On Wed, 23 Dec 2009 17:49:44 -0500, Alex Fink <[log in to unmask]> wrote:
>Indeed I'd go so far as to
>say that there aren't very many cases where mandatory anything exists,
>except the nuclear vowel, and sometimes the onset consonant -- but most
>onset-obligatory languages I've seen have /?/ as a possible onset.

According to's chapter 12, many languages have (C)V syllable-
structure but few (e.g. Hawaiian and Mba inter alia) have CV syllable-structur 
(with mandatory onsets).

Using I compiled a list of their sample-database's languages that 
they have recorded as having CV or (C)V syllable-structure, no uvular 
consonants, no glottalized consonants, and no "unusual" consonants.

Then I used UPSID to see which of those languages were recorded in UPSID as 
having a glottal stop.

9 of them do have a glottal stop; namely Andoke, Guaraní, Hawaiian, Kwaio, 
Mor, Movima, Muinane, Pirahã, and Roro.

7 of them have no glottal point of articulation; namely Aché, Bororo, Ekari, 
Kewa, Savosavo, Suena, and Yareba.

I'm not sure which of those have mandatory onsets. Of those that have glottal 
stops, I'm not sure which of them allow a glottal stop as an onset.

There may be other languages in one or the other, or even both, of 
and UPSID, which have CV syllable-structures with mandatory onsets, but 
don't have /?/ onsets. 

Those languages I mentioned are the ones in both databases that are easy to 
recognize as being the same language.  For some languages one database 
uses one name and the other database uses a different name.  (And of course 
there are languages in one database that aren't in the other or vice-versa.)