English listeners’ judgments of Japanese-accented English vowel productions: Relation to acoustic and kinematic characteristics

The present study investigates how monolingual English listeners categorize foreign (Japanese) productions of English tense and lax vowels with the aim of relating vowel identification to the acoustic and kinematic vowel space. American English front vowels (/i/, /ɪ/, /e/, /ɛ/, and /æ/) were elicited in /hVd/ context from ten Japanese adults who learned English as a second language (ESL). Listeners categorized these vowel productions in a six-alternative, forced-choice paradigm (including the choice /ʌ/). Vowels produced by two native English speakers were included in the perceptual task for comparison. Preliminary results showed higher perceptual accuracy for tense vowels than for lax vowels. As expected, vowels from English talkers were more accurately identified than those from Japanese ESL speakers. To determine whether listeners’ perceptual categorization corresponded with acoustic and kinematic patterns, perceptual data for contrasting vowel categories were compared with measured differences in formant (F1/F2) and articulatory space. Pilot data suggest a strong relation between acoustic and kinematic measures, but a weaker correspondence between these two data sources and listeners’ perceptual judgments. Additional factors, including the roles of duration, talker and listener variability, and vowel similarity between languages, will be explored.