Late Proto Albanian *3 /dz/ NOT QUITE Early Proto Romanian *3 /dz/
I found out why the reflexes of Latin /di/ doesn't exactly fit in
Romanian and in Albanian,
DESPITE the GREAT similarities that we found.
Please take a look on your timeframes [ALBANIAN(1)]
A B C D E
------ ----- ---- --- ---
*k' > *c' > *c > *T > T
*g' > *3' > *3 > *D > D ~ d
*k'w > *c'W > *c^ > *c > s
*kW /+ > *c'W > *c^ > *c > s (the rest as above)
*g'w > *3'W > *3^ > *3 > z
*gW /+ > *3'W > *3^ > *3 > z (the rest as above)
" Let's label stage A "pre-Albanian", stage B "Early Proto-Albanian",
stage C "Late Proto-Albanian" (approximately the time of close
interaction between Proto-Albanian and Balkan Latin) and stage D "Old
Albanian" (later than the first Slavic loans in Albanian). Stage E
represents the Modern Albanian state of affairs."
....so why the reflexes of Latin /di/ fit but NOT EXACTLY in
Albanian and in Romanian ?
Because the Romanian Latin Loans were borrowed EARLIER than the
Albanian Latin Loans. The Romanian loans took place in phase C on
your timeframe (when g' was 3):
Lat. di -> Rom. 3[i] (  -> means optional ) (dies -> dzi)
where Romanian 3 (later z in Romanian ) is from *3' > *3 > *D
(Rom. z) (so from *g' - Axe)
The Albanian Latin Loans took place on D-period regarding your
timeframes (when g' was D and 3^ was 3).
This explains everything :
a) Why Lat. di: -> Alb. di ---> because we are in D-period,
and we have already passed from 3 to D (on the same PIE *g' axe).
Also this explain why :
Lat. rádia- > *rádja- --> Alb. *rä3- > rreze
where Albanian 3 (later z) is from *3^ > *3 > z (so from *g'w
(is 3^ as you said but not exactly at that moment but a little
bit later as 3).
This shows us an obvious thing:
That the Romanization of Albanians started LATER (at least
the /di/(Main?) Latin wave), than the Romanization of Romanians.
This is obvious because the Romanians were fully Romanized (so
they needed more time to "can finished" this process) and the
Albanians not. (I knew this as historical fact but here I saw it only
This also shows that we have to compare regarding the Latin
Loans, the "Albanian" transformations at the T+1 moment in time (for
Albanians), with the "Albanian" transformations of the T-moment in
time (for Romanians) in order to establish if Albanian and Romanian
Phonetics Systems where ONE AND THE SAME at one moment in the past.
For instance what I found above fit very well my assumption :
Genetic link between Albanian and Romanians, and explain very well
the reflexes of Latin /di/ in both Languages.
Please check my explanation too...
P.S. Now we have to check the assumption further...
Miguel are you ready?