YAEPT: apparently bizarre semivowels

> > I agree that it's safe to say that /r/ is a semi-vowel corresponding
>with
> > /3/.
>
>Nope.  First of all, /r/  can mean too many things, so let's stick to
>phonetics. The sound [r] isn't a semivowel - although as a trill, I
>suppose [r] can be syllabic (anyone know of any examples of
>that?).  The approximant  [r\] is a semivowel, but the corresponding
>full vowel isn't [3], it's [r\=].
I'm talking about the _English_ phonemes, if I didn't make myself clear the
first time. [r] has naturally nothing to do with [3].

I admit that this analysis widens the chasm between phonemic and phonetic
structure of English (/3/ varying from [r\=] to [@r\] to [3`] to [3:] to...)
but it adds symmetry - no more of that "assorted vowels may not occur before
r" stuff, and rhotic vs. non-rhotic becomes just a question of the
realization of /3/-final difthongs.

Let's see, this gives the following vowel phonemes...

lax /I E & V Q U/
i-final difthongs /i: ei ai oi/
u-final dfT. /iu au ou u:/ (/iu/ unexpectedly being rising)
3-final dfT. /i3 e3 a3 3: o3 u3/
long monofT.s & trifT.s /A: O: iu3 / (in some 'lects also /au3 ai3/)

...Well, still asymmetric. Analyzing "ew" into /eu/  would help, but sadly
even /iu/ is a stretch already.

<yak wack rack>
/j&k w&k r&k/

<soy sow sore>
/soi sou so3/

John Vertical