Processing Double Accusatives with Korean Dative/Causative Verbs: An ERP Study

Abstract

This paper reports an ERP-based study concerning the limited productivity of the ACC + ACC subcategorization frame with Korean dative and causative verbs. This frame is compared to the unmarkedly productive DAT + ACC frame in the ERP experiment and the acceptability rating task. The results show that the double Accusatives with the two types of dative verbs expressing caused possession or caused motion recorded N400, followed by P600, while those with morphological causative verbs registered N400 only. Likewise, the double Accusatives with both dative and causative verbs were consistently rated as unacceptable in the acceptability task. We take the disconfirmed expectation of a certain Case morphology to act as an etiology of the N400 modulation. The reader expects to encounter a dative or causative verb after the DAT + ACC sequence, but the preceding ACC + ACC sequence is not compatible with such a verb after it, evoking N400 because Case encodes morpho-lexical information. Meanwhile, reduced P600 represents a severe disruption of semantic analysis, reflected by N400; dative verbs differ from causative verbs in that the former employ covert lexical feature for causation, but the latter a morphologically-overt morpheme. The upshot of this paper is that Case as an apparently grammatical relation-encoding morpho-syntactic marker serves as a cue for predicting the following word associated with it, and ERP responses to a failure in such a Case-related prediction are not confined to a late positivity but are also detected evoking negativity at posterior regions in the 250 ∼ 500 ms interval.