- 學位論文
意旁結合度、意旁表意透明度對中文閱讀的影響
Combinability and semantic transparency effects of semantic radical in reading Chinese characters
摘要
本研究的目的在探討意旁的處理對中文字辨識歷程的影響。首先在建立意旁結合度效果的基本現象。實驗一和實驗二分別使用假字和非字情境，操弄真字的字頻與意旁結合度，進行中文字彙判斷作業(CDT)。在對真字的反應時間上發現，使用假字的情境迫使受試者需要精確提取語音或語意訊息才能做出反應，得到抑制的意旁結合度效果。而使用非字的情境允許受試者依照字形上的熟悉度作反應，得到促進的意旁結合度效果。意旁結合度效果在不同實驗情境下方向相反的結果，支持意旁的處理符合二階段模式(Vitevitch & Luce, 1999)。為了進一步探討這兩個處理階段的生理指標，實驗三進行CDT作業同時收集腦電波訊號。結果發現在低頻字中，意旁結合度低的字引發比較正向的P200，反映早期字形處理的困難度。意旁結合度高的字引發比較正向的N400，反映晚期來自候選字的競爭。顯示意旁處理的兩個階段可以先後反映在P200和N400的振幅上。藉助這兩個腦電波指標，實驗四則試圖釐清意旁提示語意的功能在哪一個階段介入並且影響意旁的處理。實驗中進一步操弄意旁表意透明度與意旁結合度，進行CDT作業並且收集腦電波訊號。結果發現在P200和N400皆呈現兩變項間的交互作用，在意旁表意透明度高的情況下得到和實驗三型態相同的意旁結合度效果。綜合本研究的結果支持意旁的處理經過兩個階段，早期階段會激發一群具有相同意旁的字形表徵，這些候選字在晚期階段會互相競爭與干擾。而意旁的功能在早期和晚期階段皆產生影響，在意旁能夠有效提示整字語意的情況下，比較容易凸顯意旁處理的二階段特徵。顯示具有相同意旁的字群可能會依據其意旁對語意提示的有效度，在兩個處理階段形成次組群。
並列摘要
The aim of the present study is to explore the role of semantic radicals in Chinese character recognition. First of all is to demonstrate the property of combinability effect for semantic radicals. In experiment 1 and 2, the character decision tasks (CDT) with different no-trials, pseudo-character and non-character respectively, are used. The character frequency and the combinability of semantic radical are manipulated in real Chinese characters as materials. In the CDT with pseudo-characters, subjects have to retrieve the precise lexical representation of the character for making decision. It reveals an inhibitory combinability effect on the mean reaction time (RT) of real characters. In the CDT with non-characters, subjects could make a response relying on the familiarity with orthography. It reveals a facilitative combinability effect on mean RT of real characters. The reversed pattern of combinability effects suggest that the processing of semantic radicals follows the operation of two levels of processing (Vitevitch & Luce, 1999). In order to explore physiological indices of the two levels, event related potentials (ERPs) are recorded in CDT in the experiment 3. In low frequency characters, it shows the semantic combinability effect on P200 and N400. Characters with low combinability semantic radical show more positive P200 than those with high combinability semantic radical. Characters with high combinability semantic radical show more negative N400 than those with low combinability semantic radical. It suggests that the amplitude of P200 and N400 can index the early and late level of processing respectively. In experiment 4, the combinability of semantic radical and the semantic transparency of the character are manipulated in CDT with ERPs recorded. The interactions between combinability and transparency are revealed on both P200 and N400. The combinability effect in high transparency characters shows the same pattern as that in experiment 3. In summary, the current results suggest that the processing of semantic radicals follows the operation of two levels of processing. The functional feature of semantic radicals is involved at both the early and the late levels.