The study used event-related brain potential (ERP) measures to
investigate the processing mechanisms underlying late L2-learners'
sentence comprehension. It compared the processing of auditorily
presented sentences in native listeners as well as in three groups of
foreign language learners. Participants were native speakers of
Japanese, Russian, French, or German. Test language was German. All
L2-participants had learned German as a second language after puberty.
We presented three types of sentences. The sentences ended with a target
word that was either (a) correct ("Das Brot wurde gegessen" - The bread
was eaten), (b) semantically incorrect, i.e., violating the selectional
restriction of the verb ("Der Vulkan wurde gegessen" - The volcano was
eaten), or (c) syntactically incorrect, i.e., violating the phrase
structure ("Das Eis wurde im gegessen" - literal translation: The ice
cream was in the eaten). Participants performed an acceptability
judgment task. During sentence presentation the EEG was registered and
event-related brain potentials (ERPs) were computed off-line. In the
control group of German native speakers, semantic violations modulated
the N400 component. This component is thought to reflect semantic
integration processes. Syntactic phrase structure violations, however,
elicited a fairly different pattern in native speakers: an early
anterior negativity that was followed by a broad parietally distributed
positivity (P600). Previous experiments have revealed that the processes
reflected in the early negativity are rather automatic and might reflect
first-pass parsing processes. The P600 is thought to reflect secondary
syntactic processes of reanalysis and repair (Hahne & Friederici, 1999).
Behavioral results from the acceptability judgment showed that the
foreign language learners made more errors than native speakers but
performed clearly above chance level in all groups. ERPs on correctly
answered trials in foreign language learners, clearly differed from
those of native speakers and varied systematically as a function of
proficiency level in L2. All L2-groups showed an N400-effect. However,
the morphology and timing of the effect correlated with proficiency.
Most remarkable differences were observed with regard to syntactic
processes. A late positivity was elicited only in the more proficient
second language learners, whereas the early anterior negativity was not
seen in any of the three groups. Taken together, the data suggest that
semantic integration processes are the first to achieve a status similar
to that in native listeners. With increasing proficiency, late syntactic
processes come into play whereas the early automatic parsing procedure
was not developed in any of the L2 groups studied.