On 05-03-03 12:34, tgpedersen wrote:
> So I proposed that the American
> retroflex /r/ came from a substandard Dutch-based 17-18th c. New
> York dialect, which became the language of the new immigrants to New
> York, whence it spread as the standard dialect to the rest of the US
> (minus New England, which drops /r/'s, and Southern, which does
> likewise. Miguel objected strongly to that idea.
American /r/ varies in the phonemic continuum between "retroflex"
(sublamino-postalveolar) and "bunched" (dorso-midpalatal), the latter
variant prevailing by far, despite the popular opinion that American /r/
is retroflex. I suspect that the two alleles, which are all but
indistinguishable to the ear (except that their coarticulatory effects
may be different), naturally tend to occur in free variation. Similar
realisations are common in SW England (in the former "West Saxon" area)
and in Ireland, and I have little doubt that this is where American /r/
came from.
Piotr