兒童漢字再認的基本字形單位

The Basic Orthographic Units in Children's Recognition of Chinese Characters

本研究包含兩個實驗，旨在於探討國小五年級學童漢字再認的基本字形單位，實驗一使用「同異判斷作業」；實驗二使用「字彙判斷作業」。本研究的設計與材料取自於S.Chen和Liu（2000）以成人為對象的研究。該研究發現：成人對成對漢字的同異判斷受到字的筆畫形態數，而非筆畫數的影響（亦即單字若具有越多筆畫形態，則所需判斷時間越長），但此種反應模式僅限於對高頻漢字字對的「同」判斷反應，而對高頻漢字字對作「異」判斷反應和對低頻字作「字彙判斷」時則僅受筆畫數的影響。此結果暗示筆畫形態數的效果可能源自於對漢字字對作「同」判斷反應的特殊作業需求，而非漢字再認必然涉及筆畫形態的處理所致。本研究實驗一「同異判斷作業」的結果與成人的表現一致，但實驗二「字彙判斷作業」的結果則與成人不同：筆畫數以及筆畫形態數的效果皆達顯著，且此效果僅限於低頻字。此一結果的解釋之一是五年級學童在漢字再認歷程中，能夠依作業所需，彈性地使用筆畫或筆畫形態作為漢字處理的最基本單位。另一解釋則是，筆畫形態數的效果是由於所使用的材料字未有效控制含二單位和三單位筆畫形態兩類字的「筆畫密度分布」特性（可以單字部件的視覺複雜度代表）所致。對實驗二使用的刺激字及其結果進行一系列事後分析的結果較支持第二種解釋，此解釋亦與其他以國小學童漢字再認歷程發展的研究結果較為吻合。故本研究的結果顯示五年級國小學童漢字再認歷程中的基本字形單位較可能是筆畫，而非筆畫形態。

Two experiments were conducted to investigate the basic orthographic units used by the fifth graders in character recognition: Experiment 1 used a simultaneous 'same-different' comparison task; Experiment 2 used a character- decision task. The current study modeled after S. Chen and Liu's (2000) study with adults. The results of their study showed that simultaneous comparisons of pairs of Chinese characters were influenced by the stroke-pattems (SPs), not the individual strokes, that comprised the characters. That is, the larger was the number of SPs, the longer was the response latency; however, this pattern was restricted to the 'same' judgments for high-frequency character pairs. By contrast, in the 'different' judgments and in the character- decision task, the effect of stroke number was found instead, suggesting that the effect of SPs observed in the 'same' judgments might result from specific task-demands for making simultaneous comparisons. In the current study, children's performance on the 'same-different' judgment task (Experiment 1) showed a pattern similar to that found in adults, but their performance on the character-decision task showed a different pattern (Experiment 2). The results of Experiment 2 showed that the child participant's character decision responses to low-frequency characters were affected by the number of strokes, but a trend of the SP-number effect was also found although it was not robust-- it was significant only in the subject, not in the item analysis. One explanation of these results was that the fifth graders flexibly used either individual strokes or SPs in character recognition, depending on the testing context. However, a further inspection of the stimulus characters used in Experiment 2 suggested that the trend of SP-number effect might be apparent and resulted from mismatch in a character- based visual attribute, the visual- density distribution (as can be indicated by visual complexity of the character's right- side radical), between the stimuli used for contrasting SP conditions. Post hoc analyses of the responses to a selected subset of stimuli (matched in the visual complexity of the characters' righ side radical) revealed a pattern consistent with this proposal. The pattern of current results is thus more consistent with the view that individual strokes function as the basic orthographic units in children's recognition of Chinese characters. Implications of the current findings are discussed with respect to the existing theoretical framework for, and to the development of, character recognition processes.