Vous êtes sur la page 1sur 378

ChildhoodBilingualism:AspectsofLinguisticCognitive,andSocialDevelopment Editedby PeterHomel MichaelPalij DorisAaronson NewYorkUniversity LEALAWRENCEERLBAUMASSOCIATES,PUBLISHERS 1987Hillsdale,NewJerseyLondon iii

Copyright1987byLawrenceErlbaumAssociates,Inc. Allrightsreserved.Nopartofthisbookmaybereproducedin anyform,byphotostat,microform,retrievalsystem,oranyother means,withoutthepriorwrittenpermissionofthepublisher. LawrenceErlbaumAssociates,Inc.,Publishers 365Broadway Hillsdale,NewJersey07642 LibraryofCongressCataloginginPublicationData Childhoodbilingualism. ContainsrevisedpaperspresentedataConferenceon ChildhoodBilingualism,heldatNewYorkUniversity, June2526,1982. Includesbibliographiesandindex. 1.Bilingualisminchildren.2.Languageacquisition. 3.Childdevelopment.I.Homel,Peter.II.Palij, Michael.III.Aaronson,Doris.IV.Conferenceon ChildhoodBilingualism(1982:NewYorkUniversity) P115.2.C481987404'.2868955 ISBN0898598060 PrintedintheUnitedStatesofAmerica 10987654321 iv

Contents Preface ix PARTI INTRODUCTION 1Childhood Bilingualism: Introduction and Overview 3 PeterHomel, MichaelPalij, Doris Aaronson Introduction , 5 Language Acquisition and Processing, 5 Bilingualism and Cognitive Development , 6 Bilingualism andSocial Development , 7 Bidialectism , 7 Concluding Remarks, 8 . References, 9 2 Bilingualism and Language Policy: FourCase Studies 11 PeterHomel, MichaelPalij Canada, 11

TheSoviet Union, 14 TheUnited States, 19 ThePeople's Republicof China, 22 Some Psychological Implications, 24 References, 26 v

vi PARTII LANGUAGE ACQUISITION AND PROCESSING 29 3TheSecond Language Learnerinthe Contextofthe Studyof Language Acquisition 31 KenjiHakuta Effectsof Cognitive Maturity, 33 Effectsof Linguistic Structure: L1andL2 Similarities, 37 Effectsof Linguistic Structure: Native Language Transfer, 40 EffectsofAge, 43 Language Universalsas anIntegrated Perspective, 44 InSearchof Psychological Correspondence , 48 Conclusions, 51 Acknowledgment , 52 References, 52 4. 57 Bilingualism Language Proficiency,and Metalinguistic

Development JimCummins The Constructof Bilingualism, 58 The Constructsof Linguisticand Metalinguistic Proficiency, 61 References, 71 5TheImpactof Language Differenceson Language Processing:An Examplefrom ChineseEnglish Bilingualism DorisAaronson, StevenFerres TheStructure andMeaningof Lexical Categoriesin English, AnOverviewof ChineseEnglish Differences, Linguistic Performanceby ChineseEnglish Bilinguals, AComparisonof Sentence Processingin Bilingualsand Monolinguals, Methodsforthe RatingStudy, Subjects, Resultsofthe RatingStudy, Meaning Ratings,

75

76

77

79

86 87 88 89 96

Structure Ratings, 103 TheRoleof Verbsin Sentence Structure, 107 Comparisonof Structureand Meaning, 109 Developmental Implicationsof ChineseEnglish Differences, 112 Summary, 115 Acknowledgments , 116 References, 117 vi

vii 6Acquiring andProcessingMartinD.S. Braine Firstand References Second Languages: , Commentson Hakuta, Cummins,and Aaronsonand Ferres PARTIII BILINGUALISM AND COGNITIVE DEVELOPMENT129 7The Relationshipof Bingualismto Cognitive Development: Historical, Methodological and Theoretical Considerations 131 MichaelPalij, PeterHomel AHistorical View, 131 Methodological Considerations, 138 Theoretical Aspects, 141 Conclusions, 146 Acknowledgment , 146 References, 147 8Bilingualism, Cognitive Function,and Language MinorityGroup Membership 149 EdwardDeAvila Introduction, 149

Backgroundof Studies, 149 Sample, 153 Descriptionof Treatment Approach (Finding Out/Descubrim iento), 153 Outcome Measures, 154 Results, 158 Concluding Remarks, 163 References, 166 9Bilingualism: Cognitiveand SocialAspects 171 JosephGlick Some Comments on Bilingualism , 171 Social Natureof Bilingualism , 174 References, 179 PARTIV BILINGUALISM ANDSOCIAL DEVELOPMENT 181 10Social Psychological Barriersto Effective Childhood Bilingualism 183 DonaldM. Tuylor Multilingual Ideologies, 184 Deculturation Ideology, 186 vii

viii SeparationIdeology, 187 Assimilation Ideology Integrationist Ideology Acknowledgment , References 11TheEffectsof Bilingualand Bicultural Experienceson Children's Attitudesand Social Perspectives 197 W.E.Lambert Attitudesand Their Measurement , 200 Conclusions About Immersion Programs' Effectson Stereotypes, 207 TheFeelings andReaction Tendency Components of Attitudes, 209 Overall Conclusions, 216 References, 219 12ASocial Cognitive Perspectiveon Bilingualism: Commentson Lambertand Taylor 223 E.ToryHiggins 229 PARTV

BIDIALECTISM 13The Linguisticand Sociolinguistic PositionofBlack Englishandthe Issueof Bidialectalismin Education 231 JohnD.Roy References , 241 14 Continuities/Disc ontinuitiesinthe Functionand Useof Languageas Relatedto Situationand SocialClass 243 WilliamS.Hall, WilliamE.Nagy Situational Variation, 243 Cognitive Implicationsof InternalState WordUse, 247 Previous Researchon SocialClass Based Differencesin InternalState WordUse, 248 InternalState WordsWhat TheyAre, 249 Subjects, 253 Results, 254 Selectionof Speakersfor Analysis, 257 ResultsofData 258

Analysis, SESDifferences inCorrelations, 272 Implicationsfor theMismatch Hypothesis, 277 Acknowledgment , 278 References, 279 viii

15Copingor WilliamA. Groping? Stewart Psycholinguistic Acknowledgment Problemsinthe , Acquisitionof ix References Receptiveand Productive Competence AcrossDialects AuthorIndex 299 SubjectIndex 305 ix

[Thispageintentionallyleftblank.] x

Preface Thisvolumeisbasedprimarilyonaconferenceonchildhoodbilingualismheldat NewYorkUniversityonJune25and26,1982.Theideafortheconferencegrew outofaseriesofdiscussionsbetweentwooftheeditors,PeterHomeland MichaelPalij,whohadsubstantialinterestsinexploringthenatureofbilingual cognitionandtheeffectofbilingualismonpsychologicaldevelopment.We,the editors,werestruckbythewealthofresearchbutwereappalledbythelackof communicationbetweenresearchersin"mainstream"developmentalpsychol ogythoselookingatlanguagedevelopmentinmonolingualchildrenand researcherslookingatsimilardevelopmentalprocessesinbilingualchildren.We thoughtitwouldbeofgreatinterestandpracticalvaluetobringtogetherre searchersfrombothareasinanattempttostimulatedialogueandinteraction betweenthetwogroups. ThefirststeptowardholdingtheconferencewastakenwhenPaulDores,of SUNY,StonyBrook,gaveusacopyofarequestforproposalsfortheSociety forResearchinChildDevelopment's(SRCD)seriesofstudygroupsandsummer institutes.OurinitialproposaltoSRCDforfundingforasummerstudygroup focusedonfourareasofchilddevelopmentandhowbilingualismmightaffect eachone:languageacquisition,cognitivefunctioning,socialcognitionandcom munication,andpersonalityandemotionaldevelopment.SRCDapprovedthe proposal,addingtheissueofbidialectismanditsrelationshiptobilingualismas anotherareaoffocus. Thevolumecontainsmostofthepresentationsmadeattheconferenceand follows,withminorchanges,thegeneralorganizationoftheconference.During eachsession,two"bilingual"researchers(i.e.,doingresearchinthebi lingualism)presentedageneralreviewoftheissueswithinatopicareaandgave xi

