--- In
cybalist@yahoogroups.com, Piotr Gasiorowski
<piotr.gasiorowski@...> wrote:
> 16-12-03 23:34, alex wrote:
>
> The /i/ in "cambio" was a short "i",
> > thus *ex-cambiare > *scambeare > *scambea. Do I see it wrong?
>
> Yes. First, this short /i/ became a non-syllabic glide: /skam.bja-
/, not
> /skam.bi.(j)a-/ (dots mark syllable divisions). Secondly (if the
first
> objection didn't apply), *e from Lat. i was only diphthongised in
> stressed syllables, and the /i/ in <cambio:, cambia:re> was never
stressed.
Alex hasn't diphthongised it! He's just applied the rule that
unstressed Latin short /i/ usually becomes /e/. If he used my
encoding of the Romanian sound changes at
http://homepage.ntlworld.com/richard.wordingham/sounds/prep5d_rom.htm
, it would tell him that scambia:re should yield sc1mbeare and that
scambia:s should yield sc1mbea. I didn't have any data for the
development of /bj/. (If you scroll down the sound changes, you
will see the comment 'Yodicisation - this needs more work.')
Richard.