> But why can't there have been a contrast in Proto-Albanian between
*dz
> < PIE *g^ on one hand and *dz^ < PIE *gW/+ and Latin *di/_V on the
> other? Such a contrast might even work for pre-Romanian - Latin
> velars before front vowels did not become affricates in Proto-
Albanian.
>
> Richard.
>
In Roman Times PAlb g^ (< PIE *gW,gw/+ ) and c^ (<PIE *kW,kw/+
was already dz and ts /c/ => otherwise the Latin Loans in Albanians
would have been loaned with(or passed later to) to g^ (c^), as in
Proto-Romanian => Especially when the Proto-Romanian (<Balkan latin)
has the g^, c^ sounds (earlier or later in Roman Times doesn't
matter here).
But the PAlb outputs for Latin ci,gi are q, gj not c^, g^
showing that there wasn't any c^ and g^ in Proto-Albanian in the
same time with the c^ g^ of the Balkan Latin (so in Roman Times)....
(this indicate also how old is c^ and c^oara (<-> Alb sorrë < PAlb
c^wa:ra) 'crow' at Romanians (in fact it was already present at Pre-
Romanians (based on the above argument) in "Decebal's Times" (87-
106))
On the other hand, the PAlb dz (of Roman Times) is not only the
reflection of an older g^ but also the output of Latin dj and of the
PIE dj (see => Alb rezzë <-> Rom. ra(d)zë < Lat. radius
On the other hand the original dz of viedhull (Rom vie(d)zure)
(<PIE g^) wasn't dz in Roman Times because otherwise it would have
been passed finally to Alb. z together with all the other dz-s
described above.
Now to come back to your question:
If you will demonstrate other outputs for x and y in the
transformations dj > X > z and g^ > Y > z other that dz /3/, you
will open a new chapter in Albanian Historical Linguistic because
these transformations are largely accepted today (Example: on this
forum see Piotr's Albanian 1,2,3).
Is like in case of interviocalic d>zero where the intermediary
stages were with a great probability d > dh > h > zero
In addition Romanian Language really preserved the original (but
also the laters) dz, ts , c^, g^ of Proto-Albanian/Dacian? (I have
some doubts about the preservation of g^): because the phonetism of
Romanian words like madz&re, c^war&, dzar&, vjedzure couldn't be
explained based "on the inner Latin timeframes in Relation with
Proto-Albanian Timeframes"
This indicate that we need to introduce a chapter in Romanian
Linguistic regarding 'The Phonetic Evolutions of Romanian Substratum
before the Romanization' if we really want to obtain a coherent
picture about the phonetism of some words present today in
Romanian...
Best Regards,
Marius
Note: in fact inherited PAlb dj is considered to evolved as dj > g^
> dz > z (so is considered that has merged with g^ initially but I
think that is better to consider this merge Only at dz level dj > dz
> z => based on the Latin Loans in Albanians and based also on Proto-
Romanian that shared almost all of the phonetic rules with PAlb
(including intervocalic b, dj>dz, etc..). But this detail doesn't
matter regarding the discussed topic.