Within- and across-language acoustic variability of vowels spoken in different phonetic and prosodic contexts: American English, North German, and Parisian French

Abstract

This project investigates the acoustic variability of vowels produced in multisyllabic nonsense words /cvC1VC2(v)/ in carrier sentences by speakers of American English (AE), Parisian French (PF), and North German (NG). Variables under examination are (1) immediate phonetic context (C1=b,d; C2=b,d,p,t), (2) sentence prominence (narrow focus versus postfocus), and speaking rate (normal vs rapid). Preliminary results show that for AE vowels, phonetic context produces large differences (2–3 bark) in midpoint F2 values for mid-to-high back vowels, reflecting allophonic fronting of these vowels in alveolar context. Much smaller increases in F2 (∼1 bark) were found for both PF and NG mid/high back vowels in alveolar contexts. Most AE speakers showed little or no effect of sentence focus or speaking rate on the degree of fronting of back vowels in alveolar context. PF speakers produced more fronted mid/high back vowels in postfocused context than did NG and AE speakers. NG and PF speakers showed relatively greater F1 undershoot in low vowels in postfocus context than for AE speakers. These within- and across-language differences in the effect of prosodic and phonetic context on vowel acoustics reflect both language-universal constraints and language-specific rules for contextual warping of vowel spaces.