In
cybalist@yahoogroups.com, "alex" wrote:
> alexandru_mg3 wrote:
>
>> The issue is that we don't have in Romanian other examples where
>> Latin. e becomes Rom. u (like in the supposed lat. de post > rom.
>> dup~a).
>> As I know, the transformation rules are more important than the
>> possible cognates.
Your picture is flawed. The rules are always giving the mainstream.
Out
of phonetical rules valid for some general context, there are _always_
the so-called "accidents" which occur only on particular cases having
some specifical trait allowing them (ellision of some unstressed
vowels,
reduction of difficult consonantic groups, assimilation,
dissimilation,
hyperurbanisms, folk etymology, etc.).
For instance, in this case one could argue for some partial
assimilation
of the first front vowel to a back vowel, produced by unstressed /o/
from
the last syllable, for influence of labial consonant (as in Italian),
for
expressive reasons, or for a combination of all these. It is not to be
found in "the rules" [the knowledge of which does not ensure any guy
to
formulate ultimative sentences of the type "it is not possible", just
as
reading a book of medicine does not make one a medical expert able to
cope
with particular cases of patients].
Out of that, you (and by the same token, Alex) are too focused on
modern
Daco-Romanian form, completely ruling out crucial information
contained
in Aromanian <dipã>/<dipu>. Correctly used, this information _proves_
the
Common Romanian should have been with a front vowel in the first
syllable
and the question can't be _if_ "*d(i/e)pã" > "dupã" but _how_ did that
happen from Common Romanian to modern Daco-Romanian. My personal guess
is that the vowel shift was essentially expressive and articulatory,
and
included an intermediate stage of medial closed /1/, still present in
regional "dâpã" as quoted by George [that is /i/(/e/) > /1/ > /u/].
Of course, this is not a regular phenomenon since it's intimately
linked
with one word's particular characteristics, but it is still
unsurprising
since "accidents" of this kind _necessarily_ occur on large word
samples.
It's to be noted that being an "accident" doesn't equate explanation
lack.
For this case one has at the disposal also Italian parallel evolution
and
Aromanian dialectal forms, allowing us to write down "dupã" < "de
post",
without any "probably" attached; thus one has to look for the most
likely
explanation of /i/(/e/) -> /u/, not for other potential alien sources.
> There is no way. Despite the pseudoargumentation of Mr Iacomi here,
> there is no way of obtaining from Latin "e" an "u".
Well, you just (1) read again. If that's not enough, GOTO (1).
Have fun,
Marius Iacomi