言い間違い資料による言語産出モデルの検証(<特集>正常な発話と逸脱した発話)

抄録

The regularity and characteristics which speech errors show have been regarded as a 'window' through which one can see how unconscious language production mechanisms work in the mind. The present study examines the phonological components of two influential models based on Japanese speech error data. It is suggested that WEAVER++ adopting verification system and syllabary explains the flow from the ordering of segments in phonological encoding to the articulator in a more convincing way, whereas the interactive activation model has advantages over WEAVER++ in describing the repeated phoneme effect observed often in Japanese sound exchange errors.