Skip to main content
History & heritage

What language did the ancient Kambojas speak?

By The SaathiSutra Team 6 min read
A plain burgundy card with a thin gold border, the words “Kamboj History & Heritage” in cream lettering and “SaathiSutra” in small gold capitals beneath.

Not a sentence of their language survives. One verb, noticed by two Sanskrit grammarians, has carried the whole question, and it points west, towards the Iranian languages.

The ancient Kambojas most likely spoke an Iranian language, a relative of Avestan, the language of the Zoroastrian scriptures, rather than a form of Sanskrit. No sentence of their language survives. The case rests on one verb: the grammarian Yaska recorded that the Kambojas used the verb shavati for “to go”, and that word has no root in Sanskrit but matches the Iranian verb for going exactly. Most scholars, from Ernst Kuhn in 1904 to Rüdiger Schmitt in the Encyclopaedia Iranica in 2021, accept the Iranian reading; a few have argued for a mixed or Indian speech.

Iranian here means a language family, not Persian. Schmitt classes the Kambojas as a south-eastern Iranian people of the north-western frontier, distinct from the Persians, much as Punjabi and Bengali are both Indo-Aryan without being the same language.

One verb in Yaska’s Nirukta

Yaska’s Nirukta, an early work explaining difficult Vedic words, stops in its second chapter to make a point about dialects. Some regions, he says, use a word only as a verb and others only in a derived form. His example is the Kambojas: in Lakshman Sarup’s translation, “The verb śavati, meaning to go, is used by the Kambojas only.” The Aryas, he adds, use only a modified form of it. In the same breath he gives a second regional pair, dāti, “to cut”, among the Easterners, and dātra, “sickle”, among the Northerners.

So the Kamboja remark is one entry in a list of regional usages, the observation of a scholar who noticed how people around him spoke. The manuscripts also slip a folk etymology of the name Kamboja into the middle of the sentence, from kambala, a blanket. Sarup thought that phrase sits awkwardly and was probably added later.

Patanjali repeats it

Patanjali’s Mahabhashya, the great commentary on Panini’s grammar, repeats the remark almost word for word in its opening chapter. The setting is a reply to the objection that many forms taught in grammar are never used. Patanjali answers that the field of speech is vast and that some words live only in particular regions: shavati among the Kambojas, hammati among the people of Surashtra, ramhati in the east, while the Aryas use only the ordinary verb gam-. He does not repeat the blanket etymology.

The scholar Michael Witzel argued that because both grammarians use shavati as a stock example, it must have been familiar to them by about 500 BCE or earlier. That makes it one of the oldest recorded facts about the Kambojas.

Why the verb points to Iranian

Sanskrit has no verb shavati meaning “to go”; its inherited verb of motion with the same ancestry has a different form. Iranian has exactly this one. In 1904 the German scholar Ernst Kuhn identified the Kamboja verb with the Avestan word for “goes”, and in 1911 George Grierson, the director of the Linguistic Survey of India, made the same point in the Journal of the Royal Asiatic Society, comparing Old Persian, Avestan and the modern Persian verb for going and calling shavati a good Iranian word.

Schmitt’s modern statement is that the Kamboja form agrees exactly in sound with the Younger Avestan and Old Persian verbs, which go back to a single Proto-Iranian form. From this he concludes, with a confidence rare in this field, that the Kamboja language was Iranian and close to Younger Avestan. Whatever its classification, he adds, the Kambojas clearly spoke a language different from their Indo-Aryan neighbours.

Iranian words in Ashoka’s inscriptions

The second line of evidence comes from Afghanistan. In the north-west, some of Ashoka’s edicts were rendered into an Iranian language written in the Aramaic script, and such texts are known from Taxila, Pul-i Darunta and Kandahar. The French scholar Émile Benveniste, who studied the Kandahar finds in 1958, noticed that Ashoka twice names the Yonas, the Greeks, beside the Kambojas, and that the Kandahar inscriptions use exactly two non-Indian languages: Greek, and Aramaic carrying Iranian words. He proposed that the Iranian words belong to the Kamboja language. One of them renders Ashoka’s word for “good obedience” with a term built on the Avestan word for obedience.

Not everyone accepts the link. Gérard Fussman explained these Aramaic-and-Iranian texts as the legacy of the administration the Persian empire had set up in its Indian provinces, still at work under the Mauryas. Fabrizio Maniscalco stressed that no inscription from Kandahar names the Kambojas at all. Benveniste’s suggestion is widely repeated, but it remains an inference; the evidence is laid out in the article on Ashoka’s edicts.

The Pamir languages

Two Indian scholars looked for the verb in living languages. Jayachandra Vidyalankar argued around 1930 that the Pamir languages near the head-waters of the Oxus keep the root in the sense “to go”, and Moti Chandra suggested in 1945 that their speakers might represent the ancient Kambojas and the epic’s “further Kambojas”. A modern check bears out the first half: Yaghnobi, a small eastern Iranian language of Tajikistan, still has the verb stem šaw-, “to go”. But the same root runs right across the Iranian family, so it shows that the Kamboja language was Iranian, not where it was spoken or which language descends from it.

