1. <form id='UFnUBv'></form>
        <bdo id='UFnUBv'><sup id='UFnUBv'><div id='UFnUBv'><bdo id='UFnUBv'></bdo></div></sup></bdo>


            来源: www.zsalud.com 作者:lgg 发布时间:2018-07-11 论文字数:38741字
            论文编号: sb2018070921093521928 论文语言:English 论文类型:硕士毕业论文


            Chapter One Introduction
            1.1 Background of the Study
            In teaching English as a foreign language, communicative ability has been put onmuch emphasis. The English teaching in China has focused on cultivating students'speaking ability since the implementation of the new College English Curriculum Requirements (Department of Higher Education, 2007). Penny Ur intends to believe that“of all the four skills,,(listening, speaking,reading and writing), speaking seems intuitivelythe most important. Speaking is to express himself in the target language, to cope withbasic interactive skills. Improving speaking ability has positive effect on communicativecompetence.The assessment of students' language abilities is something on which teachers spend afair amount of class time in one way or another. Assessment is valuable for both theteachers and learners. The assessment tasks are nonthreatening and are developmental innature, allowing learner ample opportunities to demonstrate what they know and what theydo not know, and providing useful feedback both for learners and the teacher.Language assessment has an increasingly popular pastime. For many, traditional typesof "testing" have been viewed as somewhat unpopular areas of language teaching andlearning. Learners have sometimes viewed tests as unfair measures of their languageabilities and have feared that they would not perform well on them. Teachers havesometimes felt reluctant to construct classroom tests and have not been altogether satisfiedwith their results when they have engaged in test construction. On occasion, teachers havealso been suspicious of standardized when they have not been sure what these tests wereactually trying to measure. Fortunately, in the past few years there has been a growinginterest in improving the language assessment situation.
            1.2 Significance of the Study
            The situation of peer assessment on promoting EFL learners' communicationstrategies is still not fUlly developed especially in CALL environment. In order to makecontribution to the current speaking research,the author carried out the empirical study toinvestigate the effects of peer assessment for speaking skill on promoting EFL learners,communication strategies in pair work and to see what are students' attitude towards peerassessment. The empirical study aims at obtaining some theoretical and practical values toprovide more information and knowledge for the practice of peer assessment and for theteaching and learning of speaking in Chinese universities.Regular assessment of learning can provide learners with feedback about theirlanguage performance at various stages in the developmental process. While assessment istaking place, students are getting feedback on how they perform on those tasks. Once thetask has been scored and evaluated, students learn how well they did on what was assessed.Developing on the quality of the feedback and the attention the students give it, they maylearn something about their areas of strengths and also about the areas in which they areweak’ prompting further learning or review.
            Chapter Two Literature Review
            2.1 Computer Assisted Communicative Language Teaching Model
            According to the College English Curriculum Requirements (CECR)⑴(Departmentof Higher Education,2007),based on "computer-assisted and teaching model in class,,and“cultivate college students ’ English comprehensive ability, especially listening andspeaking skills”. Therefore, the teaching of oral English is not only a process of knowledgeof English, but also a process of transforming students' language knowledge intocommunicative competence. The definition of communicative competence was put forward by Hymes[6], As weknow, it is Chomsky who makes the distinction between the notion of competence andperformance. Chomsky defines competence as the perfect knowledge of the abstractsystem of rules by which a person is able to understand and produce any and all of thewell-formed sentences of his language[7]. For Chomsky, competence is obviously restrictedto the intrinsic knowledge of linguistic form and meaning[8]34. Hymes does not deny thedistinction between the knowledge that underlies linguistic behavior and actuallyobservable language use, but he points out that the dichotomy between competence andperformance. He claims that in Chomsky's theory, he ignores the socio-culture factors. Heviews the notion of competence is tacit knowledge and ability for use.
            2.2 Theoretical Framework of Communication Strategies
            Communication strategy is an approach in order to reach the communication goalsand make up the lacking communication on vocabulary or grammar under thecircumstance of grasping non-efficiency foreign language or second language. The termcommunication strategy is proposed by Selinker[i6] in his article. He referred to fiveaspects can form and developing of interlanguage. They are: mother-tongue transfer;teaching transfer; foreign language learning strategies; communication strategies;overgeneralizatioii of target language. In addition, he pointed out that communicationstrategies refer to the learners' efforts to express meanings through interaction within thelimited second language system, but he does not give the correctly definition of thecommunication strategies. Since then, several researchers begin to study thecommunication strategies. Tarone^^^' do the research reports aiming at communicationstrategies. Later, communication strategies have developed to the ones used by the secondlanguage learners with limited vocabulary and grammar to make their meaningsimderstood[i8]. Inl980,Canale and Swain[io]5G put forward communicative competencemodel,it played very important role for research of communication strategies. In 1983,Faerch and Kasper[i9] published Strategies in Interlanguage Communication aboutcommunication strategy. Bachman[3] put forward that “the function of strategiccompetence is to match the new information to be processed with relevant information thatis available and map this onto the maximally efficient use of existing language abilities".Therefore, as the resem*chers' cognitions develop, the definition of communication strategyvaries from researcher to researcher.
            Chapter Three Research Design..........21
            3.1 Research Questions......... 21
            3.2 Subjects.........22
            3.3 Teaching Contexts.........26
            3.4 Research Instruments.........30
            3.4.1 Questionnaires .........30
            3.4.2 Tests .........31
            3.5 Research Procedures......... 31
            3.6 Data Collection and Analysis .........32
            4.1 Analysis of the Students' Tests Scores......... 36
            4.2 Analysis of the Students' Communication Strategies Used .........40
            4.3 Analysis of the Students' Feedback......... 43
            Chapter Five Conclusions and Implications......... 51
            5.1 Major Findings......... 51
            5.2 Pedagogical Implications.........54
            5.3 Limitations.........56
            Chapter Four Results and Analysis
            4.1 Analysis of the Students' Tests Scores
            In order to measure if peer assessment in speaking activities has positive effects onpromoting EFL learners' communication strategies,thus to improve their oral Englishability,an experiment was carried out at a CALL-based English audio-video speaking class(EAVSC) by the author. The students' speaking tasks, in particular, their pair workdialogues were recorded and collected, and the research data for this study mainly focuseson students' pair work dialogues in pre- and post tests. In order to compare the oralperformance of subjects' in class one and class two, the form of the pre- and post tests wasNational College English Test Band Four (except composition and translation) (CET-Four)was randomly used to exam students' comprehensive language proficiency. The pre-testwas carried out in the third week, and the topic in the pre^test was: Some western holidayslike Christmas are getting more and more popular in China,what do you think of thisphenomenon? And the post test uses the similar topic as the pre-test in the same conditionto assess the learners in both classes. The post test was conducted in the sixteenth week,the topic for the post test was: What is your opinion about automobile industry in China?Because of the National GET Test and the relatively spoken topics,therefore, the validityand reliability needs no more test.
            The study is a descriptive one aiming to measure if peer assessment in speakingactivities has positive effects on promoting EFL learners,communication strategies,thusto improve their oral English ability, the teaching research was carried out at a CALL-based English audio-video speaking class. In spite of the positive contribution of the study,it also had some limitations because of limited time and resources. There were severalfactors that may have biased the results:The number of students involved was not big enough to analysis the statisticrequirements in detail. There were only fifty-seven students, twenty-six in class one andthirty-one in class two. It is only in a small scale. And a further study based on a largerange of students needs to be done. For instance, the students' different behavior,attitude,belief,and emotion, or students' gender, personality or whether they make own extraefforts after the class to practice expressing target language. These should be taken intoaccount in the future study.The criteria are not specified. A ranking grade such as “one,,or "five" without givingfurther information does not inform learning. Criteria that are specified and maketransparent to students do provide students with the knowledge about where they are inrelation to certain standards and the learning goals. At the same time,shared criteria areneeded when the assessment evidence collected in the teaching and learning process isused for accountability purposes^^^l It also remind the teacher, if let the students have peerassessment, the teacher should make sure the criteria is specific and every student has aclear idea on criteria.
            References (omitted)


            Chapter 1 Literature Review
            1.1 Definition and Characteristics of Spatial Metaphor
            Metaphor involves the comparison of two concepts. One of which is construed interms of the other. It is described in terms of a target domain and a source domain.According to different source domains, Lakoff and Johnson(1987) classify conceptualmetaphor into three categories: structural metaphor, ontological metaphor andorientational metaphor. Orientational metaphor or spatial metaphor is a kind ofmetaphorical concept that organizes a whole system of concepts with respect to oneanother. Since most of concepts have to do with spatial orientation: up-down, in-out,front-back, on-off, deep-shallow, central-peripheral, they are called orientationalmetaphors. Orientational metaphors give a concept a spatial orientation. It enablespeople to understand abstract concepts like time in terms of spatial terms, HAPPY isUP, for instance. There are expressions in English like “I’m feeling up”. Metaphoricalorientations are based on our physical and cultural experience. Take “I’m feeling up”as an example. Its physical basis is: erect posture goes with a positive emotional statewhile dropping posture goes with sadness or depression. Orientational metaphors varyfrom culture to culture. The future is in “front” of us in some cultures, whereas it is in“back” in others.There are many examples of daily expressions of spatial metaphor. For example,“GOOD IS UP”; “BAD IS DOWN”. So we have such expressions as “Things arelooking up” and “Things are at an all-time low”. Physical basis is that things such ashappiness, health, life that characterize what is good are all up. On the basis ofexamples discussed, the characteristics of spatial metaphor is given as follows(Lakoff& Johnson,1980:17-19):
            1.2 Previous Studies of Spatial Metaphor
            Most people take metaphor as a rhetoric device in poem writing or a matter ofextraordinary language rather than ordinary one. And what’s more, metaphor isconsidered as a matter of words rather than thought or action. However, Lakoff andJohnson(1987) figured in Metaphors We Live By that metaphor is pervasive ineveryday life, not only in thought and action. The concepts not only govern ourthought but also govern our everyday functioning. Our concepts structure what weperceive, how we get around in the world and how we relate to other people. Theyargue that our understanding conceptual system is fundamentally metaphorical innature. The way that we think and what we do every day is a matter of metaphor.Another researcher who makes equally great contribution to the study of spatialmetaphor is the representative of localism, John Robert Anderson. In his opinion, spatial expressions are more basic than other non-spatial expressions in that spatialexpressions are the structure model of other expressions. He brings grammaticalcategories like tense and aspect into his research. According to Anderson, tense is areferential category. There is distinct correspondence between spatial reference andtemporal reference. Relations between aspect and spatial categories are ratherapparent.
