the curious phonology of wisconsin

[Arnold M. Zwicky]

On Nov 23, 2004, at 11:55 AM, Damien Hall added to the followups on my
original New Yorker quote from Randy Quaid.

no one picked up on quaid's "more pronounced" r's, or on his theory
that wisconsinites "talk from the front of their mouths, because they
don't want to breathe in the cold."  there were some comments on the
parsing of "breathe in the cold", but mostly we've been talking about
that syllable-division thing in "Wisconsin" (which came up on this list
back in october 2003), and in other words, like "mistake".

here's a brief run-down...

1.  (leaving out some messy details...) words of the form ...V1 C1 C2
V2..., where V1 is the lax vowel of the first syllable and V2 is a
tense vowel with primary accent, tend to have three linked properties:
(a) tertiary accent on the first syllable; (b) unreduced V1; (c) C1and
C2 split between the two syllables, C1 closing the first and C2 as the
onset of the second.  as a result of (c), C1 has appropriate
syllable-final allophones and C2 appropriate syllable-initial
allophones.  examples: festoon, pontoon, monsoon, poltroon, pastiche,
Mankato, Monhegan, Matsui,...  and Wisconsin.

2.  productive, semantically transparent, "strong" prefixes, like "new"
re- (recalculate), de- (defang), mis- (miscalculate), and dis-
(distaste) come with tertiary accent, and so have the other properties
in 1 (as appropriate).  the "old", unproductive, semantically opaque,
"weak" prefixes re- (receive), de- (defer), mis- (mistake), and dis-
(disturb) are unaccented, have reduced vowels, and (as appropriate)
have C1 syllabified with the following syllable.

3.  the Oregon Effect: in very frequent, familiar words (like "Oregon"
for Oregonians and "Wisconsin" for Wisconsinites), vowels with tertiary
stress tend to lose it, with concomitant vowel reduction and (where
appropriate) resyllabification; this is what gives the Wisconsinite
pronunciation of "Wisconsin".  the same thing can happen in fast,
casual speech, so that "distaste" can be pronounced, on occasion, much
like "mistake" (though they are usually distinct).  the reverse is
possible, too, but only in hyperarticulate speech, in which each
syllable gets some accent.

so there's the quick story.

an extra twist: some prefix-like elements in names normally come with
tertiary accent and show the other properties in 1; the result is a
possible contrast like the famous one between "McAuley" (with [k]
closing the first syllable, preserving the element "Mc") and "McCawley"
(with a [k] in each syllable, or with the first deleted before another
[k]).  if these names are both familiar to you, they lose the accent on
the first syllable and become homophones.  otherwise, they *can be*
distinct, with the more infrequent "McAuley" likely to preserve
tertiary accent (and first-syllable [k]) and the more familiar
"McCawley" likely to lose the accent (and have no [k] in the first
syllable).  but they can also fall together, with tertiary accent on
the first syllable, and with the [k] of "McAuley" moved into the second
syllable to maximize syllable onsets.  The tendency to maximize onsets
is in part independent of the accent condition.

and now: "Van Ness", the name of a major street in san francisco.  the
name has a tertiary accent on the element "Van"; the first syllable can
end in [n], or have it deleted before the initial [n] of the second
syllable.  if the [n] is deleted, a hearer who knows the naming system
of   english (and of the languages that have contributed to this
system) can recognize the element "Van" ("Va" not being a possibility)
but can't tell whether the primary-accented element is "Ness" (with [n]
deleted) or "Ess" (with [n] moved into the following syllable).  so if
you don't know the street names of san francisco, you could well think
that what you're hearing is "Van Ess".  and lots of people do.  well, i
did, back in 1976.

ok, this last stuff might be impossibly intricate.  but there *are* a
lot of alternative pronunciations, and which ones you have depends on
things like word structure, familiarity/frequency, speech style, and
all that neat stuff.

arnold (zwicky at csli.stanford.edu)