I think Ray is already gone off to France, but I'll post this reply in
the hopes that he'll be able to get to it when he comes back.

On Fri, 20 Oct 2000, Raymond Brown wrote:

> At 6:05 pm -0400 17/10/00, John Cowan wrote:
> [....]
> >
> >It seems to me that there isn't any possible ambisyllabicity that can't
> >be accounted for as a covert gemination.
> What I'm finding difficult is to understand what "covert gemination" or
> "gemination in disguise" can actually mean.  To me the terms seem
> meaningless, i.e. to my simple mind either a consonant is geminate or it
> ain't.  It's quite clear to me, e.g. when I hear Welsh _hapus_ pronounced,
> that the medial /p/ is geminate.  The /p/ in English _happy_ is not
> geminate.

Well, I would hardly call your mind simple! I think the confusion
comes from our respective usages of the term "gemination." Gemination
as I've been using the term is a structural property of phonological
representations which may be realized phonetically in a number of
ways: 1) extra length, 2) "fortis" or "tense" articulation, 3)
aspiration, or 4) voicing and continuancy contrasts. There are
probably others. What all of these realizations have in common is that
they are all ways of encoding the structural property of a single
segment's content being spread over two structural positions.
Ambisyllabicity, on the other hand, is the overlapping of two
syllables such that they both dominate the same *single* structural
position. Gemination also shows properties of "inalterability" and
"integrity", which are not claimed to be properties of ambiguity.
Inalterability is the property of being immune to processes of
lenition (for example). Integrity is the inability of a geminate to be
split by epenthetic processes which would apply to other consonant
clusters in the same environment.

Gemination as you seem to be using it is the phonetic property of
extra length on a consonant. While this can be a clue to the
structural property of gemination, it isn't the only one. Because of
my training, I am used to thinking of gemination as a structural
property which may have different realizations. So it makes perfect
sense to me for a consonant to be a geminate but not to show it as
extra length.

> I can understand the argument of those who maintain that the /p/ in English
> _happy_ is ambisyllabic.  But at present I neither accept nor reject the
> theory.

I am also open to convincing arguments for ambisyllabicity. So far I
haven't seen them, though.


Dirk Elzinga
[log in to unmask]