            Chapter 2 Theoretical Framework
            2.1 The Development of Image Schema Theory
            Ancient Greek philosophers considered schema as fixed templates. Germanphilosopher Kant pointed out that schema is a bridge between perceptions andconcepts, a means of connecting concepts and substances, the necessary procedure ofconstructing images and creating meanings. Schema is related not only withperceptions, but also with imagination. Schema is not absolutely objective orsubjective. Psychological study about schema started in the 1920s and the 1930s whilegestalt was doing the research of memory.British F. Bartlett pointed out that human memories can structure kinds ofinformation and experience, form regular schema, and store them in their memories.New experience can be understood through their being compared with schemas inmemories. And that’s exactly how new information is learned.In the 1930s and the 1940s, Swiss psychologist Piaget proposed that cognitionderived from the interaction between subject and object. Object is assimilated tosubject’s schema through self- regulation. Or subject regulates schema or creates newschema to adapt to new object.Fillmore (1977) proposed that when people understand language generally, itseems that there is a schema for lexicalization. It means that lexicalizing something ispresenting it as an established category of human thought. If a lexical item exists, itmust exist as certain part of a frame and must correspond to certain part of a schema.Modern cognitive linguists developed schema theory. Langacker (2000:93)proposed: “A schema is the commonality that emerges from distinct structureswhen one abstracts away from their points of difference by portraying them withlesser precision and specificity.”
            2.2 Key Notions of Image Schema Theory
            The main notions to analyze image include: base, profile, figure, ground, trajector,landmark, and so on. The analysis and description of image need special unit. Butdifferent scholars have different opinions on defining image unit. In this section themain notions that are used to analyze images are introduced. About a century ago, Danish psychologist Rubin introduced Figure-Ground intopsychology. Later gestalt psychologists integrated it into the framework of perceptualorganization. People do not perceive everything equally. Something is more importantand people focus on them. Other things are the background, against what people maybe interested in. That is the famous phenomenon: Figure-Ground segregation.What people are interested in would be the figure. The ground would be thebackground information. When people perceive a scene, they do not perceiveeverything equally. What people are paying attention to and looking at is the figure,and the rest is the ground. The figure is the object of attention. What’s the differencebetween figure and ground? The figure has shape or form. The ground is formless. Itseems that the shared contour belongs to the figure. The figure appears to have somequalities, for example, structure and coherence, while the ground is unstructured,shapeless and uniform. The figure seems to lie in front of the ground. The groundextends continuously behind the figure. The figure is regarded to be more prominentthan the ground. It is shown that the figure is more likely to be identified andremembered. And the figure is more likely to be associated with meanings andfeelings. The figure is usually small, movable, and of simple and contact structure.The ground is usually used as the reference point.
            Chapter 3 Spatial Meanings and Metaphorical Extensions .........22
            3.1 Spatial Meanings of “Through”....22
            3.1.1 Into-out of ..........22
            3.1.2 Within........25
            3.1.3 Throughout ........26
            3.1.4 To ......28
            3.2 Metaphorical Extensions of “Through” .........29
            Chapter 4 Spatial Meanings and Metaphorical Extensions .........41
            4.1 Spatial Meanings of “Tongguo” ....41
            4.2 Metaphorical Extensions of “Tongguo” .........47 
            Chapter 5 A Comparision of “Through” in English .........53
            5.1 Differences ...........53
            5.2 Similarities ...........56
            5.3 Causes of Differences and Similarities ...........57
            Chapter 5 A Comparision of “Through” in Englishand “Tongguo” in Chinese
            5.1 Differences
            From the analyses in the previous chapter, we find many differences both inspatial meanings and metaphorical extensions between “through” and “tongguo”.Let’s look at their differences in spatial meaning first. By contrast, “tongguo” in Chinese does not have such spatial meanings. In fact,the Chinese word often used to express the above meaning is usually “chuanguo”(穿过) when LM is either too small or too large.Another difference is that in addition to the similar meaning of “into-out of”“through” in English has three more spatial meanings: “within”, “throughout”, and“to”, while “tongguo” in Chinese has only one more spatial meaning, i.e., “across”. As is mentioned in 3.1.2, when “through” is used, the trajector moves from onepoint in the landmark to another point in the landmark. The source and the goal areboth in the landmark. The trajector does not traverse the landmark. Look at anexample in 3.1.2: She searched everywhere in the office to find her wallet that was lost. Thoughher direction is not clear, she searched in the office and did not go out of the office allthe time. But “tongguo” has no such meaning.
            Different from the traditional view which regards metaphor as a matter oflanguage, cognitive linguistics views it fundamentally as a matter of thought. It iscrucial to humans cognition since it is mostly through conceptual metaphor thathumans comprehend abstract concepts. The thesis presents a study on spatialmeanings and metaphorical extensions of the English word “through” and the Chineseword “tongguo” within the framework of cognitive linguistics. After the analyses ofspatial meanings and metaphorical extensions of “through” and “tongguo”, thefollowing findings can be drawn:
            1. There are four spatial meanings of “through” in English: into-out of (TRtraverses LM.), within(TR moves in LM along a straight line.), throughout(TR movesirregularly in LM.) and to (LM is the destination of the movement of TR.).
            2. “Through” originates as a spatial concept. It has been applied to a range ofnon-spatial situations. The spatial domain of“through”is mapped to the domains oftime(LM is the period of time or time point for the movement of TR. ), cause(LM isthe cause that makes TR change.), means(LM is the means by which TR achieves acertain goal.) , sense (LM makes barrier for TR to get sensory information) and state.The domain of state includes the domain of event(LM is the event that TR finishes.)and the domain of emotion(LM is certain negative emotion that TR overcomes.).
            3. “Tongguo” in Chinese has two spatial meanings. One is similar to that of“into-out of” of the English word “through” where TR traverses the bounded LM. Theother is similar to the English word “across” where TR moves along the surface ofLM from one end to the other.
            References (omitted)


            Chapter 1 Introduction
            1.1 Research background
            The implicit-explicit dichotomy originated from cognitive psychology. Mostrelevant researches have dealt with one of the two aspects: 1) implicit/explicit as twotypes of knowledge or memory (Broadbent,FitzGerald & Broadbent, 1986; Schacter,1990; etc.); and 2) implicit/explicit as two modes of learning processes (e.g. Berry &Broadbent, 1988). Similar distinctions have been brought into many other fields,including second language acquisition (SLA) researches. In recent decades, thedichotomy of implicit/explicit language knowledge and that of implicit/explicitprocesses of developing language ability have been two important themes in SLAresearches. Language researchers and practitioners, and sometimes even the languagelearners themselves, have the feeling that what are manifested through EFL learners,performance on language tests cannot truly reflect their language proficiency. Eventhough some learners are able to gain full scores in grammar and vocabulary tests,they would probably make mistakes in natural spontaneous language production.Now it has been widely agreed on that the ultimate goal of language acquisition isto establish a native-like implicit language system that exists outside of consciousness(VanPatten, 2007: 16),whether in terms of a first language (LI) or a second language(L2). Language users can automatically produce accurate language output andperform a large variety of language tasks with the help of implicit knowledge, withoutbeing aware of the underlying structure or the existence of certain linguistic rules.However, as language acquisition is a long-lasting process, explicit knowledge alsoaccounts for a large proportion of learners' overall language proficiency in the interimstate, especially for EFL learners who experience language development underclassroom conditions. The role of explicit knowledge or explicit learning in SLA hasbeen widely acknowledged by language researchers and practitioners (Harley, 1994;Dekeyser, 2003; Saito,2011; Fordyce, 2014; etc.).
            1.2 Research aim and research questions
            The current study aims to investigate whether EFL learners' development ofimplicit language knowledge and that of explicit language knowledge are of the samepace. Implicit and explicit language knowledge cannot be directly measured, but canbe induced by referring to learners' performance on accomplishing language tasks thatcall for them to draw on their implicit and explicit language knowledge respectively.It has been the center of debate whether one can claim that certain language taskmeasures implicit knowledge or explicit knowledge, because on the one hand,sometimes both implicit and explicit knowledge are involved when performing certainlanguage tasks,and on the other hand, the processing mechanism is so complicatedthat one cannot simply owe certain performance to the influence of implicit or explicitknowledge. However, for convenience's sake,the tags implicit knowledge andexplicit knowledge are borrowed in the current study to represent the underlyingknowledge system,and the biggest effort is made to optimize the external conditionso that the separation of implicit and explicit knowledge can be achieved to the largestextent.
