Brazilian Portuguese speakers' perception of selected vowel contrasts of American English: effects of incidental contact

The purpose of this study was to examine non-English-speaking Brazilian adults' perception of American English vowels at various levels of English contact. Specifically, it addressed two two-vowel American English contrasts, /i/-/I/ and /u/-/[upsilon]/, that both occupy the vowel space of one Brazilian Portuguese category, /i/ and /u/, and one twovowel contrast, /e/-/[epsilon]/, that exists in both languages. For reliability purposes, the three contrasts were presented in two different orders to total six contrasts in all. Predictions, based on Flege, (1995) associated discrimination difficulties with this L1-L2 contrast pairing. However, previous discoveries of non-native speakers' sensitivity to subphonemic differences suggested the potential to overcome L1-related perceptual constraints (Wode, 1994). Five groups of ten participants each [aged 20-40] contributed data [N=50]. Three Austin, Texas-resident groups participated: (group A) native American English speakers, (group B) native Brazilian Portuguese speakers with high English contact, and (group C) native Brazilian Portuguese-speakers with medium English contact. Belo Horizonte, Brazil-resident groups numbered two, including native Brazilian Portuguese speakers with medium English contact (group D) and low English contact (group E.) Each participant took a same-different identification test in which the target vowels appeared within minimal and identical pairs. Within-group results for medium and low contact groups associated significantly greater difficulty with the /u/-/[upsilon]/ contrast. Between-group results found significant differences between high and low contact groups for /u/-/[upsilon]/, /[upsilon]/-/u/, and /I/-/i/; insignificant differences between high contact and native English groups appeared for the same contrasts. These overall trends suggested a degree of flexibility for non-native perception in three of four instances as well as a significant pair-wise order effect. These overall findings should not minimize the importance of individual differences. Discussion concluded with calls for greater focus on individual differences (also reflected in Bradlow et al., 1997) and greater awareness of individuals' potentials within language learning contexts.