Evidence page · corpus, checkable
Not "rarely." Not "mostly late." Three of its four terms occur one time each, in one verse, in the latest book — and the word for caste as actually lived is not in the corpus at all.
All three singletons are in the same verse: RV 10.90.12, the Puruṣa Sūkta. Brahmin from his mouth, Rājanya from his arms, Vaiśya from his thighs, Śūdra from his feet.
| Term | Sense | Total | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| bráhman | the sacred formulation | 254 | 44 | 42 | 11 | 14 | 10 | 22 | 44 | 25 | 8 | 34 |
| brahmán | the priest — a JOB | 51 | 7 | 4 | 0 | 4 | 4 | 1 | 3 | 11 | 3 | 14 |
| kṣatrá | power, dominion | 45 | 12 | 0 | 2 | 3 | 8 | 5 | 6 | 6 | 0 | 3 |
| várṇa | colour, appearance | 23 | 6 | 6 | 2 | 1 | 0 | 0 | 0 | 0 | 6 | 2 |
| brāhmaṇá | the varna | 12 | 1 | 0 | 0 | 0 | 0 | 0 | 3 | 1 | 0 | 7 |
| kṣatríya | of the ruling class | 9 | 0 | 0 | 0 | 2 | 1 | 0 | 2 | 2 | 0 | 2 |
| vaíśya | the third varna | 1 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 |
| śūdrá | the fourth varna | 1 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 |
| rājanyà | the second varna | 1 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 |
| jā́ti | caste, as lived | 0 | ABSENT FROM THE ENTIRE CORPUS | |||||||||
The two terms that do appear throughout are functional, not hereditary. bráhman is the sacred utterance. brahmán is the man who makes it — a role in a ritual, like hotṛ or udgātṛ. kṣatrá is power, not a class of person.
The terms for a birth-order — brāhmaṇá as a varna, and vaíśya, śūdrá, rājanyà — are late and rare, concentrated in Book 10, and three of them exist only inside the single verse that creates the scheme.
And jāti — the endogamous birth-group that caste actually is, in life, for everyone who has ever lived under it — is not in the Rigveda. Not once.
The scriptural charter of the caste system is one verse of one late hymn, and it introduces three words the corpus never uses again.
RV 9.112 — occupations, within one family
RV 9.112.3. In Griffith's rendering: a bard am I, my dad's a leech, mammy lays corn upon the stones — and all of us, with different plans, chase wealth like cattle after fodder.
Three occupations, three people, one household. Poet, physician, grain-grinder. The hymn's whole point across its verses is that people do different work and want different things — the smith seeks a customer with gold, the physician seeks the injured, the priest seeks a worshipper.
Occupation is not hereditary in this hymn. It is not even consistent within a single family. And the hymn treats this as ordinary — the observation is made in passing, as an analogy for how Soma flows to different ends.
Set beside this the other flexibility cases the tradition itself preserves: Viśvāmitra, a Kṣatriya who becomes a brahmarṣi and composes Book 3 — including the Gāyatrī still recited daily. Kavaṣa Ailūṣa, said to be the son of a slave woman, rejected by other priests and then vindicated. Satyakāma Jābāla, accepted as a student because he told the truth about not knowing his father.
In the earliest layer: by function. In the latest layer of the Rigveda: the birth-scheme appears, once. In the Dharmasūtras and Manu, centuries later: by birth, enforced, with penalties.
The tradition contains its own counter-evidence and never deleted it. Viśvāmitra's hymns were not removed when Viśvāmitra's mobility became doctrinally inconvenient.
The Bṛhadāraṇyaka / Chāndogya check — resolved
The five-fires doctrine and the two paths are preserved independently by two Vedic schools: the Chāndogya, belonging to a Sāmavedic lineage, and the Bṛhadāraṇyaka, belonging to the White Yajurveda. Two separate chains of custody, both carrying the same teaching, both attributing it to the Kṣatriya king Pravāhaṇa Jaivali.
Only one of them attaches caste to it.
| Term in the varna-womb passage | Chāndogya | Bṛhadāraṇyaka |
|---|---|---|
| ramaṇīya-caraṇa — pleasant conduct | 6 | 2* |
| kapūya-caraṇa — stinking conduct | 2 | 0 |
| brāhmaṇa-yoni | 2 | 0 |
| kṣatriya-yoni | 1 | 0 |
| vaiśya-yoni | 2 | 0 |
| śva-yoni — a dog's womb | 2 | 0 |
| caṇḍāla-yoni | 2 | 0 |
* the two apparent Bṛhadāraṇyaka hits are in Śaṅkara's commentary, which the standard digital file bundles with the text. In the Upaniṣad itself the count for every term above is zero.
The standard digital Bṛhadāraṇyaka file is 85,467 words. The Upaniṣad is about 17,000. The rest is Śaṅkara's eighth-century commentary, bundled with the text in the same file.