            Chapter 2 Literature Review
            2.1 Fundamentals of SLA
            Three basic questions are to be answered in the field of SLA, namely, 1) Whatdoes the L2 learner know? 2) How does the learner acquire this knowledge? 3) Whyare some learners more successful than others? (Saville-Troike,2006: 24). Differentscholars have attempted to address these issues from different perspectives, like thelinguistic,cognitive, psychological,and social perspectives, while others adoptedcombined approaches, e.g. a psycholinguistic or a sociocultural approach. Therefore,SLA has become a multidimensional and interdisciplinary subject. The predominantapproaches to SLA include the purely linguistic approach, the cognitive approaches,the sociocultural approaches and the psycholinguistic approaches (VanPatten, 2007: 8).Besides these mainstream approaches, many researchers also begin to focus on somealternative approaches in recent years, like the Complexity Theory Approach,IdentityApproach,Conversation-analytic approach (Atkinson eds,2011),Dynamic SystemsTheory approach (de Bot,Lowie & Verspoor, 2007), etc.In purely linguistic approach to SLA,researchers aim to explain what constrainsthe development of learners,linguistic systems using a particular linguistic theory-Relevant theories include Contrastive Analysis, Error Analysis,Interlanguage, UG,Morpheme Order Studies and the Monitor Model (Saville-Troike, 2006; 33-52). Thepurely linguistic approach contends that what is being acquired in SLA ranges fromunderlying knowledge of abstract linguistic principles and constraints, to ability tostructure and convey information in a second language. In terms of how SLA takesplace, some purely linguistic theories emphasize on continued innate UG capacity forlanguage learning while others focus on requirements of communicative processing(Saville-Troike, 2006: 62).
            2.2 Implicit knowledge and explicit knowledge
            Since the current study aims to investigate Chinese EFL learners' development ofimplicit and explicit knowledge on some basic grammatical forms, it is first of allnecessary to look at the definitions of implicit and explicit knowledge and theirdistinction from a theoretical perspective. Neurologically speaking, implicit linguisticknowledge and explicit linguistic knowledge are believed to be stored in two separatesystems, named respectively as implicit language system and explicit language system(N. Ellis, 2008). R, Ellis (2004: 232) cited Paradis (1994),s argument that "explicitknowledge is stored diffusely over large areas of the tertiary cortex and involves thelimbic system, while implicit knowledge memory is linked to the cortical processorsthrough which it was acquired". Thus,we have sound proof that they existindependently of each other.
            Chapter 3 Methodology...........23
            3.1 Participants...........23
            3.2 Test content............24
            3.3 Test procedure and scoring methods...........26
            3.4 Data analysis methods...........27
            Chapter 4 Results and Discussions...........29
            4.1 Synchronicity of the development of implicit and explicit knowledge...........29
            4.2 Differences in low- and high-proficiency learners' development........... 43
            Chapter 5 Conclusion........... 48
            5.1 Summary of the important findings...........48
            5.2 Implications of the study...........49
            5.3 Limitation of the current study and suggestions........... 50
            Chapter 4 Results and Discussions
            4,1 Synchronicity of the development of implicit and explicit knowledge
            The central question that the current study aimed to answer was the synchronicityof the development of Chinese EFL learners' implicit and explicit knowledge on somebasic grammatical functors. To answer the question, horizontal comparisons weremade to compare the learners' performance on certain grammatical features in theessay writing task and that in the GJT. The subjects were examined one by one, andsince the test contents for the high- and the low-proficiency groups of subjects weredifferent,the results would also be shown in two different sub-parts below. If theanalyses of the two different proficiency groups yielded similar results,the conclusionwould be deemed as more convincing and generalizable.The comparisons were made both by human interpretation and with the statisticaltool. Since every learner received a global score on his or her performance on theessay writing task and the follow-up GJT and given that the two language tasks werehighly related in content, correlational analysis was carried out. However, since asingle essay was unable to reflect a learner's knowledge on all target grammaticalfeatures, human judgment was needed to determine which grammatical functors wereinvolved in one essay and to match that with performance on the follow-up GJT.
            The current study aimed to find out whether Chinese EFL learners' developmentof implicit and explicit grammatical knowledge is of the same pace and whether thereare any differences between high-proficiency and low-proficiency learners' implicitand explicit grammatical knowledge,respectively. To fulfill these two goals,anempirical research involving an essay writing task and a follow-up GJT was carriedout among two groups of subjects representing a high-proficiency group and alow-proficiency group,respectively,and their performance on the two tasks wascompared vertically and horizontally to spot the differences.Each learner's grammatical performance on the essay writing task and thefollow-up GJT was compared to see whether there were any situations where learnershappened to correct the very or similar mistakes in the GJT that they themselves madein the essays or situations where learners managed to correct most of the items under acertain grammatical category though making a few mistakes themselves under thatsame category in the essay writing task. Then the two groups of learners' overallgrammatical performance on the GJT atid that on the essay writing task was roughlycompared with the help of statistical tools to examine the knowledge gap between thetwo proficiency groups.
            References (omitted)


            Chapter One Introduction
            Language testing is still occupied an significant part of language teaching which is amajor method of assessing teaching and learning products and examining students’ languageability such as small class quiz, middle-term examination, final-term examination, NMET,CET-4&6, TOEFL, GRE, TEM4&8, IELTS, PETS and so on. The washback effects oflanguage testing to English teaching and English learning has been the central issue of someresearchers’ studies. To test the effects of language teaching and learning, the concept oflanguage testing has risen from 1940s. Since then language testing and learning areseparated into parts.A large number of linguists have concluded the phenomenon that the complex effect oflanguage testing on teachers and students as washback. There is a definition of washbackthat “the proper relationship between teaching and testing is purely that of partnership”(Hughes, 1989:1). It also refers to “the degree to which the lead-in of an use of a testinfluences language teachers and learners to do things they will not do that boost or hamperlanguage learning” (Messick, 1996:241). It has been apparent from 1990s. Although peoplehave ever involved in the relevant new field, there is few theoretical and empirical researchon it since reformed. So more researches about washback effect are needed to be in progress.
            In China, CET-4&6 have been taken the most popularity arousing more researchers tomake relevant studies presently. In order to implement the College English Syllabus, CollegeEnglish Test Band 4 (CET-4) is a national, extensive and massive standardized EFL testdesigned for non-English major undergraduates in the name of the Ministry of Education ofChina. It was firstly put into practice in June 1987 and has devoted a lot to the developmentof College English for no less than twenty years, it has ever been reformed for several timesto be gradually suitable for more participants. There is no doubt that CET not only evaluatesEnglish proficiency of participants but also provides feedback information for teachers toimprove and polish College English teaching and learning as a criterion-related andnorm-referenced test. As Han Baocheng has said, CET-4 is also a high-wager test which hasbeen the central issue of controversial discussions all the time. On the one hand it is suitablefor social development currently, gaining common recognition, producing social benefits andcontributions to the enhancement of College English teaching and learning quality. On theother hand the misunderstanding and abuse of it also leads to some negative effects includingthe so-called phenomenon “teaching to tests”, “learning to tests”, “high scores, low abilities”, “test-oriented education” and “mute English”. Therefore we feel obliged to make someresearches about CET-4 washback in order to strengthen its positive washback effects andreduce its negative washback effects.
            Chapter Two Literature Review
            2.1 Definitions of Washback
            Washback effect is the definition of the influences of language testing to the two aspectsbetween teaching and learning. As we are all familiar with, the term “backwash” is alwaysmade use of in universal education literature and “washback” in British applied linguisticsthough there are some other terms having the same meaning such as testinfluences(Bachman and Palmer, 1996 and Baker, 1991), consequential validity(Messick,1989,1992,1994,1996), measurement-driven instruction, curriculum alignment(Sheppard,1990), feedback and bogwash. The term can hardly be defined exactly in the dictionaries. Inthis paper, we will make use of the most popular and common-used term “washback”.Washback has ever been drawn attention by some researchers. The definition “the connectivebonds between testing and learning” is given by Shohamy (1996:199). The remark ofwashback is “the use of outer language tests to influence foreign language learning in theschool extent as an impetus” by Shohamy (1992:514) . Like Buck, she had ever mentionedthat “this phenomenon is the authoritative consequence of exterior testing and the stronginfluence it has on participants”. Hughes (1989) gave another definition “the influence oftesting on both the two aspects: teaching and learning, if a test is considered significant, thenall teaching and learning activities can be commanded by preparations for it . And what thetest presents and techniques the test makes use of are with the goals of the course, then thereis a trend to be destructive backwash at divergence of views”. Bachman and Palmer (1996)believe that “the influence of utilization of the test functions in terms of a micro level, that isthe individuals who are influenced by the special utilization of the test, and a macro level,that is presented in the educational or social system”. They pointed out that “processes arethe central issues of washback, particular learning and guidance and these processesinfluence individuals, educational and social systems widely”. Gates (1995) regarded it as“the impacts of testing on teaching and learning”.
            2.2 Theoretical Working Mechanisms
            In spite of the fact that there is some research about the idea of washback, but thetheoretical mechanism of washback is still the central issue. The theory is a milestone in the history of washback researches. They started from twofundamental supposals offering 15 washback hypotheses mainly from two angles: teachersand learners, teaching and learning. Besides they are all firstly to explore workingmechanism of washback. The hypotheses are presented as following: 1,3,4,9,11 are about teaching and 2,5,6,8,10,11 are relevant to learning, 12,13,14,15 areconcerned with the strong point of washback. The previously mentioned hypotheses areimportant to some practical researches about washback. They conclude that “more researchon washback is made, for the reason that there is little evidence to be made to advocate thecontroversy that tests have impacts on testing, and classroom observations must be involvedin such research.” They believe that “simple and more discussion within the range ofconsiderations, polish and investigation are needed”. Later Alderson and Hamp-Lyonsdeeply prose into 15 hypotheses “tests will have different numbers and categories ofwashback on the roles of certain teachers and learners than on the roles of other teachers andlearners” (1996: 288).
