Spatio-temporal patterns of velopharyngeal action in phonetic and phonological nasalization.

The aim of this paper is to differentiate between effects of phonetic implementation and effects of phonological structure in the adjustment of articulatory trajectories to varying speech rate. Cross-linguistic data on coarticulatory nasalization of vowels preceding a nasal consonant at different speech rates were analyzed in American English and Spanish. The two languages show different patterns of timing, magnitude, duration, and velocity of velopharyngeal movements. In Spanish the velocity of velar port (VP) opening and closing gestures is not affected by differences in speech rate, which suggests that these trajectories reflect the default articulatory movements automatically implemented by the phonetic component. In American English, VP closing velocity is not affected by speech rate whereas opening velocity adjusts to variations in speech rate to ensure vowel nasalization across rates. The careful regulation of VP opening velocity suggests that it is centrally controlled. Interarticulatory timing data support this interpretation: In Spanish, vowels are oral for most of their duration; onset of VP opening is timed relative to the following nasal consonant and peak VP size occurs at nasal consonant onset across rates, which suggests that the motor commands for opening the velar port are part of the instructions for the nasal consonant. In American English, on the other hand, VP opening onset coincides with vowel onset and peak VP size occurs in the middle of the vowel across rates, which indicates that opening movements are part of the programming instructions for the vowel. It is argued that, in Spanish, vowels followed by a nasal consonant are targeted as oral and are nasalized as a result of a coarticulatory effect, whereas, in American English, vowels are targeted as nasalized as a result of a phonological rule.