> Alex wrote
> Latin "caepa" aka
> "kepa" gave in Alb. "qepë" and in Rum. "cepă"(without
diphtongtion).
Latin caepa gave e without diphtongation because Latin ae gave
initially an e: > ce:pa On the other hand only 'short stressed e'
gave Alb/Rom je/ja (remember also the discussion regarding Rom.
pradã that show us really an exception)
Latin caepa > Late Latin ce:pa > Romanian c^epã (I suspect via
tsepã /cep&/ based on the Aromanian form) > later Common Romanian
c^eapã
Latin caepa > Latin Latin ce:pa > Albanian qepë (with e, and not
with je/ja due to the long e:)
On the other hand we have a short e in *kepH-so and also a closed
syllable. This gave *ja both in Romanian and Albanian: PAlb./Dacian?
kjap-sa > Alb. qafë <-> Rom. ceafã that in fact is
pronounced /c^afã/.
Based on what you said seems that the word qafë is not very known in
Turkish...in Romanian (&Albanian) ceafã /c^afã/ belong to the basic
vocabulary.
The investigation should follow the source of the word Arabic
(if is an inherited word or not)
Please, if there is an Arabic speaker to help us here if possible.
Marius
P.S. Important Note: regarding the ps>f timeframe. Based on the
above derivation: e >je/ja diphtongation started before ps>f based
on the closed syllable criteria (that is confirm by the je/ja
diphtong that is really visible in Dacian words like: Diegis,
diellina, sqiarre).Also e>je/ja seems to remain active for a long
period of time (until the end of the Roman Times).
1. e>je/ja started
2. ps > f started
3. ps > f ended
----------------------
Roman Times in Balkans
-----------------------
4. e>je/ja ended
5. k/+ > Alb q <-> Rom c^ <-> Arom c
6. e > ea in Common Romanian