Dating the final IE unity, in particular the word for "horse"

[Gábor Sándi]

----- Original Message -----
From: proto-language <proto-language at email.msn.com>
Sent: Saturday, 25 March, 2000 4:10 PM

> Dear Gabor and IEists:

[ moderator snip ]

> [PR]
> I wonder how productive it is to adopt arguments from non-linguists that
> seem to have little likelihood of being accepted by linguists.

[GS]
By non-linguists, do you mean me or Gimbutas? I am a linguist - whether a
good one or not is not for me to say. Gimbutas's arguments are based on
archaeology, as well as a certain amount of theorizing that goes along with
any innovative scientific thinking. There are many linguists who accept the
Kurgan hypothesis - probably more than those who reject it outright.

> I am reasonably sure (though, subject to correction) that a process of
> deriving *H(1)ek^uo-s from *H(3)o:k^u's is simply not possible.

[GS]

Sorry, of course I meant *H(1)o:k^us, *H(3)ek^uos would have resulted in
*H(3)ok^uos.

I am not too happy about the unusual length, or even the particular ablaut
grades either. But so much of ablaut variation is unpredictable that I would
not exclude this particular development. Neither would I oppose vehemently a
semantic development like "horse-like" > "fast, quick". My main point is
that *ekwos could be the result of internal derivational processes within
PIE, and there does exist a reconstructable root that could - in theory - be
related to it.

[ moderator snip ]