Band 38, Heft 1-3

Existing models of speech production and coarticulation have failed to account for observations of real speech because they have considered timing to be a by-product of articulatory events instead of an integral organizing parameter of the speech motor plan. The model offered here considers time and timing relationships to be intrinsic to speech motor organization and the units of speech to be inherently dynamic gestures rather than static vocal tract configurations or invariant commands to the articulators.

Most work on speech motor control has been devoted to the spatial and temporal coordination of articulatory movements for successive units, segments or syllables, in the speech chain. An intrasegmental temporal domain has generally been lacking in speech production models, but such a domain is necessary at least for certain classes of speech sounds, e.g., voiceless obstruents, clicks, ejectives. The present paper examines the nature of laryngeal-oral coordination in voiceless obstruent production in different languages using the combined techniques of electromyography, transillumination and fiberoptic filming of the larynx together with aerodynamic and palatographic records for information on supralaryngeal articulations. The results suggest that laryngeal articulatory movements are organized in one or more continuous opening and closing gestures that are precisely coordinated with supralaryngeal events according to the aerodynamic requirements of speech production.

A comparison of the literatures on coarticulatory influences of a stressed vowel on consonants and unstressed vowels and on ‘compensatory shortening’ of stressed vowels in the contexts of consonants and unstressed vowels suggests that the two timing effects may be related. An experiment was conducted to provide an explicit comparison and to test a hypothesis that compensatory shortening aand coarticulation are not separate timing phenomena. The two timing effects were found to pattern very similarly and both to be predicted by Lindblom and Rapp’s compensatory shortening formula. 5 of 6 subjects showed significant correlations between coarticulation and shortening.

This research attempts to define the temporal extent of anticipatory labial coarticulation and the effect on rounding onset of a stressed nonround vowel located temporally prior to the rounded vowel. The results suggest that the onset of the labial rounding gesture is controlled in relation to the temporal extent of the nonlabial duration preceding the rounded vowel and that the onset of activity for the rounded vowel is somehow related to the stress of the preceding nonround vowel. These results argue against the classic articulatory syllable model as well as against a concept of the onset of rounding being temporally locked to the rounded vowel. They can be argued to support a look-ahead or scanning model.

Recently obtained data from X-ray microbeam experiments indicate inherently multidimensional articulatory phenomena with respect to temporal characteristics of speech. Elementary gestures in different articulatory dimensions for phonetic elements, typically representing demisyllabic transitions, constitute the content of a phrasal frame.

We studied American English speech rhythm using reiterant speech with the prosody of normal speech but without its segmental variations. Syllable duration was strongly influenced by stress and by final positions in words and phrases, but negligibly by nonfinal positions within words and phrases. Word duration increased linearly with word size, and varied with stress pattern for words of constant size. Metrical foot duration increased approximately linearly with foot size. Feet of the same type and size were of constant duration.

Two experiments were conducted to examine the manner in which utterance-final lengthening is exhibited by the durations of words containing various fricatives and vowels in different phonetic environments. In experiment 1, utterance-final lengthening was measured for words containing the fricatives /f/, /v/, /s/, and /z/ in word-initial, word-medial, and word-final positions in both phrase-final and utterance-final locations. The results showed small amounts of utterance-final lengthening for words containing fricatives in word-initial and word-medial positions, with much greater elongation when fricatives appeared in word-final position. In experiment 2, utterance-final lengthening was measured for words containing a word-medial long versus short vowel followed by a word-final voiced versus voiceless stop consonant. The results showed a greater amount of utterance-final lengthening for words containing a terminal voiced versus voiceless stop. The findings are discussed in terms of constraints on the speaker’s programming of segment durations in sentence contexts.

EMG data from the musculus palatoglossus and the m. levator veli palatini as well as velograph recordings of velic movement in the production of plosives in German are discussed. The results point to the articulation of lenis plosives being controlled by a slower raising and a faster lowering of the velum than in the case of fortis plosives, within a fairly constant time frame for the complete movement. A simple neural programme is postulated consisting of a control for the macro-timing and for two movement patterns within it.

Many perceptual and acoustic studies of word-final cognate opposition have relied mainly on isolated synthetic utterances and have usually concluded that one or another of the potential cues is primary. Recent studies of natural speech segment durations in samples of connected discourse have raised questions about the efficacy of such cues. In the experiments reported here, naturally produced tokens of the word peg were excised from two different contexts and acoustically analyzed. One representative token from each context was then edited in several ways in order to ascertain the cue values of a number of variables. The stimuli were tested in isolation, in their original contexts and in contexts from which the word peck had been deleted. The results of the experiments indicate that a number of variables can be sufficient cues to cognate opposition, and that the efficacy of a particular cue depends in a large measure both on the context from which it was extracted and on the context in which it is heard. The potentially large number of cues, their temporal distribution and their variable effectiveness suggest that future research should be focussed on discovering and describing a general articulatory strategy underlying both the production and perception of cognate oppositions.

Mechanisms of speech rate control are discussed from the point of view that unidimensional models are inadequate to explain the extensive physiological and acoustical effects observed during changes in speech rate. Evidence is presented to demonstrate that changes in speech rate can result in changes in segmental duration, articulatory displacement, atriculatory velocity, and intrasyllabic coarticulation, and that these changes represent nonlinear transformations produced by a reorganization of speech motor strategies.

This paper reviews a series of studies on the effects of variation in speaking rate on phonetic perception, in particular, on the identification of /b/ and /w/ in syllable-initial position. The major finding of these studies was that in a variety of tasks listeners adjusted for speaking rate when using the duration of the initial formant transitions to distinguish between /b/ and /w/: At slower rates of speech, a longer transition was needed to hear /w/ rather than /b/, and conversely at faster rates of speech. This effect occurred when rate was specified by the duration and acoustic-phonetic structure of the syllable containing the target consonant and, to a more limited extent, when it was specified by the duration of a subsequent syllable. Moreover, the influence of syllable duration on the processing of transition duration was also evident in young, prearticulate infants, suggesting that at least the rudiments of a system that provides perceptual constancy across rate is part of the linguistic endowment of the infant.

In two experiments, the minimal duration of silence between [s] and [lrt] required to hear ‘split’ was shown to vary. More silence was required both when the durations of the fricative and vocalic segments were increased and when the intensity fall-time of the fricative was reduced. The effect of segment duration is consistent with demonstrations of perceptual sensitivity to variations in speech rate found for other contrasts distinguished by temporal parameters. The effect of intensity fall-time can be added to the set of temporally distributed acoustical consequences of the articulation of a stop consonant to which perceptual sensitivity can be demonstrated.

For a set of formal rules for the phonetic interpretation of speech during its perception one should have data concerning a possible structure and parameters of a subjective mechanism for processing temporal information, from segmenting the speech signal to making phonetic decisions. At present such data cannot be found in the literature. An attempt has been made to find out more about some parts of the subjective mechanism of duration measurement and comparison by studying the perception of tempo changes in short sound sequences. Running or ‘momentary’ tempo is shown to be the feature of tempo change perception, and possible procedures for its detection have been tested. A computational technique for getting the quantitative parameters of the mechanism is briefly described.

Phonological questions of the sort ‘Is segment or feature [x] a surface-phonetic event or a property of the underlying mental representation?’ may be answered in some cases by considering the durations of the speech sounds in the word in which [x] appears. Here the status of the epenthetic stop [p] in words such as teamster [––t^himpstÉš] was clarified by measurement of the duration of the preceding vowel + nasal (VN) sequence: a shortened VN implies the [p] is underlying; a long VN implies that it is a surface-phonetic event.