Tongue activity in compensatory speech: The use of articulatory models to interpret acoustical data

Abstract

The question addressed was whether, in compensatory speech, the vowel following a consonant articulated with an advanced or retracted tongue placement, would also show evidence of an advanced or retracted tongue posture. Palotographic data were used to determine tongue contact region for consonants. An articulatory model of tongue configuration [Harshman, Ladefoged, and Goldstein, J. Acoust. Soc. Am. 62, 693 (1977)] was used to interpret vowel formant frequency data for vowels. Additionally, vowel formant frequencies were adjusted to account for the effects of known compensatory jaw position. With jaw position corrections included the results were that vowels following a retracted or advanced articulation of sibilant and alveolar stop consonants, showed a similar retracted or advanced set of the tongue. This was not the case for vowels following /l/. Without adjustments for jaw positon the model predictions did not indicate a similarity in tongue set for vowels and consonants for any of the CV syllable types. [Work supported by NIH Grant DE‐03631.]