The effect of rhythm on inter-gestural coupling of onset and vowel gestures and predictive timing in stuttering

The child-specific phenomenon of preferential neutralisation in initial position, which reverses a positional bias attested across adult grammars, represents a long-standing problem for formal models of developmental phonology. In a phonetically based model of phonology, child-specific phonological patterns may emerge as the consequence of physical differences between child and adult speech. This paper presents new case-study data suggesting that a child-specific pattern of fricative neutralisation in initial position has its roots in children's articulatory limitations. Coarticulated fricative and vowel gestures are shown to require independent control of the tongue and jaw, known to be problematic for developing speakers. Substitution errors affecting fricatives are analysed as a phonologised reflex of this phonetic pressure to avoid overlapping vowel and fricative gestures. The positional asymmetry emerges as the consequence of the differing degrees of gestural overlap permitted in syllable-initialvs. syllable-final position, as encoded in the framework of Articulatory Phonology (Browman & Goldstein 1986).

There is a growing body of empirical data supporting a gestural‐based view of speech production in which adjacent speech gestures temporally and spatially overlap one another. Recent work suggests that certain acoustic measures are sensitive to changes in the overlap of adjacent consonant and vowel gestures [G. Weismer et al., J. Phon. 23, 111–126 (1995); E. Zsiga, J. Phon. 22, 121–140 (1994)]. Utilizing a graded speaking rate task, eight speakers produced 160 repetitions of ten target words embedded in a carrier phrase. F2 onset frequency, measured at the consonant–vowel boundary of test syllables, was used to index the degree of spectral‐temporal overlap of adjacent consonant and vowel gestures. Regression analyses were used to evaluate the extent to which F2 onset predicted temporal variability in F2 formant trajectories due to speaking rate change. Results suggest that consonant and vowel gestures do not simply temporally slide away from each other with slowed speaking rate. Rather, modifications in the form and magnitude of vowel gestures are required to account for the present acoustic data. An empirically based acoustic model of variation in gesture overlap induced by speaking rate change is offered. Individual speaker differences also are discussed.

High vowels in Tokyo Japanese are typically devoiced between voiceless obstruents. Controversy remains over whether vowel gestures in devoiced vowels persist or are deleted. A previous ultrasound study (Iwasaki et al., 2020) showed that the lingual articulation of the /kV/ release differed by vowel context even when devoiced. This study looked at changes in the midsagittal tongue surface over time by quantifying the shape and location of the tongue within the vocal tract. Native speakers of Tokyo Japanese produced items (/kVC2e/) that contrasted in the voicing of V, which was either /i/ or /u/. Tongue contours were quantified by Discrete Fourier Transform (DFT), using the approach of Liljencrants (1971). Time series of phase and magnitude of the first Fourier coefficient were compared by vowel context (/i/ vs. /u) and vowel voicing (devoiced vs. voiced). These showed the presence of the vowel contrast whether vowels were devoiced or voiced, suggesting that some devoiced vowels retain lingual gestures. For /u/ only, we found a voicing-induced difference, consistent with reduction or deletion of this vowel gesture. DFT can be a useful tool to analyze tongue contours, supporting the analysis of tongue shape evolving over time.

Two approaches to seeking stable patterns in the gestural organization of speech are examined: local organization (individual gestures coordinated with other individual gestures) and global organization (gestures forming larger conglomerates). Articulatory evidence from American English words with a variety of initial consonants and clusters shows that syllable-initial consonants form a global organization (indexed by a metric we term the C-center) that is coordinated with the syllable's vowel gesture. For syllable-final consonants, however, the evidence suggests that a local organization is employed: The first postvocalic consonant gesture is coordinated with the vowel gesture. Implications of these different styles of organization for the perceptual and phonological structure of speech are discussed.

