alexandru_mg3 wrote:
>
>An accented *eu (as we could suppose to be the case based on rãu,
>dzãu) didn't passed 'directly' to *ãu (so there isn't such a rule)
>because we have also the Romanian word meu 'mine' < Lat. meus
>where /eu/ is well preserved.
Cf. zmeu and zmãu (but only zmeurã, and no *zmãurã). And the
example of greu & greutate, with no variants *grãu, grãutate.
I'd also take in consideration in these cases the... whims of
vox populi (in recent times of the language development), and
not only the old Latin-Protoromanian derivation rules. Many
aspects of modern Romanian are also explained by late, medieval
(or even later) transformations.
As for vras, I'd take into consideration the fact that the
cluster vr- either stays as such (as in those words which are
almost all Slavic loanwords) or has the [v] replaced by the
vowel [o] or the diphtong [oa]/[wa]. One of the most interesting
phonetic and semantic occurrence: the "doublette" oare and
vre (in vreun, vreunuia, vreunii; vreo, vreuna, vreunei, vreunele).
Oare in composita: oarecum, oarecine, oareunde, oarecare etc.
(Look them up in the DEX: the same etymology's been proposed
for both.)
This vowel happens to also replace the [v] in the Hung.
loans vár "burg" (TemeSvár > TimiSoara, SzegeSvár > SighiSoara)
and vároS "town, city" > oraS; Nagyvárad > Oradea Mare (but
in some cases the [v] has been preserved, e.g. Vãrãdia).
And one more morcel of food for thought (-: the Hungarian
particle vala ['vOlO] fits the Romanian oare in the composita
oarecare, oarecine, oarecând... as well as their synonyms build
with -va: careva, cineva, cândva, undeva, cumva... (Hung. valaki,
valamikor, valahol, valahogy etc. / In some situations,
Hung. vala is translated into English as "ever").
George