A theory proposed and withdrawn

In 1974 Fussman suggested that an undeciphered script found on Kushan-era inscriptions at Dasht-e Nawur, a high plain west of Ghazni, might record the Kamboja language. By 1994 he had withdrawn the idea himself, because examples of the same script had turned up across Central Asia and Afghanistan, far beyond any likely Kamboja homeland. A partial decipherment published in 2023 reads the script as an otherwise unknown Middle Iranian language of the Kushan period. Some summaries still list the Dasht-e Nawur theory as open, but its own author abandoned it.

The dissenting views

  • A Vedic Indian people. B. C. Law argued in 1943 that the Vedic teacher Kamboja Aupamanyava shows the Kambojas were Vedic Indians, and that shavati was an old Vedic verb they had kept. The difficulty, as Kuhn and Grierson had already shown, is that no such Sanskrit verb is known.
  • A mixed speech. Grierson proposed that the Kambojas spoke either Sanskrit with Iranian words fitted with Indian endings, or a speech that mixed the two families.
  • A north-western Indian influence. Jacob Wackernagel saw the hand of a north-western Middle Indo-Aryan language in the word’s first sound. Schmitt called this an unnecessary complication.
  • The cautious middle. The Vedic Index of 1912 allowed the Kambojas “Iranian as well as Indian affinities”. The Iranian view is the mainstream one today.

What is still debated

  • Whether the language was Iranian, mixed or Indian. Iranian, close to Avestan, is the mainstream view.
  • Whether the Iranian words in Ashoka’s Kandahar inscriptions are Kamboja words, or the old Persian chancery language of the region.
  • Whether the Pamir languages descend from Kamboja speech. The shared root proves only a family resemblance.
  • Whether Yaska counted the Kambojas among the Aryas. His sentence sets their usage against the Aryas’, and scholars read that contrast differently.

Sources

  1. Lakshman Sarup (trans.), The Nighantu and the Nirukta (1920–21), Nirukta 2.2, with introduction and notes — Oxford University Press and the University of the Panjab

    Yaska’s remark on shavati, quoted above, the Easterners and Northerners, and Sarup’s view that the blanket etymology is a later addition.

  2. Patañjali, Vyākaraṇa-Mahābhāṣya, ed. F. Kielhorn, revised K. V. Abhyankar, vol. I, p. 9 (electronic text) — GRETIL, Göttingen

    The repetition of the shavati remark, with hammati in Surashtra and ramhati in the east.

  3. E. Kuhn, “Das Volk der Kamboja bei Yāska”, in Avesta, Pahlavi and Ancient Persian Studies in Honour of P. B. Sanjana (1904), pp. 213–214 — Karl J. Trübner, Strassburg

    The first identification of the Kamboja verb with the Avestan verb for “goes”.

  4. G. A. Grierson, “The Language of the Kambojas”, Journal of the Royal Asiatic Society 1911, pp. 801–802 — Royal Asiatic Society

    The Old Persian, Avestan and Persian parallels, and Grierson’s mixed-speech proposal.

  5. R. Schmitt, “Kamboja”, Encyclopaedia Iranica (2021) — Encyclopaedia Iranica Foundation

    The exact match with Younger Avestan and Old Persian, the conclusion that the language was Iranian, Witzel’s date, Wackernagel’s view, and Benveniste’s link to the Kandahar texts.

  6. G. Fussman, “Aśoka ii. Aśoka and Iran”, Encyclopaedia Iranica II/7–8 (1987) — Encyclopaedia Iranica Foundation

    The Iranian-language versions of Ashoka’s edicts in Aramaic script, and their explanation as a legacy of the Persian administration.

  7. G. Fussman, “Dašt-e Nāwor”, Encyclopaedia Iranica VII/1 (1994) — Encyclopaedia Iranica Foundation

    The Dasht-e Nawur inscriptions, Fussman’s 1974 proposal and his withdrawal of it.

  8. S. Bonmann, J. Halfmann, N. Korobzow and B. Bobomulloev, “A Partial Decipherment of the Unknown Kushan Script”, Transactions of the Philological Society 121/2 (2023) — Wiley, record at the University of Cologne

    The reading of the script as an unknown Middle Iranian language.

  9. R. Bielmeier, “Yaghnobi”, Encyclopaedia Iranica (2006) — Encyclopaedia Iranica Foundation

    The living Yaghnobi verb stem šaw-, “to go”.

  10. Moti Chandra, Geographical and Economic Studies in the Mahabharata: Upayana Parva (1945), pp. 37–39 — U.P. Historical Society, Lucknow

    Jayachandra Vidyalankar’s Pamir argument and Moti Chandra’s own suggestion.

  11. F. Maniscalco, “A New Interpretation of the Edicts of Asoka from Kandahar”, Annali di Ca’ Foscari, Serie orientale 54 (2018) — Edizioni Ca’ Foscari

    The objection that no Kandahar inscription names the Kambojas.

  12. B. C. Law, Tribes in Ancient India (1943), pp. 1–3 — Bhandarkar Oriental Research Institute, Poona

    Law’s view of the Kambojas as a Vedic Indian people.

Our team answers every message. Reach out and we will sort it out.