使用參考工具對搭配詞學習之成效:以台灣英語學習者為例

全文

(1)國立臺灣師範大學英語學系碩. 士. 論. 文. Master’s Thesis Department of English National Taiwan Normal University. 使用參考工具對搭配詞學習之成效: 以台灣英語學習者為例. The Effects of Using Referencing Tools on Collocation Learning: A Case Study Based on Taiwanese EFL Learners. 指導教授: 陳浩然 Advisor: Dr. Hao-Jan Howard Chen 研究生: 吳依蓁. 中華民國一百零三年七月 July, 2014.

(2) 中文摘要隨著語言教學研究發展的精進，越來越多研究結果顯示搭配詞的教學不僅能幫助學習者增進單字量，還可以提升單字運用能力，進而幫助聽說讀寫各方面能力。除一般課堂教學外，學習者如能運用合適的參考工具輔助學習搭配詞，必能收事半功倍之效。因此，探討何種參考工具適合輔助英語學習者學習搭配詞有其必要性。本篇研究旨在探討台灣高中生透過字典、搭配詞辭典及搭配詞索引，學習英語動詞名詞搭配詞的成效差異。受試者為一百一十位高二普通班學生，英語程度介於中級到中高級之間，分別被指派使用朗文線上字典(Longman Dictionary of Contemporary English)、牛津搭配詞辭典(Oxford Collocation Dictionary of English)、和線上搭配詞索引 Webcollocate。測驗工具為一份由研究者設計的三十題搭配詞填空。測驗分為: 前測、後測及在使用搭配詞輔助工具的情況下填寫的測驗。實驗共為期四週: 第一週—學生分別在未使用任何輔助工具的情況下完成前測。第二週—學生利用參考工具完成測驗，並於測驗完成後，填寫一份問卷，針對學生的搭配詞背景知識，搭配詞學習型態，及使用參考工具後的感想。第四週—為了解受試者在使用搭配詞輔助工具後的學習成效，在隔了一週後，受試者接受後測。施測者接著將三組受次者的前測、後測及在使用搭配詞輔助工具的情況下填寫的測驗總分分別做比較，發現朗文組在在使用搭配詞輔助工具的情況下填寫的測驗進步最多，分數也最高，至於後測成績，三組無顯著差異，但其中 Webcollocate 組分數最高。相較於朗文與 Webcollocate 兩組，牛津組在各測驗的表現皆較不理想。造成以上結果差異的原因有三: 第一、朗文雖為一般目的字典，但它不僅提供單字定義解釋、足夠的例句顯示該搭配詞的適用情境，還有依該字的慣用搭配詞欄位，讓受試者能快速瀏覽其搭配詞，並透過其簡短解釋與例句，立即判斷該 i.

(3) 搭配詞的用法。在另一方面，牛津雖為搭配詞辭典，但因目標使用者預設為英語能力中高或高級程度的使用者，所以辭典裡沒有給單字定義解釋，此外，因意義相似的搭配詞皆列在一起，又無充足的例句供使用者分析判斷。至於 Webcollocate，雖沒有單字定義解釋，但有字頻及大量的語料例句供使用者閱讀判斷。另外 Webcollocate 因為需要較長的網路處理時間，並且花時間閱讀一筆筆語料分析決定最適合的搭配詞，對使用者而言，需耗費較多時間查詢搭配詞，因此成效雖沒有陳列明確搭配詞欄位且資料處理快速的朗文佳，但仍比缺乏字義解釋及例句的牛津好。所以於本篇研究中，相較於牛津與 Webcollocate，朗文反而較適合英文程度為中級的高中生使用。在問卷部分，學生們表示平時學習單字及閱讀時皆有注意其搭配詞的運用，但遇到搭配詞問題時，只有 30%表示會主動運用線上工具，在透過本次實驗，使用指定的輔助工具後，大多數學生普遍皆有正面回饋，表示如果未來有機會的話，願意透過此輔助工具繼續學習其他類型的搭配詞。希望藉本篇研究結果證實搭配詞參考工具之功效，促使進一步的相關搭配詞工具研究，幫助高中英文老師進行更有效的搭配詞教學，進而增進學生單字運用及聽說讀寫各項能力。. 關鍵字: 搭配詞(collocation)、搭配詞參考工具(collocation referencing tool)、字典 (dictionary)、搭配詞辭典(collocation dictionary)、搭配詞索引(collocation retrieval tool). ii.

(4) ABSTRACT Recently, more and more studies have revealed that collocation learning can enhance not only the learners’ vocabulary bank but also their ability of word use, and thus improve their overall language skills in the long run. Besides the lectures and exercises on collocations given by English teachers in class, appropriate referencing tools can also be applied to helping students learn collocations. Online referencing tools such as general dictionaries, collocation dictionaries, and collocation retrieval tools are considered useful and reachable for learners. Therefore, it is necessary to do research on investigating which type of referencing tool is more effective and how it can be applied to improve students’ collocation learning. The present research is to investigate the effect of using online/electronic dictionary, collocation dictionary, and collocation retrieval tool on Taiwanese senior high school students’ collocation learning. The participants were 110 high school students (3 classes) with intermediate level. Each class was respectively assigned to use one collocation referencing tool—Longman Dictionary of Contemporary English (LDCE), Oxford Collocation Dictionary of English (OCDE), and Webcollocate (WEBC)—to take a test containing 30 items of collocation blank-filling. The overall process of the experiment lasted 4 weeks: a pretest was conducted in the first week to check the students’ knowledge of collocation before the treatment. Three classes of participants were asked to take the test without any assistance of collocation referencing tools. In the second week, the participants could use the assigned collocation referencing tool to do a fifty-minute collocation task and a questionnaire on their perceptions of the assigned collocation referencing tool. In the fourth week, the posttest was conducted to check the effect of the referencing tools on the participants’ collocation ability. iii.

(5) Results showed the LDCE group got the highest scores in the treatment task. As to the retention, the WEBC group performed the best in the posttest, though not significantly. The OCDE group did not score as well as the other two groups in any of the tests and the task. Six vital factors were found contributing to the success rate of the LDCE group. First, a clear layout, including the bright color and the boldface type for collocations, facilitated the searching process. Second, simple but essential explanations of the collocations seemed necessary for learners of intermediate proficiency level to decide the proper collocates. Third, adequate example sentences helped the participants make sure the right usage of collocations through the context of the example sentences. Fourth, the collocation box covering the most commonly-used collocations made it fast and convenient to search for the right answers. Fifth, the numbers of collocations provided in LDCE were not too large to overwhelm the participants. Only the most frequently used collocations were presented in LDCE. Last but not least, the LDCE’s speed of searching process was high so that the users could receive the answers in a limited of time. Despite the inferior performance of the task, the OCDE group still held a rather positive attitude toward the collocation dictionary. They regarded it as a useful and authoritative dictionary with rich content. In general, most of the participants in the three groups all expressed their willingness to learn different types of collocations with the assistance of the referencing tools.. Keywords: collocation, collocation referencing tool, dictionary, collocation dictionary, collocation retrieval tool. iv.

