研究生姓名 : 施欣怡
學年度 : 101
論文題目 : The acquisition of Chinese literacy as a second language correlates with statistical learning of implicit transitional probability
英文論文題目 : The acquisition of Chinese literacy as a second language correlates with statistical learning of implicit transitional probability
指導教授 : 吳嫻
系所名稱 : 認知與神經科學研究所
論文頁數 : 92
中文關鍵字 : 第二語學習、內隱學習
英文關鍵字 : second language acquisition、implicit learning、logographic languagessecond language acquisition、implicit learning、logographic languages
中文摘要 :
近期的研究發現,統計學習能力不只和拼音語言中的母語識字學習 (literacy) 有關,且與拼音語言的第二語識字學習表現亦息息相關。然而,統計學習能力與非拼音語言作為第二語的識字學習能力之間的關係,在過去的研究中,並沒有清楚的討論。因此,本研究的目的在於探討統計學習能力這種找尋隱藏在序列訊息中之規律的能力,以及聽覺和語音處理能力是否能夠有效預測成年人學習中文為第二語時識字程度之表現。
實驗一結果顯示母語為拼音語言的受試者之中文識字表現與學習中文時間長短以及視覺統計學習能力表現呈現顯著相關。此外,透過迴歸統計分析,在排除智商、工作記憶等變項之影響後,視覺統計學習能力仍可有效預測外籍受試者的中文識字量。實驗二則是追蹤受試者之中文識字程度的發展,並進一步檢驗其與上述之可能預測因子之間的關係。實驗結果顯示,整體而言,曾經參與實驗一且參與了實驗二的受試者,於實驗二的中文識字能力表現顯著優於實驗一。儘管相關與迴歸分析結果顯示,中文識字學習能力與所有的預測因子並無顯著相關,然而進一步的分析指出,在實驗一與實驗二間隔的六個月間仍持續修習中文課程者,其進步幅度較未修習者大,此效果尤其可見於閱讀流暢度測驗的表現。實驗結果亦顯示,對於在過去六個月間並未繼續修習中文課程的受試者而言,其中文識字學習表現與視覺統計學習能力 (VSL) 呈現顯著相關。整體來說,正式的授課與教學整體而言提升了受試者中文識字能力的表現,而統計學習能力之表現對於中文識字能力表現的影響,尤其顯現於受試者沒有繼續透過正式課程學習中文的情況下。
總結本研究,統計學習能力是學習中文作為第二語之識字學習表現的重要預測因子。這結果與先前來自拼音語言之研究結果相符,亦即無論作為母語或是第二語學習,統計學習能力對於識字學習表現皆有重要影響。此外,本研究結果顯示,此種找尋內隱規則並計算項目之間關聯機率的能力對於識字學習表現之影響,不僅見於拼音語言,在中文的學習過程中,亦可見其與中文識字學習能力息息相關。
英文摘要 :
Recent findings have indicated that statistical learning ability correlates with literacy acquisition of alphabetic languages both as a first and a second language. However, it was unclear whether the correlation was demonstrated only in alphabetic languages or not. In the present study, we examined whether the same ability of detecting statistical regularities embedded in sequential stimuli, as well as the abilities of auditory perception and phonological processing, predict the literacy level in adults who learn Chinese as a second language.
Cross-sectional results from native speakers of alphabetic languages showed that the performance of Chinese literacy was significantly correlated with the time participants spent on learning Chinese and their performance on statistical learning of visual stimuli (VSL). The regression analysis further revealed that VSL accounted for unique variance of the size of Chinese characters that participants know even when the contribution from other variables, intelligence, and working memory was considered.
The follow-up research examined whether the potential factors have influence on the development of Chinese literacy. Analysis of the data from all remaining participants who were tested in both Experiment 1 and 2 did detect obvious improvement of Chinese literacy. Correlation and regression results showed no significant relationship between all the predictor variables and Chinese literacy. However, further analysis indicated that participants who continued studying Chinese during the past six months between Experiment 1 and 2 had greater improvements in Chinese literacy, especially in reading fluency. Significant correlation was also observed between the improvement of Chinese literacy indices and the performance of VSL for participants who did not continue taking formal Chinese courses. That is, formal instructions generally improved the performance of Chinese literacy, while the performance of VSL had impacts on Chinese literacy acquisition, especially when formal instruction was not available.
To conclude, the current study suggests that the ability of statistical learning is a critical predicting factor for learning Chinese as L2. These results are compatible with previous research that statistical learning ability has association with literacy acquisition both for L1 and L2. Moreover, the results demonstrated that not only in alphabetic languages, but also in Chinese, the ability to extract regularities and to compute transitional probability is greatly correlated with literacy acquisition.
 
地址:320 桃園市中壢區五權里2鄰中大路300號 服務信箱: ncu5200@ncu.edu.tw
電話:(03) 422-7151轉65200 傳真:(03) 426-3502
中大認知所 版權所有 © 2005 Institute of Cognitive Neuroscience. All rights reserved.
請以IE6.0或更新版本,以及1024 x 768以上之解析度瀏覽以取得最佳效果。