>I agree that could be an issue this one, but not so big
>as a- > �-.
I don't know the way, but I already mentioned some plausible way
in order to transform avg > *Ovg > ovg/oveg > �veg. � is a mere
further variant of �. (�<->� behave to each other in a similar
way as Turkish �<->� do. Cf. �zbek-�zbek.) Let's assume an
intermediary form was *ovag /o-vOg/. But as soon as o turns �,
the following vowel can't stay /O/. The vowel e /ae/ fits better.
Analogy: lovag /o-O/ "knight", lovaglo /o-O-o:/ "equestrian".
But leveg� /ae-ae-�:/ "air". Changes of the root vowel result
in changes of the following vowels.
> If the original form was u-ya-(g&)
It can't be: the word uiaga is the modern, today's form. We do
not know how it was 500 or 1000 years ago (if it really is so
old). We only know how its Hungarian counterpart looked like,
if we get the information from Hungarian scholars who have
studied all old Hungarian texts and know the old occurrences
of �veg in written. And I'm gonna remind now and then that
in Transylvanian Romanian uiag� means only bottle, whereas
�veg has a larger semantic scope, the primary meaning being
"glass" (which, if the Hungarian word is an Iranic loanword,
fits the "water-like" phrase in the donor's language). In
Romanian, you can't say that the window-pane is made of...
uiag�. If you say so, then the native-speakers over there
will think you've emptied a bottle of p�linc� or horinc�
(< slavic horilka), i.e. "brandy").
>we need to count for the output of a Hungarian sequence
>*u-y'a- or *u-y'e (a>e being regular at least in Hungarian loans).
>*u-y'a- or *u-y'e 'is not quite the same' with uj
There is no *uyaga word in Hungarian whatsoever. But, if the
Romanian word uiaga were the loanword in Hungarian, then
Hungarians wouldn't have to convert it to �v + -e-, since
uya is easy for them to utter. E.g. there's the word uja (in
the original spelling u and a are have accents to give /'u:-ya:/)
meaning "anew; again"; ujjak (where only the u has an accent,
i.e. is long) means "fingers; toes", ujjasz�l (u and a with accents,
long vowels) mean "to regenerate". Etc. (The (battle or hooray)
exclamation huj-huj-hajra (the second a with accent) /huy-huy-
'hOy-ra:/ is old or ancient.)
So, it is not explanable how a supposed Romanian uia- turns in
Hungarian to �ve-, and not to a uja that already exists in the
words "anew", "regenerate" and "finger". But the other way
around is okay, not only because uiaga is restricted to an
area adjacent to present day Hungary and nowhere else, but also
because it seems that in old Hungarian /v/ was pronounced /w/,
a thing that seems to be proven by Romanian conservation of
loanies where there's no /v/ where in the Hungarian variant
there is one (today). (This correspondence is there even in
an unexpected parallelism between the Hungarian particle
vala /'vO-lO/ and Romanian oare in constructions such as
oarecine=cineva, oarecare=careva, oareunde=undeva, carecand=
candva, oarecum=cumva, i.e. valaki, valamelyk, valahol, valamikor,
valahogy, corresponding to Germ. irgendwer, -welch-, -wo, -wann,
-wie. In oare- constructions, oa- reflects the /w/ sound, in
-va constructions, the Romanian languages keeps the /v/. This
is an example only to illustrate the w-oa correspondence; for
Rum. oare and Hung. vala have nothing in common directly. Perhaps
at the PIE level, given the Latin explanation for Rum. oare.)
George