Short-term Phonetic Drift in an L2 Immersion Environment

Abstract

This paper investigates the nature and time course of phonetic drift in L1 by examining the very first weeks of 20 L1 English speakers’ acquisition of Korean as L2. Acoustic analyses of these learners’ L1 and L2 production over time indicate that learning L2 stops affects the production of L1 stops (in terms of VOT and/or f0 onset) in as little as one week, with the L1 sounds approximating the characteristics of the L2 sounds to which they are most phonetically similar. These results indicate that L1 phonological categories are affected by L2 learning on a very short timescale, suggesting that the equivalence classification that gives rise to this phonetic drift may be rather low-level in nature. When we learn a second language, what happens to our native language – the language we learned first? For a long time, nothing was thought to happen, since the earliest research on the interaction of first (L1) and second language (L2) phonologies was based on two related assumptions: the assumption of a socalled “critical period” for language acquisition and the assumption of unidirectionality of cross-language influence. The classic view of the critical period (cf. Lenneberg 1967) holds that biological changes in brain development are responsible for the general decline in ability to learn another language with increasing age; children are better at acquiring language than adults, then, because they have not yet passed this critical period of neural plasticity. This idea of neural plasticity in children and, by implication, neural rigidity in adults suggests that the linguistic structures of L1 are fossilized by the time one reaches adulthood. Therefore, while the L1 may cause some interference in the acquisition of an L2, the L1 itself should not be affected. More recent work in phonetics and second language acquisition has challenged both of these assumptions. Some researchers (e.g. Flege 1987b) have pointed out numerous problems with the basic