以語料庫分析臺灣英語學習者學術寫作動名搭配詞中之少用動詞

Abstract

儘管詞彙的多樣性在論文寫作中是非常重要的，學習者的論文卻缺少這個特性。過去針對學術寫作中動詞使用的研究發現，學習者鮮少使用學術動詞，或使用之學術動詞缺乏多樣性。然而，先前的研究多半僅局限在量化分析，且研究中動詞被視為單一的獨立個體，難以觀察學習者少用動詞的語意與情境。 為補足對學習者少用動詞的理解，本文旨在以動名搭配詞為意義單位，探討台灣英語學習者學術論文中的少用動詞。本文具有三大研究目的，一為指出學習者少用而期刊作者常用之動名搭配詞，以提供學術寫作學習與教學之參考。二為分析學習者少用之動名搭配詞是否有相對應過度使用的近義搭配詞。三為詮釋少用特定學術動詞之可能原因，以增進對於學習者學術詞彙使用之理解。 為達上列之研究目的，本文採取語料庫分析之方式，比較台灣英語學習者語料庫與期刊文章語料庫。學習者語料庫涵蓋選自應用語言學與英語教學領域共494篇碩士論文，期刊文章語料庫包含1500篇相同領域之國際期刊文章。兩者語料庫各容納約一千兩百萬字。本研究運用Sketch Engine中Sketch Diff之功能，半自動化分析語料中300個高頻目標名詞，進而找出學習者碩士論文少用之動名搭配詞。 本研究共指出165個學習者少用之動名搭配詞，涵括了87個不同之少用動詞。期刊作者平均使用各搭配詞14次以上，而學習者幾乎顯少使用，亦即平均1次。其中18項為期刊作者使用20次以上而學習者幾乎不使用之顯著少用之動名搭配詞。進一步的分析後可將此165個搭配詞分為兩組：一為有相對應過度使用搭配詞的少用動名搭配詞，另一為無相對應過度使用搭配詞的動名搭配詞。 對顯著少用搭配詞的質化分析發現，學習者少用在學術寫作中帶有不同語意之動詞，且部分少用例子受到母語翻譯之影響。對比少用及語意相近的過度使用搭配詞則發現，學習者偏好使用高頻動詞，如：need, get與bring，或虛義動詞，如：make, have 與do，而迴避使用低頻、母語程度 (native-like) 之動詞，如：warrant 與yield。對其他無相對應過度使用搭配詞例子之分析，亦顯示多重語意之動詞，如：establish與permit，及動詞語意韻律，如：support對學習者學術動詞使用造成之困難。本文進一步將少用動詞歸納出六個動詞語意群組，此六群組顯示可能為學習者缺乏動詞詞彙多樣性的語意範疇。 本文研究結果顯示，不熟悉動詞的多重語義與在學術寫作中的獨特語意，為少用動詞或動名搭配的重要原因之一。母語翻譯亦影響學習者過度使用或少用動詞的現象。最後，學習者保守、不冒險的詞彙運用風格亦為其一可能之影響因素。因此，本文建議學術寫作教師將本研究發現之少用動名搭配詞融入課程中。並提示學習者部分動詞在學術寫作中常用之特定語意，與提醒學生避免以直接翻譯之方式學習學術動名搭配詞。In academic writing, the use of varied lexical items or lexical diversity is of great importance. Despite the fact, learners’ writing has been found to lack this property. Previous research on learners’ use of verbs in argumentative writing has revealed learners’ infrequent use, underuse and lack of lexical variety of academic verbs. However, these earlier studies were mostly quantitative, and verbs had been treated as independent items. It is unknown for which meaning or in what contexts the verbs were used. To address the gap, the present research set out to investigate the underused verbs for Taiwanese EFL learners by viewing verb-noun collocations as a meaningful unit. The current study aimed to first produce a list of underused verb-noun collocations for pedagogical and learners’ reference. Secondly, the present study proposed to identify the counterparts of the underused collocations. Most importantly, with qualitative analysis, this study intended to further explain some possible reasons for learners’ underuse of some academic verbs. To fulfill these research purposes, corpus analysis on the verb-noun collocations in Taiwanese EFL learners’ academic writing was performed. The learner corpus, Master Thesis corpus (MT), composed of 494 master theses by the learners from 10 programs of applied linguistics and English teaching was compared to the reference corpus, Research Article corpus (RA), composed of 1500 research articles in the same field. Each of the corpora consists of approximate 12 million words. 300 target nouns were researched, and Sketch Diff function of the Sketch Engine was employed to help extract the underused verb-noun collocations semi-automatically. A total of 165 verb-noun collocations, which covered 87 different underused verbs were identified to be underused by the learners. Published article writers used each of these collocations average 14 times while learner writers almost never used (i.e. averagely 1 time). The 18 of the collocations were identified to be salient underused verb-noun collocations, which experienced writers employed more than 20 times while learner writers barely applied. The examination of the overused verb-noun collocations further divided the 165 collocations into two categories: 53 of them were identified to be underused collocations with overused counterparts, and 112 of them without overused counterparts. Qualitative analysis on the 18 salient underused v-n collocations revealed learners’ underuse of verbs with a different meaning sense in academic writing. In some cases, learners appeared to be influenced by L1 translation. The juxtaposition of underused collocations with overused counterparts unveiled learners’ inclination to high-frequent verbs (e.g., need, get and bring) and de-lexical verbs (e.g., make, have and do) and their avoidance of low-frequency but native-like verbs (e.g., warrant and yield) in forming collocations. The analysis of underused v-n collocations without overused counterparts also indicated learners’ difficulty with the multiple meaning senses of verbs (e.g., establish and permit). In a few cases (e.g., support), learners revealed unfamiliarity with semantic prosody. In terms of underused verbs, six semantic groups of verbs, including construct, confirm, exhibit, elicit, limit and challenge, and situate were found to be the challenging for learners revealed a lack of lexical diversity. Based on the findings, some pedagogical implications and suggestions for future research were discussed and offered at the end of this research.