(6) ACKNOWLEDGEMENT. This thesis could not have been possible without many people’s help and support. First and foremost, I owe my greatest appreciation to my advisor, Dr. Hao-Jan Chen. Without his constant encouragement and patient guidance given to me all these years, I could not have finished this work. I also want to express my gratitude to the two committee members of my thesis: Dr. Chih-Cheng Lin, and Dr. Shu-Li Lai. They kindly spared their precious time out of their tight schedule to review my thesis and offered insightful suggestions to help me revise my thesis. My great thankfulness also goes to my senior high school teacher, Jenny Lin. Without her generous help and support, I could not have smoothly conducted the experiments. It is her incredible patience and inspiration that makes this work possible. I would also like to thank my dear friends: Jenny Li, Irene Hu, Roxanne Liu, Eros Chen, Linda Wu, and Madeline Liu for cheering me up and being all ears to me. Their warm company has made my study in graduate school as well as my life full of happiness. Last but not least, I want to show my deepest gratitude to my beloved family, including my parents, my brother, my sister-in-law, and my cute niece for their unconditional love and care. It is them who are always there for me, give me strength when I feel like giving up, and stand by me when I face all the obstacles during the writing process. I believe the love of my family is the greatest blessing of my life.. v.

(7) TABLE OF CONTENTS 中文摘要............................................................................................................................... i ABSTRACT .......................................................................................................................iii ACKNOWLEDGEMENT ................................................................................................. v TABLE OF CONTENTS .................................................................................................. vi LIST OF TABLES ...........................................................................................................viii LIST OF FIGURES .......................................................................................................... ix CHAPTER ONE INRODUCTION ............................................................................. 10 Background ................................................................................................................ 10 Purpose of the Study .................................................................................................. 16 Significance of the Study ........................................................................................... 17 Organization of the Study .......................................................................................... 18 CHAPTER TWO LITERATURE REVIEW ............................................................. 19 The Definitions of Collocation .................................................................................. 19 Collocations in Language Learning ........................................................................... 22 The importance of collocation learning. ............................................................ 22 Learners’ difficulties in collocation learning. .................................................... 24 Collocation Learning and Computer Assisted Language Learning ........................... 38 Electronic dictionaries in collocation learning. ................................................. 39 Corpora and concordancers in collocation learning. .......................................... 42 CHAPTER THREE METHOD ................................................................................... 62 The Overall Design .................................................................................................... 62 Participants ................................................................................................................. 63 Instruments ................................................................................................................. 63 Tools ........................................................................................................................... 64 Oxford Collocations Dictionary for Students of English (OCDE). ................... 64 Longman Dictionary of Contemporary English (LDCE). ................................. 68 WebCollocate--wikipedia edition (WEBC). ...................................................... 71 Experimental Procedure ............................................................................................. 74 Data Collocation and Analysis ................................................................................... 76 CHAPTER FOUR RESULTS ..................................................................................... 78 Results of the Tests and the Task ............................................................................... 78 Results of the pretest. ......................................................................................... 78 Results of the treatment task. ............................................................................. 79 Results of the posttest. ....................................................................................... 79 The comparison of the scores for the tests and the treatment task..................... 80 The differences in the correct rates of each item among the groups.................. 81 Results of the Questionnaire ...................................................................................... 82 vi.

(8) Results of the perceptions. ................................................................................. 82 Results of the written comments. ....................................................................... 88 CHAPTER FIVE DISCUSSIONS ............................................................................. 92 Reasons for the Differences Among Groups in the Treatment Task .......................... 92 Clear layout. ....................................................................................................... 93 Simple but essential explanations. ..................................................................... 96 Adequate example sentences. .......................................................................... 101 Explicit collocation boxes. ............................................................................... 102 The number of collocations provided. ............................................................. 104 The speed of information processing. .............................................................. 106 Reasons for the Differences Among Groups in Collocation Retentions .................. 107 Perceptions and Characteristics of the Referencing Tools ....................................... 108 The students’ perceptions of the referencing tools........................................... 108 Strengths and weakness of the referencing tools. ............................................ 112 Conclusions .............................................................................................................. 114 CHAPTER SIX CONCLUSIONS ........................................................................... 116 Summary of the Major Findings .............................................................................. 116 Pedagogical Implications ......................................................................................... 119 Limitations and Directions for Future Research ...................................................... 120 REFERENCES ............................................................................................................... 123 APPENDIX A Collocation Test .................................................................................. 132 APPENDIX B Questionnaire ..................................................................................... 136. vii.

(9) LIST OF TABLES Table 1 Empirical Studies on Collocation Learning through Dictionaries .................. 58 Table 2 Empirical Studies on Collocation Learning through Concordancers .............. 59 Table 3 Data Collecting Procedure .............................................................................. 75 Table 4 The Mean Scores of Each Group in Pretest, Treatment Task, and Posttest .... 79 Table 5 The Score gains of Each Group in the Pretest, Treatment Task, and Posttest . 80 Table 6 The Comparison of the Performances Among Groups ................................... 81 Table 7 The Comparison of the Performance of Each Item Among Groups ............... 82 Table 8 The Percentage of the 10 Items in the Questionnaire ..................................... 84 Table 9 The Average Scores of the 10 Items in the Questionnaire .............................. 87 Table 10 The Strengths and Weaknesses of the Referencing Tools ............................. 88 Table 11 The Comparison of the Three Referencing Tools ....................................... 115. viii.

(10) LIST OF FIGURES Figure 1. The search results of attention in OCDE ..................................................... 67 Figure 2. The search results of attention in LDCE. ..................................................... 70 Figure 3.The search results of attention in WEBC. ..................................................... 73 Figure 4. The search results of headline in LDCE (left) and OCDE (right). .............. 93 Figure 5. The search results of headlines in WEBC. ................................................... 94 Figure 6. The search results of awareness in LDCE (left) and OCDE (right). ........... 96 Figure 7. The search results of measurement in LDCE (left) and OCDE (right). ....... 97 Figure 8. The search results of chance in WEBC. ....................................................... 98 Figure 9. The search results of performance in WEBC. .............................................. 99 Figure 10 .The search results of tear in LDCE (left) and OCDE (right). .................. 100 Figure 11. The search results of chance in LDCE (left) and OCDE (right). ............. 101 Figure 12. The search results of duty in LDCE (left) and OCDE (right). ................. 103 Figure 13. The search results of issue in LDCE (left) and OCDE (right). ................ 104 Figure 14. The search results of issue in OCDE........................................................ 105 Figure 15. The search results of issue in WEBC. ...................................................... 106. ix.

(11) CHAPTER ONE INRODUCTION Background Over the past decades, collocation competence has been recognized as one important element in linguistic proficiency. Such competence not only helps reduce learners’ cognitive effort but also save their processing time. On top of that, learners with collocation competence acquire language available for immediate use. For learners pursuing high proficiency level, collocation competence makes them more native-like by enhancing both their accuracy and fluency. Hence, collocation competence has been regarded as one of the major factors contributing to the differences between language competence of native speakers and that of second language learners (Bahn and Eldaw, 1983; Nation, 2001; Wray, 2002). While the importance of collocation competence is highly emphasized, studies have revealed that many learners, even advanced ones, have difficulties in producing idiomatic word combinations precisely and thus impeded listeners’ comprehensibility (Cooper, 1999; Howarth, 1998; Wray, 2000). Moreover, due to inadequate awareness of collocation patterns, some learners tend to excessively rely on the synonymous words based on their L1 translations, and combined the words together for their L2 production. Others incline to make word combinations according to grammatical rules instead of lexicalized routines. Many learners fail to notice that collocations are often 10.

