ERP correlates of miniature language learning: What's learnt within five days

Müller, J.1, Hahne, A.1, Fujii, Y.2, and Friederici, A. D.1
1Max-Planck-Insitute of Cognitive Neuroscience, Leipzig, Germany; 2Humboldt-Universität, Berlin, Germany
E-mail: muellerj@cns.mpg.de

ERPs have been established as a measure of semantic and syntactic processing mechanisms in first and second language (L2) comprehension. In late L2 learners ERPs related to syntactic processing are often more distinct from native speakers than ERPs indicating semantic processing, suggesting a differential susceptibility to critical period effects.Yet some recent experiments using artificial grammars have challenged the critical period hypothesis by eliciting ERPs in trained adults similar to patterns found for native language processing. We studied sentence processing within a miniature language extracted from Japanese. Thus we assured high proficiency of the learners. Within the miniature language different rule violation conditions were implemented, namely a word category violation, a classifier agreement violation and a case violation. In three auditory ERP experiments we presented correct and incorrect sentences of the miniature grammar to Japanese subjects and to German subjects before and after training. Japanese subjects displayed a P600 for both the word category violation and the case violation. In the word category violation the P600 was preceeded by a broadly distributed early negativity with an anterior preponderance, and in the case violation by an N400. The classifier violation resulted in a left anterior negativity only. While the P600 was similar for Japanese natives and trained subjects, the N400 and the anterior negativities were not present in the learner group. The observed differences suggest difficulties in the acquisition of on-line syntactic processes and processes of thematic hierarchizing for the learners despite their high behavioural skills.