common liturgical use. The god invoked over a wedding and a conception is the god of 10.61.7 — and no liturgy anywhere cites that verse.
Griffith rendered 10.61.5 into Latin. He rendered 1.126.6–7 into Latin. He deleted 10.86.16–17 outright, writing that he could not put them into decent English. The standard English Rigveda that shaped a century of Anglophone understanding withheld all three.
Corpus: VedaWeb (Cologne), full Rigveda, 164,758 lemmatised tokens, CC-BY 4.0. Counts are of lemmas, not surface forms.
Known error: the normalisation used here strips accents, which merges distinct lemmas. It conflated ánas "cart" with anā́s "noseless" — producing 14 where the correct figure for the contested ethnic term is 1 — and double-counted the two dāsa lemmas. The category totals above are therefore approximate and, if anything, overstate hostility. The individually verified figures are the ones given in the term table.
Category membership is a judgement. Is rákṣas a demon or a people? Is ádeva an outsider or a bad neighbour? Someone else's list would give a different total. The lists are printed so they can be argued with.
And absence of a word is not absence of an attitude. A text can exclude without naming. This measures vocabulary, not ideology.
Mṛcchakaṭika is mṛd "clay" plus śakaṭikā "little cart" — "The Little Clay Cart." It is named after a toy in the play. mleccha is from an entirely different root, meaning to speak indistinctly.
A resemblance, not a derivation — the same trap catalogued on the sound-change page, and it applies here regardless of which conclusion it would have supported.
Three passages, and the three different methods used on them
āgadhitā parigadhitā yā kaśīkeva jaṅgahe
dadāti mahyaṁ yādurī yāśūnām bhojyā śatā
upopa me parā mṛśa mā me dabhrāṇi manyathāḥ
sarvāham asmi romaśā gandhārīṇām ivāvikā
Verse 7, in the woman's own voice: "Come close, press against me — do not think me lacking. I am hairy all over, like a ewe of the Gandhāris."
Romaśā means hairy. She is telling her husband her body is that of a grown woman, not the girl he apparently took her for. She is correcting him and inviting him, in the same breath, in a hymn of the Rigveda.
A woman in the oldest text in the corpus asserts her own adulthood and her own desire — and English readers were given Latin.
na seśe yasya rambate 'ntarā sakthyā kapṛt
sed īśe yasya romaśaṁ niṣeduṣo vijṛmbhate viśvasmād indra uttaraḥ
na seśe yasya romaśaṁ niṣeduṣo vijṛmbhate
sed īśe yasya rambate 'ntarā sakthyā kapṛd viśvasmād indra uttaraḥ
"He has no power whose member hangs slack between his thighs. He has power in whom, as he sits, the hairy thing swells. Indra is supreme above all."
And verse 17 is the same two clauses, exactly reversed. A matched pair, spoken by Indrāṇī, in a dialogue hymn between her, Indra and the ape Vṛṣākapi.
Griffith printed neither, and said he could not render them into decent English. The refrain viśvasmād indra uttaraḥ runs unbroken through every other verse of the hymn — so the omission is visible as a hole with the frame still around it.
madhyā yat kartvam abhavad abhīke
kāmaṁ kṛṇvāne pitari yuvatyām
pitā yat svāṁ duhitaram adhiṣkan kṣmayā retaḥ saṁjagmāno ni ṣiñcat
svādhyo 'janayan brahma devā vāstoṣ patiṁ vratapāṁ nir atakṣan
"When the father mounted his own daughter, uniting with her he poured his seed upon the earth. The well-disposed gods brought forth sacred speech, and fashioned Vāstoṣpati, guardian of the vow."
The censorship is the smaller point. The larger one is what the verse does next.
There is no punishment. There is no condemnation. The gods respond to the act by producing brahman — sacred speech itself — and by fashioning the guardian of the vow. The act is the occasion of a creation.
The judgement is added later. Śatapatha Brāhmaṇa 1.7.4 retells the same episode, calls it āgas — an offence — and has Rudra shoot Prajāpati for it. The Rigveda has the act. The Brāhmaṇa supplies the archer.
| Passage | Method | What was being hidden |
|---|---|---|
| 1.126.6–7 | Latin | A woman's statement of her own body and desire |
| 10.86.16–17 | Deletion | Explicit male anatomy, in a goddess's mouth |
| 10.61.5–7 | Latin | Incest — and the absence of any consequence |
Griffith's own notes are the evidence. He records that he is withholding, which is why the withholding is recoverable. A translator who silently paraphrased would have left no trace.
The effect on the reception is direct. For a century the standard English Rigveda presented a corpus with no explicit sexuality, no female desire, and no unpunished transgression by a god — and that sanitised text is the one on which a great deal of popular understanding was built.
The Victorian Rigveda is a different book from the Rigveda, and the difference runs in one direction.