(12) fixed and cannot be replaced by other expressions. Even advanced learners might confront such collocation challenges since their development of collocation competence did not parallel with those of their general vocabulary and grammatical competence (Skehan, 1998; Foster, 2001). As learners’ difficulties in collocation acquisition were detected, researchers have further explored the types of miscollocations language learners produced. Among all, lexical miscollocations turned out to be the most prevalent error types, and verb-noun (V-N) collocations, especially, presented the greatest difficulty for learners (Liu, 2002; Chan & Liou, 2003; Wible et al, 2003). One possible reason was that V-N collocations occupied most part of the lexical collocations. According to Cowie’s (1992) study, over 30% of collocations occurring in native speaker social context are verb-noun collocations. In addition, the interlanguage influence was found to be another possible reason for the learning difficulty of V-N collocations (Nesselhauf, 2003). How to help students overcome such collocation difficulty has thus become an urgent issue to be explored. To solve these collocation learning problems, studies have suggested that it is necessary to make language learners aware of the existence of collocations and the collocational differences in languages (Coxhead, 2008; Durrant & Schmitt, 2010; Willis, 2003). Moreover, students should be guided to produce proper collocations. 11.

(13) with the explicit collocation instructions, which could significantly improve their collocation knowledge (Conzett, 2000; Lin, 2002, Liu, 2002; Tseng, 2002). Apart from teachers’ instructions on collocation, the assistance of referencing tools also plays a key role in collocation learning. Dictionaries, among all referencing tools, seem to be the most frequently used (Laufer, 2011). BBI Dictionary of English Word Combinations (Benson, Benson & Ilson, 1997) and the Oxford Collocations Dictionary for Students of English (OCDSE), whose content is based on British National Corpus (BNC), are two primary paper collocation dictionaries. After the presence of paper collocation dictionaries, electronic collocation dictionaries have come into views. The electronic versions outperformed the paper ones in the following aspects: making the process of dictionary consultation less time-consuming, encouraging learners’ exploratory browsing, and increasing the quality of the information supplied (Laufer & Hill, 2000; Nesi, 2000). The findings of empirical studies have also shown that online dictionaries benefited language learning more than the same dictionary in the paper form (Dziemaank, 2010). Despite the attachment of the disc versions to the major monolingual English collocation dictionaries nowadays, research concerning learners’ usages of CD-ROMs is still extremely rare (Komuro, Shitara-Matsuo, Ishii, Uchida, Kawamura, & Kanazashi, 2006). Komuro et al. (2006) argued that explicit instructions in using the electronic. 12.

(14) version are necessary, for many participants in their study found the dictionaries useful, but not user-friendly enough. As the development of technology surges, data-driven learning (DDL) has been proposed by many language researchers (Flowerdew, 1996; Johns, 1991). Because the limited space of the conventional collocation dictionaries restricted the amount of information, language researchers have constructed corpora and concordancers, which contain millions of tokens of direct language use. Concordancer, particularly, is the key tool of DDL. It is a computer retrieval program used to search for words and phrases in various corpora. By displaying abundant examples of a key word or phrase, concordancers enable users to examine usable concentrated data from a large amount of authentic language (Chen, 2011). Studies have also reported that language learners benefitted from the use of concordancers (Chan & Liou, 2005; Chen, 2011; Cobb, 1999; Lee & Liou, 2003; Sun, 1999; Todd, 2001). The authenticity and richness of linguistic data create a learning environment for students to engage in exploratory learning and thus develop learner autonomy (Chamber, 2005). Some researchers, however, argued that concordancers were only useful in learning simple collocations but not difficult ones (Sun & Wang, 2003). According to the learners’ reflections in Yoon & Hirvela’s study (2004), concordancer learning could be time-consuming because it required laborious analysis, intensive training,. 13.

(15) and extensive practice. In addition, the functions and presentation of search output were also major problems. Owning to a large amount of information retrieved from the corpora, reading concordancing results might be a tough task for many learners (Kilgarriff, Husak, McAdam, Rundell, & Rychly, 2008). Hence, further investigations on learners’ usage of collocation concordancers are needed to help learners make the best use of the electronic reference tools. In Taiwan, much research has been conducted on investigating students’ collocation acquisition and the application of collocation retrieval tools in collocation learning. Liu (1999) and Chen (2002) examined the collocation errors in Taiwanese EFL learners’ writings. They found that V-N collocation was the most difficult for learners, and that the V-N miscollocations also appeared most frequently. Tien’s study (2010) also supported such results. According to his findings, many Taiwanese EFL learners lack sufficient collocation knowledge, which leads to the problematic collocation production. V-N collocations were also found especially difficult for learners. Many learners reported that they paid little attention to collocations when moving to higher level of education. Wang (2006) analyzed the V-N collocations contained in the most commonly used senior high English textbooks. She concluded that collocations received little attention in the textbooks, and thus resulted in students’ inadequate collocation awareness. It is suggested that English teachers provide. 14.

(16) explicit collocation instructions and that textbook publishers highlight more collocations in the readings to improve learners’ collocation acquisition. The application of corpora and concordancers has been promoted in much research on collocation learning. Liu (2008) examined the effect of college students’ collocation learning by applying online corpora to V-N collocation instruction. The findings showed that concordance-based collocation instruction was indeed much more beneficial than the traditional deductive lecturing. Chao (2009) also incorporated a concordancer, IWiLL, into junior high students’ collocation learning. Most of the students made significant improvement in productive collocation, and they all held positive attitude toward concordance learning. Similar results can also be found in Chen’s study (2010). Chen investigated how the consultation of an online concordancer, TotalRecall, improved senior high school students’ collocation knowledge. Students’ improvement was significant, and most of them enjoyed the benefits of the concordancer. While most students appreciated the effect of concordance learning, a few studies (Chang, 2008; Chen, 2010; Lai, 2010) have reported that some students thought it was time-consuming to find the most appropriate collocates through the extensive list of collocates generated by the corpora. Others appeared rather reserved as they struggle to concentrate on the online courses. Still others found themselves. 15.

(17) overwhelmed by the large amount of data and suffered from the tedious work. Based on these reflections, the effectiveness of applying the corpora and retrieval tools in collocation learning remains much to be investigated. Purpose of the Study In the previous studies, collocation acquisition as well as the application of collocation retrieval tools has been investigated. However, little research could be found on students’ collocation learning effects of using electronic dictionaries, especially electronic collocation dictionaries. The pedagogical utility of retrieval tools also left much to be explored. Furthermore, the comparison between the effects of applying different referencing tools, such as general-purpose dictionary, collocation dictionaries, and the collocation retrieval tools, on learners’ collocation learning remained much to be explored. Thus, the present study aims to investigate how Taiwanese high school students’ V-N collocation learning improves through the use of an electronic collocation dictionary—The Oxford Collocations Dictionary for Students of English (OCDE), an online dictionary for general purpose—Longman Dictionary of Contemporary English (LDCE), and an online collocation retrieval tool—Webcollocate (WEBC). In this study, three groups of students were asked to do a collocation translation test with the assistance of the OCDE, the LDCE, and the collocation retrieval tool. 16.

