| 摘要: | 過往以拼音文字為主要研究對象的文獻指出，語音為維持工作記憶的主要表
徵，而本實驗室過往的研究則指出：對中文刺激材料來說，因音旁相同所造成
的字形相似性也對中文短期記憶的表現有獨立的影響。在本研究中，為了更進
一步的了解中文語文工作記憶的機制，總共安排了四個系列的實驗。實驗1 主
要想釐清對於中文的工作記憶來說，共享音旁所觀察到的字形相似性效果，在
共享義旁的情況下是否仍舊可以被觀察到，因此針對義旁的位置、是否共享義
旁、以及義旁的透明度進行操弄，來闡明這三者變項與中文工作記憶之間的關
係。除了序列位置的新近效果以及語音抑制效果之外，此實驗的結果並沒有看
到任何達到顯著差異的效果，探究原因有可能是因為實驗刺激材料的操弄不夠
確實而導致。有鑒於此，實驗2 簡化了實驗1 操弄的變項，並且加強中文刺激
材料之間的的操弄，利用義旁相似性（共享義旁與否）以及義旁透明度來驗證
中文工作記憶的機制。從實驗結果中觀察到：字形相似性以及義旁透明度效果
的不穩定存在，兩者均是在沒有語音抑制的情況下才可以看到顯著差異，但是
增加了語音抑制的任務之後，兩者皆消失了。我們認為這是由於中文構字的特
殊性，使得共享義旁的字形相似性也會導致語義的相似，因而被語義相似性效
果所干擾。從此觀點出發，實驗3 則獨立操弄了共享義旁與否以及字義範疇相
似性，來探討是否中文工作記憶中存在有獨立的字形儲存系統和語義儲存系
統。且在實驗結果中同時觀察到字形相似性效果以及字義相似性效果，以及這
兩個變項與語音抑制的交互作用。進一步的事後檢定則發現，字義相似性效果
會因為同時進行語音抑制任務而增強，而字形相似性效果則不論有無語音抑制
皆被觀察到，但在無語音抑制的情況下，又和字義相似性有顯著的交互作用，
顯示這些效果間的複雜關聯。實驗4 則是確認中文工作記憶中的字形相似性是
有別於拼音文字中的視覺相似性，因此操弄了沒有共享任何中文部件的視覺相
似中文字，且安排了左右並排的中文字以及上下組成的中文字，來探討視覺相
似性以及字形組成方式與中文工作記憶的關係。實驗結果指出：對於中文使用
者來說，沒有分享任何語言表徵的部件所導致的單純視覺上的相似，有促進中
文工作記憶表現的效果，這點與拼音文字的研究發現正好相反。整體來說，本
研究四個系列實驗的結果顯示，因共享義旁所造成的字形相似性對於工作記憶
表現的影響並不穩定、也不獨立，反而會受到字義相似性、語音抑制與否等變
項的調節，和知覺層次的視覺相似性有本質上的不同。另外，字義相似性也在
有語音抑制的情況下輔助儲存短期記憶，影響對中文刺激材料的工作記憶表
現。將這些結果和因共享音旁所造成的字形相似性結果進行比較，可說明中文
部件中義旁和音旁之於工作記憶的不同重要性，以及中文工作記憶有別於拼音
文字工作記憶的獨特性。;The main research of alphabetic language (ex. English) pointed out that phonology was the main representation to maintain working memory. However, the previous research of Chinese stimuli in our laboratory had revealed that the orthographic similar caused by the shared phonetic radicals also affected Chinese verbal short-term memory independently. In this thesis, in order to further understand the mechanism about Chinese verbal working memory, four series experiments have examined. The experiment 1 focuses to clarify the similarity effect of orthography can also be found from the shared semantic radicals, not only from the shared phonetic radical. The position of semantic radical (left or right), shared semantic radicals, and the transparency of the semantic radical were manipulated in order to examine the relationship among these three variables and Chinese verbal working memory. There are not any other significant differences in addition to recency effect and the effect of articulatory suppression. One possibility was investicgated to cause this result might be because the manipulation of the Chinese stimuli was not efficient. Hence, the experiment 2 modified the manipulation of the stimuli to testified the Chinese verbal working memory mechamism between similarity of semantic radical and the transparency of the semantic radicals. It is observed from this experiment that the effect of orthographic similarity and the effect of semantic radical transparency are not stable. These two effect was observed under silent condition, but not under articulatory suppression condition. It is convinced that due to the particularity of Chinese word formation, the similarity of the orthographic by shared semantic radicals also leads to semantic similarity. Thus, the effect of similarity orthographic might be disturbed by the effect of semantic similarity. According to this, the experiment 3 independently manipulates the similarity of orthography (by shared semantic radicals) and the similarity of semantic categories in order to explore whether are there exist the orthographic and semantic storage system independently in the Chinese verbal working memory. The result of experiment 3 reveals that it is observed both the effect of orthographic and semantic similarity and the interaction between articulatory suppression and these two effects. Futher analysis tells that the effect of semantic similarity will strength under articulatory suppression condition while the effect of orthographic similarity can be observed whether under articulatory suppression condition or not. When this effect under silent condition, there is an interaction with the effect of semantic similarity. All these results can tell that there are complex relationships between the effect of seamantics and orthography. The experiment 4 aims to testify the effect of orthographic similarity in Chinese verbal working memory is different from the effect of visual similarity. Visual similarity without shard any radicals and vertical and horizontal Chinese words are manipulated in this experiment in order to explore the relationship between visual similarity effect, word formation and Chinese verbal working memory. This result points out that only visual similar (without shsre any linguistic radicals) may increase the performance of the Chinese verbal working memory, and this is different from the research about the alphabetic language. Overall, the results of these four series experiments reveals that the effect of orthographic similarity by shared semantic radical are not stable and not independent on verbal working memory. Instead, it will be adjusted by variables such as semantic similarity, with or without articulatory suppression condition, and visual similarity at the perceptual level. In addition, semantic similarity also assists the storage of short-term memory under the condition of articulatory suppression, which affects the working memory performance of Chinese stimuli. Comparing these results with the orthographic similarity by shared phonetic radicals, it can explain the different importance of semantic and phonetic radical to Chinese verbal memory, and the spaciality of Chinese verbal memory different from alphabetic ones. |
