In lieu of an abstract, here is a brief excerpt of the content:

8฀ Early฀Phonological฀Activation฀in฀ Reading฀Kanji:฀An฀Eye-Trackin g฀ Study฀ Sachiko฀Matsunaga฀ Many researchers agree that regardless of the script type, phonology play s a crucia l role i n holding informatio n i n working memor y durin g readin g (e.g., Dewey, 1996; Hayes, 1988 ; Kleiman, 1975 ; Treiman, Baron & Luk , 1981; Tzen g & Hung , 1980 ; Tzeng , Hun g & Wang , 1977 ; Zhan g & Perfetti, 1993) . However , th e timin g o f phonologica l activatio n (i.e. , whether prelexical or lexical phonology is possible) during reading remain s a controversia l issue , especially i n non-alphabetic language s lik e Chines e and Japanes e tha t emplo y kanj i (Chines e characters ; se e Leon g & Tamaoka, 1998 ; also se e Wang, Inhoff & Chen , 1999) . In bot h Chines e an d Japanese , ther e ar e studie s i n suppor t o f earl y phonology (e.g. , Cheng , 1992 ; Mizuno , 1997 ; Perfetti & Zhang , 1991 , 1995; Tan , Hoosain & Peng , 1995 ; Tan, Hoosain & Siok , 1996 ; Wydell, Patterson & Humphreys , 1993) , a s well a s thos e agains t i t (e.g. , Chen , Flores d'Arcai s & Cheung , 1995 ; Zhou & Marslen-Wilson , 1996) . Th e results fro m thes e studie s vary , perhap s becaus e th e typ e o f th e task s (e.g., naming an d lexica l decisio n a t th e word leve l or semanti c decisio n at th e sentenc e level ) use d i n thes e studie s i s not onl y differen t bu t als o unnatural. Natura l readin g involve s readin g text s fo r comprehensio n (Rayner & Pollatsek , 1989) . Thus, despite their valuable contribution , i t is difficul t t o mak e a generalizatio n abou t th e timin g o f phonolog y i n reading kanj i base d o n thes e dat a alone ; wha t i s neede d i s a n on-lin e measure, suc h a s eye tracking . 158฀SACHIK O฀MATSUNAG A฀ There ar e a few eye-trackin g studie s tha t specificall y investigate d th e timing of phonology in natural reading. In English, for example, Danema n and Reingol d (1993) , Danema n e t al . (1995 ) an d Inhof f an d Topolsk i (1994) measured firs t fixatio n duratio n an d consecutive fixation duration s on targe t word s (homophoni c o r non-homophoni c t o th e contextuall y correct words ) i n passages. The assumptio n wa s tha t i f early phonolog y is possible, then homophone effect s shoul d be observed in the first fixatio n duration (i.e. , the duratio n o f th e firs t fixatio n place d o n th e error ) a s well as in the gaze duration (i.e. , the total fixation tim e spent on the erro r before th e ey e move s off ) o n th e targe t words . Unde r thi s assumption , longer firs t fixatio n duratio n an d gaz e duratio n woul d b e observe d o n the non-homophoni c error s tha n o n th e homophoni c errors , reflectin g homophone interferenc e effect s (i.e. , difficult y wit h noticin g th e homophone errors) . If, o n th e othe r hand , onl y postsemantic phonolog y is possible, then homophon e effect s shoul d b e observed onl y o n the tota l fixation duratio n (i.e. , the su m o f duration s o f fixation s o n th e error , including late r fixation s tha t ar e th e resul t o f regressions) . Presumably , shorter tota l fixatio n duratio n woul d b e observe d o n th e homophon e errors tha n o n th e non-homophoni c errors , reflectin g homophon e facilitation effect s (i.e. , ease wit h recoverin g th e correc t meaning s fro m the homophon e errors) . The results from thes e studies, however, are not compatible. The dat a from Inhof f an d Topolski' s stud y (Experimen t 1 ) showe d homophon e interference effect s o n th e firs t fixatio n duratio n i n additio n t...

Share