Re: Alb. "vatër" ( it was (Re: [tied] Re: Albanian valle )

On 04-11-23 03:47, alexandru_mg3 wrote:

> sg.'mãr' 'apple'/pl. 'mere' < Lat. variant 'me:lum'
> (see 'me:lum' at Rosetti,Densusianu) and not 'ma:lum' because
> accented Lat. a:/a always gives 'a' in Romanian but never 'ã')
>
> Note: I will come back with this second example in another message
> because based on Lat. ma:lum you establish once (Lat. a: > PAlb o
> see Alb. mollë and implicitly the timeframe of Alb a: > o ) that is
> not true in my opinion because based on Rom. 'mãr' the form 'me:lum'
> and not 'ma:lum' circulated in Balkan Romance.

No big deal. I suggested that <mollë> could be an _old_ borrowing from
Latin (not a mainstream late Balkan Latin one), but *mah2lo- is PIE
anyway, so it may be inherited in Albanian, or it may be a borrowing
from Doric Greek.

> ---------------
> Rom. raTa /ratsa/ 'duck' - pl. raTe /ratse/ -> Alb. rosë 'id'
>
> doesn't show any alternance a/e in the root between sg. and pl. forms
> in Romanian as 'vatra'(pl. vetre) shows.

> And we have an original PAlb *a: in Rom 'raTa' viewing Alb. o
> (in 'rosë')
>
> So for sure that 'vatra' should have something different in its
> original root than a simple PAlb. *a:

Romanian /a/ for Proto-Albanian *a: is regular, so it's <raTe> that
shows the expected development. I'm not sure how to explain
<vatrã/vetre>, which looks as if the preform had been *vetr-a/e, but
this doesn't match the Albanian word, no matter how you etymologise it.
I can only suspect some kind of analogical influence in Romanian
(assimilation to the <varzã/verze> type). Is *veatrã attested anywhere
in the dialects?

Piotr