            Chapter Three Brief Introduction of CET-4 .......... 14
            3.1 Historical Review of CET-4.......... 14
            3.2 Washback Effects of Old CET-4 ............ 17
            3.3 The Analysis of New CET-4 ......... 18
            3.4 The Washback Study of New CET-4...... 21
            Chapter Four Research Methodology .......... 24
            4.1 Research Objectives.... 24
            4.2 Research Questions ..... 24
            4.3 Research Subjects ....... 24
            4.4 Research Methods ....... 24
            Chapter Five Data Analysis........ 27
            5.1 Questionnaire ..... 27
            5.2 Class Observation ....... 38
            5.3 Interviews........... 41
            Chapter Six Results and Discussions
            6.1 The Motivations to Cope With the New CET-4
            CET-4 was made reformation by The Designing Committee of National CollegeEnglish Test. According to the document of the year 2013, the goal of CET-4 is to “cultivatestudents’ strong reading ability, certain listening and translation ability, primary writing andspeaking ability, making students acquire necessary professional information with Englishtool and laying the foundation of improving their English proficiency further. Therefore thetest is mainly to test students’ competence of language use and the master of grammaticalstructure and vocabulary use.” Although students are aware of improving theircommunicative competence, they also have instrumental motivation such as finding a goodjob in the job hunting. In other words, the motivations of new CET-4 have not accomplishedas what we expect. The possible reasons may be as following:The first possible reason is from the fixed limitations of the CET-4 itself. Althoughthere are some changes of test items taken place, many students reflect that the test format isnot perfect and flexible that they can summarize some test-oriented skills in the preparationprocess. The test items distributed are still more objective and less subjective. Students putemphasis on the test items through doing the model tests or CET-4 papers in the past yearregardless of improving communicative competence and English skills. The learningmaterials are so limited that are largely connected with textbooks or some papers of the pastyear, besides they have limited time allotted for English learning, they prefer to spare moretime on their majors to prepare for the final examination in the end of the semester,sometimes they feel that the approach of CET-4 has conflicts with the final examination,their energy and time are limited and it is difficult for them to balance these examinations.
            This thesis is only a basic insight of the new CET-4 washback effects. According to theresearch results of the three instruments, the essence of the new CET-4 washback can becategorized into two sides: positive and negative. The positive effects surpass the negativeones, the positive washback effects can be listed as followed:1) The new CET-4 potentiallyforms a motivation to encourage students to study English, enlarge their volumes ofvocabulary accumulation .2) The new CET-4 largely and comprehensively reflects theirpractical English ability and English proficiency as a reliable measurement and criteria toevaluate the outputs of college English learning. 3) he new CET-4 introduces some relevantwestern culture invisibly. 4) The proportion of integrative tasks is increased, which can helpthem improve all aspects of language skills. It is the first time to make translation type as anindependent part to test students’ translation skills from sentence to passage. It is requiredthat students should not only have solid foundations of vocabulary use and grammaticalstructure, but also have abilities to divide or merge paragraphs according to context properly.It means that translation is gradually put in the important position. But there are still someproblems existed in the new CET-4 that are worthy of making further investigations andresearches in the future by more people.
            References (omitted)


            Chapter One Introduction
            1.1 Research Background
            With the development of political multi-polarization and cultural diversificationtrend of economic globalization, society puts forward higher requirements on students'English level. Students not only need to have strong reading and writing ability, but alsopossess considerable oral English and communication skills. In the face of new socialneeds, the oral English level the students have is not satisfactory at present. According tothe survey, employers are not satisfied with the decline of English comprehensive abilitycommon in oral and writing ability for college graduates in recent years. They think thatgraduates who have good oral English ability are only 5%,37% graduates are in poor orveiy poor level,someone who can qualify for participating in international meetings todiscuss is only 7% (Cai Jigang, 2012). Therefore, oral English teaching plays animportant role in English class.As the foundation stage of education, English teaching in junior school has alsocarried on the reform in our country, the importance of oral English teaching has beenconcerned widely. The new English textbooks also lay stress on the teaching of spokenEnglish. English teachers begin to pay close attention to cultivate students' oralcommunication ability to some extent. However,the progress of oral English teaching isslow,which is always unsatisfactory. The ability of students to express their thoughts andfeelings in English is limited, most of them just stay at the levels of greeting. Therefore,how to improve the oral English teaching and explore the teaching model which issuitable for middle school students to improve their oral English learning is worthexploring.
            1.2 Research Purpose and Significance
            A real problem we each English education workers must face is that speaking abilityof Chinese students is poor. The only way to solve the problem is to reform thetraditional oral English teaching mode,improve the level of oral English teaching andenhance the efficiency of college English teaching, Meanwhile, education researchersand educators is exploring and putting forward new form of oral English teachingcontinuously so as to adapt to the current English teaching. Classroom research hasalways been a focus in the study of education problems, which have a variety ofdisciplines and diverse research methods. With the advent of the era of ecology andecological trend of education development, ecology is becoming a classroom study in anew angle of view. Classroom ecology research is becoming a classroom study of a newpath as the education of ecological research of microscopic study.A key of oral English teaching reform in junior school in our country is to build anew effective teaching mode. That is using ecological research approach to analyze theteaching problems, which might be a good try. In this paper,on the basis of the scientificoutlook on development and the theory of education ecology theory as well asphilosophy and psychology, the author follows the requirements of quality education and embodies the new curriculum goal,using quantitative analysis method to explore theecological influence to the teaching of oral English class to explore an effective way ofteaching reform in spoken English.
            Chapter Two Literature Review
            2.1 Researches on Oral English Teaching
            As a form of English expression, more and more researchers pay attention to the oralEnglish teaching. There are many research achievements on oral English no matter in domestic orabroad. Last century,in the tide of global curricula reform for basic education,the oralEnglish teaching, as a part of foreign language teaching,is gaining more and moreattention.In the aspect of theory research, from the perspective of behavioral research insecond language oral fluency, Gatbonton and Segalowiz (1988:473-492) Points out thatautomatization has an important practical significance in language learning, autonomywas seen as an actual product or a kind of by-product in the process of the psychology inCreative Automatization: Principles for Promoting Fluency within a CommunicativeFramework, From the perspective of cognitive research in second language fluency,Oppenheim (2000:220-242) Mentioned that non-native speakers' fluency can be definedby selecting words and expression in The Importance of Recurrent Sequences forNonnativeSpeaker Fluency and Cognition. Oppenheim think that the sequence of someshorter cyclical repeat words can effectively embody the fluency, and he also took theexample of strength theoiy, hierarchical block theory as the basis (LaiHongling,WangJiandian, 2010).
            2.2 An Overview of the Ecological Classroom
            As a theory, ecological classroom has its own definition and characteristics whichdiffer from other teaching theories, and the mode of ecological classroom is also notsimilar with any other teaching modes. The term ecology was put forward first by the German biologist Haeckel in 1866.He is engaged in the subject that studies the relationship between organisms and theirenvironment of discipline. It is in the field of natural term. With the continuousdevelopment of ecology, the concept gets involved in many disciplines. Moreover,classroom is a kind of special ecosystem, which has a unique ecological classroomsubjects (teacher and student) and classroom ecological environment. It is an organicecological unity formed by the students, teachers and classroom environment and so on.These ecological factors are relatively independent and interact on each other. Ecologicalclassroom is to promote students' life development. Students are whole persons, whohave their own ideas, personality,needs and all sorts of living ability of individual life.Based on the development of students,ecological classroom makes the student life freedevelopment through the most effective and optimized class teaching activities andfully the potential of each student, in ecological classroom,'not only the teacher impartsknowledge and students grasp the knowledge and skills.
            Chapter Three Theoretical Basis .........17
            3.1 The Theoretical Basis of Ecological Classroom.........17
            3.2 Aspects for Oral English Measurement......... 22
            3.2.1 Definition of Accuracy .........22
            3.2.2 Definition of Fluency.........22
            3.2.3 Definition of Complexity......... 23
            Chapter Four Experiment of the Mode of Ecological Classroom ......... 24
            4.1Research Design......... 24
            4.2 Implementation of the Mode of Ecological Classroom......... 27
            4.3 Experiment Procedures......... 32
            Chapter Five Results and Discussions .........  39
            5.1 Analysis and Discussion of the Testing Results......... 39
            5.2 Analysis and Discussion of the Questionnaire Results......... 46
            Chapter Five Results and Discussions
            5.1 Analysis and Discussion of the Testing Results
            By more than three months of experiment,we compared the data collected frompre-test and post-test in the three indexes of complexity, fluency and accuracy. Bycomparing the two means compared to test whether there is a difference, and if thedifference is significant. As a result of the same subjects, we adopted paired sample t-testcorrespondingly, so that we can avoid the effects caused by participants and the influenceof other factors, which may decrease accuracy of the test results. ZhangWenzhong and WuXudong used four types in which contained twelve itemsof index of fluency, including the index of time,content, language and expression tomeasure fluency and its development of learners who studied the second language in theenvironment of foreign language classroom. When synthesized the study of Towell andLennon,ZhangLi thought that speech and timeliness of unnatural pause was the mostimportant indicators,and the paper also chose these two quantitative indicators toanalyze expression of oral English. Speech refers to the number of words that are used bythe participant every minute, unnatural pause refers to a pause more than 3 seconds.Considering the complexity of the experiment and students' oral English level in juniormiddle school,the study selected two quantitative criteria as indicators to test fluency ofstudents' oral English. The two criteria are vocabulary students spoke per minute andunnatural pauses students paused at every 60 words .Through statistical data collectedfrom pretest and posttest,we can detect whether there are differences and if thedifference is significant using the paired samples t-test. Test data is as follow.
