The interlingual identification of Spanish and English vowels.

Abstract

Spanish/English bilinguals and English monolinguals (30 each) rated pairs of vowels separated by 1.2 s on a nine-point scale. The pairs consisted of two Spanish vowels (drawn from one of three tokens each of /i/, /e/, /a/), two English vowels (tokens of /i/, /i/, /ei/, /ε/, /æ/, /Λ/, /α/), or one Spanish and one English vowel. Both auditory and phonetic factors influenced the ratings. Native Spanish subjects judged pairs consisting of two Spanish vowels to be more dissimilar than the English subjects; conversely, the native English subjects judged pairs of English vowels to be more dissimilar than the Spanish subjects. As expected, both the Spanish and the English subjects judged pairs of vowels that were distant in an F1–F2 space to be more dissimilar than pairs of vowels that were closer together. The Spanish subjects seemed not to distinguish phonetically certain pairs of English vowels (e.g., (/i/–/i/, /α/–/Λ/)) and thus to have rated them as more similar than the English subjects. The study provided little evidence that the Spanish subjects−even those experienced in English−were developing phonetic categories for English /i/ or /æ/. It did provide evidence, however, that the Spanish subjects were learning to distinguish /ε/ from Spanish /e/ (which has an [ε] allophone). As expected, given the larger number of vowels in English than Spanish, the best MDS solution involved more dimensions for the English than the Spanish subjects (three versus two). Spanish subjects who pronounced English well gave greater weight to F1 formant movement than subjects who pronounced English poorly. They also judged pairs of English vowels to be more dissimilar than the nonproficient subjects, suggesting a possible auditory perceptual basis for individual differences in L2 speech learning ability. [Supported by NIH.]