examplesoftheirownresearchwithinthiscontext.A"monolingual"researcher (i.e.,oneorientedtowardresearchinmonolingualdevelopment)thenpresented adiscussionoftheissuesraisedbythetwobilingualresearchers,indicatingthe pointsofcontactanddeparturebetweenbilingualandmonolingualresearch. ACKNOWLEDGMENTS WethankSRCDforrecognizingtheimportanceofthestudyofchildhoodbi slingualismandforprovidingfundingfortheconference.Inparticular,wethank DorothyH.EichornoftheUniversityofCaliforniaBerkeley,whowasthe ExecutiveOfficerofSRCDatthetimeoftheconference,andViolaMoulton Buck,herassistant;GrayGarwoodofTulaneUniversity,whowasthechairper soninchargeofSRCD'sstudygroupprogram;andlargaretSpencerofEmory University,whowasSRCD'sobserverattheconference. Manypeoplewereveryhelpfulatvariousstages,inbothmakingthecon ferencearealityandhelpingtocompletethisbook.WethankDickKoppenaal, ChairmanofthePsychologyDepartmentofNewYorkUniversity,forhissup portoftheconference.Wealsoacknowledgethefollowingpeopleanddepart mentsatNewYorkUniversityfortheirhelpinsettinguptheconference:Peter ChepusoftheOfficeforFundedAccounts,MichaelRobbinsoftheBudgetand FundAccountingDepartment,andSherryDauletofthePsychologyDepartment, whohelpedinhandlingtheexpensesfortheconference;JacquelineDowningof thePhysicsDepartment,andGraceSunoftheHousingOffice,whohelpedmake thearrangementsforconferencespaceandhousingfortheconferencepartici pants;theCateringServiceofNewYorkUniversityforprovidingrefreshments andmealsduringtheconference;andFelixSchererofthePsychologyDepart ment,forhistechnicalassistanceinsettinguptheequipmentfortheconference. WearealsodeeplythankfultoSallyThomasonandLauraBrighentiforboth theirmoralsupportandmaterialhelpbefore,during,andaftertheconference. Aboveall,weareindebtedtoLarryErlbaumandCarolLachmanofLawrence ErlbaumAssociates.Manydifficultiesareattendantwithpublishingabookthat triestospananareaaslargeaschildhoodbilingualism.Wesincerelyfeelthat withouttheirhelpandseeminglyinexhaustiblepatience,thepreparationofthis volumewouldnothavebeenpossible. xii

I INTRODUCTION 1 [Thispageintentionallyleftblank.] 2

1 ChildhoodBilingualism:IntroductionandOverview PeterHomel MichaelPalij DorisAaronson NewYorkUniversity In1962,PealandLambertpublishedtheresultsofastudycomparingbilingual andmonolingualchildrenonvariousmeasuresofintelligenceandachievement. Theirfindingsweresurprising,atleastinlightofcertainassumptionsthathad beenprevalentinchildpsychologyuptothattime.Theyfoundnoevidenceto indicateanysortofintellectualdeficiencyinbilingualchildren.Theperformance ofbilingualsonallmeasureswaseitherequivalentorsuperiortothatoftheir monolingualcomparisongroup.Theseresultswereinclearcontradictiontoa beliefthathadcometobeacceptedastruismbypsychologistsandlaymenalike, especiallyinNorthAmerica:Theacquisitionoftwolanguagesinchildhood. impairsintellectualdevelopmentitleadstomentalconfusionordifficultiesin coordinatinglanguageandthoughtinchildren.TheresultsobtainedbyPealand Lambertsuggestedthattherearenodetrimentaleffectsofbilingualism,andthere mayevenbesomecognitiveadvantages. PealandLambert'sstudyhadamajorimpactonatleasttwoaspectsof childhoodbilingualism.First,itsparkedarenewedinterestinthestudyof childhoodbilingualismamongpsychologistsandeducators.Second(andperhaps evenmoreimportant),itprovidedoneofthemajorjustificationsfortheestab lishmentofbilingualeducationprogramsduringthelate1960sandearly1970s, especiallyinCanadaandtheUnitedStates. Thenumber'ofstudiesdealingwithchildhoodbilingualismincreaseddramat icallythroughouttherestofthe1960sand1970s.Mostofthisresearchconcen tratedoncognitivedevelopment,basicallyreplicatingtheresultsofPealand Lamberteitherwithdifferentmeasuresofcognitiveperformanceorwithdiffer entsamplesofbilingualchildren.Afewstudieslookedatthesocialandpersonal aspectsofgrowingupwithtwolanguages.Yetanothersetofstudiesconsidered 3

thesocialphenomenacloselyrelatedtobilingualismbiculturalismand bidialectismandtheroletheyplayinthedevelopmentofthechild. By1982,researchintochildhoodbilingualismhadproliferatedtosuchan extentthatamajoreffortwasnecessarytobringtogethertheavailabledataon childhoodbilingualismandprovidesometheoreticalframeworkwithinwhichto understandthem.OnJune2122,1982,astudygroupwasheldatNewYork Universityentitled"ChildhoodBilingualism:AspectsofCognitive,Social,and EmotionalDevelopment."SponsoredundertheauspicesoftheSocietyforRe searchinChildDevelopmentandorganizedbyPeterHomelandMichaelPalij withthehelpofDorisAaronson,theaimsofthisstudygroupwere(a)to summarizethecurrentworkonbilingualismandmakeitaccessibletomain streamdevelopmentalpsychologists;and(b)toprovideresearchersinboththe bilingualandthemonolingualresearchareasanopportunitytodevelopaninte gratedmodelofthedevelopmentalprocessesoperatinginthebilingualchild. Thestructureofthestudygroupwasspecificallydesignedtoachievethese ends.Researchersinbilingualismfromanumberofdisciplines,includingpsy chology,education,andlinguistics,wereinvitedtodeliverpapersreviewing specificaspectsofchildhoodbilingualism.Thepaperswereorganizedintosec tionscoveringthefollowingareasofinterest:languageacquisition,cognitive development,socialandemotionaldevelopment,andtherelationshipofbi culturalismandbidialectismtobilingualism.Eachsectionalsoincludeda "monolingual"discussantaresearcherintheparticulararea(e.g.,language acquisition)whoseownworkhadbeendoneprimarilywithmonolingualchil dren.Thisstructureencourageddiscussionanddialoguenotonlyamongscien tistsfromvariousareasofbilingualresearch,butalsobetweenbilingualand monolingualresearcherslookingatsimilaraspectsofchilddevelopment. Itwasnotthepurposeoftheconferencetoevaluateexistinggovernmental policiesaboutbilingualeducationnortomakerecommendationsforchanging suchpolicies.Rather,theconferencewasintendedtoprovideanimpartialsum maryandsynthesisoftheresearchinchildhoodbilingualismandbilingualism's effectondevelopment.Ultimately,however,itwashopedthatprovidingsucha compilationofinformationaboutchildhoodbilingualismwouldprovetobeof benefittothoseinvolvedmakingpolicydecisionsconcerningbilingualismand bilingualeducation. ThepresentvolumeistheendresultoftheSRCDstudygrouponchildhood bilingualism.Itisintendedassomethingmorethanarecordoftheproceedings ofpapersandpresentationsgivenduringthetwodaysduringwhichthestudy groupmet.Inpreparingtheirmanuscriptsforthisbook,theoriginalparticipants inthestudygroupwereencouragedtorevisetheiroriginalpresentationsinlight ofcommentsordiscussionsthataroseduringthecourseofthestudygroup,as wellastoaddresspointsofconvergenceordivergencetheysawbetweentheir ownpresentationsandthoseoftheothers.Theresultisafargreaterdegreeof integrationamongthevariouspapersthanwouldhavebeenpossibleinapro ceedingstypevolume. 4