            Complexity,accuracy and fluency are the most important three indicators in oralEnglish study. Although these are three different concepts,there is a great deal ofcorrelation between the three concepts. Ellis thinks if we analyze the language from thecomplexity, accuracy and fluency, it can make us get more information about languagelearners comprehensively. The reason is that at the time the learners finished output taskin second language (including output in oral and written language), it will be emphasizedin the three aspects. Therefore, it can provide effective means and methods for ouranalysis of the learner's language behavior from the three aspects. Through tiie analysisof test data and survey data, we can draw the following conclusions.According to fluency indicators, fluency used the following method to determine inthis study: 1) the number of words the subjects used every minute; 2) unnatural pausetimes every 60 words. Based on the result of the two tests, the result showed that thenumber of words the subjects used every minute is on the rise, that is to say,students canexpress more words in oral ecological class. Meanwhile, unnatural pause in every 60words is a downward trend, which means that the mean of unnatural pause times fell inthe post-test. Table 5-2 indicates that two pairs variables collected before and after theexperiment have a correlation. 
            References (omitted)


            在学习者对英语的学习和使用中,通常会遇到这样的问题。在某个与动词有关的表达中,不知道一个特定旳与这个动作相关的一个对象应该放在句子的什么位置,以及使用什么样的功能词来借助表达。如例(1)关于动词赌(Bet)的例子:(1)我用一顿饭做赌注和张三打赌中国队能够在本次世界杯预选赛出线。表达上述这句关于赌的意义,涉及到几个与赌这个动作相关的对象,也即动词的表达不同语义关系的论元(以下简称语义论元):谁在赌(Better),和谁赌(Co-better),用什么做赌注(Amountofbet或Stake),以及赌什么(Theme)等。在使用英语表达上述论元时,比如”我和张三赌”,因为母语的影响,学习者可能会不知道是否表达为“Ibet with ZhangSan”,还是“I and ZhangSan bet”,还是两者都不是。“用一顿饭赌”是否表达为“using/by/with a meal”。对于“赌什么”是否能把相关论元直接与谓词连起来表达,如“bet China soccer team's winning”等,类似汉语的表达习惯:我赌中国队赢。例(2)为同一个意义的英语的例子,可以反映英汉语在表达这些语义关系的句法要求的不同:在语言表达中,表述不同的语义关系对相应论元存在不同的句法要求。这反映了一个语言群体的语言使用习惯,并且这个习惯因语言而异。因此,这往往成为了外语(如本文研宄的英语)学习者所容易混靖的学习和使用难点。这即为本文研宄的动因所在。我们旨在设计和建立一个有效的检索系统,从而满足学习者对相关语言知识的而安。
            本文的贡献在于:第一:在型式语法的基础上,提出了整句型式,即把型式元素的构成范围扩展为句子中动词的左右两边,并把型式按照动词语态做了区分,同时加入了包括短语中心词和小品词作为型式元素的候选类别。此为对型式语法理论做了一定的扩展和补充。第二:在型式和意义的关系方面,提出了按照标准语义角色类别划分框架对动词型式元素进行语义分类和描述,为研究词汇的语义倾向性提供新的角度。第三:提出了一个联合完成型式提取和自动语义角色标注的软件模型和基于Map/Reduce的并行计算方案。第四:显著提高了对SRL核心论元的标注效率。对SRL核心论元的识别和语义角色标注的准确率提高了 2.8个百分点。第五:建立了一个按照语义角色索引的动词型式检索系统,尝试为英语学习者对型式的学习和使用提供一个更简便和直接旳工具,在机器辅助型式学习方面探索了新的思路。系统并可作为相关研宄的辅助工具。文章组织结构如下,第二章为相关研究综述,第三章讨论研究设计和方法依据,第四章为计算相关部分,将详述包括型式提取算法,自动语义角色标注,软件构建,并行计算模型等。第五章为对结果的初步评估。第六章为结论部分。
            传统语法对语言结构的描述采用诸如主语+动词,动词+宾语,或者动词+补语的方式。随着语料库语言学的发展,人们对搭配现象的认识逐渐深入。自上世纪以来,在Firth, Halliday禾n Sinclair等的研究椎动下,研究者通过难于大量语料的研宄,对词汇周围环境的特征有了更深入的理解。语料库语言学界逐渐趋向于认为语言的选择性倾向表现在词汇和句法的两个方而。基于句法形式的检索系统也括StringNet (Wible 2010),以及Wu (2010)和Grasp(Huang and Chen 2011 )等。StringNet从大型语料库中自动识别和提取了词汇的句法型式,并在此基础构建一个检索系统。同时,StringNet还识别了语言表达型式之间的包含关系,用以反映语言的句法构造(Construction)结构和依赖关系。其研宄目的在辅助语结构的关系的研究,但其最终研宄目标为辅助在二语教育(SLE)领域的教学:Grasp的研究通过提取搭配词周围的型式用来辅助相关学习。(Wu, 2010)从Google Web N-Grani提取两种型式类型:第一种和StringNet的型式相似;第二种提取包含人称代词的特定型式,用来辅助人称代词有关的语言学习。这些系统缺陷是明显的。
            型式在语言学习中扮演着关键角色。Hornby (A Guide to Patterns and Usage inEnglish, Hornby 1954/1995)对型式做了较为全而的总结和论述。这本书对句法型式的研宄起了重要的推动作用。作者在书(Hornby 1995: V)中说:“关于如何把词语组^r到一起的知识,比关于词语意义的知识来得更加重要。学习者除非知道这个知识,否则将不会知道如何运他/她的词汇”。“学习者在学到一定程度之后,语法分析是有肋益的,但是一个学习者更需要知道如何建构一个句子”。Sinclair把句法型式放在搭配的概念下进行了研宄,并对型式研究的发展做出了突出的贡献。Sinclair (1991)指出,“外语学习过程屮,对目标语中的常用词及其主要用法型式和典型搭配的学习是最主要的”。Willis (2003: 71 )在他对动词型式的研宄中认为,“学习者需要搞清楚指定动词的型式,也要能够把指定动词恰当地应用到型式当中”,又说“对句子结构学习的要问题在于动词的型式”。对型式的正确使用也是衡量一个语言学习者准确和流利地运用语言的重要标志。型式作为更广泛的搭配现象的部分,体现了语言在结构上的约定俗成的习惯,也体现了词汇搭配上的选择倾向性。“地道和流利的语言组织在于对词汇化和习俗化的结构型式的广泛知识”(Pawley and Syder 1983)。而这种倾向性和习惯上的约定,“不能依据语义和句法规则来预测出” (Cruse 1986)。并且,从形态句法和语义两个方面,型式结构和型式构成成份的挤配用词都不可被其他类似词替换,或者只能被有限的一部分类似词替换。这种句法和词汇的特定性,不同的语言有不同的体现,在不同语言之间存在有显著区别,“里然所有语言都可以用来达到基本交流的目的,然而他们采用的方式却各异,尤其在句法,语义和语用之间的交互方式上的差异”(VanValin Jr2005: 1),正如本章例(1)中的句子所示。
            3.1型式类型表示......... 17
            3.2 型式元素及其标签设定......... 23
            3.3检索系统设计.......... 27
            第4章检索系统的构建......... 30
            4.1型式提取......... 31
            4.2句法树解析......... 33
            4.3语义角色自动标注的计算模型......... 36
            4.4型式提取流程......... 39
            4.5分布式的型式提取和语义标注的一体系统......... 43
            4.6检索系统......... 46
            第5章结果与讨论......... 47
            5.1型式提取......... 47
            5.2型式元素识别和语义角色标注模型评估......... 54
            型式元素识别和语义角色标注两个自动分类模型的评估按照ConLL-2005对语义角色标注模型的设置进行。其中训练集为Perm Treebank il的WS. l(Wall Street Journal)新闻数据的第02到第21部分,测试集为第23部分,开发集(development set)第24部分。由于型式元素识别和语义角色标注两步前后承接,存在依赖关系,研究者一般釆用综合评价的方式,对总体结果做出评估(实际上,这也是标准任务竞赛的评估方法)模型的总体效率如表5.3和5.4所示。总体的效率,在Punyakanok (2008)的基础上准确率提高了 2.8个百分点。而要注意的是,Punyakanok (2008)的结染为使用6种句法树(包括 Charniak Parser top-50, top-100 等 5 种和 Collins Parser 的输出)的组合而得到的效率。


            近十年发展起来的新型的科技形式Web 2.0,相较于Web 1.0来说,升级建立了交互式的网络环境,例如微博,百科,论坛等等。发布帖子已经不仅仅是后台网络运营者独白式的单向表达,而是人人能够发布自由意志、更新网络内容的交互式双向表达。这就创建了多方协作,多元化的平台环境。新媒体就有适应新媒体的表达方式,网络语言己初步形成了自己的交际规则施春宏(2010)。而笔者认为,评价语言作为系统功能语法的一大分支,如同林大津(2003)所说“能够通过评论体现信息在实际言语事件中的传递”。而在此环境下产生的网络书评,是一种新型的评论语言,赋有新的人际交际意义,具有网络新型评价语体及评价语域特征,也应该与传统的报刊书评有评价语言特征的差异。而这些评价语言差异能够代表新的评价思维及价值观,亟待研究。除此之外,司莉(2005)曾写到:“书评具有通报、鉴定、激励、导读等职能,其中,通报职能是书评的基本职能,鉴定主要是评论,评论才是书评的灵魂与主体,不然书评则等同于书目提要”。而笔者则认为书评的价值,在于提供多元的客观的信息来培养合格的读者,使其具有更加良好的格调和鉴赏能力,防止单纯以商业化经济收益来评论书籍的好坏,帮助大众深度了解书籍隐含的深刻意义及艺术造诣。因此笔者企图通过评价语言特征差异,来探究网络书评是否能够化解报刊书评的弊端,弥补报刊书评的不足,为大众提供一种更为全面、更完善的书评方式,与报刊书评一起增强书评的评论职能。因此,作为新技术媒体下创造出的新型书评文学体裁,是否与传统书评有极大的话语评价差异,且其是否能够增强书评的评论职能亟待探索和研究。
            本研究在此选择网络书评作为切入点,搜集5万余网络书评语料及1万余报刊书评语料,运用Martin & White评价理论体系,旨在分析发现两种体裁的语言评价差异及特征。笔者认为,报刊书评指发表在报纸、刊物类传统纸质媒体上,对畅销书籍的内容、幕后内容及周边新闻等进行报道及评价的文学体裁。司莉(2008)认为可将网络书评分为两类:一是印刷版书评的网络版,二是在线产生的数字书评,包括网络书店、个人网站中的读者书评以及学术交流网络中的书评(在线学术书评)。而笔者认为,网络书评应该是指发表在网络上,针对任何书籍的自由的个人评论的集合。印刷版书评的网络版在笔者认为属于报刊评论。因为其属性特点是依赖于纸质媒体。报刊新闻性质的语言特征己经有许多的分析,而基于Web 2.0新型科技形式下的书评,其语言特征是否也会完全区别与传统的报刊书评,又有怎么样独特之处,其又表达了怎样区别与报刊书评的思想、价值观、社会观亟待研究与探讨,这也是本研究探索的重点。
            Martin & white的话语评价体系(The APPRAISAL Systems)是评估语言评价能力的体系,其评价系统是系统功能语言学中语篇语义学的一部分。所谓评价,就是评估篇章中所写上的各种态度,所涉及的感情的强度,以及作者是以什么方式把价值归属于源头,是一个人际意义系统(Martin & White 2005: xvii)。因为语言研究只从结构和功能出发是不够的,还需要赋值语义的研究(semantics ofevaluation),即研究说(对)话者通过语言赋予语言对象的价值意义(王振华2001)。正如林大津(2003)所说,功能语法把自然语言看作是一种语用现象,语言形式是作为意义的载体而存在的,而意义的存在是为使信息在实际言语事件中能够在表达者和接受者之间相互传递,因此,形态句法是为语义服务的,而语义是为语用服务的。评价理论体系从语义的角度,对单词的修饰属性以及显性或者隐形的表达思维方式,价值观念,人际意义提供了明确、战略性的分类角度及标准,为从语义角度评判社会关系及语言特征提供了依据。本文研究中,笔者需根据子系统的定义对评价类单词进行分类,其分类标准及界限具有一定的主观性的成分在里面,作者也借鉴了一些学者对分类的方法。Read & Carroll (2012)为子系统的分类及定义提供了一个较全面的标注区分的方法,对本文有极大的借鉴性。刘世涛(2009)的情感评价意义结构研究也为本文提供了一定的参照。
            4.4鉴赏语类质性分析......... 22
            5结论 .........30
            5.1研究发现......... 30
            5.2研究不足......... 33