The transition portion of the speech signal has been identified as critical to the perception of both consonants and vowels. Furui (1986) demonstrated that for adult Japanese speakers a 10 ms segment of the transition centered on the area of maximum spectral movement contained the most critical information for joint consonant and vowel perception. The distribution of phonemic cues contained within consonant-vowel (CV) non-reduplicated (NRB) syllables of infants ages 6 months to 15 months was compared to the distribution observed in adult productions to examine the extent of coproductive overlap of infant consonant and vowel gestures. Truncated versions of infant CV syllables in which initial and final portions of the syllable were systematically deleted were presented to adult listeners for syllable identification. Perceptual critical points (truncation point at which syllable identification fell below 60% for the first time) were determined for both initial and final truncation conditions, and critical intervals were calculated. Results revealed significant differences in the temporal distribution and strength of phonemic cues within the infant NRB CV syllable. First, perceptual critical points could only be determined in approximately 2/3 of the 60 NRB syllables included in the perceptual study, and fewer than 1/3 of the infant syllables demonstrated normal, adult-like identification functions under both initial and final truncation conditions. Obviously, the infant syllables did not manifest the level of cue redundancy or cue constancy seen in adult syllables. Second, infant critical intervals were on the average 2 1/2 times longer than those of adult Japanese speakers reflecting an overlap of the separate consonant and vowel gestures that is less compact than in adult syllables. Results also failed to identify a clear developmental progression towards either more adult-like critical interval durations or identification functions in the NRB syllables studied.

The effects of voicing, manner, and number of consonants on the duration of previous vowels in English can be created in three ways: by planned expansion or contraction of the entire vowel, by changing the relative timing of consonant and vowel gestures, or by modifying the duration of the closing movement for the consonants. An X-ray microbeam data base of 2 speakers of American English reciting monosyllabic words, which differed in their final consonants and in their position in the accentual structure of the utterance, was studied to evaluate these three accounts. In this data base, vowel duration differences due to the voicing of the following consonants were amplified by the presence of accent. Also, many of the duration differences associated with the following consonants are localized in the later portion of the opening movement, suggesting that subjects often create vowel duration changes by initiating the consonant closing gesture at an earlier or later time relative to the opening gesture.

This paper presents examples of consonant‐vowel interactions collected from a group of Edinburgh children presenting with developmental phonological disorder. Three types of interaction are described: vowel conditioning of consonant production, consonant conditioning of vowel production and use of consonantal material to maintain vowel contrasts. Evidence of context conditioning in both consonant and vowel error patterns highlights the potential importance of assessing the child's sound system as a whole. It also raises the question of how far coarticulatory constraints play a role in phonological disorder. In this paper, the data are interpreted, where possible, in terms of production constraints and the degree of compatibility between consonant and vowel gestures. This approach represents one particular perspective and is not intended to preclude other perceptual and/or phonological accounts of the data.

Some questions concerning the underlying nature of the so‐called P‐center phenomenon (the acoustic anisochrony of rhythmically regularly produced/perceived sequences of monosyllables) are discussed in the light of recent EMG data from experiments with rhythmically produced /Cak/ syllables (with /C/ = /p/, /pf/, /pl/, /pt/, /pfl/). The differences between /p/ vs /pf/ syllables resemble the timing differences in the microbeam data of Browman and Goldstein [Haskins SR‐93/94, 85–102 (1988)]. Whereas they found a stable timing relation between the “articulatory mean” of the initial consonants (termed C center) and the acoustical offset of the following vowel, in terms of their task dynamic description, however, the phase relations between consonant and vowel gestures should remain the same. This would result in a reduced stiffness of the vowel gesture for the syllables containing consonant clusters. The observations of the present study would also correspond to a reduction of stiffness for the vowel but the numerical data seem to contradict a stable phasing relation. In connection with the psychoacoustic interpretation of the P‐center phenomenon, an argument is made against a strictly intrinsic timing interpretation (based on articulatory regularities) of the effects of rhythmical production/perception.

