Non-economical Verbal Information Processing Driven by a ``Look-ahead'' Strategy Under Poor Availability of Structural Information

  • SOSHI Takahiro
    Department of Humanity and Literatures, Faculty of Letters, Tokyo Metropolitan University National Center of Neurology and Psychiatry
  • HAGIWARA Hiroko
    Department of Language Sciences, Graduate School of Humanities, Tokyo Metropolitan University Department of Humanity and Literatures, Faculty of Letters, Tokyo Metropolitan University Research Center for Language, Brain and Genetics, Tokyo Metropolitan University

Abstract

Natural language demonstrates discontinuous relationships between related words. Neurophysiology has reported that two types of brain electrical activities are related to discontinuous dependency. Sustained anterior negative potentials (SAN) are considered to be short-term memory storage for dislocated words. A late positivity (P600) appears concurrently with SAN decay, and is interpreted as the integration cost of a discontinuous word with a related word. Discontinuous verbal processing through SAN and P600 is considered to be unfavorable because of redundant neural resource consumption. However, SAN and P600 may reflect a global prediction-based strategy, which rather actively consumes memory resources to comprehend an overall meaning. We thus prepared four types of sentences (noun 1/adverb/noun 2/verb), which were modulated by grammatical information (case marker) and word order factors, and recorded brain potentials from Japanese participants performing a sentence comprehension task. Consistently with our prediction, P600 appeared and SAN disappeared upon the presentation of second nouns, but only in the canonical order including first subject words without a case marker. Hence, discontinuous verbal processing and its neural correlates should be re-considered in the context of interactions between local memory costs and global prediction strategies.

Journal

References(44)*help

See more

Related Projects

See more

Details 詳細情報について

Report a problem

Back to top