Second International Conference on Spoken Language Processing (ICSLP'92)
Banff, Alberta, Canada
This experiment is an extension of earlier work, [1,2] aimed at understanding the relationship of the principal muscle forces underlying vowel tongue shape. The experiment was done in two parts. A single male talker, TB, was recorded for both speech samples. Stimuli for both parts of the experiment were multiple tokens of disyllables of the form / pVp/, with the vowels /EE, IH, AY, EH, AE, A, AW, OOH, UW/. For Part One of the experiment, electromyographic recordings were made from the principal extrinsic muscles of the tongue using hooked wire electrodes. For Part Two, x-ray microbeam recordings of tongue body, lip, and jaw were made on the same subject for the same inventory. In general, tongue X and Y position correlated for front pellet and front vowels, but the situation is more complicated for rear pellets. While the correlations of extrinsic tongue muscle activity with tongue pellet positions were as one might expect, no vowel in the set can be considered as a lax equivalent of any other, as Wood's  model suggests.
Bibliographic reference. Harris, Katherine S. / Vatikiotis-Bateson, Eric / Alfonso, Peter J. (1992): "Muscle forces in vowel vocal tract formation", In ICSLP-1992, 879-882.