Re: YAEPT alert! [Re: Not phonetic but ___???]

Henrik Theiling wrote:

> Hi!
>
> "Mark J. Reed" <markjreed@...> writes:
>
>>...
>>The symbol [r\] represents the usual American English "r"
>>sound, which is an approximant ...
>
>
> Does it?  Is that not a simplification of the correct [r\`], i.e.,
> retroflex?  An *alveolar* approximant sounds different to me.  I know
> at least one (small) region in Germany were they seem to pronounce
> German /r/ as [r\], but that is not [r\`].  It may be close, but not
> the same.  For 'cure', it should be [kjr\`=], no?  And for 'error', it
> may be [Er\`=:].
>
> If this starts yet another YAEPsubT, so be it.  I want to know! :-)
>
> **Henrik
>
I used to think that American English "r" is retroflex (and notated it
that way on my Ljoerr-teg page), but I'm not sure about that any more.

I think it might be as much velarized and labialized as retroflex, but
it's probably not any single feature that makes the "American r" sound
distinct from a plain alveolar approximant. A syllabic retroflex
approximant is more like Chinese "i" in "shi" than the American "ure" in
some pronunciations of "sure". The "r" in Chinese "er" sounds closer to
the American "r". Since there isn't a convenient way to notate the
combination of features that makes the American "r", [r\] is close
enough for practical purposes. The syllabic version differs from the
consonant in the same way that [i] differs from [j] (i.e., hardly at
all, but enough to make a difference). So "error" is actually [Er\r\=]:
the vowel and consonant are distinct like the /j/ and /i/ in "yield" or
"yeast". (Phonemically, I'd write /Er@r/, with the /@r/ realized as
[r\=] in some dialects, [@] in others, etc.)