Volume 51, Issue 1-3

In this paper, a deductive approach to the study of vocal tract characteristics, vowel, vowel-to-vowel and vowel-consonant-vowel productions is presented. The distinctive region model, deduced from acoustic theory, is used to develop this approach. The model is characterized by simplicity and efficiency in its handling of the articulatory-acoustic relation. Vowel systems can be explained by means of this approach. A tentative study on the predictability of vowel-to-vowel characteristics is also presented. The obtained results suggest that the human communication system is driven by physical laws of the real world.

With a few exceptions, EMG data are interpreted with reference to the intended output, such as the phonetic description of utterances spoken by speakers. For a more rigorous interpretation, the data should also be analysed in terms of the displacement of the articulators and the acoustic patterns. In this paper, we describe our attempts to calculate the formant patterns from EMG activity patterns via an articulatory model. The value of the model parameters, such as the tongue body position or tongue body shape, is derived from the EMG activities of the specific pairs of antagonistic tongue muscles. The model-calculated F₁-F₂ patterns for 11 American English vowels correspond rather well with those measured from the acoustic signals. What strikes us is the simplicity of the mappings from the muscle activities to vocal-tract configurations and to the formant patterns. We speculate that the brain optimally exploits the morphology of the vocal tract and the kinematic functions of the tongue muscles so that the mappings from the muscle activities (production) to the acoustic patterns (perception) are simple and robust.

Measures of inferred subglottal air pressure, glottal airflow waveform characteristics, sound pressure level (SPL) and the acoustic spectral slope were studied for individual speakers with normal voices. Combinations of different levels of sub-glottal air pressure and varying glottal configurations could result in the same SPL. Relatively high air pressure levels were associated with a steep spectral slope, reflecting a more sinusoidal glottal waveform and a relatively abducted membranous glottis, which would result in damping of F₁. Data suggested that the interarytenoid glottal opening could vary without systematically affecting SPL or voice quality. The results indicate that the principles of production-related economy of effort and physiological, acoustic and perceptual constraints may apply to voice production.

This paper presents the results of an auditory and instrumental investigation into glottal stops and glottalization as boundary markers in German colloquial read speech of a North German non-dialect variety. It also reports on the occurrence of glottal stops and glottalization, outside the domain of word and morpheme boundaries, as reduction phenomena of supraglottal stop articulations. Finally, these sound features are discussed as elements in an articulatory phonology of connected speech in German, providing insights into the general theory of human speech production.

This study examines vertical and horizontal tongue body movements in VCV sequences, where the consonant is a voiced or voiceless velar stop. The movement data were recorded using a magnetic transduction technique in two subjects. Consistent with studies of lip and jaw kinematics, the duration of the tongue body raising movement towards closure for the consonant was longer for the voiced stop. In contrast to lip and jaw movements, peak velocity and amplitude of the raising movements were consistently higher for the voiced stop. The larger displacement of the closing movement for the voiced stop was due to a lower starting position of the movement during the preceding vowel. Examination of the tongue body lowering movement for the vowel preceding the velar stop showed it to be longer when the following stop was voiced. Also this lowering movement had a higher peak velocity and amplitude in the voiced environment. These results thus suggest that both the lowering and raising movements in the VC sequence are affected by the voicing status of the consonant. In addition, the second vowel in the VCV sequence showed reliable influences on tongue body movements for the first vowel and the consonant.

Although the schwa sound is by far the most frequent vowel in Dutch, it has up to now been phonetically the most neglected. We used an existing database of vowel sounds from focus words in spontaneous speech and in lexically the same text, read aloud by one male speaker, to analyse durational and spectral characteristics of schwas, and we compared the results with data on schwa diphones used in Dutch text-to-speech synthesis. It turned out that, contrary to what is usually thought, lexical schwa sounds in natural continuous speech are considerably shorter than other short vowels, that there is no strong consonantal influence on schwa duration, that schwa sounds display a spectral spread larger than any other vowel, and that surrounding consonants seem to play a role with respect to the midpoint formant distribution of the schwa within the whole vowel system. In no way can the schwa be considered as the ‘bench-mark’ of a speaker’s vowel system.

Audio recordings of lively conversational speech produced by 3 Swedish, 4 Finnish and 3 Estonian speakers were analyzed for durational correlates of quantity distinctions. The data suggest that duration contrasts are maintained more consistently by Finnish and Estonian than by Swedish speakers. This is attributed to the unusually complex structure of the Finnish and Estonian quantity systems, and to the fact that Finnish and Estonian, in contrast to Swedish, do not use vowel quality or diphthongization as correlates to quantity distinctions.

This paper reports some production and perception experiments challenging the idea that the rate of articulation of interpause intervals of speech is solely conditioned by internal phonological factors. The results reported demonstrate that, at least for one professional newsreader, rate of articulation in interpause intervals is conditioned by context and that deviations from the intended rate are easily noticeable and affect perceived naturalness negatively. It is also demonstrated that a context-conditioned rate of articulation does not affect interpause intervals uniformly but is unevenly distributed. These results are interpreted in terms of speaker adaptation to the listener’s running access to signal-independent information.

