--- In cybalist@..., "Piotr Gasiorowski" <gpiotr@...> wrote:
> The language of the Indus script (or, shall we say, the "lingua
franca" of the Harappans) may well have been Dravidian, though one
can hardly be "sure" of that at present. The argument for its having
had Dravidian-like grammatical structure looks sound to me. But it's
highly likely that the linguistic situation in the Indus Valley
civilisation was rather complex, with more than one family
contributing to the make-up (as in Mesopotamia, for example) --
perhaps with Para-Munda dominant in the Panjab, Dravidian in Sindh,
and still more mysterious languages near the Arabian Sea coast.
Substrate studies will perhaps throw some light on these enigmas, but
as far as I know very little work has been done on subjects like
Sindhi or Gujarati words without Indo-Aryan etymologies.
I agree with you Piotr. Para-Munda dominance is the key, together
with the Munda tradition of metal-working (dawn of bronze age and all
that stuff). There is also a possibility that in this time-period
(ca. 3500 to 1500 BCE), Dravidian and Munda might also have been
dominant characteristics of what Emeneau calls the 'linguistic area'.
Lot more work needs to be done on Gujarati, Sindhi, Punjabi substrate
studies. After all these are the languages of the area circumscribed
by the Sarasvati Sindhu river valleys -- the core region, the
substratum of the civilization.
As regards the 'dravidian-like grammatical structure' hypothesised by
Parpola and Mahadevan, it looks totally UNSOUND to me. Look at the
statistics: about 400 'signs' [one sign, 'narrow-necked jar'
dominating the inscriptions], about 100 'pictorial motifs or field
symbols', corpus of ca. 3000 inscribed objects, an average number of
five 'signs' per inscription on an object. These are not constituents
of a grammatically constructed sentences or words of ANY language (be
it Burushaski, be it Tamil). The blunder committed by Parpola and
Mahadevan are due to the assumptions that (1) each 'sign' HAS TO
represent an alphabet or a syllable; and (2) somehow a name or a word
(religious) has to be represented. Considering the occurrence
of 'numeral strokes' even on inscribed weapons and inscriptions on
many duplicates of copper plates and many duplicates of 'tablets',
the writing system is more likely related to rebus representation of
OBJECTS -- and hence the writing system is more likely to be related
to trade of such objects (as evidenced by the occurrence
of 'sealings' or burnt clay with multiple 'seal impressions'].