Re: [tied] Albanian

On 2006-07-18 19:08, Andrew Jarrette wrote:

> There's something I've always found very remarkable about Albanian: the
> fact that Indo-European *s in initial position before vowels evolved to
> the Albanian consonant <gj>, which I understand is a voiced palatal
> plosive. To me, it is perhaps the most unlikely phonetic change I have
> ever heard of,

Haven't you ever heard of *dw- > erk- in Armenian? Or *wik^- > z- in
Albanian? _Those_ are strange!

> yet it occurred (I actually sometimes wondered whether
> Albanian words with <gj-> truly originated from IE words with *s-).

The equations are impeccably solid. "Unlikely" changes turn out to be
less unlikely if you break them up into stages. The change of *s > gj
didn't happen overnight. PIE *s was first voiced, giving *z before a
stressed vowel. This *z was then retracted to *z^ (in parallel to the
change of *s > s^). Finally, palatoalveolar *z^ was "hardened" into a
palatal stop. The whole thing isn't much stranger than Lat. [j] ending
up as Fr. [z^] and It. [3^].

> Can
> anyone offer a reasonable hypothesis as to how a voiceless alveolar
> sibilant can evolve to a voiced palatal plosive? It would be very
> reassuring and satisfying to me if someone had an explanation for this
> change which to me seems to defy the laws of nature!

Piotr