Can 'Semantic P600' effects be the evidence for the existence of semantic-processing stream?

Recently a growing number of studies have found that a specific type of semantic/thematic anomaly would elicit a P600 effect, which has been considered as an index of syntactic-processing difficulty, suggesting (i) that there exists a semantic-processing stream which operates independently from syntactic-processing stream and (ii) that the output of semantic processing stream can, at least under certain circumstances, challenge the output from syntactic-processing stream and (iii) that the conflict between those outputs results in a so-called “Semantic P600” effect. The present study, however, points out that Semantic P600 effects can be task-relevant components and examined whether a Semantic P600 effect would be elicited even when participants are not asked to do a secondary task like plausibility judgment task. We recorded event-related brain potentials (ERPs) when participants passively read semantically/thematically-reversible Japanese sentences (e.g., literal translation: the leaves of eucalypts-NOM the koala on a tree-ACC took-a-bite-into) in the absence of a plausibility judgment task. We found a larger N400 effect but no P600 enhancement in response to the target words. We argue (i) that the plausibility-judgment task might motivate readers to actively anticipate how the incoming sentence will develop and (ii) that the P600 effects observed in the previous studies might reflect the processing load associated with reanalysis processes resulted from the mismatch between the tentative thematic-role assigned to the argument(s) and the argument structure of the verb which was actually encountered. We conclude that the so-called Semantic P600 effects are task relevant and that they cannot be the evidence for the existence of semantic-processing stream.