Search

The purpose of this article was to discuss the role of phonological processing on word recognition of the second language that has been indicated in the first language. In the first section, the history of bilingual memory representation researches were reviewed. Some problems were described based on the criticism of the previous researches reviewed and the words type effects investigated in cross-languages. Then, after referring to the role of phonological processing on word recognition of the first language, the neglected roles of phonology on bilinguals' lexical processing were discussed based on introducing the previ-ous researches. Finally, some proposals for the phonological processing researches on bi-lingualism were discussed.

The purpose of this study was to investigate whether the processing routs of Japanese Kanji words for the Taiwanese learners of Japanese were different between cognates and non-cognates in Japanese-Chinese. Two experiments were performed by using a semantic-decision task. In experiment 1, involving cognates, the interference was shown similar in both the reaction times and the error rates only when the targets were orthographically. This result indicated that the meaning of Japanese Kanji word was accessed directly from orthography or phonology of the Chinese. The phonology of Japanese did not contribute to the accessing of Japanese Kanji word. In experiment 2, involving non-cognates, a significant interaction was observed in both the reaction times and the error rates. This result indicated that the phonology of Japanese did contribute to the accessing of Japanese Kanji word as well as the orthography. It is suggested that the processing routs of Japanese Kanji words varied according to the lexical relationship between Japanese and Chinese.

本研究の目的は,Tipper&Driver(1988)の処理モデルが妥当であるか否かを検討することであった.Tipper&Driver(1988)のモデルは,検索段階が終了した後に選択段階に移るというように時系列的に各処理が進むというモデルである.このモデルに従えば,妨害情報の抑制が行われるのは,適切情報の検索が終了した時点より後になると予想される.実験の結果,促進は反応遅延間隔O msから200msの間で観察されたが,反応遅延間隔400ms以降では生じず,適切情報の検索処理は刺激提示から400msが経過するまでに終了することが分かった.また,干渉は反応遅延間隔Omsで観察されたが,反応遅延間隔100msから800msの間では生じなかった.この結果は,妨害情報の抑制は刺激提示から100msが経過するまでに終了することを示している.これらの結果は,妨害情報の抑制が終了する時点は適切情報の検索が終了する時点よりも前にあることを示しており,Tipper&Driver(1988)の直列的なモデルは頑健なものでないことが分かった.