The view that individual speaker characteristics reflect morphological and functional differences in the speech apparatus has been accepted as self-evident despite the lack of convincing empirical data. A recent study on the issue of individual differences based on x-ray microbeam data from ten American English and ten Japanese speakers (five males and five females in each group) has pointed to a significance of morphology in accounting for speaker characteristics of vowel gesture. The midsagittal form of a speaker’s orofacial structure was parametrized by several landmarks on x-ray scan images. A few points and lines were chosen to form a quadrilateral framework so that they define a geometrical ‘‘articulatory space’’ where articulation of the tongue–jaw complex takes place. Pellet data for vowels were mapped onto speaker’s articulatory space to observe the effect of morphological difference on vowel articulation. Preliminary results have shown the extent of morphological variation of the speech apparatus and adaptability of vowel gesture to different forms of the space. The aspect ratio of the articulatory space was judged as the primary causal factor of articulatory variation. Acoustic manifestation and perceptual contribution of the variation will be discussed from a viewpoint of auditory-articulatory linkage.

In Tokyo Japanese, high vowels /i/ and /u/ are frequently devoiced when they are surrounded by voiceless obstruents. Controversy remains over whether vowel gestures in devoiced vowels are retained or instead deleted. Both static (Iwasaki et al., 2020) and dynamic (Iwasaki et al., 2022) ultrasound data have indicated that vowel-specific lingual gestures can persist even when vowels are devoiced. This study focuses on the lip articulation of devoiced vowels by examining lateral lip aperture, where lower values index a greater degree of the rounding of vowels. Native speakers of Tokyo Japanese produced nonce word pairs with the form of /C1VC2V2toko/. V1 was either /i/ or /u/. C1 and C2 were either voiced or voiceless, which determined the voicing of V1. Lateral lip aperture during the first mora was calculated by identifying facial landmarks using OpenFace 2.0 (Baltrusaitis et al., 2018), and compared across vowel quality (/i/ vs./u/) and vowel voicing (devoiced vs. voiced). Preliminary results show that lateral lip aperture is larger for /i/ than for /u/ in both devoiced and voiced environments, indicating that these vowels maintain their labial specifications even when devoiced, providing additional evidence that devoiced vowels can retain their articulatory gestures.

To examine the generality of Strange's Dynamic Specification Theory of vowel perception, two perceptual experiments investigated whether dynamic (time-varying) acoustic information about vowel gestures was critical for identification of coarticulated vowels in German, a language without diphthongization. The perception by native North German (NG) speakers of electronically modified /dVt/ syllables produced in carrier sentences was assessed using the "silent-center" paradigm. The relative efficacy of static target information, dynamic spectral information (defined over syllable onsets and offsets together), and intrinsic vowel length was investigated in listening conditions in which the centers (silent-center conditions) or the onsets and offsets (vowel-center conditions) of the syllables were silenced. Listeners correctly identified most vowels in silent-center syllables and in vowel-center stimuli when both conditions included information about intrinsic vowel length. When duration information was removed, errors increased significantly, but performance was relatively better for silent-center syllables than for vowel-center stimuli. Acoustical analyses of the effects of coarticulation on target formant frequencies, vocalic duration, and dynamic spectro-temporal patterns in the stimulus materials were performed to elucidate the nature of the dynamic spectral information. In comparison with vowels produced in citation from /hVt/ syllables by the same speaker, the coarticulated /dVt/ utterances showed considerable "target undershoot" of formant frequencies and reduced duration differences between tense and lax vowel pairs. This suggests that both static spectral cues and relative duration information for NG vowels may not remain perceptually distinctive in continuous speech. Analysis of formant movement within syllable nuclei corroborated descriptions of German vowels as monophthongal. However, an analysis of first formant temporal trajectories revealed distinct patterns for tense and lax vowels that could be used by listeners to disambiguate coarticulated NG vowels.

