Abstract: There has an amount of researches on the question of whether the speed and accuracy with which a word is identified is affected by the existence of other, orthographically words in alphabetic writing systems. However, it is unknown about the impacts of orthographic neighborhoods in the Chinese characters. Orthographic neighborhoods were defined as the characters with the same structure and the same semantic radicals or phonetic radicals as targets in the present study. Characters with the same semantic radicals composed the semantic radical family, while characters with the same phonetic radicals composed the phonetic radical family. There are inconsistent findings regarding the effect of semantic radical family size on phonogram recognition in the character decision tasks (CDT). The effect of high frequency phonetic radical family members which were mixed with the effect of the phonetic radical family size in the past researches should be examined independently. According to the multiple read-out model, orthographic neighborhood effects should vary as a function of the difficulty of the lexical decision task. The present study aimed to investigate whether the effects of family size, high frequency phonetic radical family members and phonological information of phonetic radical on phonogram recognition were different in the CDT with pseudo-characters or with non-characters.
A total of 100 undergraduate students participated in present study. All of them were native speakers of Mandarin Chinese. The experiment was run on IBM computer using E-Prime software. A fixation point “+” was presented on the screen for 500ms, and then the target characters were displayed. Subjects were asked to perform the CDT. Both reaction time and error rate were analyzed with repeated measure ANOVA.
The results indicated that:⑴ Characters included in large size semantic radical family were responded to more rapidly than characters included in small size family. The responses to characters included in large size phonetic radical family were faster than responses to characters included in small size family. ⑵ In the CDT with pseudo-characters, the lexical decision latencies to words with high frequency phonetic radical family members were slower than those to words without those members, whereas there was no inhibitory effect of high frequency phonetic radical family members in the CDT with non-characters.⑶ The effect of phonological information of phonetic radicals was not salient in the CDT with non-characters, but a pronounceable phonetic radical inhibited recognition of characters included in small size semantic radical family in the CDT with pseudo-characters. Multiple read-out model was tested and modified according to the data. These results were explained by the interactive activation model.
The conclusions are directed from this study: A facilitatory radical family size effect and an inhibitory effect of high frequency phonetic radical family members on phonogram recognition were observed. These effects were mediated by the difficulty of the CDT. The effect of the phonological information of phonetic radicals was affected by the semantic radical family size and the difficulty of the CDT. It revealed an inhibitory effect of the phonological information of phonetic radicals on characters included in small size semantic radical family in the CDT with pseudo-characters

Key words: semantic radical family, phonetic radical family, high frequency phonetic radical family member, the multiple read-out model, interactive activation model