            主题词通常能够直接反应出文本的核心内容及关键点。因此作者对报刊书评和网络书评二者的主题词进行提取和对比分析,来揭示报刊书评和网络书评在核心内容上和评论焦点上的异同。作者利用AntConc软件,选择布朗语料库的CROWN语料库作为对照语料库,分别将报刊书评语料库和网络书评语料库进行了主题词的提取,得到的主题词如表所示。首先,非常明显的一点差异是,在网络书评中,I作为主题词非常凸显,而在报刊书评的主题词中第一人称代词作为主题词的位置相对靠后。作者搜索网络书评中1单词词族之后发现,如图5.1所示,网络书评中“我”连接最多的单词是接纳单词,例如think,否定单词,例如don’ t,正面情感词以及助动词“是”。这充分印证了网络书评更多的是主观想法的公布和分享,强调评论者的个人意见,其次,网络书评主题词中出现了 unrealistic、believable, intelligent,而报刊书评中没有这方面含义的主题词出现。研究发现,网络书评中对小说人物的描写是否符合实际情况或客观规律进行了讨论,即对男女主人公的言谈举止与年纪不符的质疑,而在报刊书评中,评论者并没有关注此方面的问题,因而没有相关内容的主题词出现。


            I Introduction
            1.1 Research Background
            Polysemy is a very common semantic phenomenon in all languages. At the same time,it is the research object of many disciplines, such as linguistics, logic and philosophy, etc.The word polysemy, originating from Greek, is composed of two parts-polys and sema,which refer to “many meanings”. As its formation suggests, polysemy means a word thathas more than one meaning. We can get a more detailed and precise understanding ofpolysemy from some linguists’ definitions. Taylor considered polysemy as “the associationof two or more related meanings with a single linguistic form”. (Taylor, 1995) AndFromkin defined polysemy as “a word that has multiple meanings related historically orconceptually”. (Fromkin, 1998)Aristotle in ancient Greece did the earliest study on polysemy from the perspective oflogic. In the following researches on polysemy, many learners and scholars have madegreat contributions from different perspectives, such as psychology, philosophy, linguistics,semantics, etc. But most of these studies focus on the level of words, the result being thatthe polysemy research is mainly confined to semantics.Some linguists have tried to adopt several methods to explain this phenomenon. Katzand Fordor put forward componential analysis which makes an attempt to explain a largequantity of underlying meaning contrast. (Katz, J. J. & Fordor, J. A. 1963) This theorycould help measure whether a phrase is meaningful or is acceptable or not. But it can notmake a explanation on principle to polysemy phenomenon. Componential analysis thenwas challenged by semantic field theory that was first put forward by the German linguistJ. Trier. (J. Trier, 1931) J. Trier held a point of view that semantic field means that a part ofwords under the control of a certain general word or a concept have characteristics thatcould distinguish themselves from other words. According to this theory, every wordshould be put into semantic field to be researched. A single word can only have meaningwhen is put into “semantic field”.
            1.2 Significance of the Research
            We make effort to reveal the properties and nature of polysemous verbs through aseries of analysis from the frame semantics point of view. The most important thing for usis to construct the relations between frames and polysemous meanings. Frame is usuallyused to refer to the background knowledge or conceptual structure needed inunderstanding the concept represented by words. Word meanings depend on speaker’sexperience and presupposed social institutions of discourses. If we do not consider theframe structure which acts as holistic knowledge, then we can not identify meanings ofwords and discourses. Once English learners construct the relations between frames andmeanings of polysemous verbs, a knowledge network connecting the language world andcognitive domains will come into being in their minds. Frame semantics theory derivedfrom the empirical semantics tradition also provides an effective way for teaching ofEnglish polysemous words. Polysemous word learning has always been a difficult task forEnglish learners. Frame semantics emphasizes the understanding of word meaningdepends on its corresponding frame. It makes learners avoid memorizing meanings ofpolysemous word mechanically. This research aims to elaborate the relation betweenframe and English polysemy, and try to find out how an English verb get its severalmeanings from the perspective of frame semantics.
            II Literature Review
            2.1 An Overview of Polysemy
            The word polysemy, originating from Greek, is composed of two parts-polys andsema, which refer to “many meanings”. Aristotle in ancient Greece did the earliest studyon polysemy from the perspective of logic. From then on, lots of linguists made greatcontribution to polysemy research. It has been a long course of research on polysemy phenomenon in linguistics circle.For example, Taylor held an opinion that polysemy phenomenon is a single linguisticpattern which has two or more than two types of closely related meaning aggregations.(Taylor, 1995: 99) McCauley considered that it makes no sense to arrange all the relevantmeanings that are related to a phonetic form together. One entry can only have onemeaning, one form and one syntactic category. (McCauley, 1968:126, quoted from Hopper& Traugott, 1993: 69) According to this point of view, there should exist an most ideallanguage in the world that every meaning of the word in this language is highly evident.Otherwise, this “the most ideal language” is not practicable in that it is obviouslyimpossible for a human’s brain to remember every lexical form that contains so manymeanings. (Hopper & Traugott, 1993: 71) Lyons also pointed out that it can be virtuallysure that there is no its existence and the phenomenon that one word corresponds to onemeaning in one natural language has never existed before.
            2.2 Frame Semantics
            The foundation of frame semantics marks that Fillmore’s linguistic theory starts an important cognitive diversion after case grammar into a new developing stage. Throughthe changes of syntactic frame, case frame and so on, frame has become a cognitiveconceptual tool which constructs and explains human experience. In the mid 1970s, when case grammar faced a lot of problems and got itself into apredicament, Fillmore changed his research direction about linguistics. Following hisinterest in semantics in case grammar, he began to pay more attention to the meaning oflanguage and the problems of its understanding. Fillmore once wrote an article The Casefor Case Reopened, which responded the outside world’s doubt to case grammar.(Fillmore, 1977) This article mainly used scenes-and-frames paradigm of early framessemantics and the notion perspective in cognitive psychology to explain the choosingAgents of grammatical relationship like subject, object, etc. in sentences, and highlightedthe research direction of integration of syntax and semantics which was emphasized byFillmore. (Fillmore, 1970) So it meant that this article formally marked the detachment ofFillmore from case grammar under the frame of generative grammar.
