语料库研究基本方法.ppt

语料库研究基本方法,中国外语教育研究中心梁茂成,主要内容,语料库语言学的性质几个常用术语语料库研究的基本方法,语料库语言学的性质,理性主义与经验主义Rationalism:IthinkthereforeIam.Empiricism:Mymindisablankslate.Seeingisbelieving.,语料库语言学的性质,theWaxArgument:Heconsidersapieceofwax;hissensesinformhimthatithascertaincharacteristics,suchasshape,texture,size,color,smell,andsoforth.Whenhebringsthewaxtowardsaflame,thesecharacteristicschangecompletely.However,itseemsthatitisstillthesamething:itisstillapieceofwax,eventhoughthedataofthesensesinformhimthatallofitscharacteristicsaredifferent.,语料库语言学的性质,theWaxArgument:Therefore,inordertoproperlygraspthenatureofthewax,hecannotusethesenses.Hemustusehismind.Descartesconcludes:“AndsosomethingwhichIthoughtIwasseeingwithmyeyesisinfactgraspedsolelybythefacultyofjudgmentwhichisinmymind.,语料库语言学的性质,Empiricism:Empiricismemphasizesthoseaspectsofscientificknowledgethatarecloselyrelatedtoevidence,especiallyasdiscoveredinexperiments.Itisafundamentalpartofthescientificmethodthatallhypothesesandtheoriesmustbetestedagainstobservationsofthenaturalworld,ratherthanrestingsolelyonreasoningandintuition.,语料库语言学的性质,Scienceisconsideredtobemethodologicallyempiricalinnature.Corpuslinguisticsisempiricalinnature.,语料库语言学的性质,语言研究中的数据类型内省数据（introspectivedata）:rationalism实验数据（experimentaldata）:empiricism真实数据（anthenticdata）:empricism,语料库语言学的性质,语料库语言学提倡真实数据我们不排斥其他数据类型,语料库语言学的性质,即便在语料库语言学阵营之中Corpus-driven:minimumtheory-reliance.ExclusiverelianceoncorpusdataforalltheoriesCorpus-based:Relianceoncorpusdataforhypothesis-testingCorpus-referenced/informed:Occasionallyresortingtocorpusdataforillustrations,语料库语言学的性质,我们坚决反对不顾语言事实的任何论断Nointrospectioncanclaimcredencewithoutverificationthroughreallanguagedata(Teubert2005).,几个常用术语,CorpusCorpuslinguistics,几个常用术语,Token,type,lemmaThelittleboylookedattheotherboys.,几个常用术语,Collocationisdefinedasasequenceofwordswhichco-occurmoreoftenthanwouldbeexpectedbychance.abigsmokerastrongsmokerahardsmokeraheavysmokerafurioussmoker,几个常用术语,Itisquitepossible,infact,todescribeawomanashandsome.However,thisimpliesthatsheisnotbeautifulatallinthetraditionalsenseoffemalebeauty,butratherthatsheismatureinage,haslargefeaturesandacertainstrengthofcharacter.Similarly,amancouldbedescribedasbeautiful,butthiswouldusuallyimplythathehadfemininefeatures.,几个常用术语,Colligationisdefinedasasequenceofgrammaticalcategorieswhichco-occurmoreoftenthanwouldbeexpectedbychance.,几个常用术语,SemanticprosodyisinstantiatedwhenawordsuchasCAUSEco-occursregularlywithwordsthatshareagivenmeaningormeanings,andthenacquiressomeofthemeaning(s)ofthosewordsasaresult.Thisacquiredmeaningisknownassemanticprosody.(Stewart2010),语料库研究的基本方法,Corpus-basedapproach:ahypothesis-testingapproachCorpus-drivenapproach:withas“fewpreconceivedideas”aspossible,“keepingtheamountoftheory-reliancetoaminimuminordernottohindertheprocessofdiscoveringnewphenomena”(Rmer2005),语料库研究的基本方法,Bothapproachesalmostalwaysinvolveacomparionofsomekind.,语料库研究的基本方法,Sizesofcorporaincomparison(Rayson2003)SmallbigEqualsizes,语料库研究的基本方法,TypesofcomparisonAcrossgenresAcrossusersAcrossdifferenttimesAcross(varietiesof)language(s),语料库研究的基本方法,Corpuscomparability,语料库研究的基本方法,LinguisticfeaturesincorpuscomparisonLexicalLexico-grammaticalSyntacticDiscoursal,语料库研究的基本方法,StatistictestsincorpuscomparisonSimple:Relationship(correlation,etc)Difference(chi-square,loglikelihood,etc.)Complicated:regressionanalysis,factoranalysis,clusteranalysis,correspondenceanalysis,语料库研究的基本方法,语料库,研究问题,软件,词汇短语搭配语义韵类联接句式等,内容5,Thankyou.,