Re: [tied] Re: crows and the glottalic theory

With my experience as a teacher of practical English phonetics in Poland I'd like to add that learning to _avoid_ the full voicing of lenis obstruents in post-pausal positions is an important step in achieving a plausible near-native pronunciation, be it Received Pronunciation or General American-like. Polish is like French & al. in this respect, having fully voiced obstruents after a pause, and they don't work very well for English.

Like German, Polish has no voicing contrasts before a pause (due to obligatory devoicing in this position), but Poles who have mastered the English contrasts tend to overdo the voicing and again have to practise the partial devoicing of /g/ in <dog> in order to avoid unnatural "declamatory" pronunciations. The partial devoicing effects, being subphonemic, may be hard to detect by ear _especially_ to native speakers, whose faculty of hearing was trained to take them for granted in the process of first-langauge acquisition, but they are real enough to anyone familiar with laboratory phonetics. For details, those interested may consult any good handbook of English phonetics that discusses positional variants of phonemes.

Finally, I am amazed to hear that anyone may be entirely unfamiliar with varieties of English that have glottalised stops in words like <cat>, <cap>, <back> or <catch> ([PCR:] "I have never heard a dialect of English with glotallization in this position."). Most British people, including speakers of what can be regarded as the standard educated accent, use prepausal and preconsonantal glottalisation either consostently or at least occasionally.

Piotr

----- Original Message -----
From: Miguel Carrasquer
To: cybalist@yahoogroups.com
Sent: Sunday, November 10, 2002 11:33 AM
Subject: Re: [tied] Re: crows and the glottalic theory


On Sun, 10 Nov 2002 05:19:42 -0000, "ikpeylough"
<ikpeylough@...> wrote:

>I double checked (including using MCV's procedure, next post) myself,
>and I'm certain that full voicing consistently occurs before the
>_release_ of occlusion. After voiceless consonants, voicing tends to
>begin immediately before the _start_ of occlusion, even. Of course,
>when sentence-initial, the sequence of start of voicing and start of
>occlusion is irrelevent.
>
>Perhaps I'm in the minority.

Perhaps so, as I don't think Peter Ladefoged, Scottish as he may be,
would have used "most speakers of American English" lightly, being
professor of Phonetics at UCLA.

It appears you have _partially voiced_ stops in the position after
pause, with some vocal chord activity during the stop closure. Now
compare this with /#b-/, /#d-/ or /#g-/ in languages which have fully
voiced stops after pause, such as French, Spanish or Dutch. The
difference should be clear, which explains why English b-, d-, g-,
even if partially voiced, sound as (unaspirated!) p-, t-, k- in their
native language to speakers of such languages, and why their
(spelling) pronunciation of b-, d-, g- in English (amongst other
things) is a major component in being perceived as "having a [French,
Spanish, Dutch] accent" in English.

In fact, in Ladefoged & Maddieson's "The Sounds of the World's
Languages" there is some discussion that seems to imply that English
so-called voiced stops ("phonologically voiced stops") are not even
fully voiced in a _voiced_ environment (such as between vowels).

Pages 50-51:

"The physiological position for modal voice can be regarded as one in
which the arytenoid cartilages are in a neutral position for speech,
neither pulled apart nor pushed together (Stevens 1988). The vocal
folds would be very slightly apart, if there were no air flow. We
assume that the same position as occurs in ordinary voiced vowels and
in voiced continuant consonants such as nasals is normally maintained
in stops that are phonologically voiced. It is well known that in some
languages, English being a familar example, the vocal folds may not
vibrate throughout the closure for a voiced stop. Even when surrounded
by other voiced sounds, such as vowels, the vocal fold vibration often
ceases shortly after the closure is made and only resumes shortly
after the closure is released. Most English speakers appear to leave
the vocal folds in a constant position throughout such a sequence, but
passive devoicing occurs as the supralaryngeal pressure builds up
behind the oral closure. There area number of maneuvers that can be
made to assist the continuation of vocal fold vibration during an oral
stop closure by expanding the size of the cavity behind the location
of the closure; these include a relaxation of the cheeks and other
soft tissues around the oropharyngeal cavity so that the pressure will
passively expand the volume, as well as active gestures, such as
moving the articulatory constriction forwards during the closure,
moving the root of the tongue forwards, lowering the jaw, or lowering
the larynx (Hudgins and Stetson 1935, Bell-Berti 1975, Ohala and
Riordan 1979, Keating 1984c). Some English speakers utilize such
gestures to a sufficient degree to produce vocal fold vibration during
their voiced stop closures (Westbury 1983), but similar gestures are
often executed by speakers producing intervocalic phonologically
voiced stops without sustained vocal fold vibration (noted by Kent and
Moll 1969). Flege (1982) has shown that the variation in the time at
which vocal fold vibration starts near the release of
utterance-initial voiced stops in English does not depend on how long
before the release the vocal folds are adducted. The target for voiced
stops in English can therefore be said to include the maintenance of a
position of the vocal folds appropriate for voicing, but not to
require the employment of other strategies to sustain vocal fold
vibration.

In contrast to English and several other Germanic languages, a
considerable number of languages have voiced stops which require more
energetic efforts to produce sustained vocal fold vibration. Such
languages include well-known ones such as French and Thai, as well as
more obscure ones such as Ilwana. In languages of this type, the
target in the production of voiced stops must be defined as including
the presence of actual vocal fold vibration through the articulatory
closure period. Figure 3.1 shows the word budda 'pelican' from Ilwana.
This word contains an initial voiced stop and an intervocalic geminate
voiced stop which are both produced with full vocal fold vibration.
This occurs despite the fact that these are both positions in which
sustaining voicing requires particular additional effort, as has been
shown by Westbury and Keating (1986).

In some of the languages in which sustained vocal fold vibration is
part of the target for voiced stops, the downward movement of the
larynx and the other cavity-enlarging movements used are sufficient to
actually rarefy the air in the oropharynx. On the release of the oral
closure, some inward airflow occurs. That is, there is a continuum
between fully voiced stops and implosives. We will discuss this latter
type of stop more fully below; here we merely want to note that
implosives and fully voiced stops are not the same thing."