On Mon, 02 Dec 2002 08:05:44 -0000, "altamix"
<
alexmoeller@...> wrote:
>2)let alone the toughts latin di,de>zi,ze in romanian. There are many
>examples which shows that latin "di/de" did not became zi/ze in
>romanian.
First you have to know what you're talking about. There is no rule
that Latin di, de > Romanian zi, ze. The rule is that before Eastern
Romance i (vocalic /i/ or consonantal /y/, including the diphtong
/ie/), d > dz > z, t > ts and s > S. This does *not* mean that any
Latin e(:) or i(:) after d (t, s) provokes the soundshift. In
particular, it is *not* caused by:
1) any short /i/ (> ERom. /e/)
2) any long /e:/ (> ERom. /e/)
3) an unaccented short /e/ ~ /ae/ (> ERom. /e/)
4) an accented short /e/ ~ /ae/ before a nasal (m, n) + other
consonant (> ERom. /e/).
It *is* caused by:
1) long /i:/ (> ERom. /i/), if not elided.
2) i + vowel (consonantal /i/) (> ERom. /y/)
3) stressed short /e/ ~ /ae/ *not* followed by a nasal cluster (>
ERom. /ie/).
Furthermore, the palatalization does also not occur in words
introduced into the language *after* the working of the soundlaw was
over, whether they be neologisms, borrowings from Slavic, Hungarian or
French or learned borrowings from (Neo-)Latin or Greek.
Now go through your list of so-called counterexamples, and scratch any
word that had short /i/ in Latin (such as digitus), long /e:/ in Latin
(such as the prefix de:-), unaccented short /e/ in Latin, accented
short /e/ before -nC in Latin (such as dentem), or that is not
inherited (such as diarrhoea). You'll find that all your examples
evaporate.
=======================
Miguel Carrasquer Vidal
mcv@...