以眼動儀和測驗探討聽讀法對第二語言英文閱讀理解與認知歷程的影響

Abstract

聽讀法（Reading-while-listening）已被指出對初學者的語言學習和閱讀有益，也在第二語言（second language）的研究中越來越受到關注。針對該新興議題，本研究以測驗和眼動儀探討聽讀法對第二語言學習的閱讀理解與認知歷程有何影響，且是否會因讀者有不同英文能力而有不同影響效果。此外，由於解碼（decoding）效率長期以來一直被視為影響閱讀理解的重要因素，其中詞頻效應（the word frequency effect）為字詞辨認受到最多關注的焦點之一，因此，本研究也納入詞頻為變項，探討二語習得的閱讀歷程。最後，該實驗也會探討讀者的情意因子，了解不同語言程度的學習者對於聽讀法的態度與感受。本研究招募87位大學生英語學習者，根據其自陳的英語證照考試結果，分為高能力組（CEFR C1級）與低能力組（CEFR B2級），兩組分別有43名和44名參與者。本實驗一共設計三種閱讀情境—聽讀（reading-while-listening）、有時間限制的閱讀（reading-only with time limitation）、沒有時間限制的閱讀（reading-only without time limitation），受試者被隨機分配到這三種情境中的兩種進行閱讀，讀兩篇不同的英文文章。讀文章前，每位受試者都需完成一份詞彙程度的預測測試（vocabulary levels test），而閱讀文章時，讀者的眼動歷程會被記錄，之後完成英文閱讀理解後測。最後，參與聽讀組別的受試者需要填寫一份閱讀經驗的偏好問卷，以了解他們對聽讀的態度。研究結果有四項主要的發現。首先，閱讀情境和英語程度有交互作用，高英文能力的讀者自己讀的閱讀理解表現顯著優於聽讀的情境，可用冗餘效應（redundancy effect）來解釋此結果，外來的語音對於高語言能力讀者處理文本可能反而是干擾；相反地，低英文能力讀者則是在聽讀情境的閱讀理解表現最好，此結果可以雙路徑理論（dual-route theory）和字型—語音整合理論（print-speech integration）來解釋，提供語音可能可以幫助語言能力低的讀者快速連結形音對應，進而促進閱讀理解表現。第二，眼動資料的分析結果顯示無論英文能力高或低的讀者，在沒有時間限制的閱讀情境下，他們讀英文文章的總閱讀時間（total reading time）和總凝視次數（total fixation count）都比在另外兩種情境來的長和多。而英語程度則影響單一平均凝視時間（mean fixation duration）和平均跳移距離（mean saccade length），高能力組在任何閱讀情境下都有較低的單一平均凝視時間和較高的平均跳移距離，代表高能力者解碼英文的速度較快且處理英文文本的知覺廣度較大。第三，本研究發現詞頻效應在二語閱讀也很明顯，低頻詞的被凝視次數、重讀時間（rereading time）和總凝視時間（total fixation duration）都顯著高於高頻詞。第四，雖然低英文能力者在聽讀情境的閱讀理解表現較佳，但問卷資料顯示無論英文能力高或低的讀者皆自陳其偏好單純閱讀的情境。綜上，本研究結果的發現除了支持二語閱讀的理論架構，也提供語言學習與教學的實務建議，針對不同英文能力者是否要提供語音應有不同的做法。

Reading-while-listening has gained growing attention in second language (L2) reading, and it has been demonstrated to have beneficial effects on language learning. However, research to date has yet reached the consensus over its influence on reading comprehension, and little is known about L2 eye-movement behavior during reading-while-listening. In response to this emerging research focus, this study aimed at exploring how the additional audio would impact the cognitive processes and comprehension results of L2 readers with different English proficiency levels. Besides, since decoding efficiency has long held as an essential factor of reading comprehension, the word frequency effect, one of the most well-established areas regarding word recognition, would also be investigated through the eye-tracking technology. Lastly, readers’ perception towards different reading experiences would be examined to understand L2 reading-while-listening in a comprehensive manner.A total of 87 adult English learners were recruited. Based on their self-report English certification results, they were assigned to either the high-proficiency group (CEFR C1 level) or the low-proficiency group (CEFR B2 level), with 43 and 44 participants respectively. These learners were then be randomly assigned to two of the three reading conditions – reading-while-listening, reading-only with time limitation, and reading-only without time limitation. During the experiment, each participant had to finish a vocabulary levels pre-test, read two English articles with their eyes monitored, and complete a comprehension posttest right after reading each text. In the end, those participating in reading-while-listening were required to fill out a questionnaire about their preference towards the reading experiences.The results showed four major findings regarding the reading-while-listening condition. Firstly, both the reading condition and the English proficiency level had effects on the comprehension posttests. That is, the high-proficiency participants performed significantly worse on the comprehension posttests after reading-while-listening, while the low-proficiency counterparts gained higher scores, compared with those after the two reading-only conditions. Hence, the redundancy effect can be supported through the results of the high-proficiency group, whereas the dual route theory and print-speech integration can be supported by those of the low-proficiency group. In addition, the reading condition had effects on the total reading time and total fixation count, meaning that all participants had higher total reading time and total fixation count during reading-only without time limitation than the other two conditions. By contrast, the English proficiency level influenced the mean fixation duration, and mean saccade length, implying that the high-proficiency group had lower mean fixation duration and longer mean saccade length than the low-proficiency group across the three reading conditions. Moreover, the word frequency effect was larger in L2 reading, with the fixation count, rereading time, and total fixation duration of low-frequency words significantly higher than those of high-frequency words. Lastly, even though the low-proficiency group was benefited from the auditory inputs, all participants regarded reading-only to be the preferable reading condition. Through these findings, not only the theoretical framework of L2 eye movements can be strengthen, but language learners and educational practitioners could have clearer idea about the implementation of reading-while-listening to facilitate the reading process and comprehension results.