"di?nt" (with glottal stop)

[Laurence Horn]

At 5:01 PM -0500 11/16/04, Benjamin Zimmer wrote:
>On Mon, 15 Nov 2004 22:14:46 -0800, Arnold M. Zwicky
><zwicky at CSLI.STANFORD.EDU> wrote:
>>
>>just a warning...  the spelling <di'nt> (or similar things) is often
>>used to code a pronunciation in which the intervocalic voiced tap is
>>simply deleted.  not the same thing as a pronunciation with an
>>intervocalic glottalish bit.
>>
>>i suspect that ben zimmer's examples include some with an intervocalic
>>glottal stop and some with no intervocalic consonant at all.  this is
>>not to deny that some of them have glottal stops, possibly from a
>>catchphrase.
>
>At least as far as the catchphrase is concerned, the exaggerated
>glottalization seems to require something along the lines of [dI?In?] or
>[dI?En?] (with at least secondary stress on the second syllable), as
>opposed to [dI?@n?] or [dI?n-?] (where [n-] represents syllabic [n]).  I
>believe this is what the pronunciation spellings of "di-int" and "di-ent"
>are supposed to represent.

Right; that's what makes it perceptually a lot like the "Connecticut"
[but not just Connecticut] rendering of "(New) Britain", "kitten",
"mitten", etc., all with a secondarily stressed -[In].

>
>Even when the intervocalic consonant remains a voiced tap, the vowel of
>the second syllable in the stressed form of "didn't" often sounds like [I]
>or [E] when spoken by young East Coasters.
>
exactly--but let me throw out a purely speculative WAG.  (This isn't
inconsistent with the factors Arnold brings up but might reinforce
them.)

It's been suggested (by Bolinger, and more recently me in a Linguist
List thread a few years ago) that it's not an accident that medial
glottal stops occur in both "uh-oh" and "unh-unh", where the latter
rendering is intended to represent the open-mouth version of the
disagreement marker (the counterpart of the agreement/assent marker
"uh-huh", which crucially has no medial [?]).  The closed-mouth
version of the disagreement/denial/rejection marker, which I won't
even try to represent, also contains a medial glottal stop, and (like
uh-oh and unh-unh) high tone-low tone.  The generalization seems to
be that medial intervocalic [?] (not, of course, the [?] showing up
before initial vowels) is associated with negation, especially in
conjunction with the high/low tone sequence.  (Note that the open-
and closed-mouth affirmative/assent markers, uh-huh and m-hm, both
have low-high tone sequences as well as medial voiceless vowel (i.e.
[h]) or voiceless nasal in place of [?].)  (Arguably, although this
is even more speculative, it could be claimed that both the [?] and
the high-low sequence are partially iconic representations of the
semantic effect of negation.)   Now "didn't" is pretty negative too,
which may (I did say this was speculative) have motivated speakers to
assimilate it to the pattern of those voiceless "kitten" words that
have the medial [?], even though the voiced [d] means it "should"
(ceteris paribus) pattern with words like "hidden" and "wooden",
which as Arnold and I have noted don't transform to -[?In] in the
same way.

Any buyers?

larry