Voicing of English coda stops
|From:||Mark Jones <markjjones@...>|
|Date:||Saturday, December 15, 2007, 8:55|
David Peterson is quite right to say that the vowel before a final phonologically
voiced ([+voice]) plosive (i.e. <e> in 'bed') is longer than before a final
voiceless one (<e> in 'bet'), but the difference is a few tens of millisconds.
If you try and produce this consciously as a learner, you're likely to
introduce a sufficient increase in duration that a *phonologically* long vowel
(closer to what you might get in 'bait') is heard. If you're a native Dutch
speaker, you too will have extra vowel duration before a phonologically voiced
plosive, but English is supposed to have exaggerated the durational difference
relative to other languages.
Interestingly it doesn't matter what the actual phonetics of voicing are, whether [+voice]
is fully voiced or voiceless unaspirated.
I'm not sure what Gary Shannon means when he says you can't prolong the closure of
a voiceless /t/ - of course you can, ask an Italian or Finn to say fatto 'fact'
or katto 'roof'! The closure of a /t/ is actually longer than the closure of a
/d/ generally, which may be driven by the fact that voicing during full closure
of the vocal tract is difficult without special measures. This is because
you're trying to vibrate the vocal folds by forcing air through them, but the
closed vocal tract above the glottis results in a rapid equalisation of
pressure above and below and a rapid cessation of the airflow.
So, that's the theory, how to help! Well, it may be that there's no problem for
English speakers to understand you because you may well have some extra
duration to the vowel before *phonological* /d/ etc. anyway, being Dutch. If
it's just style you're worried about, conscious practice may not help. We're
talking very fine-grained details here. One thing you could do is to learn to
pre-glottalise your final voiceless stops. In my accent of (reasonably
standard) Southern British English (SBE), it's not [bEt] but [bE?t], with a
glottal closure at oral closure for the /t/. That makes the final /t/ etc.
stand out from a final /d/, and also tends to help shorten that vowel
(pre-fortis clipping). For long vowels like 'beat' you can do the same. This is
normal for speakers of SBE. It's an empirical question that as far as I know
has not been addressed yet, but I'd guess that it's as much that voice quality
and F0 change that cues voicelessness as anything else for speakers of SBE and
many varieties of American English too these days.
Hope that helps,
Mark J. Jones, PhD
British Academy Post-doctoral Research Fellow in Phonetics
Department of Linguistics
University of Cambridge
Who's friends with who and co-starred in what?