Evidence page · onomastics
Parse the names and the roster splits in two. One group are transparent Sanskrit descriptions — bad grain, the drier, the obstructor, the one who does not let go. The other group have no Sanskrit etymology at all. Only the second group holds forts.
| Name | Parses as | What the text says |
|---|---|---|
| Kúyava | ku- "bad" + yáva "barley" = BAD GRAIN | His two wives bathe in milk and are to be drowned (1.104.3). Named beside Śuṣṇa in the harvest verses |
| Śúṣṇa | √śuṣ "to dry" = THE DRIER, the scorcher | Griffith renders the epithet as "foe of harvest," "bane of crops" (2.19.6, 4.16.12, 6.31.3). Killed for Kutsa, in a fight for cattle |
| Vṛtrá | √vṛ "to cover, obstruct" = THE OBSTRUCTOR | Blocks the waters. Released, not defeated |
| Vala | √vṛ = THE ENCLOSURE | One occurrence in the whole Rigveda. A cave, not a person |
| Námuci | na-muc = "NOT LETTING GO" | The one who will not release. Beheaded four times over |
| Vyàṃsa | vi-aṃsa = "SHOULDERLESS" | Smashes Indra's jaws before being killed (4.18.9) — the god loses the first round |
| Dhúni | √dhvan/dhū = THE ROARER | Associated with floodwater and with the Yadu clan |
Bad Grain. The Drier. The Obstructor. The Enclosure. Not-Letting-Go.
Every one is a common Sanskrit word doing the job of a name, and every one names a condition rather than a people: failed harvest, drought, blocked water, withheld release.
You cannot be conquered by drought. You can only survive it — which is what the hymns ask Indra for.
So a substantial part of the enemy roster is personified agricultural catastrophe, and reading it as a record of conquered peoples over-reads it. That correction runs against this platform's own preference and it is the correct one.
| Name | Etymology | What they have |
|---|---|---|
| Śámbara | OPAQUE. No accepted Sanskrit derivation | Ninety-nine forts, of metal. Called a Dāsa. His forts are given to Divodāsa |
| Pípru | OPAQUE. Reduplicated shape, no Sanskrit parse | Called māyín — possessor of magic — and dāsa. Holds forts. Handed to Ṛjiśvan |
| Cúmuri | OPAQUE. The least Sanskrit-looking name in the corpus | Called Dasyu. Killed asleep, for Dabhīti |
| Arbudá | Uncertain | Trodden underfoot; head cleft |
| Varcín | varcas "lustre" + -in — semi-transparent | A hundred thousand sons. Explicitly Dāsa Varcin |
The names that parse as Sanskrit describe natural disasters. The names that do not parse hold fortified settlements, possess magic, have wives and sons, and are called Dāsa.
Śambara, Pipru and Cumuri are the three strongest candidates for real foreign names in the Rigveda — precisely because Sanskrit cannot account for them. An opaque name in a corpus that names everything else transparently is evidence of an outside source.
The Bharata king for whom Śambara's ninety-nine forts are destroyed is called Divodāsa — "dāsa of heaven." And the great Bharata king of Book 7 is Sudās — "good dās."
The conventional reading takes these from dāś, "to worship, to give" — so "worshipper of heaven" and "good giver" — and that derivation is sound.
But the homophony is exact, and it is not a small thing. The word applied to the enemy sits inside the name of the king the enemy is destroyed for. Whether that reflects a shared root, an older neutral sense of dāsa as simply "man/people" — matching Old Persian dahyu — or coincidence, it means the word was not always and only a slur.
Every figure the hymns give for enemies killed or forts taken. Log scale. These are praise-poetry claims and there is no independent check on any of them.
A hundred thousand of Varcin's. Sixty thousand at once. Fifty thousand "swarthy." Thirty thousand Dāsas put to sleep.
These cannot be literal. No Bronze Age polity in the region fielded such numbers, and the figures are the standard inflation of commissioned praise poetry — the same genre that reports sixty thousand cattle paid to a single poet.
What they do establish is the frame. The composers thought in terms of populations destroyed and settlements counted, and considered both fit subjects for celebration. The number ninety-nine forts recurs across separate hymns for separate patrons — it is a formula, which is itself informative: fort-taking was the standard achievement the genre existed to record.
| 2.15.9 | With sleep whelming Cumuri and Dhuni, thou slewest the Dasyu, keptest safe Dabhīti |
| 6.20.13 | Thou sentest Dhuni and Cumuri to sleep and slumber. Dabhīti lit the flame for thee and worshipped with fuel, hymns, pressed Soma |
| 6.26.6 | Thou sentest Cumuri to his sleep, to please Dabhīti — and slewest at once the sixty thousand |
| 7.19.4 | Thou sentest in swift death to sleep the Dasyu, both Cumuri and Dhuni, for Dabhīti |
| 4.30.21 | The thirty thousand Dāsas he with magic power and weapons sent to slumber, for Dabhīti's sake |
Dabhīti appears in every one. And 6.20.13 states the transaction outright: Indra sends the enemy to sleep, and Dabhīti lights the fire, brings fuel, presses soma and sings. Service rendered, payment made, in a single verse.
Whether "sleep" is a euphemism for death, a description of a night attack, or a magical incapacitation, the text does not say — and it does not need to, because there is no discomfort in it to resolve. The method is stated as a deed worth praising.
On whether this is a wider Indo-European motif: I don't know, and this page does not assert a parallel I have not checked. Sleep-slaying occurs in Norse and Irish material, but establishing a shared inheritance requires more than a shared plot device.
Pipru: 11 mentions. Cumuri: 5. Varcin: 4. Arbuda: 2. Vala: 1. And every one of them falls to 0% in the Popular stratum, the youngest layer.
They are not in the epics. They are not in the Purāṇas. They have no temples, no festivals, no descendants, no lineages claiming them. Ṛjiśvan and Dabhīti — the patrons who paid for the killings — are equally gone.
The named enemies belong to the world of forts. When the villages appear, the enemies stop having names.
What replaces them is a category. Rakṣas — the generic demon — has its heaviest concentration at 25% Popular, its largest share of any stratum. The corpus moves from named opponents in named conflicts with named beneficiaries to an abstract class of evil.
That is the moment the material stops being a record and becomes usable as mythology. A conflict with names, numbers and a paying client is a historical claim that can be checked. Rakṣas cannot be checked, which is why it survived and Cumuri did not.