Re-exploringO.HenrysShortStoriesA Corpus-Based Pilot Study【摘要】本文试图采用语料库的方法从文体学视角分析欧亨利小说集四百万。研究揭示，通过语料库软件计算出的总体统计数据，为有关欧亨利小说广泛认同的文学阐释提供了更为具体的描述基础。在探讨小说场景和基本主题方面，重现序列的搭配及频数信息发现了前人并未关注过的语言学特征。 Abstract：Thisarticleattemptstoapplycorpus-basedmethodtoastylisticinterpretationofO.HenrysshortstorycollectionTheFourMillion.Itisshownthattheoverallstatisticscomputedbycorpussoftwarehasprovidedamoredetaileddescriptivebasisforwidelyacceptedliteraryinterpretationsofhisstories.Intermsofstorysettingsandgeneralthemes,thecollocationandfrequencyinformationofrecurrentsequencescanidentifyvaluablelinguisticfeatureswhichliterarycriticsseemnottohavenoticed. Keywords：corpus;O.Henry;setting;theme 1.Introduction HenrywascalledtheAmericanGuyDeMaupassant.Bothauthorswrotetwistendings,butO.Henrysstoriesweremuchmoreplayfulandoptimistic.AmongtheformerstudiesonO.Henrysshortstories,thereisconsensusthatO.Henrysworksaregenerallybrandedwithsuchfeaturesassurprisingendings,useofcoincidenceorchancetocreatehumor,ingeniousandexquisitelayouts,smile-in-tearsironyandsoforth.Despitethedetailedliterarydiscussion,littleworkhasbeendonetorevealitslinguisticstyles.Noristhereworkwithquantitativedataasconvincingevidence.Intermsoftheestablisheddescriptionofhisstyle,itseemsunlikelythatthecorpus-basedmethodcanfindanythingoriginal.However,thestylisticanalysisinthepresentpaperaimstoillustratethevalueofcorpusempiricalmethodinexploringtheliterarystyles.Ontheonehand,statisticdatahelptoconfirmthecanonicalviewonO.Henrysshortstories;ontheotherhand,stylisticsisrelatedtolinguisticfeaturesofhisworks. 2.DataandMethodology ThispaperisdevotedtoinvestigatingthelinguisticstylesofO.Henrysworksinanempiricalway,applyingbothquantitativeandqualitativemethods.Thestudyadoptstwocorpora.OneisO.HenrysbookTheFourMillion(acollectionofstories),publishedin1906,containsaseriesofshortstorieswhichtookplaceintheNewYorkCityintheearlyyearsofthe20thcentury.Thecomputerreadableversionsavailableontheinternetareusedtosetupaminorworkingcorpusforinvestigation(http:/www.literaturepage.com/read/thefourmillion.html).TheotheroneisBrowncorpususedasareferencecorpus. ThecorpusconcordancesoftwareusedinthisstudyisWordsmithtools.Wordsmithcanundertakemoredetailedanalysesoffrequenciesofconcordanceitemsandextractcollocationalinformation.Byuseofcorporasoftware,wordswithsignificantkeynessinthebookTheFourMillionwillbesortedoutfirst,andthenconcordancelineswithakeywordanditscollocateswillbeextracted.Thecorporadatawillbeprocessedbystatisticalinstruments. 3.OverallStatistics Theoverallstatisticsareoneessentialstartingpointforasystematiccorpus-basedtextualanalysis.WordsmithToolsareusedtoprovidetheoverallstatisticsofthetwocorporaandcomparisonismadeasshowninTable3.1. Table3.1ComparisonofOverallStatisticsbetweentheTwoCorpora textfile intext types(distinctwords) type/tokenratio(TTR) standardisedTTR standardisedTTRbasis meansentencelength(inwords)meanwordlength(incharacters) wordlengthstd.dev. OverallofMini 52,770 8,251 16 46.97 1,000.00 15 4 2.24 OverallofBrown 1,390,505 47,146 4 39.07 1,000.00 23 5 2.52 Intermsofthenumberoftokens,thecorpusforthepresentstudyseemsrathersmallcomparedwiththatofBrown.However,theobservedstandardizedtype-tokenratio(TTR)intheminorworkingcorpusishigherthanthatofBrown.Thehighertheratiois,themorelexicalvariationitreflects.Theresultisnotquitesurprising,forthefactthatO.Henrywasnotableforwritingstorieswithadiversityofexpressions.ThemeansentencelengthshowsthatO.Henrywroteshortersentencesthanthoseingeneralwrittentexts.Ineachofhisshortstories,severalincidentsandcluesarecondensedintomerely750wordsorso.Naturally,sentencesinshorterlengthcandelivernewinformationinaflashoftime.Ittakeslesstimeforreaderstofollowupthedevelopmentofthedramaticplot.Fundamentallyaproductofhistime,O.Henry#39;sworkprovidesoneofthebestEnglishexamplesofcatchingtheentireflavorofanage.Whetherroamingthecattle-landsofTexas,exploringtheartofthegentlegrafter,orinvestigatingthetensionsofclassandwealthinturn-of-the-centuryNewYork,O.Henryhadaninimitablehandforisolatingsomeelementofsocietyanddescribingitwithanincredibleeconomyandgraceoflanguage. 4.KeywordAnalysisforSettings Settingisanindispensableconstituentforstorytelling.Settingbasicallyincludesseveralcl