| 摘要: | 過去以單一語言刺激材料為主的短期記憶研究，已支持語音的訊息是短期記憶中最主要的編碼方式。然而，鮮少有短期記憶研究探討非語音訊息的影響，更遑論使用雙語刺激材料來研究相關議題的研究。本研究設計四個行為實驗，利用探針序列回憶作業的特性，首先探討雙語刺激材料中的非語音的表徵，是否可以在短期記憶表現中彰顯出母語優勢效果。接著，我們使用中、英刺激材料依序交錯呈現的字串，去探討母語優勢效果的穩定性；同時，我們也進一步的去探討在雙語使用者的短期記憶中，是否存在著以語言類型為基礎的組織準則。在前兩個實驗中所使用的雙語刺激材料乃近乎同音，藉此方式以控制中、英配對刺激材料之間的語音複雜度，企圖去驗證當語音沒有辦法有效幫助受試者記憶時，能否依舊能觀察到短期記憶表現中的母語優勢效果。實驗一和實驗二的結果指出，即便在使用咬音抑制的狀況下，仍能觀察到短期記憶中顯著的母語優勢效果，故支持非語音表徵對雙語使用者的短期記憶表現有顯著的貢獻。實驗三和實驗四，藉由使用雙語刺激材料交錯呈現的字串，企圖檢驗雙語使用者的短期記憶中，母語優勢效果的穩定性。過去的研究者曾利用混和字串的特性，即，由兩種不同類型的刺激材料組合成的字串（如：中、英文刺激材料），並將之與探針序列回憶作業做結合的方式，發現短期記憶中的組織法則。我們也效法這樣的設計，希望能在實驗三和實驗四中去探討在雙語使用者的短期記憶中，是否存在著以語言類型為基礎的處理準則。與前兩個實驗不一致之處是，實驗三的結果並無顯示出中、英文刺激材料的短期記憶表現有顯著的不同。我們推測，實驗三的資料無法發現顯著的母語優勢效果，乃因雙語字串的使用增添了記憶作業的難度，導致受試者採取使用語音訊息的方式記憶刺激材料。在實驗四中，我們使用四個字為一組的中、英文同韻刺激材料，取代近乎同音的中、英刺激材料，是為了提高每組字串中刺激材料間的語音相似度，進而降低受試者依賴語音的程度。我們測試了兩組分別在七歲前(英文較流利)以及七歲之後(中文較流利)到達美國中、英雙語使用者。這兩組受試者的資料顯示出，短期記憶中有顯著的優勢語言效果存在，重複驗證了前兩個實驗中得到的母語優勢效果。更重要的是，這兩組資料都證實，當探針與目標皆屬同語言的狀況下，受試者的記憶表現會優於當探針與目標分屬不同語言而有語言轉換的狀況下。換句話說，即便當探針或目標屬於受試者的優勢語言的狀況下，並不會促使受試者於此狀況下的記憶表現顯著優於當探針及目標同屬受試者的非優勢語言的狀況下的記憶表現。此結果支持不同語言的刺激材料以獨立的方式被保留在雙語使用者的短期記憶中。總體而言，我們從探針序列回憶作業獲得的結果，支持非語音表徵確實會影響短期記憶中的母語優勢效果；另外，本研究結果也顯示，中、英雙語使用者的短期記憶中，確實存在以語言類型為基礎的組織/處理法則。Memory research has indicated the prominent role of phonology for retaining verbal materials in short-term memory in monolinguals. However, there has been limited research addressing how non-phonological representations play roles in verbal short-term memory (VSTM), and even less is known in the context of bilingualism. The present study first aims to investigate whether the non-phonological representations give rise to the effect of first language (L1) advantage on a probed recall task. Additionally, we aim to explore the robustness of the L1 advantage effect by employing bilingual lists with mixed language modes. Moreover, we further investigate whether language-based chunking/organizing principles exist in VSTM for each language in Chinese-English bilinguals. To achieve these aims, four experiments were conducted. In the first two experiments, we manipulated the phonological complexity between pairs of approximately homophonic Chinese and English materials to examine whether the L1 advantage effect would be observed, even when the phonological information was not helpful. The results indicated that there was a significant non-phonological contribution to bilingual VSTM by demonstrating a significant L1 advantage effect in the two experiments even under articulatory suppression. In Experiments 3 and 4, we examined the robustness of the L1 advantage effect by employing bilingual lists. We also took advantage of the bilingual lists, which were composed of stimuli in two different classes (i.e., Chinese and English materials) that have been employed in the probe recalled task to reveal the organization of short-term memory, to investigate whether there are language-based chunking/organizing rules in bilingual VSTM. Inconsistent with the findings from Experiment 1 and 2, the results of Experiment 3 did not show significant difference between the recall of Chinese and English materials. We suspected that the non-significant L1 advantage resulted from the fact that participants mainly relied on phonology to retain the materials. That might be due to the increasing task difficulty by employing the bilingual lists. In Experiment 4, we employed tetrads of rhyming Chinese and English materials, instead of homophonic ones, to increase phonological similarity among the materials within each trial to avoid the participants heavily relying on phonology. We tested two groups of Chinese-English bilinguals who have immigrated to the US before or after age 7. The two groups of early- and late-arrival bilinguals have better proficiency in English and Chinese, respectively. The findings from both groups revealed the robust advantage of dominant language, which replicated the L1 advantage in previous experiments. More important, both groups demonstrated better memory performance when the probe and the target were within the same language class than when they switched from one language to the other. That is, having either the probe or the target in participants’ dominant language did not result in better recall accuracy than when both the probe and the target were in the non-dominant language. Such results suggested that materials in different languages were retained separately in VSTM. Overall, the current findings from the probed recall task support the contribution from non-phonological representations to the effect of L1 advantage, and reveal the chunking/organizing principles based on language classes in Chinese-English bilinguals’ VSTM. |
