Brain potentials and syntactic violations revisited: no evidence for specificity of the syntactic positive shift.

One of the current issues in the investigation of language by means of event-related brain potentials (ERPs) is whether there is an ERP effect that can be specifically related to the processing of syntactic information. It has been claimed that a late positivity (P600 or SPS-syntactic positive shift) occurring to syntactic violations or ambiguities qualifies as such an effect. In the present investigation we compared ERPs elicited by morphosyntactic (case inflection errors), semantic, and orthographic (misspelled words) violations in a group of young German subjects. All three types of violations gave rise to late positivities having the characteristics of the previously described P600/SPS. In an earlier time window, however, semantic violations were associated with a centroparietally distributed N400 component, whereas syntactic violations gave rise to a negativity of smaller amplitude that had a frontocentral distribution. In light of the present experiment, the view that the P600/SPS as a whole reflects specific syntactic processes appears to be untenable and an alternative interpretation is proposed. The different distributions of the late positive shifts merit further investigation.