(18) WEBC respectively. By comparing the test results of the test, the researcher analyzed the effectiveness and user-friendliness of the three electronic reference tools. It was expected that the findings of the study not only provide pedagogical implications for English learners as well as teachers, but also offer some suggestions for future development of electronic collocation dictionaries and online collocation retrieval tools in language learning. The research questions of the present study are addressed as follows: 1.. Which group performs better in the collocation tests? The group with online general-purpose dictionary, the group with disc-version of collocation dictionary, or the one with the online collocation retrieval tool?. 2.. What are the students’ perceptions of using the electronic dictionaries and the online collocation retrieval tool? What are the strengths and the weaknesses of these referencing tools?. Significance of the Study By conducting the present study, the researcher intended to make contributions in the following perspectives. First, it is hoped that students of English could learn how these electronic referencing tools assist them to produce correct collocations accurately and efficiently. Second, the results of the present study enabled language researchers and English teachers to better understand the effectiveness of these referencing tools, helping students reduce the rate of producing miscollocations and 17.

(19) achieve native-like linguistic competence. Last but not least, the strengths and weaknesses of the referencing tools reported by the students could be useful for collocation retrieval tool developers and dictionary compilers to make the interfaces of the contents more effective and user-friendly. Organization of the Study This study consists of five chapters. Chapter one is an introduction of the background, rationale, and focus of the study. The significance of conducting this research is also included. Chapter two includes the definitions of collocation, the importance of collocation learning, learners’ difficulties in collocations, verb-noun miscollocations, and the application of referencing tools in collocation learning. In Chapter three, the overall design of the study is given. The participants, instruments, experimental procedure, and methods of data analysis are described. Chapter four presents the results of the current study, including the students’ performances of the collocation tests and their comments on the collocation searching. The findings are further discussed in Chapter five. Possible reasons for the performance differences and the students’ perceptions were analyzed. In Chapter six, the conclusions of the research questions are made, followed by the pedagogical implications, limitations of the present study, and directions for future research.. 18.

(20) CHAPTER TWO LITERATURE REVIEW This chapter includes three sections. The first section presents the definitions of collocation given by different scholars. The second section discusses the importance of collocation learning and learners’ difficulties in collocation learning, especially difficulties in learning verb-noun collocations. The third section reviews studies investigating the integration of collocation learning and Computer Assisted Language Learning (CALL), particularly the implication of electronic/online dictionary and concordancers in collocation learning. The Definitions of Collocation The definition of collocation can be discussed from two different aspects—the frequency-based and the phraseological-based. In the frequency-based aspect, collocations are viewed as units consisting of co-occurring words within a certain distance of each other, and a distinction is often made between frequently and infrequently co-occurring words. Pioneering works within this aspect were conducted by Firth (1952, 1956, 1957), Halliday(1961, 1966), and Sinclair (1966, 1987, 1991, Sinclair et al., 1970). Firth (1952, 1956, 1957) viewed collocations as sequences of co-occurring words, where the length of sequence varied greatly from two words up to fifteen. He stated that one should judge a word by the company it keeps because part of the meaning of 19.

(21) a word could be established by collocation. In other words, the meaning by collocation can attribute to a contextual approach, instead of conceptual one, to word meaning. As to Halliday’s studies, collocation is considered a syntagmatic association of lexical items that could be quantified textually in terms of their probability of occurrence at a certain distance from one another. The researcher further introduced the terms “node,” “collocate,” and “span” to respectively refer to the item under investigation (node), the co-occurring item (collocate), and the specified environment in which the node and the collocate may co-occur (span). These terms have established a prominent research foundation for subsequent collocation studies. Expanding Halliday’s concepts of collocation, Sinclair defined collocations as “the occurrence of two or more words within a short space of each other in a text.”(1991:170) The probability of co-occurrence within a certain distance could be calculated as a span of four locations to the left and to the right of the node respectively, and one collocation might be retrieved from a corpus based frequency measure. Compared to the frequency-based approach, the phraseological approach was more interested in word combinations, the degrees of fixedness, and the commutability of the word elements in these combinations. Investigations within the. 20.

(22) phraseological tradition can be found in the studies done by Cowie (1981, 1988, 1991, 1994, 1998), Benson (1989, Benson et al., 1986, 1997) and Hausmann(1989). According to Cowie, collocation is an association of two or more lexemes occurring in a specific range of grammatical constructions. The researcher further categorized collocations into “free combination,” “restricted collocation,” “figurative idioms,” and “pure idioms.” A specific feature of collocation was found that while collocations are lexically variable in most cases, they are also characterized by arbitrary limitations of choice to some extent. Benson et al. further developed a different typology between grammatical and lexical collocations. In a grammatical collocation, a dominant word (noun, adjective or verb) was combined with a preposition or grammatical structure. On the other hand, lexical collocation represented word combinations consisting of nouns, adjectives, verbs, and adverbs only, without any function words. Furthermore, since phraseological approach required words in each collocation to. be. syntactically. related,. Hausmann. classified. collocations. into. six. types—noun+noun, noun+verb, verb+adverb, verb+noun, adjective+noun, and adverb+adjective. The current study will also adopt this classification because the aim of the current study is to investigate how high school students acquire verb-noun collocations with facilitators.. 21.

(23) Collocations in Language Learning As more and more EFL/ESL research and studies have been presented, collocation learning is found to be an important part of language acquisition. In this subsection, three aspects of collocation in language learning were discussed as follows: the importance of collocation learning, learners’ difficulties in collocation learning, and verb-noun miscollocations. The importance of collocation learning. As collocation have increasingly drawn ESL and EFL researchers’ attention, its importance in language learning and its pedagogical implications have been underscored. McCarthy (1995) pointed out that the role of collocations is fundamental in vocabulary learning. Learning a word with its collocations can not only reduce learners’ burden of memorization but also expand learners’ vocabulary bank (Nattinger & DeCarrico, 1992). Collocations help learners comprehend better and faster because learners can understand the meaning of text as a whole without attending to the meaning word by word (Hunston & Francis, 2000). In addition, collocations are essential for learners to enhance their fluency through reducing the processing effort in communication (Aitchison, 1987; Pawley & Syder, 2000; Partingson, 1996). Learners can also reach higher level of language proficiency if. 22.

(24) properly trained to notice words with appropriate collocations (Lewis, 2000). Nesselhauf (2003, p.223) also stated, “collocations are of particular importance for learners striving for a high degree of competence in the second language but they are also of importance for learners with less ambitious aspirations, as they not only enhance accuracy but also fluency.” The results of following empirical studies have stated how crucial collocation knowledge is in language acquisition. Zhang (1993) gave a fill-in-the-blank collocation test and a pen-and-paper writing test to 60 college freshmen, 30 native and 30 non-native speakers of English, respectively. In the study, native speakers significantly outperformed their non-native counterparts in both tests. Zhang concluded that collocational knowledge was a source of fluency in written communication among college freshmen, and that quality of collocations in terms of variety and accuracy was indicative of the quality of college freshmen writing. To examine the relationship between the knowledge of English lexical collocations and the general English proficiency of EFL students, Al-Zahrani (1998) conducted three tests—a collocation test consisting of 50 verb-noun lexical collocations, a writing test, and an institutional version of paper and pencil TOEFL test—to 81 Saudi EFL college students. The results showed that students’ knowledge of lexical collocation strongly correlated with their TOEFL scores. In addition, the. 23.

