>"Arama, ca si celelalte
>forme romanice (it. rame, log. ramine, engad. aram, v. fr. arain,
>prov. aram, port. arame), reproduce un *aramen, cu asimilarea a-a <
>ae-a (aeramen)" --> "<arama>, as well as other Romance forms [...]
>reproduces an <*aeramen> with assimilation a-a < ae-a" -- I gave the
>full quote in order to point out that Rosetti didn't made any claim
>about "ae-a > aa-a"].
It's true Rosetti said only "a-a < ae-a" and I apologize if it was
any misunderstanding on what Rosetti said via my presentation ....
So to present only my opinion:
Latin caepa gave Latin ce:pa with long e so to expect an
assimilation to a long vowel a: in this case is at least possible
=> not for my pleasure => but in order to can explain next the lost
of intervocalic -d- in Albanian pre => because we have also Alb.
hedhim a bi-syllabic word where intervocalic d is visible having the
same stressed pattern as Latin praeda....=> with no explanation
until now regarding Piotr's theory that assert: 'the Albanian
intervocalic -d- was lost everywhere'
So we need to make a model in order to explain all the forms ...
A Balkan Latin *pra-a-da really explain all the forms 'by rules':
1. Romanian pradã pl. prãzi
2. Aromanian pradã pl. prãdz
3. Albanian pre as in Albanian pyll (Rom pã-du-re) (with
intervocalic -d->zero in tri-syllabic words in opposition with Alb
hedhim)
I would be happy to can remove the assimilation ae <-> a: regarding
praeda and to still can explain 'by rules' all the above words...
Marius