alan little's weblog

more vedic

13th July 2004 permanent link

I’ve been reading a 1999 paper by Michael Witzel, Early Sources for South Asian Substrate Languages, which comes to some very interesting conclusions about what languages were spoken, when and where in northern India in Vedic times. I have no idea what, if any, reaction to this paper there might have been from linguists; but Witzel is a serious historical linguist, not a biologist with some stats software and a bright idea, so I assume his ideas are at least worth taking seriously.

Witzel believes the vedic scriptures, because of their unbroken tradition of precise oral transmission as sacred texts, are the closest thing we will ever have to a tape recording of an ancient language, far better than any written evidence. Also that even within the earliest group of texts, the Rg Vedas, it is possible to clearly identify a sequence of composition in at least three different periods that also appear to have originated in different areas and with different substrate languages (*).

Very broad-brush summary:

• Geographical features identifiable in the texts at different periods show a clear geographical progression, from the north west into the Punjab and from there into the Ganges basin.
• There are no Dravidian loan words in the earliest parts of the Rg Veda. On this basis he completely rejects the idea that the previous inhabitants of the area, and the Indus Valley civilization (Mohenjo Dara/Harappa), were Dravidian speakers.
• Instead he identifies three probable substrate languages:
  1. One found in in the form of about 300 apparently non-Indo-European loan words in the earliest Rg Vedas, that he identifies as an Austro-Asiatic language related to the present day Munda language family. These languages are currently spoken in central and eastern north India, and there is no other evidence of them exisitng this far west. He believes this was the language of the northern Indus Valley Civilization.
  2. A different but related language encountered later and further south in Sindh, and possibly much further west along the Iranian coast into the Gulf. This may have been the language of the southern Indus Valley Civilization – he believes there are discernable differences between the (undeciphered) scripts of Mohenjo Dara (south) and Harappa (north).
  3. A third “Language X” already identified by previous researchers, encountered later and further east in the Ganges basin, that might be an even older substrate of the Munda-related languages.
• Dravidian borrowings are first encountered after the earliest Rg Vedas and further south in Sindh, but are not the same as substrate language [2] from that area. Witzel believes the Dravidians were also relative newcomers to the area at that time, having arrived by a different route further south – central Iran-Baluchistan-Sindh. Dravidians appear to have had horses – dravidian horse words unrelated to Indo-European – but not chariots.
• Both Vedic and Dravidian appear to have borrowed most of their vocabulary for agriculture and crafts from the pre-existing substrate languages, strongly suggesting that their speakers were pastoralists until they arrived in India and came into contact with more settled cultures.

Substrate language = a language that has died out in an area, but left traces in the form of borrowed vocabulary in the language(s) that replaced it.

related entries: Yoga

all text and images © 2003–2008