>    I have rarely been informed of others' view of my origin. When I
>have been, I was led to believe that I have very little discernable
>accent. [I was, briefly, a radio announcer in college.] I do, however,
>pronounce certain indicative words in ways which reveal several of my
>prior residences. Others may not realize that. [I would have liked to
>be one of the subjects of the expert (name forgotten) who had a radio
>show on which he performed like a carnival weight-judger, identifying
>people's origins from their pronunciation of a short list of "key"
>words like "greasy", "either", "pen", "Mary", "merry" and "marry". He
>usually pegged it exactly, even within a few miles.]

That would be Henry Lee Smith, of Trager-Smith fame.  I don't know
where he was a professor, but he was a lot more involved in public
education than most linguistics professors.  Besides the radio show,
he made a series of educational movies in the 60's that got shown in
all kinds of places; there might still be some copies lying around in
school libraries and media centers -- they could be very interesting.
Though they weren't really about dialectology, but general

Charles Ferguson, who worked with Smith, once told me that he [Smith]
took it for granted that every linguist he dealt with could do what he
did; he once asked Ferguson to give something to "the secretary from
North Dakota", assuming it would be obvious.

And a friend of mine actually was a subject of Smith's, and he got
pinned down like a butterfly - Northern Alabama with an overlay of
the Western Reserve of Ohio.  Which was a synopsis of his boyhood

 The "Trager-Smith" system of representing English phonemes uses a
/dh@ treyg@r-smith sist@m @v repr@zenting inggl@sh fohniymz yuhz@z @/

 bare minimum of special symbols and has considerable latitude for
/ber minim@m @v spesh@l simb@lz @n haez k@nsid@r@b@l laetitud f@r/

 dialect representation, but it was originally designed for typewriter
/day@lekt repr@zenteysh@n, b@t it w@z @rij@n@li d@zaynd f@r taiprait@r/

 usage and has largely fallen out of use in linguistics.  However,
/yuhs@j @n haez larjli fol@n aut @v yuws in linggwistiks.  hawev@r,/

 it's known to all MacinTalk and Speech Manager users, as well as
/its nown tuw ol maek@ntok @n spiych maen@j@r yuhz@rz, @z wel @z/

 others.  Though usually not by that name.
/@dh@rz.  dhow yuhzh@li nat bai dhaet neym./

 -John Lawler                     More grammar
  Linguistics Program   University of Michigan
 "..and, who knows? Maybe the horse will sing."