(25) writing test was a better predictor of the students’ knowledge of lexical collocations than the TOEFL score. Likewise, Hsu (2007) found a significant positive correlation between the holistic score given by Criterion 7.1—a web-based application evaluating students’ writing skills with diagnostic scoring reports—and the frequency and variety of lexical collocations in the essays written by 62 Taiwanese college students. In the research, Hill (2000) proposed nine reasons why collocations are important in terms of the lexical nature of a language, and Nation (2001) summarized Hill’s nine reasons as follows: Language knowledge is collocation knowledge; collocation knowledge is important for developing both fluency and accuracy; knowing a word involves knowing its set of its collocates (chap. 9). Learners’ difficulties in collocation learning. Recognized as a crucial element for students to achieve language proficiency, collocation has caused much trouble to learners because it is not rule-governed and can vary from language to language. The difficulty of collocation learning, long been underestimated, actually prohibited learners from reaching higher proficiency (Granger, 1998; Howarth, 1996; Shei & Pain, 2000). Throughout the vocabulary learning, learners might know many individual words, but it was the collocate of certain words, rather than their dictionary definition, that. 24.

(26) revealed their different shades of meaning. Learners with inadequate collocation competence could still produce grammatically acceptable language, but they did not achieve native-like production, thus making comprehension more difficult for their hearers (Conzett, 2000). Research showed collocation errors constituted a high percentage of errors committed by L2 learners, even for advanced ESL student. (Marton, 1977; Arabski, 1979; Bahns & Eldaw, 1993). Granger (1998) indicated that though learners did use collocations, they underused native-like expressions and tended to use atypical word combinations instead. Wray (2000) also suggested that even the most proficient non-natives might be unable to avoid making grammatical but non-idiomatic sequences. In Skehan’s (1998) and Foster’s (2001) research, non-native speakers, compared to native speakers, who used more conventional phraseology to express meanings, tended to produce language based on grammatical rules. The results reflected the fact that learners often acquired words individually which caused them to combine words that did not normally go together. As a result, these unexpected combinations could make learners less effective in communication and hinder their ultimate goal of native-like proficiency. The difficulties of collocation learning could be found from the following empirical research.. 25.

(27) Farghal and Obiedat (1995) selected 22 common collocations related to specific topics to examine Jordanian ESL students’ collocation knowledge by employing an English fill-in-the-blank test and a translation task. The researchers found that the participants did well in tasks involving collocations with equivalents in Arabic but had problems when they faced English collocations that did not exist in their native language. In other words, the incongruence of the structure of collocations between their native language and the target language resulted in a great difficulty to learners. Based on the students’ performance of the study, the researchers summarized four principles used by the participants when they did not know certain collocations: synonymy, avoidance, L1 transfer, and paraphrasing. Gitsaki (1996) applied several writing tasks to measure Greek EFL learners’ collocation knowledge. The researchers noted that learners’ collocation knowledge did not increase alongside their grammatical competence. Learners mentioned that compared to simple grammatical collocations, lexical collocations are much harder for them to acquire. Gitsaki concluded the study with four factors influencing collocation acquisition: degree of L1-L2 differences, complexity of the collocations, frequency in the input, and the order of collocation parts. Siyanova and Schmitt (2008), in a series of studies on Russian ESL learners’ L2 production and processing of adjective-noun collocations, discovered that around 45%. 26.

(28) of learner collocations were appropriate collocations. However, the high percentage of appropriate collocations did not mean that non-native speakers necessarily developed fully native-like knowledge of collocation. Instead, they demonstrated poorer intuition than native speakers regarding the frequency of collocations, and they were also slower than native counterparts in processing collocations. Durrant and Schmitt (2009) proposed three major findings in their study focusing on the extent to which non-native writers made use of collocations in comparison to native speakers’ norms. The first finding was that native writers used more low-frequency combinations than non-natives. Secondly, non-native writers made as much use of collocations as natives do. However, non-native writers tended to repeat certain favored collocations, and overused those collocations in comparison to native norms. The third finding was that non-native writers significantly underused collocations when the collocation types between non-natives and natives were compared. Since learners showed a high preference for frequent collocations, it was suggested that non-native writers might work in a bottom-up direction from words to phrases. The results also implied that language learners did acquire high frequency phraseology of their L2 effectively; however, strongly associated but less common collocations might take longer to acquire. The researchers thus concluded that non-natives’ phraseological production. 27.

(29) differed from that of natives not because learners avoided formulaic language but because they overused high-frequency collocations, and underused low-frequency but strongly related collocations. The absence of those low-frequency but strongly associated pairs might create the feeling that non-natives’ production lacked idiomaticity. As a result, if language teachers intended to help learners develop an authentic native-like phraseology, they could draw their students’ attention to those low-frequency but strongly related collocations. Viewing such learning difficulties in collocation, Wray (2002) provided a possible explanation—L2 learners relied heavily on creativity and thus made “overliberal assumptions about the collocational equivalence of semantically similar items.” In other words, learners tended to assume that two synonymous words were equally appropriate in a collocation. Ying and O’Neill (2009) summarized three main reasons why many EFL learners performed rather poorly in producing proper collocations: First, due to a lack of collocations necessary to express precisely, learners tend to produce longer phrases and utterances. Second, learners’ production of English may sounds odd and foreign because of their wrong assumption that English word combination is the same as that of its translation equivalent in their mother tongues. Third, learners may overuse of a few general items and results in an oversimplified style.. 28.

(30) Marton (1977) and Biskup (1992), on the other hand, indicated that the reason for difficulties in EFL learners’ collocation use might be that collocations generally did not result in comprehension problems, and thus were largely neglected in the process of foreign language teaching. Since collocation learning had been proved to be a cumulative process that involved a great deal of time and effort more than rote memorization, students with limited study time would not learn appropriate collocations unless they were deliberately selected and incorporated into language materials (Swan, 1996). To improve learners’ collocation learning, providing rich exposure has been suggested. Willis (2003) found that many learners were not consciously aware of collocation or of the importance of fixed phrases. Researchers thus noted that noticing language chunks should be emphasized because it was a crucial initial process in dealing with comprehended input. Durrant and Schmitt (2010) also revealed that any shortfall in non-natives’ knowledge of collocational associations between words was because of inadequate input, rather than a non-native-like approach to learning. Hun and Beglar (2005) observed that rich exposure of collocations could be enhanced through conscious attention to, and recycling of, frequent collocations and the collocates of words that learners already knew well. Similarly, Coxhead (2008) found that by reading extensively, students would become more sensitive to noticing and. 29.

