Re: [tied] Re: Dog

From: Piotr Gasiorowski
Message: 30004
Date: 2004-01-25

25-01-04 19:43, Richard Wordingham wrote:

> I'm not convinced of the vocalisation. The Proto-Germanic is
> *xundaz, which implies *k^untós. *k^wn.tós would have given
> *xwundaz, though I suppose that might have simplified to *xwundaz -
> I can't find any examples or counter-examples.

I'd say that *xun- is the normal develoment of both *kWn.- and *k^wn.-
in Germanic, just as *kum- is the normal development of *gWm.- (the nil
grade of *gWem- 'come'), cf. Ger. (Zu)kunft etc. In other words, I'd
support *xundaz < *k^wn.tos.

> until in Albanian <kand>,
>> thanks to dispalatization of palatal, followed by nasals.
>
> I'm not persuaded of the development either. Piotr gives (at
> http://groups.yahoo.com/group/cybalist/message/29908 ) the
> development
> dhjetë '10'< *Die(T&)t& < *diäcata < *dek^m.t (+ *-a:)
> The development k^m. > th& better explains the haplology (semi-
> haplology?) than your implicit proposal kam / kan (Geg kâm / kân?)
> does.
>
> My 24-year old notes say PIE *m. > Geg âm, ân, Tosk ë, PIE *n. > ân,
> but I suspect the difference reflects an inaccuracy in my notes or
> possibly an error in the thesis on the development of Albanian I
> made the notes from. The development of *n. should have been
> parallel to that of *m. Alas, I was not well enough trained to make
> a note of the author. I do remember that the author said he could
> find no find trace of PIE *s.

Syllabic nasals give Alb. a (> e in umlauting environments). If you need
a good example, take *septm. (extended with *-t- + *-a:) > shtatë, or
(një-)zet '20 < *wi(:)k^m.tih1 (with umlaut). It's possible that *n.h
(with a tautosyllabic laryngeal) developed into *na: > Mod.Alb. no, cf.
njoh 'know' < *gna:sk- < *g^n.h3-sko:.

> Doubtless there will be a fuller exchange of view when Piotr
> addresses the development of the syllabic consonants. It seems to
> me that your proposal derives a Geg _qen_, whereas the Tosk would be
> something like *thët. But a Geg _qen_ could just as well be from
> Latin! We await more expert commment.

*k^wn.to- would have developed into *c'Wat- > *c^at > Mod.Alb. *sat (in
Geg or Tosk alike)! Lat. canem was borrowed as *kan/*k'en(i), with the
umlauted plural stem <qen> generalised, as in gjel 'cock' <
*gal/*g'el(i) <-- gallum, another "singularised plural" of Latin origin.

Piotr