Articulatory adjustments in initial voiced stops in Spanish, French and English

This work reports cross-languages differences in the voicing of utterance-initial voiced stops, and in the use of active maneuvers to achieve closure voicing, using correlated aerodynamic and acoustic data. Oral pressure, oral and nasal flow, and acoustic data were obtained for utterance-initial /b d p t m/ for 10 speakers of Spanish, 6 speakers of French and 5 speakers of English. Voiced stops were first classified as prevoiced or devoiced. Then they were classified by shape of the oral pressure pulse and/or occurrence of nasal flow or oral flow during the stop closure in an attempt to relate aerodynamic data to motor adjustments to facilitate voicing. Such adjustments were found to be related to (i) language-specific differences in the use of glottal vibration as a cue to the voicing-distinction, (ii) place of articulation, and (iii) speaker dependent variation. Voiceless stops showed no such active maneuvers except nasal leak (i.e. nasal closure following oral closure). Comparison of the timing of oral-velic closure in voiced and voiceless stops showed that nasal closure took place later in voiced than in voiceless stops. The longer nasal leak in voiced compared to voiceless stops is argued to be related to voicing initiation and maintenance. Finally, we seek to find acoustic evidence of articulatory adjustments to lower oral pressure for voicing. A correlation is found between oral pressure and voicing amplitude during the stop closure in the three languages: as oral pressure rises, voicing amplitude decreases. Thus the time course of voicing amplitude during the stop closure allows us to infer whether (any) motor adjustments to keep a low oral pressure for voicing are present but not specifically which ones.