(31) retrieving multiword combinations. However, Ellis and Larsen-Freeman (2006) concluded in their study that knowledge of collocation was not entirely determined by input frequencies. Likewise, Nesselhauf (2005) also contended that increased exposure to the target language did not seem to increase collocational use by learners. It is suggested that providing direct feedbacks, giving learners a high degree of autonomy, increasing learners’ motivation, and reducing possible fears of making mistakes could assist learners’ collocation acquisition more. Besides rich exposure, awareness raising is also proposed as a possible solution for. collocation. difficulty.. Yang. and. Hendricks. (2004). explored. how. collocation-awareness-raising (CAR) tasks helped postgraduate students improve their collocation use in drafting and redrafting essays. The students reported that this approach helped them become more aware of the collocations in their readings and develop their use of collocation in other language tasks. Instructions on collocations were recommended as well. In Hsu’s study (2002), the researcher investigated whether teaching of lexical collocations in a business English workshop would enhance Taiwanese college EFL learners’ development of language proficiency and collocational proficiency. In the classroom, the researcher highlighted the target collocations and made students aware of them. The results. 30.

(32) demonstrated that the direct emphasis on lexical collocations indeed assisted students to acquire more collocations both in the written and spoken discourses. In another Hsu’s research (2010), he studied the effects of collocation instruction on the reading comprehension and vocabulary learning of Taiwanese college English majors at three academic levels. The experimental group received collocation instruction while one of the two control groups received only single item vocabulary instruction and the other received no instruction at all. The results indicated that direct lexical instruction improves the learners’ collocation learning. Liu (2008) also conducted a study to examine the effects of verb-noun collocation instruction by using online corpora. Participants were sophomores of non-English majors in China, and they were divided into three groups. The first group took courses with the aid of on-line concordancer to summarize the frequently occurred verbs that collocated with the key nouns. The second group received collocation instruction with paper worksheets prepared by the teacher. Learners in the third group were given traditional deductive teaching. The results of the study revealed that students in the first group learned much more than the other two groups. Thus the researcher concluded that concordance-based collocation instruction is much more beneficial than the traditional deductive lecturing, and that the use of concordance-based worksheet could be applied in collocation instruction.. 31.

(33) Based on the previous research, the factors influencing collocation acquisition might be—the incongruent structure of collocations between L1 and L2, the complexity of collocations, the frequency in the input, and the order of collocation parts (Farghal & Obiedat, 1995; Gitsaki, 1996). To help learners overcome the challenge of collocation learning, researchers have proposed many approaches, including awareness-raising, rich exposure, and direct instructions on collocation. Providing direct collocation instructions, particularly, has been further investigated. Verb-noun miscollocations. According to Howarth(1996), Over 5000 verb-noun (V-N) combinations in a written 240000 word corpus, over a third of the combinations were found to be collocations. And among all types of collocations, V-N collocations are not only the most frequent but also the most difficult for learners (Lombard, 1997; Biskup, 1992; Howarth, 1996). Moreover, V-N collocations are particularly important because “they tend to form the communicative core of utterances where the most important information is placed.”(Altenberg, 1993:227). Bahns and Eldaw (1993) examined German advanced EFL learners productive knowledge of 15 V+N collocations through a written translation task and a fill-in-the-blank test. The researchers discovered that although verbal collocates made up around 23 % of all lexical words, those collocates accounted for more than 46 % of. 32.

(34) all errors. Thus, the researchers contended that collocation competence did not develop in parallel with general vocabulary knowledge. The deficiency in V-N collocation might be because the verb collocates are more restricted and difficult to be paraphrased. In addition, since collocations have been largely neglected in EFL instructions, learners may be unaware of collocations as a potential problem. Nesselhauf (2003) analyzed the use of verb-noun collocations by advanced German speaking learners of English in free written production in order to investigate the influence of the degree of collocation restriction and of the learners’ L1 on the production of collocation. The findings revealed that the verb collocate in a collocation has am restricted sense, and such restriction makes its correct use more difficult as learners cannot fully distinguish subtle differences between the verb candidates. Moreover, the researcher observed that learners’ L1 has a greater influence on V-N collocations learning of L2. If learners cannot find congruent expressions of word combinations in their L1 and L2, those word combinations may be far more difficult for them to acquire. Nesselhauf, in 2005, further investigated the use of collocations involving verbs in a learner corpus of German students. Among over 2000 verb-noun collocations in the corpus, nearly a quarter, 500, of V-N collocations were found to be wrong or questionable. Factors affecting the acquisition of V-N collocation included non-congruence of the structures between students’ L1. 33.

(35) and L2, restriction of collocations, and exposure to the target language. In addition, the results also showed that learners tended to use free combinations where collocations could be used. Wu (1996) investigated Taiwanese college students’ lexical collocation competence by using a fill-in-the-blank test. The test contained 25 items, and each item consisted of a V-N collocation or an N-V collocation. Participants were divided into two groups (non-English majors and English majors) and were asked to complete the test in class without any help and then at home with the help of dictionaries. The results revealed that both groups were not able to generate proper collocations competently. Wu argued that, to improve learners’ collocation competence, collocation instruction should be given in the early stage of language learning. The results also showed that intermediate students are capable of learning V-N collocation structures. In addition, L1-L2 contrastive analysis can be a collocation teaching aid because some English collocations without direct Chinese translation equivalents tend to be unpredictable and thus more difficult for learners. As the findings showed that lexical collocations can seldom be acquired inductively, collocation dictionaries are also suggested by Wu to assist non-native learners. Liu (1999) examined the collocation errors in 127 Taiwanese college students’ essays, and found that V-N collocation errors appeared most frequently. Those V-N. 34.

(36) miscollocations mainly resulted from three factors: the false concepts of de-lexicalized verbs, which contain little content meaning and can be combined with other words with little restriction (e.g. *make the cooking instead of do the cooking), the misuse of synonym (e.g. *carry on an experiment instead of carry out an experiment), and negative transfer (e.g. *eat medicine instead of take medicine). Among the three causes, negative transfer was found to be the major source of miscollocations. The researcher suggested that teachers provide collocation instructions and should focus on collocations without direct translation equivalents, synonyms involving different collocation restrictions, and de-lexicalized verbs. In Liu’s another study (2002) on miscollocations in students’ writings, 87% of the lexical collocation errors were V-N miscollocations and 93% of them were due to the misuse of verb collocates. Liu further examined the lexical semantic relations between the improper collocates and the right ones. 32% of miscollocations resulted from inadequate knowledge of synonyms (e.g. *carry out my goal instead of achieve my goal), 15% were because of troponym relations (e.g. *break the foundation instead of damage the foundation), and 12% were due to hypernym relations (e.g. *create songs instead of compose songs). While 59% of miscollocations were caused by lexical semantic relation, 39 % of misused verbs were found to be attributed to learners’ L1 negative transfer. Liu thus concluded that learners’ V-N miscollocations. 35.