Thebookisdividedintoseveraltopicareas:(a)languageacquisitionand processing;(b)cognitivefunctioning,style,anddevelopment;(c)socialand emotionaldevelopment;and(d)bidialectismandbilingualism.Followingthe structureoftheconference(atwhichmostofthesepaperswereoriginallypre sented),eachtopicareahastwoorthreechapterswrittenbyresearchersin bilingualismandadiscussionchapterbyaresearcherwhosemainworkhasbeen inamonolingualcontext.Thefollowingisabriefoverviewofthesechapters. INTRODUCTION TheaccompanyingchapterintheIntroductorysectionisbyPeterHomeland MichaelPalijanditprovidesasocialandhistoricaldescriptionofbilingualism andlanguagepolicyinfourcountries:Canada,theSovietUnion,theUnited StatesandthePeople'sRepublicofChina.Intheirconcludingsection,Homel andPalijdiscussthefutureofbilingualismandlinguisticdiversityineachcoun try,aswellassomeofthepossiblepsychologicalrelationshipsbetweenchild hoodbilingualismandthesocialcontextinwhichitoccurs. LANGUAGEACQUISITIONANDPROCESSING ThefirstchapterinthissectionbyKenjiHakutafocusesontheprocessesin volvedintheacquisitionofasecondlanguageandhowtheseprocessescontrast withthoseinvolvedinfirstlanguageacquisition.Hakutaexaminesthesepro cessesandhowtheyareaffectedbysuchfactorsascognitivematurity,similarity inlinguisticstructureofthefirstandsecondlanguage,transferfromthefirst languagetothesecond,andageeffects.Heconcludesbyarguingthatthebest waytoguidefutureresearchinfirstandsecondlanguageacquisitionistoadopta conceptualframeworkthatidentifieslanguageuniversalsandtypologies(i.e., categoricalmembershipfeaturesthatidentifyhowonelanguagesystematically differsfromanother).Withinthisframework,researchonsecondlanguageac quisitionisseentobecomplementarytoresearchonfirstlanguageacquisition insteadofbeingseparatefromortangentialtoit. Next,JamesCumminsexaminestheinterrelationshipsamongbilingualism, linguisticproficiency,andmetalinguisticawareness.AccordingtoCummins,it iseasytomisperceivethesefactorsasbeingcategorical(saying,forexample, thatachildiseitherbilingualornot),therebyglossingoverthefactthatthese factorsconstitutecontinuathattheperformanceofindividualchildrenmay varyconsiderablyalonganyoneofthesefactors.Cumminsprovidesatwo dimensionalschemeforunderstandingtheinterrelationshipamongthesethree factors:onedimensionreflectsthedegreetowhichthereis"contextual"support forunderstandingacommunication(bycontextCumminsmeansthesociocul 5

turalsettinginwhichthecommunicationisbeingmade);theseconddimension reflectsthedegreeofcognitiveinvolvementforthetasktobeperformed. Thefirstdimensionrangesfromoneextreme,whichcanbereferredtoas contextembeddedwhereacommunicationisembeddedinanappropriatesit uation,acontextinwhichtounderstandthecommunicationtotheotherex tremeofbeingcontextreduced,wherethereareveryfewcontextualaidsin interpretingthecommunication.Thedimensionofcognitiveinvolvementranges fromthosetasksthatrequirelittlecognitiveprocessingtothosethatarevery demandinginprocessingdemands.ThismodelallowsCumminstocharacterize anumberofdifferentstudiesonbilingualproficiencyandmetalinguistic awareness. ThechapterbyAaronsonandFerresexaminessomeofthedifferencesthey havefoundinEnglishlanguageprocessingbynativeEnglishspeakersandChi neseEnglishbilinguals.Strikingdifferencesbetweenthetwogroupsseemtobe directlyattributabletodifferencesinthestructuresoftheEnglishandChinese languages.Oneofthemostintriguingconclusionsdrawnfromtheseresultsis thattraditionalgrammaticalcategoriesfoundinEnglishmaynothaveexact counterpartsinChinese.Differencesinlinguisticperformanceappeartoberelat edtothebilingual'sknowledgeandexperiencewiththedifferencesinboth languages,especiallywhenthelanguagesderivefromdifferentlanguagefam ilies. MartinBraineprovidesadiscussionofthesethreechaptersfocusingonthe implicationsofeachfortheorybuildingandfutureresearchbymonolingual researchersinlanguageacquisitionandprocessing. BILINGUALISMANDCOGNITIVEDEVELOPMENT PalijandHomelexaminethequestionofhowbilingualismaffectscognitive developmentinChapter7.Thischapterisdividedintothreesubsections:(a)a historicalreviewofstudiesrelatingcognitivedevelopmentandprocessingto bilingualism,(b)theoreticalissuesinvolvedindirectingresearchinthisarea,and (c)methodoligicalproblemswithpastandpresentstudiesandtheuseofcontem porarystatisticaltechniquesinconstructingmorecomprehensiveandvalid models. ThenextchapterbyEdwardDeAvilaexamineshowintelligenceandcog nitivestyle,interestandmotivation,andeducationalopportunityandaccessall interacttoinfluenceschoolbehaviors.DeAvilaarguesthatthepooracademic performanceseeninmanyschoolsituationsisnotdirectlyrelatedtostudents beingbilingual,orevendirectlyrelatedtootherfactorsthatarerelatedtobeing bilingual.Instead,itistheinteractionofthethreepreviouslystatedfactorsthat givesrisetothepooracademicperformance.Forschoolperformancetoim prove,thesefactorsmustbefacedandeffectivelydealtwith.DeAvilareviewsa

studythatclearlyidentifiesthesefactorsandsuggestsonemeansofimprovingstudents'school performance. JosephGlickconcludesthissectionwithadiscussionofthechaptersbyPalijandHomel,and DeAvila.Heraisesissuesregardingtheroleoftraditionalgoalsineducationandhowthemethods forimplementingthemoftenoverlookthespecificneedsofstudents.Thisbecomesparticularly importantintheconsiderationofclassroomgoalsandperformanceofstudentsfromdifferent socioculturalbackgroundsandethnolinguisticgroups. BILINGUALISMANDSOCIALDEVELOPMENT Inthissection,Chapter10byDonaldTaylorfocusesonsocialpsychologicalfactorsthatpromoteor inhibittheacquisitionofasecondlanguage.Taylorstressestheimportanceofintergrouprelations, thesocioculturalgoalsofeachgroup,andhowthesefactorsinfluencetheacquisitionofsecond languagebychildrenfromdifferentgroups.Hedescribesapossiblemodelfordepictingsuch intergroupsituations:a2X2classificationschemewhereonedimensionreflectseitherpositiveor negativerelationsamonggroups,andtheotherdimensionreflectswhetheragroupdesiresto maintainitsowncultureandlanguage.Taylordetailsthesocialandpsychologicalconsequencesthat followfromeachofthesepossibleconditionswithinthisschemeandprovidesexamplesfrom contemporarysociety. WallaceLambert'schapterisconcernedwithhowexperiencesinbilingualandbiculturalsettings affecttheattitudesandperspectivesofthedevelopingchild.Hegoesontoshowhowtheseattitudes andperspectivestheninfluencelanguagelearningandthedevelopmentofbilingualism.Lambert presentsexamplesofsomeofthehistoricalforcesthathaveaffectedbothsocialattitudesand researchinCanada.HealsoreviewsthefindingsoftheCanadianlanguageimmersionprogramsand howattitudesandlanguagelearningwereaffectedwithinthem. E.ToryHigginsprovidesthediscussionforthissection.Hetiestogetherthethreadscommontothe severalchaptersandindicateshownewresearchontheroleofsocialcognitionmayprovide additionalinsightsintotherelationshipbetweensocialrealityandcognitivefunctioning. BIDIALECTISM Althoughthedistinctionbetweenwhatconstituteslinguisticvariationandwhatconstitutesdialectal variationmaybecontestable,bilingualismitselfmaybedescribedaslanguagevariationatthe interlanguagelevelandbidialectismasthestudyoflanguagevariationattheintralanguagelevel. Dialectsrepresentsystem 7

