基於語料庫的動詞學習困難之研究: 及物性、語意韻、近義詞

Abstract

許多研究已經證實學習者電腦語料庫(computer learner corpora)有助於第二語言的學習。然而因電腦語料庫的建構不易，導致過去研究在語料數量上頗受限制。本研究採用臺灣師範大學陳浩然教授所建立約150萬字的臺灣學習者英語語料庫(TLEC)與中國學習者英語語料庫(CLEC)107萬字，探討動詞的學習困難。主要研究學習動詞困難之三個面相:(1)探討外語學習者動詞及物性錯誤分析；(2)外語學習者與母語人士在語意韻使用的差異；(3)外語學習者在近義詞使用的錯誤。 本研究先以AntConc為語料分析工具，找出臺灣與大陸學生最常使用的200個動詞，再加上動詞及物性、語意韻與近義詞之相關文獻資料分析，以列出研究動詞為(1)Arrive, Agree, Care, Dream, Listen, Reach (2) Cause, Happen (3) Happen/Occur, Say/Tell, Understand/Realize。然後進行錯誤分析(error analysis)與中介語對比分析(contrastive interlanguage analysis)，至於母語語料庫之參考，則以美國當代英語語料庫(COCA)為基準。 研究結果有三項顯示：(1)外語學習者較易將不及物動詞當成及物動詞使用,以Arrive的錯誤率最高(37.2%)，Listen次之(29.4%)，至於及物動詞Reach的錯誤率則最低(9%)。錯誤原因可能是母語影響(L1 influence)，因中文並沒有不及物動詞的使用情形。(2)過去研究顯示Cause與Happen帶有負面或消極語意韻，常與偏負面詞類並列，就學習者語料庫分析，學生多能正確使用，但仍出現正向積極的詞彙，造成語意上混淆。(3)關於近義詞使用，參考字典多未能清楚說明兩個近義動詞的區別，在Occur使用上出現很多搭配詞錯誤，原因可能是學生在英文寫作上仰賴逐字翻譯(word-by-word translation)。至於Say/Tell，學生亦習慣於中文翻譯方法而出現”say to the teacher”(跟老師說)的錯誤，對這兩個動詞結構，如Tell須加直接受詞，而學生無法正確使用。Realize在語料中出現很多中介語搭配詞(interlanguage collocation)，在母語語料庫中相當少用，如 *realize the theory, *realize the reason。 同時，本研究結果也發現台灣與大陸語料庫，儘管學習背景、學生年齡、英文程度不相同，但高頻動詞使用的相似度達94%，顯見學生在英文寫作上多仰賴這些高頻動詞。 透過語料檢索工具，研究者得以處理與分析大量語料，本研究提供許多實證資料，以了解學生的動詞學習困難，尤其搭配詞的錯誤，在各個面相中均大量出現，建議英語教學者能針對本研究所出現的搭配詞錯誤以設計教材，並且提供學生真實語料，讓學生在英文寫作上可以正確使用。Many studies have confirmed the benefits of using computer learner corpora in SLA. Yet, few of the studies were conducted in Taiwan due to a lack of availability of large computer learner corpora. The present study applied 1.5 million-word Taiwanese Learner English Corpus (TLEC), created by Professor Howard Chen at National Taiwan Normal University, and 1.1 million-word Chinese Learner English Corpus (CLEC) to explore EFL learner difficulties in verbs. The purposes of the study were focused on (1) exploring transitivity errors (2) differences of semantic prosody between native speakers and EFL learners (3) errors in near-synonyms produced by EFL learners. The study first used AntConc to generate top 200 verb list from the EFL learner corpora. Based on a review of related studies, the selected verbs for analysis were (1) Arrive, Agree, Care, Dream, Listen, Reach, (2) Cause, Happen, (3) Happen/Occur, Say/Tell, Understand/Realize. The methods included error analysis and contrastive interlanguage analysis with COCA as the reference corpus. The results were divided into three parts. (1) Learners tended to misuse intransitive verbs as transitive ones. The highest error rate of transitivity was Arrive (37.2%) followed by Listen (29.4%); the lowest error rate is for Reach (9%). The possible source of error may be attributed to L1 influence since there is no intransitive verb usage in Chinese. (2) The past studies showed that Cause and Happen carry a negative semantic prosody. The present corpus-based analyses revealed that most students were able to use semantic prosody appropriately with Cause and Happen. However, there were instances where positive words collocated with the two selected verbs and created some confusion of meaning. (3) As for near-synonyms, the consulting dictionaries failed to provide relevant information to differentiate the two competing words. The corpus-based analyses showed that there were collocation errors with Occur and Happen. Students seemed to rely on word-by-word translation strategy in English writing. As for Say/Tell, students were unable to identify the correct structure with each verb, such as a direct object after Tell. Many errors were also resulted from L1 direct translation. In the last group of near-synonyms, there were many interlanguage collocations with Realize:*realize the theory, *realize the reason. Those lexical combinations were rarely found in the native corpus. Despite different learning backgrounds, students’ age and language proficiency, the study also found that students in TLEC and CLEC depend on similar high frequency verbs in their writing (94% of similarity in the top 200 verb list). With the help of corpus tools, we were able to process and analyze large amounts of corpora. The present study provided empirical evidence on learners’ interlanguage, learners’ errors in using verbs. Particularly collocation errors were identified in almost every search verb. We suggest that language teachers design teaching materials based on the errors and provide students with an authentic corpus to improve their writing.