(37) might not be arbitrary but presented with certain patterns. Lin. (2010). investigated. Taiwanese. and. Chinese. learners’ verb-noun. miscollocations in detail by the method of semi-automatic data extracted from Taiwanese and Chinese’ English corpora. Both data were mostly compositions and examination essays by English and non-English major college students and senior secondary school students in Taiwan and China. Topics of the compositions and essays were various. As for the reference corpus, a large-scale English native speaker corpus—British National Corpus (BNC) was adopted. In the results of Lin’s research (2010), 210 Taiwanese and 268 Chinese learners’ common verb-noun miscollocations were found, and these identified miscollocations occurred at least three times in the corpora. Of all types of errors, the one occurring most frequently was wrong choice of verb. It was found that Taiwanese and Chinese students shared about ten percent of completely identical verb-noun miscollocations. It was also found that both Taiwanese and Chinese students had a tendency to misuse the following 14 collocates: know, learn, catch, make, get, use, develop, hold, search, cause, take, transfer, keep, and bring. However, the most commonly misused collocates of these two groups of learners were different. The Taiwanese students were found to misuse the following 10 collocates most frequently: watch, receive, memorize, improve, read, lose, write, accept, conduct and offer; whereas Chinese. 36.

(38) students were found to misuse the following 19 collocates most frequently: realize, meet, achieve, beat, win, own, draggle, pay, follow, notice, judge, put, cost, need, advocate, exchange, overcome, give, and control. The analysis of the misused collocates also showed that wrong choice of verb and L1 transfer were the most frequent type and cause respectively for both sample populations. Besides Liu’s and Lin’s findings, Chang, Chang, Chen, and Liou (2008) also observed that many Taiwanese EFL learners misused verb collocates because the misused ones often shared the same translation with the suggested correction in their first language (e.g. the suggest verb ‘increase’ and the verb ‘add’ in a learner’s sentence ‘add more knowledge by using computers’). The researchers further indicated two concepts related to L1 interference: split category and direct translation. Proposed by Stockwell, Bowen, and Martin (1965), the notion of split category described a situation when two words in one language were covered by only one term in another language. For instance, due to the shared translation between ‘create’ and ‘compose’ in Chinese, Taiwanese learners who lacked full understanding of collocation restrictions might resort to their L1 translation and therefore produced an inappropriate V-N collocation “create a song,” instead of “compose a song.” On the other hand, direct translation meant a condition where learners misused. 37.

(39) collocations because they think in Chinese first and translate their ideas into English directly. The production of ‘write homework’ instead of ‘do homework’ might be an example of direct translation. In Chinese the verb collocate xie3 (‘write’) co-occurred with gong1 ke4 (‘homework’); nevertheless, the English equivalent of xie should be ‘do’ rather than ‘write.’ This might be a reason why Taiwanese learners often trapped in Chinese translations and produced V-N miscollocations. To sum up the above studies, V-N collocations, compared to other types of collocation, seem much more difficult for EFL learners to acquire. There are mainly three reasons for such learning difficulty. First, many learners lack knowledge of collocation restrictions in semantic relations. Second, L1 interference often results in negative transfers. Third, misuse of de-lexicalized verbs can easily lead to V-N collocation errors. Collocation Learning and Computer Assisted Language Learning Computer Assisted Language Learning (CALL) has a number of inherent advantages in language learning, such as providing direct feedback, giving learners a high degree of autonomy, increasing motivation, and reducing possible fears (Nesselhauf & Tschichold, 2002). Thus, in this subsection, two aspects were discussed subsequently:. electronic. dictionaries. in. concordancers in collocation learning.. 38. collocation. learning,. corpora. and.

(40) Electronic dictionaries in collocation learning. Over the last two decades, electronic dictionaries have been referred to as “a valuable tool in the quest for knowledge” (James-Catalano, 1996). Research showed that the tool was indeed valuable in many aspects. It made the process of dictionary consultation less time-consuming and encouraged exploratory browsing, which resulted in a large number of words looked up (Aust, Kelly, and Roby, 1993; Guilot & Kenning, 1994; Roby, 1991). Besides, electronic dictionaries were appreciated by users because of the speed and ease of consultation as well as the quality of the information supplied (Laufer & Hill, 2000; Nesi, 2006; Taylor & Chan, 1994; Tono, 2000). Kilgarrif et al. (2008) noted that dictionary users appreciated example sentences. If a dictionary entry included an example which was a good match for the context where the user had encountered a word, or for the context where they wanted to use it, then dictionary users could generally get what they wanted quickly and straightforwardly. But in paper dictionaries, the opportunities were limited since examples took up space, and space was limited. Therefore, electronic dictionary have become increasingly popular with language learners. However, it was still not certain whether the fast search and a large number of dictionary consultations stimulated by the tool had long-lasting advantages (Nesi, 2000; Tono, 2000). Some researchers held positive attitudes towards the long-term. 39.

(41) benefits of electronic dictionaries. Komuro et al. (2006) and Roby (1991) believed that looking up more words might be conductive to vocabulary acquisition. Guillot and Kenning (1994) observed that the lateral browsing encouraged by the electronic dictionary in their study “enabled students to leave no stone unturned.” While much research findings supported the advantages of electronic dictionaries, the benefits of speed alone were still open to question. Sharpe (1995) pointed out that the short time needed to retrieve information might not enhance the retention of the information for language learning. Nesi (2000) claimed that the most easily found information might require least thought, and would be forgotten the soonest. Laufer and Hill (2000) also found that the number of times the word was looked up during a learning session had no relation to its retention. Considering the risk of shallow processing of the information supplied by electronic dictionaries, many researchers called for monitoring long-term effects of electronic dictionary consultation on language learning (Laufer & Hill, 2000; Nesi, 1999, 2000). Many investigations had been done on the development of electronic collocation dictionaries and students’ learning effects of electronic collocation dictionaries. Han (1999) mentioned in his study that many Korean engineering researchers as well students in IT (Information Technology) sector usually relied on various computer programs as their writing aids. However, finding appropriate collocations could be the. 40.

(42) most time-consuming and difficult during the writing process because of inadequate knowledge of collocations. Han argued that this problem might not be solved by conventional dictionaries which only provided limited collocation examples and definitions. Thus, an electronic English collocation dictionary specialized in IT sector was designed, contributing to the efficient retrieval of necessary collocation information of ITU (International Telecommunication Union) terms from large-scale language database, BNC, and texts of ITU reports. In this way, not only Engineering majors but also EFL learners could benefit from the aid of the electronic collocation dictionary. Dziemianko (2010) compared the usefulness of a monolingual English learners’ dictionary in electronic and paper forms in receptive and productive tasks and assessed the role of dictionary form in the retention of meaning and collocations. She compared the two versions of COBUILD 6, which offered the same information in the entries of target items. The results revealed that the COBUILD online was more useful in both tasks. The unexpected retention test also proved the online version to be a better learning tool as well. The findings in the study could support Laufer and Hill’s (2000) claim that what matters to word retention was greater attention during the consultation process rather than the effort of time put into dictionary search. Instead, it appeared that the visual impact created by the electronic dictionary and the. 41.

(43) prominent position of a key word on the computer screen could attract more attention than a printed page. It was possible that the form of presentation on the computer screen was more captivating and less distracting than the view of headwords on a page in a paper dictionary. These might be the reasons why the electronic version online was regarded as superior to the paper version in the process of remembering the meaning of new words and learning collocations. The findings were informative to English learners and teachers, who still quite often relied on dictionaries available in bookstores, which usually came in paper form rather than on electronic version at the cutting edge of computer technology. Corpora and concordancers in collocation learning. The use of authentic linguistic examples was suggested to help language learners better than that of invented or artificial examples (Johns, 1994). However, in the past, especially in EFL contexts, where most language teachers were non-native speakers, students used to be presented with unauthentic examples made up by teachers themselves. These examples were not only poor in authenticity but also limited in the numbers. In the end, students often had difficulty exploring the underlying patterns or meanings from the poor examples. To solve such learning problem, researchers suggested that one effective way to learn languages was through observing vast amounts of recurring patterns of concrete examples in texts (Hill, 2000).. 42.

