分佈表徵記憶與語言間翻譯促發效果

DISTRIBUTED-MEMORY MODEL AND TRANSLATINGASYMMETRY PRIMING EFFECTS

本研究嘗試以連結主義中的分佈表徵記憶及激發組型轉換理論，來解釋中英文雙語使用者的詞彙與概念表徵型態，以及語言間翻譯促發不對稱現象。實驗一以詞彙判斷作業來比較英文與中文的語言內相關字促發效果，結果發現母語（中文）語言內相關字的促發效果顯著地大於外語（英文）語言內相關字。實驗三與實驗三則一致地發現，前向翻譯（中文－英文）促發效果顯著地大於後向翻譯（英文－中文）促發效果，而且無論是在促發項後有無遮蔽刺激或無關字中介刺激的狀況下皆是如此。上述發現可藉促發項與目標項間的激發組型轉換來加以解釋。本研究另一發現為，無論在前向或後向翻譯時，低頻詞所獲得的促發效果大於高頻詞。此點與分佈表徵的預測不符合。研究者除提出激發組型轉換論的限制外，亦提出以頻率論來解釋語言間翻譯促發效果不對稱現象的可能性。

Concepts of distributed representation memory and activation-pattern shifting are used to explain bilingual memory and between-language translating-asymmetry effects. In Experiment 1 using lexical decision task, Chinese participants demonstrated greater within-language semantic-priming effects in native language than in second language (English). Experiments 2 and 3 consistently demonstrated that, for Chinese native speakers, priming effects of forward-translation (Chinese-English) was greater than those of backward-translation (English-Chinese). This result occurred both when there was no mediating stimulus and when semantically-irrelevant words or random-dot mask mediated between prime and target items. These findings were consistent with the prediction of distributed-memory model. Experiments 2 and 3 found that low-frequency words produced greater priming effects on performance than high-frequency words did, both in forward and in backward-translation. This finding was inconsistent with the prediction of distributed-memory model. The translating-asymmetry priming effect are discussed with word-frequency influences.