Around 300 BCE, Megasthenes, a Greek ambassador sent by Seleucus I to the court of Chandragupta Maurya, wrote an account of India — the Indika. The book itself is lost, but substantial fragments survive in later Greek writers, and they contain the oldest description of Indian caste by an outsider. It is worth reading for two opposite reasons: what it confirms, and what it gets wrong.
What he saw
McCrindle’s public-domain translation preserves the core passage:
The whole population of India is divided into seven castes, of which the first is formed by the collective body of the Philosophers… The second caste consists of the Husbandmen, who appear to be far more numerous than the others.
He goes on through Herdsmen, Artisans, the Military, the Overseers, and the Councillors — seven groups. And crucially, he records the two features that make it caste rather than mere occupation:
No one is allowed… to exercise any calling or art except his own: for instance, a soldier cannot become a husbandman, or an artizan a philosopher.
[Documented] Hereditary occupation, and no crossing between groups (he notes elsewhere that none may marry outside his own caste). An independent foreign witness, around 300 BCE, describing the fixed, birth-locked social order — more than two thousand years before British rule. On the antiquity of caste, Megasthenes and Al-Biruni agree across a gulf of thirteen centuries.
The warning built into the source
But this witness carries his own caution, and this site is obliged to flag it rather than just bank the confirmation.
The “seven” is almost certainly wrong. India’s classical order had four varnas, not seven — and Megasthenes’ own translator notices the problem. In a footnote McCrindle observes that Herodotus had earlier divided the people of Egypt into seven castes, and that Megasthenes “may therefore have taken it for granted that there were seven castes in India.” [Debated] In other words, the outsider very likely imposed a Greek template onto what he saw — counting India through Herodotus’s Egypt rather than reporting it plainly.
The lesson generalises. Outside observers document caste and distort it, filtering an unfamiliar order through categories they brought with them. It is exactly what the colonial census would do on a vastly larger scale two millennia later — and it is why this site treats even its own favourite foreign witnesses (Megasthenes, Al-Biruni) as evidence to be weighed, not gospel.
The tell that proves the point: the slavery he wished away
There is a sharper error, and it is decisive. Megasthenes reports, admiringly, that in India:
the law ordains that no one among them shall, under any circumstances, be a slave.
This is simply false — and we can prove it from an indigenous source of the same era. Kautilya’s Arthaśāstra, the Mauryan manual of statecraft, legislates slavery in detail, grading the very fines for enslaving each varna. So the Greek visitor, idealising a distant civilisation, imagined away an institution that his hosts’ own law book carefully regulated.
[Documented] That contradiction is the whole method of this site in miniature. A foreign account said “no slaves”; the local statute book said otherwise; the statute book wins, because it is closer to the ground. Read your outside witnesses for what they independently confirm — here, the antiquity and rigidity of caste — but check them against the record made from the inside, and never let the flattering version stand when a harder source contradicts it.