            III Theoretical Foundation .......... 10
            3.1 About Frame Semantics........ 10
            3.1.1 Frame........ 10
            3.1.2 Frame Elements ...........11
            3.2 Cognitive and Semantic Properties of Frame........... 13
            3.2.1 Cognitive Properties of Frame...... 13
            3.2.2 Semantic Properties of Frame ...... 14
            3.3 Relation between Frame and Word Meaning ........... 16
            IV Verb Polysemy Within a Frame ...... 18
            4.1 Simple Frame vs Transferable Frame..... 18
            4.2 Simple Frame Analysis......... 19
            4.2.1 Transitive-based Simple Frame Analysis ....... 19
            4.2.2 Intransitive-based Simple Frame Analysis ..... 21
            4.2.3 Theoretical Extraction ......... 24
            4.3 Transferable Frame Analysis ......... 25
            V Conclusion ...... 32
            5.1 Major Results and Findings........... 32
            5.2 Limitations of the Research........... 32
            5.3 Suggestions for the Future Study ........... 33
            IV Verb Polysemy Within a Frame
            4.1 Simple Frame vs Transferable Frame
            From the previous part, it can be seen that word meaning could not be understoodwithout structural background or the participation of frames. Fillmore once said that wordsare tools that are utilized by language users, the function of which is to evoke or projectthe frame in language users’ minds. (Fillmore 2006: 613) This study is an exploration ofthe English verb polysemy on the basis of frame. There are two cases in which verbpolysemy happens. In one case, one frame corresponds to several grammatical meaningsof the verb. In another case, one frame corresponds to one lexical meaning of the verb.The former frame can be called simple frame. A simple frame refers mainly to one inwhich the prominence of the frame elements in different syntactic positions may producean effect on the grammatical meaning of a certain verb. The latter frame can be calledtransferable frame. A transferable frame refers to one in which the original frame activatedby a verb is likely to shift to other frames via metaphor in accordance with the Principle ofLeast Effort so that the original meaning of the verb is correspondingly changed.
            Under the guidance of frame semantics, we have made a relatively detailed andsystematic study on verb polysemy through the analysis of some certain typicalpolysemous verbs picked up from magazines, periodicals and dictionary. This study is anew attempt to apply frame semantics theory to the research of English verb polysemy inthe hope of helping Chinese learners of English to learn English verbs. There are threemajor findings that we get from the analysis.First of all, we did a quite detailed study on frame semantics theory and then find thatthe frame semantics theory includes some important concepts, such as frame, frameelements that do facilitate the understanding of meanings of verb polysemy.Secondly, in this thesis we divided frame into two kinds of frames which are closelyrelated with meanings of verb polysemy: simple frame and transferable frame. A simpleframe refers mainly to one in which the prominence of frame elements in differentsyntactic positions may produce an effect on the grammatical meaning of a certain verb. Atransferable frame refers to one in which the original frame activated by a verb is likely toshift to other frames via metaphor in accordance with the Principle of Least Effort so thatthe original meaning of the verb is correspondingly changed.Thirdly, when analyzing transferable frame, we find that the obtaining of differentlexical meanings of verb polysemy can be explained by frame shift, namely metaphor,even though metaphor degrees are different.
            References (omitted)


            1.1 Objective of the Study
            The relationship between verbs and nouns has always been a hot issue in the fieldof linguistics and computational linguistics. (Roark & Sproat, 2007: 116) With naturallanguage study emerging, the importance of this topic is increasingly exposed. Englishis a language of morphologic changes which usually signifies its part of speech and theuse of a word. (Baker & Hengeveld, 2012: 221) Deverbal nominals are a kind of wordscharacterized by the properties of both of verbs and nouns. Morphologically speaking,they carry the feature of nouns; semantically speaking, their senses can be divided intotwo types: concrete and abstract. The former is quite close to the sense of nouns, but thelatter is closely associated with that of verbs. The above situation is represented as infigure 1.1: In the above example, deverbal nominal “occupation”, in morphology, ends withthe suffix -tion, which indicates its noun feature. In semantics, the word is loaded withabstract sense and concrete sense. The former “占据、占领” expresses an action or abehavior itself, which indicates its verb feature; however, the latter “职业、工作”expresses something that takes up most of one’s time, which indicates its noun feature.Differences between morphologic and semantic feature and those between abstractsense and concrete sense usually become Chinese students’ dilemma in their Englishstudy. At the same time, these differences are a huge challenge to compile the meaningof English-Chinese dictionary. Therefore, objectives of this study are: on one hand, tofigure out internal relationship between the concrete sense of deverbal nominals andthematic roles associated with their root verbs. This relationship will be explained fromperspectives of syntactic and cognitive foundation; on the other hand, to exploretranslation methods or strategies of deverbal nominals.
            1.2 Method of the Study
            This thesis studies semantic connection and translation methods of deverbalnominal which ends with suffixes like -(a)tion, -sion, -ment, -ure. Detailed methods ofthis study are as follows:The first step is target word search, which aims to obtain eligible deverbalnominals with regular expression, such as *tion, *sion, *ment, *ance and *ure, inCorpus of Contemporary American English. Then come the choice of the first 1,000words of high frequency and the deletion of words that are not eligible like motion,procession, segment etc. Part of results searched for is shown as table1.1.The second step is semantic description. This study will describe the classificationof concrete sense of “-(a)tion-typed” deverbal nominals chosen. Because the abstractsense of “-(a)tion-typed” deverbal nominals is similar to the meaning of their root verbs,this study focuses on their concrete sense. In the stage of semantic description, thesemantic connection of “-(a)tion-typed” deverbal nominals is described from theperspective of thematic roles and syntactic argumenta activated by their root verbs.
            II Literature Review
            2.1 Studies on English Nominalization Abroad
            To some extent, studies on English nominalization can date back to Plato andAristotle in ancient Greek. Although they did not clearly propose the notion ofnominalization, they divided part of speech into two parts: nominal constituent andverbal constituent, discussing the transformation from one to another. (刘国辉, 2000: 5)Nominalization is a kind of nominal constituent. According to the definition of ModernLinguistic Dictionary, nominalization refers to a process of forming a noun derivedfrom other parts of speech or a process of forming a noun phrase derived from a basicclause. (戴维 克里斯特尔, 2000: 240) Trauth & Kazzazi (2000: 327) also held the viewthat nominalization is the derivation from other parts of speech such as verbs (feeling),adjectives (redness) and nouns (boyhood). In other words, nominalization itself is aprocess of derivation or zero transformation. (Carstairs-McCarthy, 2002: 49) This thesisfocused on nominalized words which are derived from verbs, especially transitive verbs,which are termed deverbal nominals in the study.“The deep understanding of nominalization studies abroad is reflected upon fourmajor linguistic schools: ‘structuralistic linguistics’, ‘transformational-generativelinguistics’, ‘systemic-functional linguistics’ and ‘cognitive linguistics’. ” (刘国辉,陆建茹,2004: 17)
            2.2 Studies on English Nominalization at Home
            With the advancement of society and the development of thought, the applicationof English abstract nouns is widely spread, which is a characteristic in modern English.However, English abstract nouns are difficult to master because of their abstractmeaning and complicated use. Compared to nominalization studies at broad,nominalization studies at home have lagged behind.Throughout the publication of foreign books before 1990s, in general, booksconcerning many parts of speech and important grammar phenomena in English arepublished. However, there are few monographs and articles of English nouns study. Themost typical is Number of Nouns in English written by Bo Bing in 1990. Nouns arecompiled to a book, which indicates the importance of nouns. In 1995, a huge changetook place in English nouns study. This change is the A Study of Abstract Nouns inEnglish published by professor zhang Jin and Liu Guangyao of Henan University,which fills in gaps in noun study at home. In this book abstract nouns are divided intofour types according to their source: abstract nouns of action, abstract nouns of quality,abstract nouns of status and inherent abstract nouns. Abstract nouns of action arederived from verbs, such as amazement, preparation, removal. These abstract nounsmainly express the meaning of “behavior or action” of their corresponding verbs andthat of “state”, “result, manner, etc.” Abstract nouns of quality are derived fromadjectives, such as greatness, freedom, possibility. Because adjectives are to attributenouns, abstract nouns derived from adjectives mainly express the meaning of “quality”.Abstract nouns of status are derived from nouns, especially concrete nouns, such aspupilage, membership, and childhood. This kind of abstract nouns can express status,qualification, industry etc.
            III Theoretical Framework..........14
            3.1 Thematic Roles ..............14
            3.2 Syntactic Arguments......15
            IV Semantic Connection of “-(a)tion-typed” Deverbal Nominals with ThematicRoles..............17
            4.1 Connection with External Argument .........17
            4.1.1 Connection with Percept ..........17
            4.1.2 Connection with Agent ......17
            4.2 Connection with Internal Argument ..........19
            4.2.1 Connection with Direct Argument..........19
            4.2.2 Connection with Indirect Argument .......22
            V Syntactic and Cognitive Foundations for “-(a)tion-typed” Deverbal........26
            5.1 Syntactic Foundation for Deverbal Nominals’ Semantic Connection........26
            5.2 Cognitive Foundation for Deverbal Nominals’ Semantic Connection.......27
            VI Research Significance Viewed from the Perspective of Translation
            6.1 Collocative Meaning Analysis
            According to American translation theorist Nida and Taber, the four-phase processin translating consists of analysis, transfer, restructuring and test. Among them,“analysis” is the most complicated and critical, but the core of “analysis” is meaning.(Nida & Taber, 1969) Thus the translation of English deverbal nominals should be basedon the understanding of their meaning in sentences or contexts. Famous English linguistLeech holds the view that there are seven types of meaning. One of them is collocativemeaning. Leech claims that collocative meaning is the combination of associationacquired the deverbal nominal connected with other words neighbored (Leech, 1983). Inother words, collocative meaning refers to the meaning which must collocate with othermeanings. The so-called “association” is actually the concrete sense of the deverbalnominal which collocates with other words neighbored. Leech indicates that collocativemeaning can play a part when other meanings fail to explain the meaning of one word.Nevertheless, it is held that collocative meaning is last but not least when dealing withdeverbal nominal translation, because the meaning of some English deverbal nominalscan be figured out only by the analysis of their collocative meaning.
