Re: languages of pre-I.E. Europe and onwards

On Wed, 2009-01-21 at 21:21 +0100, Jörg Rhiemeier wrote:

> Yes.  One cognate set is no cognate set.  You can always find
> chance similarities.  In order to prove a relationship between
> languages, you need at least about 100 cognate sets displaying
> regular sound correspondences.
There are a lot of relationships accepted on far less than that.
Consider the words vhraterei "to the brother", leuzorophos "children",
and zonasto 'he gave". If they were all that survived of Venetic, it
still wouldn't be too difficult to guess that it was Indo-European.

For me, the Etruscan mi "I" is a pretty good start towards classing it
as Nostratic. Consider the probabilities. If the average language has 25
consonants, any consonant should occur for "I" in about 6000 / 25 = 240
languages, randomly distributed. The facts that m- and k- are absent
from South America, and p- and r- are absent everywhere shows that the
distribution is non-random, and hence significant. When we also see an
oblique form mini, comparable with Turkic and Uralic min-, it becomes
even more difficult to take the Etruscan forms as coincidental. Then we
have  Etruscan -c "and", ati "mother", ta "this", tur- "give", tul
"stone". One could be coincidence, like the Australian language that has
the word dog, but the odds against half a dozen similarities is very
large.

Replies