Re: Armenian, Sumerian, Burushaski, and Turkic languages

On Jun 6, 4:53 am, "John Atkinson" <johna...@xxxxxxxxxxx> wrote:
"Brian M. Scott" <b.sc...@xxxxxxxxxxx> wrote...

Darkstar <darkstar...@xxxxxxxx> wrote:

[...]

Moreover, Lithuanian has /berzas/

bér¸as

According to Pierarskas and Svecevic^ius's dictionary, Lithuanian has
<bérz^as> (where <z^> is z-hat, viz /Z/).

which indicates that /z/ in Slavic */berz/ is ancient,
and probably not from /g/.

From *g^, actually. Derksen makes the PIE *bHerhg^-o-
(where *h is an unspecified laryngeal) and reconstructs
Proto-Balto-Slav. *bér?®-o-, *bér?®-a? and PSl *bèrza.

/Z/ is the expected outcome of *g^ in Baltic. Which I'm sure is what
you're saying, except it's not what's coming through to me.



[...]

Brian- Hide quoted text -

- Show quoted text -

Speaking of Italo-Celtic "betula" again. This seems to be one of those
interesting examples where it is difficult to precisely relate all of
the phonemes to each other despite their obvious similarity. We could
postulate a borrowing, but there's no evidence for that. We could also
postulate a secondary, folk-type etymology on "bright white". But all
of these would just be ad hoc hypotheses.

This is one of those frequent cases where precise phonological link
cannot be found. And there are hundreds of those instances, if you
look hard enough. So instead of postulating many ad hoc conjectures
for each specific case, I'd simply suggest only once that phonetical
laws are not always entirely regular and they can sometimes be broken
by a number of factors. That's just deterministic vs. indeterministic
behavior of phonemes...

.