资源描述：
《Unit 10 Corpora and language studies1》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。
1、Unit10Corporaandlanguagestudies10.1IntroductionWehavesofarintroducedmostoftheimportantconceptsandpracticesincorpuslinguisticsrelatedeithertokeyissueslikecorpusdesign,markup,annotation,andthemultilingualdimension,ortoancillaryissuessuchasmakingstatisticalclaims,u
2、singready-madeandDIYcorporaandcopyrightclearanceincorpusbuilding.ThefinalunitofSectionAconsidershowcorporacanbeusedinlanguagestudies.AccordingtoLeech(1997b:9),corpusanalysiscanbeilluminating‘invirtuallyallbranchesoflinguisticsorlanguagelearning’(seealsoBiber,Con
3、radandReppen1998:11).Oneofthestrengthsofcorpusdataliesinitsempiricalnature,whichpoolstogethertheintuitionsofagreatnumberofspeakersandmakeslinguisticanalysismoreobjective(cf.Biberetal1998;McEneryandWilson2001:103;thoughseeunit12foradiscussionofthisclaim).Inthisun
4、itwewillconsidertheuseofcorpusdatainanumberofareasoflinguistics.Units10.2–10.8areconcernedwiththemajorareasoflinguisticswherecorporahavebeenusedwhileunits10.9–10.14discussotherareaswhichhavestartedtousecorpusdata.Inunit10.15,wewillalsodiscussthelimitationsofusin
5、gcorporainlinguisticanalysis.10.2LexicographicandlexicalstudiesCorporahaveproventobeinvaluableresourcesforlexicographicandlexicalstudies.Whilelexicographers,evenbeforetheadventofmoderncorpora,usedempiricaldataintheformofcitationslips(e.g.SamuelJohn’sOxfordEnglis
6、hDictionary),itiscorporathathaverevolutionizeddictionarymakingsothatitisnownearlyunheardoffornewdictionariesandneweditionsofolddictionariespublishedfromthe1990sonwardsnottobebasedoncorpusdata.Corporaareusefulinseveralwaysforlexicographers.Thegreatestadvantageofu
7、singcorporainlexicographyliesintheirmachine-readablenature,whichallowsdictionarymakerstoextractallauthentic,typicalexamplesoftheusageofalexicalitemfromalargebodyoftextinafewseconds.Thesecondadvantageofthecorpus-basedapproach,whichisnotavailablewhenusingcitations
8、lips,isthefrequencyinformationandquantificationofcollocationwhichacorpuscanreadilyprovide.Somedictionaries,e.g.COBUILD1995andLongman1995,includesuchfrequencyinformati
