Keidel, J. L. 1 , Mora, J. C. 2 & Flege, J. E. 3
1 Bangor University
2 Universitat de Barcelona
3 University of Alabama at Birmingham
Research on the acquisition of L2 phonology has focused largely on late learners’ inability to acquire contrasts in their new language, with special attention paid to interference effects from L1 phonology. Numerous recent studies have examined difficulties that non-monolingual speakers of Catalan experience in perceiving the close and open mid-vowels /e-E/ and /o-O/. These results are often taken as evidence of a vowel space acquisition mechanism that is operative predominantly during the first year of life.
Yet converging sources of evidence suggest that these distinctions are not optimal for studying interference between the L1 and L2 systems. First, linguists as far back as 1969 noted that these distinctions were already dying out in everyday speech. Second, these distinctions carry a very low functional load, with few minimal pairs. Most importantly, the linguistic environment of Barcelona provides a vast amount of conflicting information about these sounds’ distribution in the Catalan lexicon.
To assess the current state of these distinctions in Barcelona, we recruited a representative sample of 82 bilinguals varying on their language history and daily use of Catalan. Participants completed tests of identification and discrimination of high-mid and mid-mid continua of front and back vowels. A large effect of vowel height was observed, such that participants were much more consistent in identifying vowels along the high-mid continua than the mid-mid continua. Interestingly, however, the effect of native language was extremely small: many participants who were very nearly monolingual in Catalan performed surprisingly poorly on the Catalan mid-vowels, while some participants who were nearly monolingual in Spanish performed quite well. Given such a lack of correspondence between language background and perceptual performance, we suggest that it is extremely difficult to draw general conclusions about the time course of language learning based on studies of these specific contrasts.