High vowels in Tokyo Japanese are typically devoiced between voiceless obstruents, but controversy remains over whether vowel gestures persist when devoiced or are instead deleted. A previous ultrasound study (Iwasaki et al., 2020) showed that the lingual articulation of the release burst of /kV/ differs by vowel context even when the vowels are devoiced. This study uses tongue surface contours derived from midsagittal ultrasound images to investigate the effects of vowel devoicing on changes in quantified tongue shape over time. Native speakers of Tokyo Japanese produced word pairs (/C1VC2e/) that contrasted in the voicing of V, which was either /i/ or /u/. Tongue shape was characterized by Fourier transforming tangent angles along each contour (Dawson et al., 2016). Time-normalized trajectories over the /C1VC2e/ sequence were compared by vowel context (/i/ versus /u/) and voicing environment (devoiced versus voiced). Preliminary results show that the real component of the first Fourier coefficient is sensitive todetecting evolving shape differences between the two vowel contexts over the sequence, not just when the vowels are voiced, but also when they are devoiced. Based on these results, the high vowel contrast persists even in the devoiced environment, suggesting that devoiced vowels are not deleted.

The distribution of phonemic cues contained within consonant–vowel (CV) nonreduplicated (NRB) syllables of infants ages 6 to 15 months was compared to the distribution observed in adult productions [S. Furui, J. Acoust. Soc. Am. 57, 1016–1025 (1986)] to examine the extent of coproductive overlap of infant consonant and vowel gestures. Truncated versions of infant CV syllables were presented to adult listeners for syllable identification. Perceptual critical points (truncation point at which syllable identification fell below 60% for the first time) were determined for both initial and final truncation conditions, and critical intervals were calculated. Results revealed significant differences in the temporal distribution and strength of phonemic cues within the infant NRB CV syllable. Fewer than 1/3 of the infant syllables demonstrated normal, adult‐like identification functions under both initial and final truncation conditions, and infant critical intervals were on the average 21/2 times longer than those of adult Japanese speakers. Results did not reveal a clear developmental progression towards either more adult‐like critical interval durations or identification functions in the NRB syllables studied.

Korean lenis stops have been reported to become voiced intervocalically (Kagaya, 1974) and are often reduced to sonorants in casual speech with different degrees of reduction depending on the place of articulation (Lee, 1995). Velar lenis stops are more often reduced than the labials and coronals, which makes sense from the point of view of articulatory phonology (Browman and Goldstein, 1990). The oral gesture for a velar stop is on the same vocal tract tier (tongue body tier) as the neighboring vowels, so it should be more affected by overlap and blending with the vowel gestures than the closure gestures for other stops. This study investigates in what vowel environments lenis stop closure shows the most reduction. /VCV/ tokens in carrier sentences were recorded and acoustic analysis was done. The results showed that generally, reducing lenis stop closure is most in the environment of preceding and following low vowels, and least in the environment of high vowels, where the tongue body gestures of neighb...

Previous acoustical research [e.g., C. Fowler, Phonetica 38, 35–50 (1981)] has suggested that consonants and vowels are not produced strictly sequentially but rather overlap in time. By this account, the degree of co-production of underlying segments affects the pattern of the resulting acoustical durations. In the present experiment, two subjects produced utterances of the form CVC(C), where the vowels were either /ε/ or /æ/ and the following consonants were /p/, /k/, /s/, /ps/, /ks/, /sp/, or /sk/, Movements of the lips and jaw were monitored with a Selspot system. Laryngeal movements were measured by means of transillumination, and the accompanying acoustic signal was recorded. Analysis of the acoustic durations revealed that vowel duration was shorter when followed by a cluster than by a singleton in both vowel contexts. However, the following consonant durations were longer for the vowel /ε/ than for /æ/. Analyses of the lip and jaw kinematics suggested that the vowel shortening in the cluster contexts was produced in part by greater temporal overlap of the underlying segments. The influence of the intrinsic vowel duration on the following consonants may reflect a syllabic constraint on articulatory timing. The results will be discussed in terms of the articulatory structure of phonetic segments and the sequential coordination of consonant and vowel gestures. [Work supported by NINCDS.]