aticandcoherentlinguisticsystemsthatoperatewithinalargermonolingualframework.For example,BlackEnglishhasspecificfeaturesthatidentifyitasabonafidelanguagesystemthat alsousesmanyStandardEnglishgrammaticalformsandwords. WilliamHallandWilliamNagyexaminehowdifferencesincommunicationpatternsbetween blackandwhitechildrencanbeattributedtodifferencesinthechildren'sculturalbackground. HallandNagyreportthatblackchildrenusestatewordslike"think,""know,""happen,""see," and"want"muchlessoftenintheirclassroomsthantheydoathome,wherethelevelofusageof suchwordsiscomparabletothehomeusagelevelbywhitechildren.Apparentlyblackchildren experiencesomesortofdiscontinuitybetweentheirhomeandschoolenvironmentsthatresults inreducedusageofstatewordsinschool.Thisfindingcontradictsthenotionthatblackchildren comefromlinguisticallydeprivedbackgroundsandindicatesthattheproblemsthatblack childrenencounterinschoolmaybeduetofactorsthatarefarmoresubtlethanhasbeen previousconsidered. Next,JohnRoy,inhischapter,reviewsthedevelopmentofBlackEnglishandcontrastsits developmentwiththatofbilingualismbyimmigrantgroupswhohadnotforciblybeenbroughtto Americanshores.HebeginswiththedevelopmentofBlackCreoleanddiscusseshowvarious socialprocessescausedittogiverisetothemorefamiliarBlackEnglishofcontemporarytimes. Hepointsoutthatthispatternofdevelopmentapparentlydiffersfromthatofotherdialects, particularlyregionaldialects.BlackEnglishrepresentsaconvergencetowardStandardEnglish fromBlackCreole,whereasotherdialectsusuallyrepresentadivergencefromStandardEnglish totheirpresentform.Royconcludesbyexaminingthefactorsthatmakeitimportantforteachers ofEnglishtobesensitivetothedialectalbackgroundoftheirstudents. WilliamStewartprovidesthediscussionchapterforthislastsectionanddescribessomeofthe linguisticandpsychologicalimplicationsofcrossdialectalcommunication. CONCLUDINGREMARKS Thisvolumeisintendedtoserveadualfunction.Ononehand,forthoseunfamiliarwith bilingualresearch,itprovidesacomprehensivesummaryofpastworkinthisarea.Wefeelthat therearemanyaspectsofbilingualresearchthatcancastlightonresearchdoneinotherareasof developmentalpsychology,andviceversa.Forthosefamiliarwithbilingualresearch,thisbook shouldserveaheuristicfunction,providingasourceofideasforfutureinvestigation.Manyof thechapterspresentedherehighlighttheneedtotakeintoaccountthemediatingroleofsocial andculturalfactors;othersdescribepossibleresearchdesignsandstatisticalproceduresthat mightbeusedtohandlesuchmultivariatesituations. 8

Wehopethatthisbookwillstimulatefurtherresearchintothecomplexrelationshipbetween bilingualismandpsychologicaldevelopmentandprovideamorecomprehensiveviewofthe linguistic,cognitive,social,andemotionalprocessesinvolvedinthedevelopmentofthebilingual child. REFERENCES Peal,E.,&Lambert,W.E.(1962)."Therelationofbilingualismtointelligence".Psychological Monographs,76,123(No.546). 9

[Thispageintentionallyleftblank.] 10

BilingualismandLanguagePolicy:FourCaseStudies PeterHomelMichaelPalijNewYorkUniversity Inthischapter,weexaminethelanguagepoliciesoffourcountries:Canada,theSovietUnion,the UnitedStates,andthePeople'sRepublicofChina.Inparticular,wetrytoindicatethedifferent perspectivethateachofthesecountrieshastakenwithregardtolinguisticdiversityandbilingualism andhowthisisreflectedinthemannerinwhicheachcountryapproachesbilingualeducation. Wefirstpresentageneraloverviewofeachcountry,includingadescriptionofthegenerallinguistic andethniccompositionofthecountry,aswellassomeofthepastandpresenttrendsinpolicyofthe particularcountrytowardminoritylanguagesandbilingualism.Wethendiscusssomeofthe implicationscertainsocialpoliciesmayhaveforthepsychologicaldevelopmentofbilingual children. CANADA Canadaisofficiallyabilingualcountry,withEnglishandFrenchenjoyingequalstatusasthe languagesofgovernment.Ofatotalpopulationofapproximately24millionin1976,67%ofall CanadiansreportedEnglishastheirfirstlanguageand26%reportedFrench(Beaujot& McQuillian,1982).TheFrenchspeakersareconcentratedprimarilyintheprovincesofQuebec (87%ofthepopulationoftheprovince)andNewBrunswick(34%).InadditiontoEnglishand French,programsforthemaintenanceoflanguagesspokenbyNativeIndiangroupsandthe Inuktitut(Eskimo),aswellasthosespokenbymajorimmigrantgroups(German,Italian, Hungarian,andUkrainian),arealsosupportedbytheCanadiangovernment. ThetotalrateofbilingualisminCanadais13%.Thebreakdownis33%for 11

FrenchCanadians;8%forEnglishCanadians.Onecauseofsuchdifferentratesofbilingualism amongFrenchandEnglishspeakersappearstobethegeographicaldistributionofbilingualism. Approximately57%ofallbilingualsinCanadaliveinQuebecprovince.Infact,35%ofall bilingualsinCanadaliveintheMontrealarea(Beaujot&McQuillian,1982). TheEnglishandFrencheachestablishedcoloniesinCanadainthe17thcentury.Bythemid18th century,thenumberofEnglishsettlershadincreasedenormouslyascomparedwiththeFrench.In 1763,afterdefeatbyEnglandintheSevenYears'War,Francewasforcedtocedeallofher territoriesinCanadatotheBritish. Overtheyears,anumberofofficialconcessionsweremadetoFrenchspeakingCanadians.The QuebecActof1774recognizedQuebecasaFrenchspeakingareaandallowedtheFrenchthereto maintaintheirownreligiousandpublicinstitutions.Inparticular,theCatholicChurchremainedin controloftheeducationalsysteminQuebec.Theseconcessionsweremaintainedunderthe ConfederationActof1867,whichalsogaveallCanadianstherighttopoliticalparticipationatboth 1 thenationalandprovinciallevels. Unfortunately,theofficialrightsaccordedtheFrenchfailedtooffsetwidespreadsocialand economicdiscriminationthattheyexperiencedfromtheEnglishspeakingmajority(Whitaker,1984 ).OneofthemajormeansofcontrolovertheFrenchwastheuseofEnglishinalmostallaspectsof government,commerce,andhighereducation.Addedtothiswasthegenerallyconservativeroleof theCatholicChurchinFrenchCanadiansociety,encouragingthepassiveacceptanceofthestatus quoamongtheFrench. EveninQuebec,withitsmajorityofFrenchspeakers,theEnglishspeakingcommunitystill succeededinmaintainingpoliticalcontrolatboththelocalandtheprovinciallevelsbymeansof theireconomicpower.Theyownedmostofthebusinessesandfactoriesandtendedtoshowfavor eithertoothernativeEnglishspeakersortothoseFrenchwhowerefairlywellassimilatedintothe Englishspeakingculture. Bythe1960s,however,therewasagrowingmovementinthemajorFrenchspeakingareascalling fortheCanadiangovernmenttoshowagreaterrecognitionofthelinguisticandpoliticalrightsof Frenchspeakers.ItwasinresponsetothisthataRoyalCommissiononBilingualismand Biculturalismwasconvenedbetween1965and1968tolookintotheseproblems.Onthebasisofthe suggestionsofthecommission,theOfficialLanguagesActwasadoptedin1969,declaringthat EnglishandFrenchweretobetheofficiallanguagesofCanadaandthattheypossessequalstatusin termsoftheiruseinallaspectsofgovernment. Theprimarypurposeoftheactwastoencouragebilingualismataninstitu ____________________ 1Kaalt(1977)hasalsosuggestedthat,indeferencetotheFrenchCanadianswhowouldhaveseen suchanattemptasthefirststepinitsownassimilation,attemptstoassimilateothernonEnglish speakinggroupswhocameasimmigrantswerenotasstronginCanadaastheywereintheU.S. 12

