The partial incremental argument interpretation model: Real time argument interpretation in simplex sentence comprehension of the Japanese language

Sentence comprehension is an essential cognitive function inhumans. Sentence comprehension studies to date have focused onlanguages where Subject-Verb-Object word-order is foundational, such asEnglish. Less attention has been paid to verb-final languages whereSubject-Object-Verb word-order is key, such as Japanese. Hence, areview of psycholinguistic and neurolinguistic simplex sentencecomprehension studies for Subject-Object-Verb word-order languages(also known as verb-final languages) and the proposal of a model basedon the integration of these recent findings is of value to furtherunderstanding of sentence comprehension. Based on a consideration ofthe understanding of simplex sentence comprehension in the Japaneselanguage (a Subject-Object-Verb word-order language) that has beengained from psycholinguistic and neurolinguistic findings, this paperproposes The Partial Incremental Argument Interpretation Model as amodel of argument interpretation in Japanese simplex sentencecomprehension. In Japanese, since a verb, or head, is put at the end of asentence, arguments are input before the input of the verb, or head. Thismodel postulates that, promptly after the input of both the first andsecond arguments, an incremental interpretation of these arguments isprocessed based on the relative relationship between the two arguments.In addition, this model asserts that arguments marked by nominative andaccusative cases, but not by the dative case, are semantically interpretedbefore the input of a verb, or head. In order to process these argumentsincrementally, it is also assumed that three pathways of information areused--case particles, animacy information in the case of nouns, and theword order of nouns--with a priority in terms of their processing order.Based on the information given by the interpreted arguments, a head isanticipated before it is input. The above four assumptions in combinationcomprise a new characteristic which we have incorporated into theJapanese simplex sentence comprehension model proposed herein. In thispaper, we describe previously proposed Japanese sentencecomprehension models and their problems, then propose our modifiedmodel. Details of the architecture of the model proposed herein areprovided, then what motivates the sub-processes in this architecture isexplained using findings reported in previous psycholinguistic andneuroimaging studies. The processing procedure in this model isillustrated by and simulated with several examples, and compatibilitybetween this model's predictions and actual behavioral findings reportedin the literature to date are discussed. The modified model was able tocorrectly predict most previous findings. Finally, the implications of thismodel for simplex sentence comprehension models in general arediscussed.