Word Order and Clause Boundary Insertion: Behavioral and ERP Studies

Abstract Processing scrambled word order requires extra cognitive cost as revealed by previous studies. Questions remain as to at which position the cost emerges, and why. We conducted two self-paced reading experiments using Japanese double-scrambling constructions. We observed reading time slowdowns at the nominative NP and at the verb in scrambled sentences, and we show evidence that the slowdown at the nominative NP reflects the parser's expectation to encounter a verb, and the slowdown at the verb arises because the parser anticipated an intransitive verb. Finally, we will provide some preliminary results from an ERP study