Re: [tied] *g'(h)- > d as aberrant outcome

On Sun, 09 May 2004 01:25:08 +0200, alex
<alxmoeller@...> wrote:

>I understand you was talking about Albanian change here. I brought the
>example of "gjendër" because we have had as target an another word,
>namely "ngjesh" which I see to be cognate with Rom. "inghesui"; both
>compositum (ngjesh & inghesui) but too "gjesh" and Rom. "ghes"; the
>outgoing point was the derivation of Alb. word which was supposed to
>be from the root *h1en-)joh3s-

I assume that stands for *(h1en-)joh3s-?

>and from this root I cannot see any
>modality to get Rom. "ghes" or "inghesui".

Why not? *(h1en-)joh3s- > pre-Alb. *(n-)gjes- regularly.
From this are derived Alb. (n)gjesh- (s > sh) and Rom.
(în-)ghes-.

>As for the derivation from Latin glandula > ghindura.. there is
>nothing againt derivation from IE *ghend as well; the suffix "-urã"
>makes the job: *ghend +ura > ghendura > ghindurã.

There's lots of things against it. For starters, you
probably don't mean *ghend- "to grab", but *ghendh-
(*gh(&)ndh-) "ulcer, tumor" (Grk. kanthúle:, konthe:laí
(Hes.), Goth./OE/OHG gund and that's it), a rather obscure
root which is not attested in e-grade anyway. Secondly,
ghíndurã obviously does not contain the suffix -úrã, which
always carries the accent. In the third place, gh- should
have palatalized in Romanian before a front vowel (or do you
think that PIE *gh and Romanian <gh> mean the same thing?).
In the fourth place, why invent such a fantastic scenario
when a much better, easier and more logical solution is
offered by Latin glandula?

=======================
Miguel Carrasquer Vidal
mcv@...