Volume 39, Issue 1

A scaling experiment using dyadic comparisons of vowel pairs was performed to determine if significant perceptual variation existed on an individual level and whether such variation could be correlated with differences among subjects in their production of vowels. Subjects were required to scale six separate sets of vowel stimuli. Each stimulus set consisted of the 36 possible dyads from the vowel set /i, I, ε, æ, a, Λ, o, Æ, u/. Analysis of the resulting data indicated that significant individual perceptual variation was present which could be seen as a function of perceptual structure differences between the subjects and not merely error variance. In addition, such variations were related to aspects of each subject’s acoustic vowel space – indicators of tongue height and advancement during vowel articulation. These results were interpreted as suggesting a link between perception and production.

The point of departure for this investigation were some generally accepted impressionistic statements about durational differences between German and Dutch vowels. Acoustic measurements revealed that differences can best be described with a phonological rule (to be applied in German but not in Dutch) rather than be regarded as inherent. Furthermore it appeared that second-language performance with regard to vowel duration is not necessarily predictable from the duration of pronunciation training.

Statistical relationships among F₁ F₂ and F₃ in vowel segments in continuous speech were investigated with 909 vowel samples from Japanese sentences uttered by 4 adult male speakers. Determined by minimizing the square errors from the measured F₃, two separate planes corresponding to front and back vowels are situated in the F₁F₂F₃ space to obtain the F₃ values estimated from F₁ and F₂. Estimation errors are found to be comparable to difference limens in hearing. Also, uniform/nonuniform formant frequency scaling is discussed in terms of physical differences in the planes for individual speakers.

Initial prevocalic labial, apical and velar stops, as produced by eight 9- to 12-month-old Swiss-German infants, were examined spectrographically for voice onset time (VOT). Results indicate that VOTs for each place of articulation cluster in the 0- to + 30-ms voicing lag range. Comparisons of the present infant data with available results from adult Swiss-German stop productions show the VOT distributions to be very similar. The present infant data for apical stops are also compared with available data from 12-month-old American-English and Lebanese-Arabic infants. Since all three infant populations show similar VOT distributions, one might conclude that the forces of nature rather than nurture determine manner of infant apical stop production.