On Mon, 26 Jun 2000 21:21:12 -0400, Richard Fontana
<re...@columbia.edu> said:

[In a discussion of the vowel symbols used in an article by William
Labov at <http://www.ling.upenn.edu/phono_atlas/ICSLP4.html>.]

Some day soon I will probably journey to a local university library,
which is only two or three miles away from here, to see if they have
some of Labov's books.  I hope to find that in at least one of them he
has a table defining his symbols in articulatory terms.

I might buy one of his books, but I have little doubt that it would
tell me a great deal more about vowel shifts than I really want to
know.

In that last sentence lies the great, fundamental difference between
your interpretation and mine.  I would be astonished to learn that he
does not continue to describe the unshifted sound with the same symbol
as before, and I would fully expect him to say that after a sound has
shifted he will represent it by a different symbol corresponding to
its new position.

Yes, he says things like that at the Web site I've mentioned.  But I
feel certain that what he means is that some words previously
pronounced with /o/ are now pronounced with /ae/, while some words
that were previously pronounced with /ae/ are now pronounced with a
vowel that is represented by yet another symbol.  It takes fewer words
and less effort to say that /o/ has shifted to /ae/'s former space,
but it can be quite misleading, if my interpretation is correct.

From looking at his examples I gather that he uses /ie/ to represent
the sound of ASCII IPA /i/ and /i/ to represent the sound of ASCII IPA
/I/.  (It doesn't help much when he says that /ie/ stands for the
vowel in 'idea', since 'idea' contains at least three vowels (four,
counting the two in the diphthong of 'i'), but I think he must be
referring to the *stressed* vowel in 'idea'.)  

Incidentally, I question the accuracy of your implication that all of
the vowels that have moved are 'short'.  Although I dislike using the
terms 'short' and 'long' in the Thistlebottomian sense, I would say
that in that sense ASCII IPA /i/ (the vowel of 'feet') is long, and
ASCII IPA /I/ (the vowel of 'fit') is short, and they are both
involved in the Northern Cities Vowel Shift.  

But only part.  As I understand vowel shifts, including the one that
reportedly screwed up English spelling in Caxton's time, all of the
vowels involved move in order to preserve former contrasts.

I feel quite certain that when he speaks of the merger of /o/ and /oh/
he does *not* mean that sounds formerly represented using /o/ and /oh/
can now be represented using either of them, as you seem to be
assuming.  He must mean something like:  

   Some words that were formerly pronounced with /oh/ are now
   pronounced with /o/, and some words that were formerly
   pronounced with /o/ are still pronounced with /o/.

When the AUE FAQ says that /A/ and /O/ have merged, it doesn't mean
that the two *symbols* have merged, it means that some words that were
formerly pronounced with /O/ are now pronounced with /A/, while some
words that were formerly pronounced with /A/ are still pronounced with
/A/.

No, it can't mean that.  It must mean that while some other US
speakers use different vowels in 'cot' and 'caught', Western speakers
use the same vowel in 'cot' and 'caught'.  While other speakers
pronounce 'cot' and 'caught' /kAt/ and /kOt/, respectively, Western
speakers pronounce them both /kAt/.  But from Labov's map we now know
that this generalization is too sweeping, although it applies very
well to my speech.  

Actually, the statement 'Western speakers have merged "cot" and
"caught"' is quite misleading.  It suggests that Western speakers
formerly pronounced words one way and now they pronounce them another
way.  I think that any actual merger probably occurred in places like
Scotland -- or in places the Scots came from -- many generations
before I was born, and that from the time I first began to speak I
learned to pronounce 'cot' and 'caught' with the same vowel.

I have no trouble at all hearing the difference between /kAt/ and
/kA.t/ when the speaker is fully rounding his lips to pronounce /A./.
A difference between my *speech* and that of Markus Laker is that he
pronounces many words with /A./ that I pronounce with /A/, while he
pronounces many words with /A/ that I also pronounce with /A/, and I
have no trouble hearing which vowel he's using.

I understand now why you would say something like that, but as I've
explained I don't think it's true.  I think his symbols probably stand
for vowel sounds that are as fixed as are the sounds corresponding to
symbols in IPA.  It's the sounds that have moved, not the symbols.

It would be great if William Labov himself or one of his disciples
could be persuaded to comment on this question.  It couldn't hurt to
send him e-mail asking for a clarification.  I think the worst he
could do would be to ignore us.

It's possible that his symbols do move around to correspond to
different articulatory realizations, but I would be disappointed to
learn that they do.  It would make his symbols seem worthless to me.

With some uneasiness I'm going to crosspost this to sci.lang.  Often
when I poke a limb into sci.lang I pull back a bloody stump, but maybe
this time will be different.  It would be quite helpful if someone
over there could provide a table of definitions of Labov's symbols in
articulatory terms, or could point to a specific place where such a
table can be found.  I would also hope to evoke some statements of
informed opinion about the meaning of statements like '/oh/ has
moved'.