            This thesis mainly discussed the semantic connection of “-(a)tion-typed” deverbalnominals based on thematic roles and syntactic arguments. Major discoveries wereidentified as follows:To start with, it was found that the meaning of “-(a)tion-typed” deverbal nominalsis closely associated with their syntactic arguments and thematic roles activated by theircorresponding verbs. Broadly speaking, if the argument structure of a verb is morecomplex, or more thematic roles are attracted humans’ attention, there are moresemantic items of its deverbal nominal. Besides, “-tion-typed” deverbal nominals arecharacterized by semantic connection. Specifically, they can connect their meaning withexternal arguments including percept and agent. They can also connect their meaningwith direct arguments including product, theme and patient and with indirect argumentsincluding quantity, location and instrument. The knowledge of their semanticconnection contributes a lot in reasonable prediction, accurate understanding and correcttranslation of the meaning of “-(a)tion-typed” deverbal nominals.In addition, semantic connection of “-(a)tion-typed” devebal nominals could beexplained by syntactic and cognitive foundation. In other words, their semantic meaning,to some extent, depends on the relationship between their root verbs and otherconstituents and human’s cognitive law. This phenomenon can be explained bysubject-object asymmetry of generative grammar and attentional view of cognitivelinguistics respectively.
            References (omitted)



            索引行十分便于人们观察语言中的词汇复现。早在1262年,Hugh of St Cher召集了 500多名僧人把圣经的拉丁武加大译本(Vulgate Bible)做成了索引行。后来诸多学者先后把不同语言版本的圣经和名家作品做成了逐词索引(Wisbey1962: 161-162)。20世纪初,索引技术由国外传入中国。其后不久,民国时期海军将领、外交家蔡廷干在1922年编成了《老解老》一书,即老子《道德经》的语词索引,这是我国最早以“堪靠灯”(concordance,语词索引)的方式研究《道德经》(王雅戈、侯汉清2007: 81)。但受制于当时的技术手段,上述词语索引的制作通常由多人长年累月的不懈努力才得以完成。计算机诞生之后,索引行制作发生了革命性的变化。以John Sinclair为代表的早期语料库语言学家意识到“同时观察大量语言时,语言会显得格外不同”(Sinclair 1991 : 100 )。这些语言学家开始对索引行进行人工解读。Sindair( 2003 )更是根据扩展意义单位理论,对索引行的解读方法进行了系统的归纳,形成了一整套系统的分析方法。然而,时至今日,自然语言处理技术已经获得了极大的发展,但索引行解读仍然基本靠人工来完成,这种方法耗时费力、效率低下。特别是随着大数据时代的到来,真实语料大幅增加,上亿词次的语料库已屡见不鲜,基于检索词所析出的词语索引常达上万行,人工解读所有的索引行己经成为不可能。在这种背景下,一些学者开始尝试对索引行进行抽样,然而这种方法所带来的直接结果是数据浪费,使得大型语料库建设的意义大打折扣。基于以上现状,本研究尝试采用文本分类中的聚类方法,把每个索引行视作微型文本进行自动聚类,最终实现英语动词型式的自动识别与提取。
            本研究的理论意义主要体现在以下三个方面:第一、推动短语学思想和习语原则的发展涉及重复词语共现的概念或术语繁多,语言学界莫衷一是,出现了诸如语料库语言学中的扩展意义单位与型式,自然语言处理中的多词表达、N元组及词丛,以及心理语言学中的预制语块、程式语等。虽然以上术语采用的视角不同,但均与语料库语言学界的短语学思想和习语原则契合。本研究基于型式语法和聚类算法,实现动词型式的自动识别与提取,有助于型式后续研究的开展。所以,本研究的发现有助于推动短语学思想和习语原则的进一步拓展。第二、推进词汇语法不可分的相关实证研究,回归词汇为中心的研究范式传统语法认为,词汇和语法在语言学中地位不同,语法处于支配地位,词汇则处于边缘地带,是语法研究的附属品。实际语言使用中,语法限制确实存在,但却无法解释一定语境内词汇间的共现现象。其实,上述现象反映了语法限制的同时,又反映了词汇间的相互限制作用,语法限制和词汇限制并存,两者同时制约着人类语言的使用。如eat可以出现在Vn型式中,具体语言实例包括 eat breakfast, eat an apple 等,但几乎不出现 eat glasses 和 eat table 的词语组合》上述实例反映出eat作动词使用时,其后可以跟名词,但并非所有名词均可出现在动词eat之后。这种现象既反映了语法限制的存在,又体现了词汇限制的作用。事实上,英语动词型式列表及其实例(Francis 1996/2000)同样揭示了语法与词汇的密不可分。
            型式(pattern)由Homby (1954)在语言学界使用以来已有60年历史。由于其《英语型式用法指南》一书的目的在于为高级英语学习者提供帮助,属于教学语法系列,不属于语言学方面的理论著作,因此该书没有对型式进行定义。书中仅对型式中的组成元素进行了部分解释说明,然后列举了许多型式实例。但是“Homby最重要的贡献或创新在于开始关注语言的结构化或型式化问题,并用简洁的方式将词汇使用中的型式进行归纳总结后提供给学习者。此外,他还发现每个动词可以使用在数量相对固定的一系列句法结构之中。通过对线性结构的语言现象分析,建立了动词型式列表,使得原本杂乱无章的语言现象有了秩序” (Hanks2008:89)。下表所示为其书中所列的25种动词型式:Hornby (1954)同时指出,“动词型式最为重要,但语法书或词典没有提供此方面足够多的信息,因而(学习者)在动词型式的学习上应该投入更多的时间”。Hanks (2008:90)将Homby和Palmer在学习者动词句法及词汇选择方面研究的贡献总结为以下三点:1)语言是高度结构化或型式化的,每个单词与少量的句法型式相结合。2)日常话语由少量高频词汇型式构建而成,并且每个词汇只使用在少量的型式结构之中。同时,语言使用中还包括大量的词语或是结构,但是其中一部分的使用频率很低。3)动词是小句的轴心。
            型式语法是本研究的理论基础,是指导本研究模型构建的实践语法。正如Hunston & Francis (1998:62)所言,型式语法是第一部运用语料库数据编写的教学语法,.编写的过程中将语法与词汇当做整体看待,其目的在于揭示传统if法中宾语、补足语等语法范畴,这主要是由于上述语法范畴对具体动词的使用行为解释力不强。型式是词汇和语法的交汇点,是描述英语的基本单位,同时也是语料库驱动研究中的核心概念。“型式语法的目的是描写英语中所有的型式。如果这个目的实现了,整个英语语法也就描述出来了”(Hunston & Francis 2000: 32)。型式语法基于大型语料库数据对词汇及其经常共现的特定短语和语法型式进行了详细的描写。在其描写语言的过程中,采用语料库驱动的方法,试图摆脱先前语言学理论的束缚来描写语法结构。因而,“作为第一部语料库驱动的英语描写语法,型式语法可谓是开创性的,它颠覆了以往语言学家看待语言、获取语言知识的传统,给语言研究者展现出了一个全新的视角”(陈功2012:51)。型式语法理论的提出,目的是多方位的,如服务于语言描写、型式与意义关系的分析等。例如Hunston (2001:15)指出,型式语法的目的在于使用词汇途径描述语法(Francis 1996, 1998; Hunston & Francis 1998, 2000);型式间差异与型式的意义有密切关系(Sinclair 1991; Levin 1995),型式可用来解决一词多义现象(Sinclair 1991: 53),具有同样型式的单词有可能会具有相似的意义(Francis 1993, 1995; Hunston & Francis 1998, 2000)。综上,型式语法的理念及观点体现在如下几个方面:1)对短语学思想及习语原则的继承与发展;2)词汇与语法不可分割,应该以词汇为研究中心;3)型式与意义联系紧密。下面几个小节将针对上述观点展开详细的描写。
            3.1型式识别标准........ 41
            3.2型式总结的必要性........ 42
            3.3型式的识别 ........43
            3.4现有的语言型式网络平台数据库 ........51
            3.5小结........ 55
            4.聚类分析........ 56
            4.1文本表示........ 57
            4.2特征选择及其权重........ 58
            4.3相似度计算 ........59
            4.4聚类算法........ 63
            4.5聚类质量评价指标........ 66
            4.6聚类在本研究的应用理据........ 67
            4.7小结........ 68
            第五章文本预处理与特征集的建立及转换........ 69
            5.1研究概述........ 69
            5.2语料选取........ 69
            5.3研究工具........ 70
            5.4语料预处理流程........ 71
            5.5动词型式中的必要元素及其转换方法........ 73
            5.6 小结........ 104
            本研究在相似度计算的基础上,对索引行进行自动聚类,进而实现了英语动词型式的自动识别与提取。研究流程包括以下五个环节,1)数据的准备;主要指析出包含核心词项的索引行或句子,并进行赋码。2)动词型式列表中型式元素特征的归纳总结;特征选择是聚类环节中最重要的一环,直接影响聚类结果的优劣,因此本环节主要对动词型式列表中的具体词项、相邻结伴的词项组合、词类、语义类及句法类等元素特征进行归纳,进而建立特征集。3)语言特征的转换;主要将索引行中相关语言信息转换为型式代码。4)索引行间相似度计算与自动聚类;主要涉及特征权重计算、特征位置权重计算、特征-索引行矩阵的生成、相似度计算和索引行自动聚类。5)索引行实现聚类后,进行型式的自动识别与提取。上述五个环节环环相扣,尤其是后四个环节是模型的核心部分,任一环节出现问题,均会对模型最终的输出结果产生影响,因此对上述环节中涉及的诸多细节进行了不断调整与尝试(详见7.2和7.3节。本研究目的在于通过对索引行的自动聚类实现英语动词型式的自动识别。研究中提出了两个研究问题:1)影响索引行聚类的因素有哪些?如何确定索引行聚类中的分组数量? 2)型式自动识别与提取模型的准确率及召回率如何?影响因素有哪些?本节将针对上述两个研究问题,总结及归纳本研究的主要发现。