(44) Now with the development of computer technology, electronic stored corpora and concordancers had been developed and used widely (Houston & Francis, 1998). It was argued that concordancers were superior to traditional grammar books, dictionaries and coursebooks, because they allowed easy access to huge amounts of real language use, fostered the learners’ analysis capacities, promoted their explicit knowledge of the L2, facilitated critical language awareness, and supported the development of learner autonomy (Gabel, 2001). A concordancer was a sophisticated computer retrieval program with a large amount of information in the form of a computer language corpus accessible to encourage data-based inductive learning. Such a program could display many examples of a key word or phrase and thus enabled learners to examine useable concentrated data from a great deal of authentic language (Chan, 2005; Lewis, 2000). When language was presented in an authentic context, learners were able to examine a key word in the context of a string of sentences which could exemplify the use of that particular word (Kaur & Hegelheimer, 2005). Several studies reported the use of concordancing for the ESL and EFL acquisition, such as collocation learning (Kita & Ogata, 1997), lexical acquisition (Aston, 1998; Cobb, 1999, Murphy, 1996; Thuratun, 1996), grammar (Tribble, 1990), writing (Mills, 1994; Todd, 2001; Weber, 2001), and syllabus design and evaluation. 43.

(45) (Ma, 1993). Research revealed that computerized concordances benefited non-native speakers in defining words and assisted them in the transfer of word knowledge to novel texts (Cobb, 1999). It was observed that the concordancing approach was more effective compared to conventional methods of teaching vocabulary to second language learners (Gan, Low, & Yaakub, 1996). With a corpus-based concordancer, second language learners had an easy access to various domains of writing and media. They could search for language features and patterns commonly occurring in everyday speech or written discourse in real world situations, rather than “myths and distortions that are too easily perpetuated from one generation to another of dictionaries, grammar and coursebooks” (John, 1994). Gabel (2001) also contended that the conscious and systematic exploration of the target language would support and supplement acquisition processes. He further indicated the concept: language awareness or conscious raising. By raising learners’ awareness of L2, their ability to produce functionally appropriate and formally correct texts could be improved. For example, when ESL learners wanted to describe someone as very beautiful, in a concordancer, they could find several alternatives, such as really, truly, stunningly and incredibly, which sprang quickly to the mind of a native speaker and were usually preferred. With rich, systematic and open-ended supply of data, concordancing was a tool. 44.

(46) that encouraged learners to explore and discovered the language patterns. Thus, language learning “becomes authentically heuristic” (Butler, 1990), inspiring learners to become more independent in selecting the most appropriate vocabulary required in a given situation. Stevens (1995) also proposed three main reasons for using concordancing in language learning: authentic, learner autonomy and data-driven learning. Furthermore, based on constructivist learning theory, Kaur and Hegelheimer (2005) suggested, “learners are more likely to transfer knowledge gained through such experience”. Some researchers believed that the finding of a solution by an individual becomes an integral part of the individual; however, knowledge transmitted by others had little connection to the individual’s personal experience (Gruender, 1996). Hence, the combination of corpora and concordancers provided a promising future in the field of language teaching and learning by allowing learners to discover patterns or meanings and adjust their misconceptions by observing extensive naturally occurring examples in real texts (Hill, 2000). Much research had shown that electronically-based corpora and concordancers could provide a more facilitative, efficient and effective way for language learners to learn collocations and for teachers to prepare teaching materials (Woolard, 2000). Some collocation-learning tools had currently been developed on the web. For example, Word Sketch Engine (Kilgarraff et al. 2004) was a concordancer that. 45.

(47) automatically summarized a word’s grammatical and collocational behaviors, while services. such. as. TANGO. (http://www.tango.me/),. WORD. NEIGHBORS. (http://wordneighbors.ust.hk/), and WORDSMITH (http://wordsmith.org/) provided a means of collocation finding and collocation correcting. These tools typically accepted only one querying word and retrieved sentences with it or words co-occurring more frequently than usual. Sun and Wang (2003) conducted a study to investigate the relative effectiveness of inductive and deductive approaches to learning collocations by using a concordancer. They assumed that English teachers in Taiwan tended to favor the deductive approach because the teachers regarded the inductive approach as more time-consuming for both students and teachers. The result showed that the subjects in the inductive group had significantly more improvement than those in the deductive group. They suggested that concordancers, with ample authentic language usuage, created effective discovery learning possibilities for language learning and teaching. Wu, Franken, and Witten (2010) contended that concordance data allowed learners to analyze collocations. But they argued that vocabulary learning in a computer environment often made use of exercises that isolated target vocabulary items and removed them from their original context, and thus paid scant attention to the need for learner to learn and manipulate the form and contexts of words. They cited Peachey’s. 46.

(48) (2005) statement that ‘concordancers are primarily linguistic research tools. Almost all have been designed with the sophisticated researcher in mind.’ They stated that, when using a concordancer, learners potentially were overwhelmed by the vast number of collocations returned when searching for common words. Thus, teachers might find it hard to identify sets of useful collocations for their students from large collections of text. The researchers of the study (Wu et al., 2010) argued that with limited study time, students would not learn appropriate collocations unless they were deliberately selected, prioritized and incorporated into language materials. Based on this conception, they designed a digital library, a system to provide a pedagogically enriched collocation-learning environment. To avoid overwhelming students, the teachers controlled the collection size by importing the right amount of material into the library. It was noted that digital libraries could provide authentic, focused on material that was carefully selected and organized, exposing learners to contemporary language usage. Preliminary experience with student users indicated that the digital library already provided a new and engaging way of enriching their knowledge of collocations. This system might be appropriate for EFL/ESL learners as well, except that learner assignments could not be included in the collection due to lack of authenticity.. 47.

(49) However, selecting the collocation patterns and organizing the reading materials would be very time-consuming and challenging for teachers. The research also revealed that the British National Corpus (BNC) and the web both had limitations, but they were complementary. The BNC provided far few examples, the number declining rapidly for longer collocations. In many cases, there were none at all—even for items that occur reasonably frequently on the web. On the other hand, the web text was often unclean, incomplete and repetitive—but the examples it provided were authentic and contemporary. Chan and Liou (2005) investigated the influence of web-based concordancing instruction on EFL students’ learning of verb-noun collocations with the help of a bilingual concordancer TotalRecall. The participants in the study were thirty two non-English major college freshman students in Taiwan. Three sets of tests with identical items were used for the pre-test, immediate post-test and delayed post-test. The results showed that students’ overall V-N collocations increased significantly after learning from the five online units. Their awareness of collocation was raised significantly in the immediately post-test and retained in the delayed post-test two months later. The gain scores comparisons revealed that the scores of the items taught via concordancing were evidently higher than those taught without it. The delayed post-test also showed similar results.. 48.