Is "Sanskrit" Dravidian ?

George Thompson thompson at JLC.NET
Thu Jun 10 20:01:07 UTC 1999

Since Hans Henrich Hock has taken the leap, perhaps it will useful to
belabor a few more points as well. They will be obvious to many.

In order to respond adequately to Samar Abbas's assertion that '30-50 % of
the words of Sanskrit are of Dravidian etymology anyway', we would have to
ask ourselves exactly what stage of the language is meant. Insofar as
Sanskrit was/is a living language, it was/is a changing language, of course.

The percentage of non-IA words in Sanskrit at 1000 BCE will be different
from that at 1 BCE or at 1000 CE. Or for that matter at 1999 CE. It is also
reasonable to suppose that there will be not only temporal variation but
also geographic variation. So we need to ask whether there is a greater
amount of non-IA words in the North or the South, or in the East or West,
etc. This is not a new idea.

I think that we can come up with meaningful figures only by setting limits
like this, and restricting our assertions within such limits.

For the RV, we can be fairly specific, thanks to the work done by previous
scholars. Kuiper has listed 383 non-IA words -- that is, borrowed words --
in that text. Kuiper's list is explicitly provisional, and by no means is
it overly restrictive. For example, the three terms cited in my previous post:

1. bRba'duktha = hapax at RV 8.32.10, epithet of Indra.

2. bRbu'H = name of a Vedic patron [cf. his dAnastuti at RV 6.45.31-33].

3. bR'bUkam = another hapax, at RV 10.27.23; meaning unclear, possibly of
Munda origin [thus Kuiper].

are cited by Kuiper as three separate entries on his list [nos. 280-282],
in spite of the fact that there is a very good chance that two or all of
these might be related.

Now Kuiper carefully demonstrates how he comes up with a percentage for
non-IA words in the RV. I won''t go into it. But, very conservatively, he
estimates that this percentage might be well above 5 %. But this figure
includes not only possible Dravidian borrowings, but also possible Munda
ones. And clearly, my examples above show that he has included some that
may be good old Indo-Iranian words. These may have been borrowed from some
unknown language at a pre-Vedic Indo-Iranian period. Or possibly, as I am
willing to speculate, they may be borrowings into Vedic from Iranian.

Finally, this figure of 5 % includes borrowings from one or more other
languages that are otherwise unknown to us [see the 'language x'' posited
by Masica, accd to whom possibly one third of the Hindi lexicon may be
borrowed from this unknown source]. So it is fair to state that, even if we
do not go into great detail examining individual etymologies, the
percentage of specifically Dravidian borrowings in the RV is going to be
quite a bit less than 5 %. Maybe 2. Hindi, however, stands at the other end
of the spectrum, much closer to the figure cited by Mr. Abbas. What does
this fact suggest?

Will such figures apply to epic Sanskrit? No. Nor to later classical Skt,
nor to Middle IA, etc. It would be useful to get precise figures from as
many different sources as possible [i.e., different times and places].
Maybe we could identify trends or patterns.

In any case, then our statements would be meaningful and verifiable. Or not.

Hoping that I haven't wasted anyone's time.

George Thompson

More information about the INDOLOGY mailing list