tionallevelinanefforttoprovideequalsocialandgovernmentalservicesforbothEnglishand Frenchspeakers.Coupledwiththis,however,wereofficialeffortsinsupportofeducational opportunitiesforminoritystudents,aswellastheestablishmentofprogramsofbilingual educationandsecondlanguageinstructionforbothFrenchandEnglishspeakers. AccordingtoGrosjean(1982),theresultsoftheOfficialLanguagesActappeartobe encouraging.Forexample,censusresultsindicatethatagrowingnumberofNativeEnglish speakersespeciallyinQuebecarelearningandusingFrench.Ontheotherhand,however,about athirdoftheFrenchchildrenoutsideofQuebecprovinceapparentlystilldonotreceive instructionintheirnativelanguage. OneofthefearsexpressedamongFrenchCanadiansisthatbecausetheyconstituteaminority groupwithinCanada,thegeneralencouragementofbilingualismmightresultinanincreased tendencytowardassimilationofFrenchCanadianspeakersintothedominantEnglishspeaking community.ThiswouldcompoundthelossofFrenchspeakersthatisalreadytakingplacein QuebecasaresultofthelowbirthrateamongtheFrench,themigrationofFrenchspeakersto other,nonFrenchspeakingareasofCanada,thepreferenceofnewimmigrantssettlingin QuebectolearnEnglishratherthanFrench,amongotherfactors(Beaujot&McQuillian,1982). Thelate1960sandearly1970switnessedthegrowthofaseparatistmovementamongtheFrench inQuebec.ThePartiQuebecoiscameintopowerinQuebecprovincewithaplatformcallingfor thepreservationofaFrenchspeakingQuebecandagreaterdegreeofautonomyoftheprovince fromtherestofEnglishspeakingCanada.In1977,undertheadministrationoftheParti Quebecois,anactwaspassedmakingFrenchtheonlyofficiallanguageoftheprovince. BusinesseswererequiredtoadoptFrenchasthelanguageofeverydayaffairs,childrenof immigrantswerecompelledtolearnFrenchinschools,andchildrenofEnglishspeakingparents wereallowedtobetaughtinEnglishonlyiftheirparentscouldprovethattheythemselveshad beentaughtinEnglishintheprovince. ThefederalgovernmentofCanadagenerallymaintainedapolicythatsoughttocounterthe separatistmovementamongtheFrenchinQuebecwhileatthesametimecontinuingto encourageabilingualCanada.In1982,anewCanadianconstitutionwasproclaimedthat essentiallycontainedalltheprovisionsoftheoriginalActofConfederationof1867,aswellasall oftheamendmentsthathadbeenmadetoitovertheyears.Inaddition,itcontainedcertain proposalsthathadbeenworkedoutwiththeleadersofthevariousprovinces. AmongthesewasaCharterofRightsandFreedomsthatcontainedaprovisiongivingparentsthe righttochooseeitherEnglishorFrenchasthelanguageofinstructionfortheirchildreninany provincewherethenumberswarrantedit.ThischarterservedasthebasisfortheCanadian SupremeCourt'sdecisionin1984tostrikedowntheprovisionoftheQuebecActthatrestricted English 13

instructiontochildrenofparentswhohadbeeneducatedinEnglishinQuebec. Ontheotherhand,theCanadiangovernment'sactionswithregardtotheprovinceofManitoba seemedtounderscorethegovernment'sdedicationtothecauseofbilingualism.Although ManitobahadenteredtheCanadianConfederationasabilingualprovincein1870,theprovincial governmentsubsequentlyrescindedthelanguagerightsofitsFrenchspeakingminority.In1979, theCanadianSupremeCourtorderedthattheserightsberestored.After4yearsofdelayby Manitoba,abillwaspassedintheCanadianParliamentinsupportoftheoriginalcourtdecision torestorebilingualisminManitoba. Insummary,Canadaappearstohaveestablishedalongtermcommitmenttoencouragingand maintainingbilingualismatboththenationalandtheprovinciallevels.Practically,thismaybe viewedasanefforttoensurethatbothFrenchandEnglishspeakingCanadiansenjoyequalaccess tosocialservices,business,andeducation.Moreimportant,however,thismaybeviewedasa solutiontothegeneralproblemofreconcilingthedemandsofnationalunitywiththeneedsofits multilingualmulticulturalsociety. Oneindicationofthesuccessofthispolicymaybethefactthatethnicpolarizationhasbecome lessofanissueinrecentyears.ThisnotionhassomesupportintheapparentdeclineoftheParti Quebecoisduringtheearly1980s,whichculminatedinthedefeatofthepartyintheprovincial electionsof1985.Ontheotherhand,Whitaker(1984)hassuggestedthatthedeclineoftheParti Quebecoismayhavebeentheunintentionalresultofitsownefforts.Becauseofitslanguage programs,itmayhavesucceededinstrengtheningtheFrenchidentityofQuebec,thusrelieving theveryanxietythathadoriginallycompelledFrenchspeakerstosupportthePartiQuebecois.In anycase,bilingualisminCanadaappearstobeakeyelementinitsnationalpolicy,andthereare noindicationsatthistimeofanymovementawayfromthatposition. THESOVIETUNION TheSovietUnion(theUnionofSovietSocialistRepublics),withapopulationofapproximately 262millionpeople,isaconstitutionalfederationconsistingof15memberrepublics.Thelargest istheRussianSocialistFederalistSovietRepublic(RSFSR),whichservesasthecenterofthe Sovietgovernment.Theremainingrepublicsarereferredtoas"national"republics.They correspondmoreorlesstothetraditionalhomelandsofthemajornonRussiannationalorethnic 2 groupsthatmakeupatleasthalfthepopulationoftheSovietUnion. ____________________ 2Thepopularpracticeofusingtheterm"Russia"torefertotheSovietUnion(whicheven Sovietsarepronetodo)thusrepresentsafailuretoappreciatetheextentofethnicand linguisticdiversityinthatnation.Itparallelstheuseoftheterm"America"inreferringtothe UnitedStatessomethingthathasoftenbeencriticizedbytheotherpeoplesofbothNorthand SouthAmerica. 14

