Evidence page · corpus-verified

The fifth varṇa entered Sanskrit as one man's guess about an ambiguous phrase.

The Rigveda refers to pañca janāḥ — "the five peoples" — fourteen times, and never once says who they are. Centuries later, in the oldest surviving work of Sanskrit etymology, Yāska records two competing explanations. One is a list of mythological beings. The other is the four varṇas with the Niṣāda added as a fifth — and Yāska attributes it, by name, to a single earlier commentator. The etymologists then supply a meaning for the word: "evil is settled in him." This page shows the passage.

What the Rigveda actually says.

fourteen occurrences, no definition

Machine-counted in the Rigveda: the phrase "five peoples" and its derivatives occur 14 times — pañca janāḥ, pañcajanā, pāñcajanya. They are invoked, addressed, and praised. They are never enumerated.

Reference	What it says
RV 1.89.10	Aditi is the sky, the mother, the father, the son; all the gods; the five peoples.
RV 6.11.4	The five peoples anoint him with reverence, bringing oblations.
RV 6.51.10	May Indra, Earth, Pūṣan, Bhaga, Aditi and the five peoples strengthen us.
RV 8.32.21	Come past the three distances, come past the five peoples.
RV 10.53.4	"Five peoples, take pleasure in my offering." — the verse Yāska glosses.

The corpus test

If the five peoples were the four varṇas plus one, the Rigveda might be expected to say so somewhere. It does not. Across the whole text, pañca and varṇa co-occur within a 600-character window exactly once — at RV 2.34.13, in a passage about a shining colour, with no social sense at all. The identification of the five peoples with the varṇa scheme is nowhere in the Rigveda.

Where the fifth comes from.

Yāska, Nirukta 3.8

Yāska's Nirukta is the oldest surviving Sanskrit work of etymology, and here it is doing exactly what it exists to do: explaining a phrase whose meaning was already unclear to its own tradition. Glossing pañca janā mama hotraṁ juṣadhvam from RV 10.53.4, Yāska reports what others have said.

gandharvāḥ pitaro devā asurā rakṣāṃsi ity eke.
catvāro varṇā niṣādaḥ pañcama ity aupamanyavaḥ.
niṣādaḥ kasmāt? niṣadano bhavati, niṣannam asmin pāpakam iti nairuktāḥ.

"Gandharvas, ancestors, gods, asuras, rakṣases — so say some.
The four varṇas with the Niṣāda as fifth — so says Aupamanyava.
Why 'Niṣāda'? He is a settler; evil is settled in him — so say the etymologists."

Nirukta 3.8, glossing Rigveda 10.53.4

Read the structure of that passage rather than its content. Yāska is not stating doctrine. He is reporting a disagreement. One reading makes the five peoples supernatural classes — gandharvas, ancestors, gods, asuras, demons. The other makes them the social order. The second is attributed to Aupamanyava, a named earlier authority, with the marker iti — "so says."

Not doctrine. A disagreement, with the social reading credited to one man by name.

Then the etymologists supply the insult.

niṣannam asmin pāpakam

Having introduced the Niṣāda as a candidate fifth class, the passage immediately asks why the word means what it does — and the answer given is not historical or linguistic in any modern sense. It is a pun.

The root is ni-ṣad, "to sit down, to settle." The nairuktas — the etymologists — derive niṣāda from it twice over: he is niṣadana, "one who settles," and niṣannam asmin pāpakam — "evil is settled in him."

What has just happened in three lines

An ambiguous Rigvedic phrase is given a social reading on one commentator's authority. A community's name is attached to it as the fifth term. And the name is then explained by a wordplay that locates evil inside the people it names — as a property of who they are, not of anything they have done.

This platform has documented the same operation in the capture of ordinary words and in manufactured genealogies. Here the mechanism is visible at its smallest scale and earliest date: a folk etymology doing the work of a moral judgement.

The word is not in the Rigveda.

a false positive, checked and discarded

A raw search for niṣāda in the Rigveda returns one hit, at RV 3.19.5. Read in context, it is not the noun:

…yat tvā hotāram anajan miyedhe niṣādayanto yajathāya devāḥ…

"…when the gods anointed you as hotṛ at the sacrifice, seating you for the offering…"

Rigveda 3.19.5 — niṣādayanto is the causative participle of ni-ṣad, "seating." Not a people.

The social term Niṣāda does not occur in the Rigveda at all. The word that Yāska's source proposes as the fifth of the five peoples is absent from the text those five peoples appear in.

Where the term actually lives.

354 occurrences across 758 texts

Location	Count	What kind of text
Rigveda	0	The single apparent hit is the verb "seating."
Yāska, Nirukta	2	Both in the passage above.
Mīmāṃsāsūtra with commentary	19	Ritual-eligibility law — whether a niṣāda-sthapati may have a sacrifice performed for him.
Brahmasūtra commentaries	32	The adhikāra debate: who is entitled to the Veda.
Rāmāyaṇa (three recensions)	41	Narrative — Guha the Niṣāda chief, and others.
Manusmṛti	12	Mixed-caste origin and prescribed occupation.
Purāṇas	62	Origin myths.
Music-theory texts	15	Niṣāda is also the seventh note of the scale — 4% of all occurrences.

The distribution tells its own story. The term is densest in the literature that adjudicates entitlement — Mīmāṃsā, Brahmasūtra commentary, the law books. It appears where someone is deciding what this category of person may and may not do. It is a term of administration before it is a term of description.

Limits.

what is and is not claimed

Machine-verified: the 14 pañca janāḥ occurrences; the single pañca/varṇa co-occurrence; zero social niṣāda in the Rigveda; the 354 occurrences across 758 texts and their distribution. Corpus and method below.
NOT claimed: that Aupamanyava invented the fifth-varṇa idea. He is the earliest named attribution surviving in this passage; the idea may be older, and the Nirukta reports rather than originates.
NOT claimed: that Niṣāda communities did not exist before the term is attested. Absence from a text is a detection limit, not evidence of absence — and the record here was kept by those doing the classifying.
NOT claimed: a date for the Nirukta. Its dating is contested and nothing here depends on it; the argument is about the structure of the passage, which is the same whenever it was composed.
Translation: the renderings above are literal and the Sanskrit is given so a reader can check them. Niṣannam asmin pāpakam is the phrase carrying the weight, and it is short enough to verify directly.
On the counts: 4% of niṣāda hits are the musical note. Those are identified by source text and excluded from the social figures above.

Reproduce this

Corpus: GRETIL plain-text mirror tokushige-koyasan/gretil-corpus, 784 files; 758 scanned above 200 tokens; 18,777,671 tokens.

Query: NFC-normalised regex ni[ṣs]ād[aā] across all files with per-file counts; pañca\s*jan|pañcajan|pāñcajany in 1_veda/1_sam/sa_Rgveda-edAufrecht.txt; co-occurrence of pañca and varṇa within ±300 characters. Nirukta passage located in 1_veda/5_vedang/sa_yAska-nirukta.txt.

Gap: no lemmatisation, so inflected forms outside the pattern are missed and the counts are a floor. The musical-note sense is separated by source text rather than per-hit reading. The Śaunaka Atharvaveda is absent from GRETIL entirely.