【正文】Abstract：Vocabulary is central to language and crucially important for L2 learners. The pattern of a word is all the words and structures which are regularly associated with the word and which contribute to its meaning. In the current study, our aim addresses the utilization of Corpus-based learning in the instruction of verb patterns in English verb pedagogy via a case study of take.
　　Keywords：Corpus    the instruction of verb patterns     language features
　　I. Introduction
　　Vocabulary is central to language and crucially important for L2 learners (Thomas M. Segler 2001). As evidence from error-tagged large learner corpora suggests, lexical errors are the most common among L2 learners2（Wang Guozheng 2004）. There are indeed abundant researches on the acquisition of L2 vocabulary (MearaandIngle1986, Fox1987, Olshtain1987, Laufer 1990a). However, most of these studies focused on the acquisition of individual words or the size of vocabulary in general, i.e. the breadth of vocabulary knowledge, while little has been done to investigate the development of verb pattern teaching.
　　The pattern of a word is all the words and structures which are regularly associated with the word and which contribute to its meaning (Hunston and Francis 2000). Therefore, in the current study, our aim addresses the utilization of Corpus-based learning in the instruction of verb patterns in English verb pedagogy via a case study of take. In other words, it is not our task to distinguish verb collocations in any way.
　　II. Corpus data
　　Corpus Linguistics studies language features based on large databases of authentic language samples stored on computer. Because its automated quantitative analysis provides novel and refreshing insights into real language use, the corpus-based approach has spread into many language-related research such as SLA (Second Language Acquisition) and EFL (English as Foreign Language). Considering large databases of authentic English, students can gain concrete evidence and a wider perspective.
　　In this current study, the pre-tagged corpora-British National Corpus (BNC) of 100 millions words-is used for investigation. The work here utilizes mainly its subset BNC-EE text of 31417 words. And the retrieval is processed on the Corpus analysis tools Antconc3.2.0w.
　　III. Methodology
　　There are hundreds of verbs and thousands of verb collocations in English. The verb patterns are dependent upon the first verb. There are 25 normal types of pattern in English (New Century English Teaching & Learning Dictionary 2000:3). When we learn a new verb, we need to learn the verb pattern. Unfortunately, some verbs have more than one possible pattern! Sometimes, if a verb has two patterns, each pattern can have a different meaning or use. Here, we just use take as an example to help teaching verb pattern. 
　　In order to help observe the pattern of the verb take, we first search take and construct a minifile for the following study. From corpus BNC-EE, we calculated 1084 frequencies of verb take (the lemma of take). Figure 1 is a sample randomly selected from the concordance lines of take in BNC made by Antconc3.2.0w.

















　　The second step is to mark the verb pattern. Corpus BNC-EE is pre-tagged, we can see that from Figure 1, with Basic (C5) Tagset, for example, VVB means the finite base form of lexical verbs and AVP Adverb particle (Details of the BNC Basic (C5) Tagset is in appendix 1). According to New Century English Teaching & Learning Dictionary and the New Standard English Textbook for junior students, we get 10 usual patterns for take. We make formulae as follow for marking the corpus. 
　　1. %TANP% -->Take +NN/ PN
　　2. %TANV%-->Take+ NN/ PN+AV
　　3. %TANR%-->Take +NN/ PN+PR
　　4. %TANN%-->Take +NN/ PN+NN/ PN
　　5. %TANT%-->Take +NN/ PN+TO(infinitive)
　　6. %TANJ%-->Take +NN/ PN+AJ
　　7. %TA00% -->Take+0 (0 means nothing but punctuation follows take)
　　8. %TAAV% -->Take+ AV (including adverbs and adverb phrases
　　9. %TAAJ% -->Take+ AJ
　　10. %TAPR% -->Take+ PR
　　While marking, we made several rules: (1) We only utilize the first two letter of the pre-tag signal so that we can category the pattern easily and they can suit junior students for understanding, for example, both ?<w NN1>liberty? and ?<w NN2>degrees? belong to “NN”,  ?<w AV0>else? and ?<w AVP>? belong to “AV”. (2) We use the tag of the key word in the phrase to represent the whole phrases. (3) Clauses are not in our consideration. (4) We only consider to the 4th word on the right side of take.
　　Then we analyze every concordance line in view of the verb patterns and mark it according to our standard. The following is a sample selected from the result:
















　　IV. Results and discussion
　　Now, we can use Antconc to show the patterns one by one and guide students to classify the verb patterns. We take Pattern TAAV as an example. 
　　First, we can hide the tags and let students to observe the raw authentic materials:







　　Then, we can show the concordance with tags.









　　Next, students can try to form the pattern and concordance with %TANT% can be shown out:









　　With the authentic sentences, it may be better for students to understand verb patterns and help learn verbs so that vocabulary may be a little easier for them to learn.
　　According to the result, a table can be made to help students to observe and summarize the uses of take with respect to verb Patterns by counting the occurrence of each Pattern. 







　　From the table, we can find some patterns, such as TANR and TANP, are used frequently, but some, such as TANV and TANN, are used very little, and the patterns like TAAJ and TANJ are almost not used. Teachers may pay attention to the frequent?patterns to make teaching and learning more effective.
　　Take is only one familiar verb to get students understand verb patterns and facilitate them to learn. There are only 9 patterns of take (No TAAJ pattern is found in our minifile) . But Students may have a picture of verb pattern construction. The primary purpose of this corpus-based study of verb pattern is to contribute to what Read ( 2004) describes, “fresh insights for vocabulary learning and for language teaching may yet transform our understanding of vocabulary and the way it is taught?(Dee Gardner and Mark Davies 2007)
　　Reference:
　　[1] 何安平. 语料库语言学与英语教学[M]. 北京：外语教学与研究出版社，2004.
　　[2] 汪国征.基于语料库的中国英语学习者动词句型习得研究[D]. 洛阳：解放军外国语学院，2004.
　　[3] Thomas M. Segler. Second Language Vocabulary Acquisition and Learning Strategies in ICALL Environments[D]. Computer Assisted Language Learning , 2002，4（15）：409 – 422. 
　　[4] Geoffrey Leech. The state of the art in corpus linguistics: three book-length perspectives[J]. English Language and Linguistics 8.1: 103?119.