Therearesome130distinctlanguagesspokenwithinthebordersoftheSovietUnion(Comrie, 1981;Isayev,1977).SomeideaoftheextentoflinguisticdiversityintheSovietUnioncanbe gottenfromthefollowinglistofmajorlanguagefamiliesspokenintheSovietUnion,aswellas someprominentexamplesofeachfamily: 1. IndoEuropean,includingtheSlaviclanguages(Russian,Ukrainian,andByelorussian), Balticlanguages(Lithuanian,Latvian),Iranianlanguages(e.g.,TadjikandKurd), Moldavian(aromancelanguagesimilartoRomanian),Yiddish,andArmenian. 2. Altaic,includingtheTurkiclanguages(e.g.,Uzbek,Kazakh,Azerbaijani),Mongolic languages(BuryatandKalmyk),andtheTungusManchulanguages. 3. Uralic,includingtheFinnoUgariclanguages(e.g.,Estonian,Karelian,andMordovian)and theSamodicgroup. 4. IberianCaucasic,includingtheKartvelianlanguages(e.g.,Georgian),alongwiththe AbkhazAdyghe,theNakh,andtheDaghestanilanguages. 5. PaleoAsiatic,includingtheChukchiKachatdalandEskimolanguages. IntheSovietcensus,adistinctionismadebetweennationalorethnicidentityandnative language(NarodnoeKhozjajstvoSSSR19221982,1982).Forexample,Russianisthedeclared languageof58.6%ofthetotalpopulationoftheSovietUnion.Thisgroupcanbefurtherdivided intoethnicRussianswholivewithinthebordersoftheRSFSR(approximately114million accordingtothe1979census);ethnicRussianswholiveintheotherrepublicsoftheSoviet Union(24million);andnonethnicRussianswhodeclareRussianastheirnativelanguage(13 million,or5%ofthetotalpopulation). Othermajorlanguages(presentedinorderofpercentageofspeakers)spokenintheSovietUnion include:Ukrainian(14%),Uzbek(4%),Byelorussian(3%),Kazakh(2.2%),Tatar(2.2%), Azerbaijani(1.8%),Armenian(1.3%),Georgian(1.3%),Lithuanian(1.1%),Moldavian(0.91%), Tadjik(0.86%),Chuvash(0.61%),Latvian(0.58%),Kirghiz(0.58%),andEstonian(0.37%), Bilingualsmakeup21.5%ofthetotalpopulationoftheSovietUnion(Comrie,1981).Among nonRussianstherateis42.6%;amongRussians,itis3.1%.EvenamongthoseRussianslivingin nonRussianareas,therateofbilingualismstilltendstobefarlowerthanthatforthenon Russiansintheparticulararea. ThelinguisticandethnicdiversityoftheSovietUnionisadirectresultofitsprerevolutionary past.TheRussianempirewasformedastheresultofaseriesofmilitaryconquestsbetweenthe 16thand19thcenturies.Duringthecourseofthisperiod,whathadbegunastherelativelysmall principalityofMoscowSuzdalexpandedwestasfarasPolandandtheCarpathianmountains, southasfarastheBlackSeaandtheCaucasusMountains,northasfarasFinlandandtheArctic, andeastasfarasChinaandthePacificOcean. 15

ThelanguagepolicieswhichthetsaristgovernmentadoptedtowardindividualnonRussian minoritiesintheseconqueredareaswerebasedprimarilyonpoliticalconsiderationsspecifictoeach group.ArelativelyliberalapproachtolanguagepolicywasadoptedforEstoniaandFinland.Atthe timeoftheirannexationbytheRussianempire,theseareasalreadyhadhighlevelsofcultureand industrycomparabletothoseinwesternEuropeancountries.Moreover,bothregionswerealready underthepoliticalandeconomicdominationofnonindigenousminorityethnicgroupsethnic SwedescomposedtherulingeliteinFinland,withGermansholdingpowerinEstonia.These minoritiesmaintainedtheirdominantstatusbyservingasoverseersandgovernmentofficialsforthe Russians.TheRussiangovernmentaccordedtheseminoritygroupsalimiteddegreeofpoliticaland linguisticfreedominreturnfortheirloyalty. ThecaseoftheUkrainiansrepresentsamoreextremepolicy.TodiscourageUkrainianseparatism, animperialdecreewasissuedin1876banningthepublicuseofUkrainian(Savchenko,1970).It prohibited,amongotherthings,theteachingofUkrainianinschools,thepublicationoforiginal worksandtranslationsinUkrainian,andthepublicperformanceofplaysandsongsinUkrainian. ThegeneralbanagainsttheuseofUkrainianremainedineffectuntiltheRevolutionof1905. FortherestofthenonRussianlanguageswithintheRussianempire,discouragementwascarriedon throughapolicyofsocialneglectratherthanofrestrictions.Inthoseregions,noschoolingwas allowedinthenativelanguageasidefromthatconnectedwithreligioustrainingormissionarywork. MoreablenonRussianstudentswereencouragedtolearnRussianandtoassimilateintoRussian cultureinordertosucceedingovernmentorbusiness. Despitetheeffortsofthetsaristgovernment,nationalisticmovementssteadilydevelopedwithin severalethnicgroupsduringthe19thcentury.Asaresult,anumberofthesegroupsattemptedto formtheirownindependent(and,insomecases,socialist)governmentsduringtheperiodofthe BolshevikRevolution.However,facedwiththepotentiallossofrawmaterialsandfoodfrommany oftheseareas,theBolshevikstookimmediatemilitaryactionagainstthesenationalistgroups.At thesametime,theBolsheviksadoptedpoliciesdesignedtogainfavoramongthenonRussian minorities. Thus,oneoftheaimsofthenewSovietstate,asoutlinedbyLeninin1917,becamethefulland equaldevelopmentofallethnicandlinguisticminoritiesintheSovietUnion(Comrie,1981; Kreindler,1982).TherewastobenoofficiallanguageforthenewSovietstate.AllSovietcitizens weretohavecompletefreedomtousetheirnativelanguagesinprivateandinpublic.Publicusage includedtherighttouseone'snativelanguageforaddressingpublicmeetings,correspondingwith thegovernment,andgivingtestimonyincourt.Moreover,allSovietcitizenswereguaranteedthe righttoreceiveaneducationinandtohaveaccesstoliteratureandculturalmaterialsintheirnative languages. 16

Inaddition,theestablishmentofasystemofnationalrepublicsgavethemajorethnicgroupsat leastsomedegreeofautonomyandselfgovernment,althoughprimarypoweralwayslaywiththe centralgovernmentinMoscow.Eachnationalrepublicwasallowedtouseitsnativelanguageas theofficiallanguageofgovernment.Eachnationalrepublicwasgivencontroloverlocalaspects ofeducationalpolicyandoverthedevelopmentofthenationalculture. The1920switnessedanactivecampaignaimedatencouragingthedevelopmentofnonRussian languagesandethniccultures.Forlanguageswithanalreadyexistingwrittenlanguageand literarytradition,programswerecreatedforsettingupschools,publishingnewspapersand books,andsoon.Forthoselanguageswithoutawritingsystemorforwhichthealreadyexisting systemsofwritingwerecumbersomeandunsuitedtoeasyacquisition,linguistsweresentto studythelanguagesanddevelopwritingsystemsforthem. UnderStalin'sleadershipinthe1930s,however,ageneralretrenchmentoccurredinSovietpolicy withincreasinglygreateremphasisbeingplacedontheneedtounifythenationanddevelop centralizedcontrolofthestate.ThisledtoagrowingencouragementoftheuseofRussianasa commonlanguageforcommunicationamongmembersofdifferentethnicgroups. ManyofthenonRussianwriters,intellectuals,andscientistswhocameintoprominenceduring thepreviousdecadeeitherperishedduringthepurgesofthe1930sorwereforcedtoconformto moreideologicallyacceptable(andlessnationalistic)topicsofwritingorresearch(Luckyj, 1975;Simirenko,1969).Yetanotherexampleofretrenchmentwasthediscontinuanceofthe publicationofmaterialsinsomelanguages(e.g.,Lapp,Karelian)ontheexcusethattheyhad veryfewspeakers,mostofwhomwerebilingualinother,moreprevalentlanguages. AsaresultoftheemphasisonnationalunityduringtheSecondWorldWar,the1940ssaw increasesinthegreaterprominencegiventoRussianthantothenonRussianlanguages.This trendwasexemplifiedbyStalin'svictorytoastof1945,whereinhepubliclyreferredtoRussiaas thenationthatservedasthe"leadingforceoftheSovietUnion"(Bilinsky,1964).Russianwas proclaimedasthelanguageofhighculture,aswellasthelanguageofsocialism.Asaresult, linguisticreformswereeffectedinvariousnonRussianlanguagestobringthemcloserin appearancetoRussian.Thesereformsconsistedprimarilyofchangesingrammarand orthographybasedonRussianpatterns,andtheintroductionofmanyRussianloanwordsor calques,whichoftenreplacedwordsthathadalreadybeenwellestablishedintheparticular language. AfterStalin'sdeathin1953,therewasaperiodofrelaxationintheSovietUnionthatlasteduntil theresignationofKhruschevin1964.Anofficialacknowledgmentwasmadeofthecontribution ofthenonRussiannationalitiestothelifeandcultureoftheSovietUnion:Thedevelopmentof aninternationalculturewasnottoresultinthelevelinganddisappearanceofnationaltraditions. 17

