Breath-Stream Dynamics of Simple-Released-Plosive Production
By Martin Rothenberg
[Bibliotheca Phonetica, No.6, Karger, Basel and New York, 1968 (117pp.)]
Breath-Stream Dynamics is a slightly revised version of the author's 1966 doctoral dissertation of the same name at the University of Michigan. It is an attempt to define a large class of stop consonants in quantitative physiological terms as a set of coordinated gestures at the respiratory, laryngeal, articulatory and velopharyngeal levels.
During the decade preceding the publication of Breath-Stream Dynamics, the quantitative literature in phonetics increasingly emphasized the description and classification of the sounds of speech in terms of the acoustic properties of the realizations of the phonemes of various languages. The publications increasingly used acoustic terms such as formant, formant trajectory, aspiration interval, aspiration spectrum, silence duration and closure voicing interval. This tendency toward acoustically based analysis undoubtedly reflected the post-WWII introduction of powerful tools for acoustic analysis, especially the sound spectrograph at Bell Laboratories and the acoustically based speech synthesizer at the Haskins Laboratories. As a result of this trend, a new scientific discipline developed - namely, acoustic phonetics. Some of the seminal publications from this era were republished in book form as Readings in Acoustic Phonetics (ed. by Ilse Lehiste, MIT Press, 1967)
In the analysis of vowels, the previous physiological descriptions, often in terms of binary oppositions (as open vs. closed, front vs. back, rounded vs. not rounded, and tense vs. lax), continued in parallel with acoustic descriptions. However, in the analysis of stop consonants, the attempts at quantitative physiological description made by pioneers such as Hudgins and Stetson were largely replaced by acoustic descriptions, using measures such as voice onset time, formant transitions, and aspiration spectrum.
Though acoustic models can result in a certain degree of descriptive adequacy
and be quite useful for many practical applications, for stop consonants they
do not provide the explanatory power of a physiological dynamic model. This
is because the production constraints are primarily physiological, with the
caveat that in order to be learned, the phonetic distinctions predicted by the
model must be perceptually differentiable in addition to being physically producible.
(The primacy of production constraints after the initial learning phase is illustrated by the fact that post-lingually deaf speakers retain their speech ability to a great degree, even though they needed their hearing in the initial learning phase.)
When the physiological limitations for the above-mentioned four coordinated gestures are considered, both for the gestures themselves and for the ability of the speech mechanism to consistently coordinate the various gestures, it is found that the phonetic structure of the stop consonants of various languages arises as a natural result. Examples of this explanatory power are given from English, Korean and Indic languages.
To measure certain parameters of the model and to show the application to various languages, a number of innovative techniques were employed, including the measurement of subglottal pressure from the interpolation of supraglottal (intraoral) pressure during the closure of an unvoiced stop, and the bypassing of the articulatory closure by a small tube as a minimally invasive method ascertaining the state of the glottis during the period of articulatory closure.
* * * *