Incremental Sentence Production: Observations from Elicited Speech Errors in Japanese

Abstract

This chapter proposes that in Japanese sentence production, a secondary processing procedure that utilizes frequent correspondences between thematic roles and case particles may be at work in the absence of yet-to-be retrieved verb lemma, enabling incremental sentence production processes. Speech error data (slips of the tongue) among native speakers of Japanese are provided to demonstrate that Japanese speakers are likely to be using the common correspondence between Patient role NPs and the accusative particle o (and between Location and dative/locative particle ni). In experimentally elicited data, errors of o are observed when the Patient role is expressed in subject NPs in sentences such as passive sentences and intransitive sentences that require Patient as the subject and errors of ni are observed when Location is expressed as the sentence-initial NP followed by verbs of activities, which require the particle de for the location of activities. Such case particle selection based on the mapping between the thematic role of the NP and the case particle enables Japanese speakers to produce sentences fluently (and accurately, most of the time) since they do not always need to wait until the verb selection is finalized to produce sentences.