NonRussianminoritiesweregrantedgreaterconcessionsineducationalpolicyandinliterary andintellectualfreedomofexpression. Sincethelate1960s,however,theSovietgovernmentappearstohavereturnedtoatacitpolicy encouragingthestatusofRussianastheofficiallanguageoftheSovietUnion.Itisconsideredto betheonlylanguagewiththestatusofalinguafrancawithintheSovietUnionandistheonly languagethatcanbeusedincommunicationwithindividualsfromothernations. Therehasalsobeenageneralacceptanceoftheeventualconsolidationofsmallerethnicgroups andlanguagesintolargerones,withtheeventualgoalofdevelopingaunifiedSovietPeoplewith acommonSovietlanguageRussian.Bilingualismappearstobeencouragedaspartofthis gradualincorporationofsmallerlanguagegroupsintolargerones.AccordingtoIsayev(1977): "...bilingualismshouldbeviewedasatransitionalstagetomonolingualismwhichwillbe reachedbythesmallerethnicgroupswhentheirassimilationintothecorrespondingnationsis complete"(pp.199200). EvidenceindicatesanactivepolicyofpromotingboththeassimilationofnonRussianminorities andtheincreaseduseofRussianvisvisthenonRussiannationallanguages.ThenonRussian republicsintheEuropeanportionoftheSovietUnionhaveshownsteadydecreasesinthe percentagesoftheirnativeethnicpopulationswithacorrespondingincreaseinthepercentagesof personsdeclaringthemselvesasbeingethnicallyRussian. Similarly,therehasbeenasteadydeclineinthenumberofcopiesandnumberoftitlesofbooks andpublicationsinnonRussianlanguages,withacorrespondingincreaseinthenumberof imprintsinRussian(Lewis,1972).Scientificandtechnicaljournalsthatwereformerly publishedinnonRussianlanguagesarenowpublishedinRussian,presumablytomakethem moreaccessibletoreadersbothwithinandoutsidetheSovietUnion. YetanothertrendisindicatedbythefactthatnonRussianparentshaveincreasinglybegunto sendtheirchildrentoRussianlanguageschoolsratherthantonativelanguageschoolsinorderto increasetheirchildren'schancesofsuccessinenteringinstitutionsofhighereducation(Comrie, 1981;Kreindler,1982).Eveninthenationallanguageschools,Russianisacompulsorysubject intheearlygrades,andeffortshavebeenunderwaytointroduceitasearlyaskindergartenand preschoolclasses. Inshort,itappearsthatbilingualismiscurrentlybeingviewedasatransitoryphenomenoninthe SovietUnionanecessarypartoftheprocessofassimilatingnonRussianminoritiesintoa RussianspeakingSovietnation.Partofthejustificationforthismaybedemographic.Brunner (1981)reportsthatbirthratesamongethnicRussians,aswellasamongthenonRussian nationalitiesintheEuropeanportionsoftheSovietUnion(Ukrainians,Byelorussians,Estonians, etc.),havebeendecliningconsiderablyoverthepast2decades.Overthesameperiod,birthrates amongtheMuslimTurkicnationalities(Uzbeks,Kazakhs,Azerbaijanis,etc.)haveincreased. 18

Onthebasisofhisdemographicdata,Brunner(1981)hasestimatedthat,atpresent,ethnic Russiansprobablyconstituteonly49%ofthetotalofalldraftagemalesandthatby1995they willconstituteonly46%.ConsideringtheSlavicgroupsasawhole,heestimatesthattheynow are67%ofalldraftagemalesandthatby1995thatfigurewilldeclineto62%.Ifthesefigures arereflectiveofgeneraltrendsintheSovietpopulation,ethnicRussiansmaysooncometo constituteaminoritygroupwithintheSovietUnion.Currenteffortstoinculcateamongthenon RussianminoritiesasenseofidentitywithaRussianspeakingSovietculturemaybeanattempt tooffsetsuchfuturedemographicchangesintheethnicRussianpopulation. InherreviewoflanguagepolicyintheSovietUnion,Kreindler(1982)commentedthatitis paradoxicalthatthecountrywhichwasoneofthefirsttochampiontherightsofminority populationstodevelopandmaintaintheirownlanguagesshouldnowbemovingawayfromthat originalpolicy.ShepointstothefurtherironythattheSovietUnionappearstobedoingsoatthe sametimethatothernationsseemtohaveawakenedtothelinguisticandeducationalneedsof theirownminoritygroups. THEUNITEDSTATES TheUnitedStatesisoftenthoughtofasalinguisticallyandethnicallyhomogenousnation. Althoughthereisnolegalbasisforit,Englishhascometoassumetheroleoftheofficial languageoftheUnitedStates.ThevastnumbersofnonEnglishspeakingimmigrantswhocame totheU.S.overthepast200plusyearspresumablywereallassimilatedintotheEnglish speakingmainstream,leavingbehindtheiroriginallanguagesandethnicties.Asingle "American"culturethusdevelopedfromtheamalgamationofmanyimmigrantgroupsintoone national"meltingpot."Thiscultureperhapsshowsinfluencesfromthemanydiversegroupsthat wentintoitsformation,butitstillconstitutesamonolithicentityoflanguageandculture. ThisnotionofthehomogeneityoflanguageandcultureintheU.S.cameundercriticalscrutiny andreappraisalduringthe1960sand1970s(Glazer&Moynihan,1970;Greer,1974;Novak, 1977).Asaresult,thecommonlyaccepted"meltingpot"imageoftheUnitedStateswasfound tobeonlypartiallycorrect.AlthoughtheU.S.isanationthatprimarilyusesEnglishasa languageofcommunication,andthereappearstobeadistinctU.S.culture,theU.S.isneither linguisticallynorculturallyhomogenous. TheHarvardEncyclopediaofAmericanEthnicGroups(Thernstrom,Orlov,&Handlin,1980) estimatesthatthereareatleast106distinctethnicgroupsintheU.S.andcautionsthatthisisa conservativeestimatebecauserelatedgroupswereoccasionallycollapsedintoasinglecategory (e.g.,theentryforAmericanIndiansactuallycomprisessome170differentgroups). 19

