Revisiting the critical period: age of onset or length of exposure?

In this study, we re-evaluate the debate on the impact of age on second language learners' attainment, and expand the discussion by two dimensions: first, we use non-experimental, spontaneous speech data; second, we show to what extent exogenous factors can override the influence of age on learners' attainment. Statistical analysis of spontaneous speech data collected from Bengali learners of English living in London shows that a prolonged stay in England has a much stronger effect on respondents' L2 English proficiency than age of acquisition onset. In particular, a multiple regression analysis reveals that exposure accounts for 53% of variability in L2 proficiency, compared to a mere 1% for age. We argue that particularly if L2 acquisition takes place in a naturalistic setting with little formal instruction but through interaction with native speakers, length of residence and prolonged exposure to the TL may extenuate the negative effect of late acquisition onset.