Evidence for Schwa + r in phonology

[Aaron E. Drews]

On Tue, 8 Feb 2000 Dfcoye at AOL.COM wrote:

}But it would seem to me that writing the 'bird' vowel
}to indicate it's a diphthong (schwa plus r) when the vast majority of
}Americans use a monophthong is deceiving and confusing.    DARE uses the
}monophthongal hooked schwa, and we all should follow suit.

Where does the [r] come from when there is metathesis with hooked-schwa?
By the same token, where does the [r] come from in intrusive-r in Boston,
Australia, RP...?  There has to be an r encoded somewhere. Is it the hook
in rhotic speech, or the I in Brooklyn?  What about in Australia
(heraliding back to the original question) Is it
phonemic/underlying/lexical?

At some level, at least in comparative studies, hooked schwa has to be
described as two sounds, either as an underlying representation or through
some phonological process.  If I'm not mistaken, DARE, but its very
nature (the R), has to be "comparative".


Donald M. Lance wrote:

}Do a vowel chart of the possible English vowels before spelled -r and see
}what you happens.  Only a few distinctions are possible in the
}phonology/phonetics of both "/r/-keepers"  and "/r/-losers."

(@ = schwa, &=hooked schwa)


Historically, syllable-rhyme /r/ cause the epenthesis of a schwa after a
tense vowel, and also lax vowels were merged before syllable-rhyme /r/.
/ir/ -> [I at r]   /ur/ -> [U at r]   /Ir/
/er/ -> [E at r]   /or/ -> [O at r]   /Er/
                /ar/ -> [a at r]   /Ur/    -> [@r]
                                /^r/



This is shared by both RP and American (what I'll call GA).

(I think this is in Wells 1982 and E. Dobson.  This doesn't apply to those
that have a three way distinction with Mary-merry-marry.  Then again, the
tense-lax distinction becomes a bit cloudy with low vowels)

RP lost its r's.  There is still a residue of the /r/ that is hidden in
the schwa (or, if the schwa has disappeared, in the full vowel), and comes
out before a vowel.  The loss of /r/ has two major results: the
phonemicization of the centring diphthongs, and the merger of [@r] with
just [@].  The phonemicization of the diphthongs caused a secondary system
to form.  The @ off-glide is being lost in the low vowels in RP, causing a
further merger (caught and court).  This has happened in Australian to
some extent, where long lax vowels represent what was a tense vowel plus
/r/, and short lax vowels represent short lax vowels.


}Not a simple matter, by any means.

You ain't kiddin'.


Rudolph Troike wrote:

}Phonemically, it makes no sense to
}consider /schwa-r/ a single segment, creating an odd phoneme with a very
}restricted distribution. If it is done nevertheless, then it is more
}plausible to treat it as syllabic /r/, rather than a peculiar "hooked
}schwa". (Will we also treat the unstressed syllables of <bottle>,
}<cotton>, <bottom> as unique phonemes?)

Not to be critical, but unstressed vowels after /t/ are not good examples.
In order to be tapped, /t/ has to be released, which implies a vowel (even
with a duration of 5ms) or a change in manner of articulation. /t/ can't
be released homorganically, so something else happens before /n/ (in
short).

Back to schwa and r.....

The thing with schwa+r is that it can appear in stressed syllables as well
as the unstressed syllables _murder_, _burger_.  IMHO,  this
sound is not an odd phoneme with a restricted distribution.  schwa+r/& is
distributed in the same was as any other tense vowel.

When RP lost its r's, did GA lose
them, too, but maintain retroflexion of the schwa?

The merger of [@r] and [@] into a monophthong [@] in RP readily explains
intrusive-r. The resultant sound is the off-glide of centring
diphthongs.  As with RP, [@r]/[&] became the off-glide for the
GA centring diphthongs.  Did they form a secondary system, like RP?

The 'loss' of rhoticity in GA, the maintenance of retroflexed off-glides
and a secondary vowel system is what I propose.  At least for comparative
purposes, a *very* large part of my work.  At any rate, it helps explain
my data.

Is [&] one sound or two?  Phonetically, one.  When trying to account for
metathesis, it has to be two, somewhere in the phonology.  If you're
talking about a variety where [r] doesn't hop into a syllable-onset
somewhere, the choice is yours.

--Aaron


________________________________________________________________________
Aaron E. Drews                               The University of Edinburgh
aaron at ling.ed.ac.uk                  Departments of English Language and
http://www.ling.ed.ac.uk/~aaron       Theoretical & Applied  Linguistics

"MERE ACCUMULATION OF OBSERVATIONAL EVIDENCE IS NOT PROOF"
        --Death