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

11฀ Frequency฀and฀Position฀Effects฀of฀ Component฀Combination฀in฀ Chinese฀Character฀Recognitio n฀ Buxin฀Han฀ The morphologica l constituen t o f Chines e characte r ha s bee n name d a s 'radical' o r 'component' . A radical is the structural uni t o f character use d for sortin g o r retrievin g relate d characters , a s use d i n dictionarie s (e.g. , TK, ? , , m ); whereas a componen t i s th e structura l constituent s o f a character (e.g. , / % !],/* , x) . Therefore , al l radical s ar e include d i n a component set , but a component i s not necessaril y a radical (Wang , Cui , &c Chai , 1997) . A componen t ha s bee n show n t o b e a functiona l uni t (Han, 1994a ; Pen g &C Wang, 1997 ; Yu, Cao , Feng, &c Li , 1990 ; Zhan g &C Feng , 1992). Another structura l unit , the component combination , ha s been ignored fo r a long time. A component combinatio n consist s o f tw o serial neighbourin g component s i n on e character , suc h a s 1L/ P i n character fP . A s a structura l uni t interva l betwee n a componen t an d a character, i t can b e shared b y a group o f characters . For example , 3/ 7 P is shared b y the character s nP , p'J , Rn, # etc. , and hold s som e cognitiv e property a s a commo n featur e amon g thes e characters . A s a possibl e functional uni t o f Chines e character , a componen t combinatio n ca n b e used t o grou p character s an d investigat e th e effec t o n th e characte r recognition proces s (Han , 1995) . As an orthographi c structura l uni t betwee n th e whole characte r an d stroke, a componen t i s mor e suitabl e t o th e principl e o f cognitiv e simplicity. Ove r 80 % o f character s contai n component s (radicals ) whic h more o r les s indicat e th e pronunciatio n o f th e whol e character . I n on e 208฀BUXI N฀HAN฀ word, a componen t mean s mor e tha n a stroke , a s bot h morphologica l and phonologica l constituent s o f th e character . However , th e frequenc y effect o f component s i n the visual recognitio n o f Chines e character s ha s been a rar e topi c i n earlie r studie s o n cognitiv e processin g o f Chinese , partly becaus e o f the complexity of its position i n characters an d th e lac k of possibl e manipulatin g index . The Frequenc y Databas e o f Chines e Characte r Component s an d Component Combinatio n (Han , 1994b ) wa s develope d t o provid e a statistical basi s for futur e studies . The databas e include s bot h toke n an d type frequenc y o f 567 component s an d 7,58 3 componen t combinations , which reflecte d th e statistica l feature s (e.g. , strok e number , frequency , phonological an d morphologica l component , etc. ) o f 6,76 3 simplifie d Chinese character s include d i n GB2312-8 0 (Han , 1993) . The database s are usefu l fo r furthe r stud y o n linguisti c computin g o f Chines e (La i &c Huang, 1997) , Chinese information processin g by computers (Han , 1998 ; Han &c Chen, 1993 ; Han &c Ren, 1996) , in addition t o its application o n the stud y o f Chines e cognition . There ar e thre e kind s o f componen t frequencies . Type frequency i s the numbe r o f character s containin g a component . Fo r example , typ e frequency o f j £ i s two becaus e it was used i n two characters, i.e., fl! an d ]£. Token frequency i s th e summar y occurrenc e o f thos e character s containing a componen t i n th e Chines e media . Fo r example , toke n frequency o f $ i s 1.9952% o summed fro m 1.9734% o ({£) an d 0.0218% o (j£). I...

Share