Neural changes associated with grammatical competence in late acquisition

International audienceTheories abound as concerns the acquisition of a “second” language, i.e. a language that is acquired in either a natural or scholastic setting some time after early childhood. For some, grammatical processing is qualitatively different in a second and native language (Shallow processing hypothesis: Clahsen & Felser 2006); whereas for others, the fundamental mechanisms of L1 and L2 grammatical processing are the same (Dekdtspotter, Schwartz & Sprouse 2006; Herschensohn 2000; Hopp 2007; Schwartz & Sprouse 1996), with processing differences being attributable to factors such as reduced automaticity (slowness) of L2 processing (Frenck-Mestre 2002; 2005), the influence of L1 properties (Prévost & White 2000; Foucart & Frenck-Mestre 2011; Frenck-Mestre et al. 1997; Hopp 2007) and limited computational resources (McDonald 2006). In the present talk we will show that the adult learner indeed goes through developmental stages in the acquisition of the target grammar, which can be tracked via the neural response to grammatical and ungrammatical structures. We will show that during early stages of L2 acquisition, the automatic neural response elicited during online syntactic processing is qualitatively different than that found for either more advanced learners or native speakers. At more advanced stages of learning coupled with immersion, however, the differences in the neural response observed between native and non-native speakers during syntactic processing are quantitative rather than qualitative in nature. We will argue from these data that, indeed, full acquisition of grammatical competence is possible even for languages learned “late” in life, while not undermining the various factors that impinge upon processing capacity