Hi!

I have a question of what you'd think would be the most likely status
in phonotatics of nasal+stop clusters after a specific sound shift.
I will explain:

As I mentioned last week already, I intend to drop initial unstressed
/i/ in my new romlang.  This changes the set of existing initial
clusters, of course.  Let's assume for simplicity that we will deal
additionally with nasal+stop clusters as in 'mperie' < IMPERIUM.

Now, I will also allow for the preposition 'in' to collapse to only
'n' before vowels and before stops, so that we get the same additional
clusters here, too: nasal+stop.  E.g.:

   in Rome          - in Rome
   n'Talie          - in Italy

My question is: how would the syllable structure most likely behave?
Will 'mperie' be two syllables: /mpe.rje/ ({i} in {ie} looses one
syllable anyway, that's not my question)?  Or will it be three:
/m=.pe.rje/?  What this be different if the nasal stems from the previous
word as in 'n'Talie': /n=.ta.lje/ or /nta.lje/?

And if a vowel precedes a word with an initial nasal+stop, will the
nasal be moved to the previous syllable: 'ku mperie' /kum.pe.rje/ or
/ku.mpe.rje/ or /ku.m=.pe.rje/?  The decision is influential since
open and closed syllables select different vowel allophones:

   /kum.pe.rje/   =  [kUm.pe:.rje:]
   /ku.mpe.rje/   =  [ku:.mpe:.rje:]
   /ku.m=.pe.rje/ =  [ku:.m=.pe:.rje:]

In case you think that /mpe/ is one syllable and will stay so if a
vowel precedes, what effect would you think does this have on medial
clusters?  Would the syllable structure be changed by analogy and
e.g. 'kanta' would not be /kan.ta/ anymore but /ka.nta/?

I don't know what has happened in Sicilian.  I only see that they
*write* initial nasals -- I have never heard it, nor have I seen any
descriptions of this level of detail.  So if you know more, please
share!

Bye,
  Henrik