Across and within languages voiced sibilants tend to be disfavored relative to voiceless ones. This paper explores the claim that voicing more adversely affects the distinctive acoustic properties of sibilants than those of nonsibilants. One prediction associated with this claim is that voicing differentially lowers the amplitude of frication noise for sibilants and non-sibiliants so that amplitude differences between the two classes are reduced. Acoustic measurements confirm this prediction. A second prediction is that voicing has a greater negative effect on the identification of sibilants than nonsibilants. Perceptual results from this and previous studies are somewhat variable, but averaged data support this prediction. The findings suggest that voiced sibilants are disfavored in part for perceptual reasons.

When listeners’ identifications of speech sounds are influenced by adjacent sounds, is it only the quantitative phonetic characteristics of these neighboring sounds that matter, or could their qualitative linguistic identity play a role? We tested this by inducing subjects to ‘restore’ a noise-obliterated medial consonant in VCe utterances by first presenting them with several prior utterances where this medial consonant could be heard clearly and was consistently the same, either a /b/ or a /d/. Included as V were synthetic steady-state vowels from the /i-u/ continuum. More /u/’s were identified out of this continuum in the environment of physically present /d/’s than /b/’s. Restored /d/’s had the same effect, thus indicating that the influence of context need not operate only via physical phonetic features. These results suggest that strict phonetic invariance of phonological units may not be necessary.

The present study extends the phonological realm of locus equations to include nasals, fricatives, approximants and unas-pirated voiceless stops. The question of interest was to determine if locus equation slopes and y-intercepts could serve as phonetic descriptors for place of articulation across varied manner classes. Fourteen speakers, 7 male and 7 female, produced CVC tokens with 10 medial vowels and initial consonants varying across manner classes. F₂ onset and F₂ vowel frequencies were acoustically determined and plotted as locus equations. Fairly distinct clustering of slope and y-intercept coordinates were found for labial, dental/alveolar and velar places of articulation. Approximants (/w, j, r, l/) formed a unique cluster of points characterized by slope = 0 and contrasted solely by y-intercepts – correlates of steady-state onset frequencies of approximant productions. The consistent linearity observed in all locus equation plots across all consonant-vowel tokens is discussed according to the orderly-output constraint that hypothesizes that the robust linearity of the acoustic F₂ transition onset-to-offset relationship exists to best serve neural encoding and perception of the speech signal.

When informally singing memorized instrumental or vocal themes, we tend to invent nonsense texts, e.g. duda duda dudaa. This article presents a first attempt to test the assumption that in this type of singing the choice of syllables is guided by certain principles which reflect musical structure. In one experiment, 11 members of a professional orchestra sang 6 melodical excerpts, and their choice of syllables was analyzed. With regard to the phonetic properties, this choice showed a certain coherence. In a second experiment, one of the subjects showed by articulatory markings such as slurs, dots and dashes his preferred performance of the same 6 excerpts. Comparisons with his syllable choice suggested that the syllables are musically meaningful. In a third experiment, this subject’s performance suggestions were compared with the occurrence of micropauses in rule-generated performances of the same 6 excerpts; the result suggested that the phonetic characteristics of the syllables in this kind of singing are used for conveying information about grouping.

Our understanding of phonological acquisition has benefited immensely from cross-linguistic investigations which allow researchers to separate biological and learned factors. To date, most cross-linguistic studies have focused either on differences in phonetic inventories or on differences in frequency of occurrence of particular phonetic and phonological properties in the adult language. This paper describes a third type of study: comparisons of segments that occur in two (or more) languages but differ in their phonetic properties. We present perceptual and acoustic analyses of adult and child productions of word-initial alveolar /t/ in American English and dental /t/ in Swedish. Results showed that listeners’ perception of place of articulation was strongly associated with language (alveolar: American English, dental: Swedish) for both adult and child tokens, and was effective in assigning individual speakers to language groups. Three acoustic measures, voice onset time, burst intensity and burst spectral diffuseness correlated with language for both child and adult tokens; the latter two measures correlated with perception as well. The findings suggest that American and Swedish children at 30 months of age have acquired some language-specific phonetic aspects of/t/ phonemes.

Phonological structure may be seen as emerging in ontogeny from the combined effects of performance constraints rooted in the neuromotor and perceptual systems, individual lexical development and the influence of the particular ambient language. We review here the nature and origins of the earliest ambient language influences. Global effects within the first year of life include both (1) loss of early appearing phonetic gestures not supported by the ambient language and (2) positive effects, reflecting infant attention to prosody and to cues available in the visual as well as the auditory modality. In the course of early lexical development more specific effects become manifest as individual children pursue less common phonetic paths to which the ambient language provides ‘sufficient exposure’.

Speech signals contain various types of information that can be grouped under the headings phonetic, affective, personal and transmittal. Listeners are capable of distinguishing these. Previous theories of speech perception have not considered this fully. They have mainly been concerned with problems relating exclusively to phonetic quality. The theory presented in this paper considers speech signals as the result of allowing conventional gestures to modulate a carrier signal that has the personal characteristics of the speaker, which implies that in general the conventional information can only be retrieved by demodulation.

Although it would seem fundamental to the understanding of human beings, there has been virtually no attempt, in modern science, to develop a theory regarding the sound pattern of the first language. As a step towards such a theory, I suggest that ontogeny may recapitulate phylogeny. I propose that the pattern of open/close alternation of the mandible underlying the syllable which first appears in infants at the babbling stage is the missing link in the evolution of speech from ancestral primate vocal signals. In addition, I suggest that the conjoint set of sounds and sound patterns favored in babbling and early words, and in the world’s languages, constitutes, in effect, the fossil record of true speech, and I provide a tentative enumeration of this set.