> AFAIK Southwestern England today has retroflex r's including
> syllable-final, and I believe Shropshire does as well.
Those sites I've seen present this 'retroflex r' in initial position.
The interesting retroflex r is the syllable-final one which produces
the preceding vowel to be r-colored. Do you have a reference on such a
phoneme on the British.
> There may have
> been more areas of England that were rhotic in the past, from which
> emigrants could have gone to America.
The r of which then would have to be retroflex and r-coloring to make
your theory stick. We know there is a retroflex, r-coloring r in
Leids, we know the Dutch colonized parts of the American East Coast
which later became very important, and that they must have made up the
most important non-English-speaking element there, perfectly situated
to deliver elements to pull the incipient American language away from
the British they had freed themselves from. What is the problem? Why
would you rather assume influence from a non-documented and
non-documentable assumed dialect in England than from one we know was
actually there?
Torsten