Re: Two part verbs (Why They Shouldn't Make Me Wait)

On Fri, 8 Dec 2006 13:39:11 +0000, Christopher Bates
<chris.maths_student@...> wrote:

>Not at the moment I'm afraid. I don't have a microphone or anything to
>make such recordings... and I'm not convinced my own pronunciation is
>perfect either. :P However, some natlangs are at least as bad (think
>Bella Coola etc) so I think it's feasable... and since the transcription
>is, afterall, phonemic rather than phonetic I can possibly allow some
>kind of epenthesis if the clusters turn out to be too problematic. Since
>/@/ is a phoneme, though, the epenthetic vowel would have to be
>something else.
Not necessarily; Chukchi can be analysed to have both epenthetic and
underlyingly present [@].  So for instance there are nouns /mem@l/ [mem@l]
'seal' and /miml/ [mim@l] 'water'.  They can be told apart because they
select different allomorphs of the plural: /mem@lte/ [mem@lte] gets the
allomorph /-ti/ that follows VC[CORONAL], while /mimlt/ [miml@t] gets the
elsewhere variant /-t/.
(This from  http://privatewww.essex.ac.uk/~spena/Chukchee/chapter2.html .)

In fact I'd be a little surprised if there were a [@] and it wasn't the
regular epenthetic vowel.  This is one thing that's always struck me as a
little unfortunate about Lojban, that /@/ is taken and so the epenthetic
vowel has to be something else.

Alex