The Bṛhadāraṇyaka is also the fourteenth kāṇḍa of the Śatapatha Brāhmaṇa, which circulates without commentary. Re-running the check on that:
ramaṇīya 0 · kapūya 0 · caṇḍāla 0 · śvayoni 0 · sūkara 0 · brāhmaṇayoni 0 · kṣatriyayoni 0 · vaiśyayoni 0
Zero across every term, in the uncontaminated text.
And the two-paths passage is present and complete, at ŚB 14.9.1.18–19. Those who know go by the flame — day, waxing fortnight, the northward sun, the world of the gods, the sun, lightning, and "for them there is no return here." Those who win worlds by sacrifice and giving go by the smoke, reach the moon, become food, and come back down through space, wind, rain and earth.
And the third place, in full: atha ya etau panthānau na viduste kīṭāḥ pataṅgā yadidaṃ dandaśūkam — "those who do not know these two paths become insects, moths, and biting things."
Insects, moths and biting things. That is the whole of the third destination. No wombs, no ranks, no Caṇḍāla.
The Chāndogya's version of the same doctrine, from the same teacher, sorts that third group by birth: a brahmin, kṣatriya or vaiśya womb for pleasant conduct; a dog's, a pig's or a Caṇḍāla's for stinking conduct.
Two recensions. One doctrine. One of them added the sorting.
ChU 5.10.7, in the Sanskrit: those of pleasant conduct reach a pleasant womb — brāhmaṇayoniṃ vā kṣatriyayoniṃ vā vaiśyayoniṃ vā. Those of stinking conduct reach a stinking womb — śvayonim vā sūkarayoniṃ vā caṇḍālayoniṃ vā: a dog's, a pig's, or a Caṇḍāla's.
The Bṛhadāraṇyaka's version of the same doctrine has none of this. Its two paths end at the third place — the small creatures, born and dying — with no womb hierarchy at all.
The doctrine is shared by two schools. The caste is one school's addition.
This matters because ChU 5.10.7 is the earliest clear statement that birth-rank is earned by prior conduct — the machinery that makes caste deserved rather than merely inherited. If it were integral to the five-fires teaching, both custodians would carry it. One does not.
And the caste terms are absent from the Chāndogya's own source claim. Pravāhaṇa tells Uddālaka that this knowledge never passed to a Brahmin. The passage the Brahmins added to it is the one that sorts human beings by womb.
Yes, and it is mainstream scholarship, not speculation
V. S. Sukthankar, editor of the Mahābhārata Critical Edition, argued in 1936 that the Bhṛgu clan of brahmins inserted large quantities of Bhārgava material — Paraśurāma, Cyavana, Jamadagni — into what had been a Kṣatriya war epic. Robert Goldman developed the argument at length in Gods, Priests and Warriors (1977).
The epic states its own growth: 8,800 verses → 24,000 → roughly 100,000. It does not hide that it was expanded; it advertises it.
And the didactic books are where the caste material lives. The degradation lists naming Drāviḍas, Kaliṅgas and Pulindas as fallen Kṣatriyas sit in the Anuśāsana Parva — a book that is almost entirely brahmanical instruction attached to a story about a war.
The Kurus and Kauravas are Kṣatriyas, not brahmins — Kuru is the dynasty, and both the Pāṇḍavas and Duryodhana's brothers descend from it. So the epic is a Kṣatriya family story, narrated by a Sūta, redacted by brahmins. Three parties, and only one of them was doing the editing.
Brahmins are the only group in the epic who are both outside the fighting and indispensable to it. Droṇa, a brahmin, teaches every prince to fight and commands an army. Paraśurāma, a brahmin, exterminates the Kṣatriya order twenty-one times and trains Bhīṣma, Droṇa and Karṇa. Kṛpa, a brahmin, survives the whole war.
A brahmin teaching weapons violates varna on its face. The likeliest reading is not that the tradition remembers a fluid past — it is that the redactors placed brahmins at the origin of the martial skill the story is about.
The insults come from Kṣatriyas. Bhīma and Duryodhana's circle call Karṇa sūtaputra. Not brahmins.
But the category is brahmanical. Sūta is defined in Manu as Kṣatriya father, Brahmin mother — pratiloma, against the grain. The offence in the definition is a higher-status woman marrying down.
So: the taxonomy is written by brahmins; the enforcement is by whoever gains from it locally. Kṣatriyas police the ranking among Kṣatriyas because that is where their own standing is decided. That pattern is the general one — a hierarchy is maintained not by its authors but by everyone with something immediately above them to defend.
And it is the most Manu-like passage in the Gītā: the men die, the women are corrupted, varṇasaṃkara follows, the ancestors fall for want of rice and water.
There is no manuscript evidence of interpolation, and this page will not claim any. The Critical Edition does not excise it.
What is observable is what happens next. Krishna's reply does not address the caste argument. He goes to the imperishability of the self, then to duty, then to devotion. The stated reason Arjuna gives for refusing to fight is answered by changing the subject — and by 3.24 the premise is affirmed in passing rather than argued for.
Five verses raise it. Nothing in the text defends it. And it became the part that mattered.