Moreevidencecomesfromthe1980censusdata(U.S.DepartmentofCommerce1980Census, 1983)whichreportsthatofatotalof210millionpeopleintheU.S.,25millionreportedspeaking alanguageotherthanEnglishathome.About7millionofthesenonEnglishspeakerswere undertheageof17.ThemajornonEnglishlanguagesappeartobe:Spanish(11.1million speakers);Italian(1.6million);German(1.59million);French(1.55million);andPolish(820 3 thousand). TheBilingualEducationActof1968providedfederalsupportfortheestablishmentandfunding ofbilingualeducationprogramsinthoseareaswherenonEnglishspeakingchildrenfromlow incomefamiliescomprisedalargeportionoftheschoolpopulation.Theoriginalintentionofthe actwasthatbilingualeducationshouldremedysituationswhereinnonEnglishspeakingchildren were,asaresultoftheirinabilitytospeakEnglish,deprivedoftheopportunityforanadequate education.ItwasexpectedthatEnglishwouldbeintroducedwiththechild'shomelanguageand itsuseexpandeduntilthechildcouldfunctionadequatelyinEnglish.Thenthechildwouldbe transferredtoaregular,Englishspeakingclassroom. However,becausebilingualprogramswereadministeredatthestateandlocallevels, interpretationofpolicywasoftensubjecttotheneedsanddesiresofthelocalcommunity.Actual implementationofbilingualeducationoftenvariedfromitsoriginalintent,sothatitisdifficult todescribebilingualeducationintheU.S.asaunitaryprogramorphenomenon.Someschool systemsonlyofferedEnglishasaSecondLanguage(ESL)programs,whichwereessentially remedialprogramsinwhichtheprimarylanguageofinstructionwasEnglish.Othersystems offeredinstructioninbothEnglishandthechildren'snativelanguagebutmadetheuseofthe nativelanguageonlytransitional,buildingupthechildren'scompetencesufficientlyforthemto attendregular,Englishspeakingclasses.Othersystemsdevelopedprogramsaimedat maintainingthechild'snativelanguageandatteachingEnglish.Childreninsuchprograms receivedinstructioninboththeirnativelanguageandEnglish,andoftenreceived,instruction aboutthehistoryandcultureoftheirparticularnationalorethnicgroupaswell. Effortstoconsolidatefederalpolicytowardbilingualeducationwerestimuatedbytwoevents thatoccurredin1974.InthecaseofLauv.Nichols,(1974),theSupremeCourtaffirmedthatthe failuretoprovideeducationalassistanceprogramstononEnglishspeakingchildrenwasin violationofTitleVIoftheCivilRightsAct,whichforbadediscriminationonthebasisofrace, color,ornationaloriginintheoperationofanyfederallyassistedprograms.Inresponseto ____________________ 3WHilethefigureof25millionasthetotalnumberofpeoplespeakingalanguageotherthan EnglishintheU.S.isquiteimpressive,eventhatmaybeanunderestimate.Fishman(1980b) haspointedoutanumberofproblemsinvolvedincollectingcensusdataandhassuggested thatthenumberofnonEnglishspeakersintheU.S.maybeashighas40million. 20

this,theDepartmentofHealth,Education,andWelfareissuedguidelinesforlocalschoolsystems recommendingbilingualeducationasaspecificsolutiontorectifyingthissituation. TheBilingualEducationActwasamendedbyCongressin1974tomakeitmoreexplicitinintent anddesign.Bilingualeducationwasdefinedas"instructiongivenin,andstudyof,Englishandto theextentnecessarytoallowachildtoprogresseffectivelythroughtheeducationsystem,thenative languageofthechildrenoflimitedEnglishspeakingability"(pp.186187).Theactstipulatedthat bilingualeducationshouldincludeinstructioninthechildren'sownlanguageinadditiontoEnglish andshouldofferinstructiondealingwithaspectsofthenativeculturalheritageofthechildrenas well.Finally,itallowedforbilingualinstructionforchildrenotherthanthosefromlowincome families.Congressrenewedthisactin1984;itremainedessentiallyunchangedexceptfora provisionallowingfor4%ofthefundsallocatedtolocaleducationalsystemsforbilingualeducation tobeusedfor"alternativeinstructionalmethods"oftheirownchoosingmethodsplacingmore stressonEnglishastheprimarylanguageofinstruction. Theconceptofbilingualeducationasameansofmaintainingthelinguisticandculturalidentitiesof minoritychildrenhasmetwithconsiderableoppositionfromvariousindividualsandgroupswho stillsupporttheideaofalinguisticallyandculturallyhomogenousU.S.Areflectionofthisopinion isseeninrecentattemptsbytheDepartmentofEducationtorescindmanyoftheregulationsforthe conductofbilingualeducationprograms,inparticular,theregulationsspecifyingbilingual educationastheonlyapproachtoprovidingadequateeducationfortheminoritylanguagechild (Fiske,1985;U.S.Departmentof'Education,1984).Someofthealternativessuggestedincludea returntomoretraditionalESLprograms(Keefe,1985). SuchattemptsbytheDepartmentofEducationareintendedtogivegreaterautonomytolocalschool systemstousewhatevermethodofinstructionisbesttoteachminoritylanguagestudentseither bilingualinstructionorsomemeansofremedialinstructionconductedinEnglish.Critics,however, havechargedthatsuchapolicyismerelyaploytoweakenandeventuallydoawaywithbilingual education(Garcia,1985;Howe,1985). Althoughquestionsabouttherelativeusefulnessofdifferentapproachestoeducatingchildrenof nonEnglishspeakingbackgroundappeartobeunderconsiderationintheU.S.government,the needforsuchprogramsappearstobemorecrucialthanever.Demographicprojectionsofchanges inthenumberofnonEnglishspeakingbackgroundindividualsintheU.S.indicatethatthis populationmayincreaseto34.7millionin1990and39.5millionbytheyear2000.Thenumberof personsinthiscategoryundertheageof14willincreaseto5.4millionby1990andto7.7million bytheyear2000.Thefuturedispositionanddevelopmentofthelattergroupwillbeheavily dependentonthedecisionsconcerningbilingualeducationnowbeingmadebytheU.S. government. 21

THEPEOPLE'SREPUBLICOFCHINA ThePeople'sRepublicofChina(PRC),withoveronebillionpeople,isthemostpopulousnation intheworld.TheofficiallanguagepolicyofthePRCissimilartothatoftheSovietUnion,anot surprisingsituation,consideringthecloserelationshiptheChinesecommunistsenjoyedwiththe SovietUnionatleastintheirearlyyears.Therightsofminoritynationalitiesareguaranteedby theconstitutionofthePRCadoptedin1954(Lehmann,1975).Discriminationagainstnational minoritiesandtheirlanguagesisprohibited,andallnationalminoritiesaregrantedfullrightsto useanddeveloptheirwrittenlanguages. ThelanguagesofthePRCarepredominatelySinoTibetan.Han,theofficialdesignationforthe Chineselanguage,isspokenby95%ofthepopulation.Hanitselfisdividedintoninemajor 4(spokeninBeijingandintheprovincesofGanshu,Shaanxi, sublanguagegroupings:Mandarin Sichuan,Hubei,andYunan),withabout665millionspeakers;Xiang(spokeninHunan province),with50million;Gan(spokenlargelyinJiangxiprovince),withabout25million; Huizhou(spokeninsouthernAnhuiprovince),with4million;Wu(spokeninShanghaiand Zhejiangprovince),with85million;Minnan(spokeninsouthernFujianprovince),with10 million;Minbei(spokeninnorthernFujianprovince),with12million;Hakka(spokenthroughout southernChina),with40million;andYue(orCantonese),with55million(Moser,1985). TheuseofthetermdialecttorefertodifferentsublanguagesofHanissomewhatmisleading;the differencesbetweenthesegroupingsmaybeasgreatasthosebetween,forexample,Spanishand French.Moreover,Mandarinitselfcanbefurthersubdividedintoatleastfivedifferentdialect groups,eachwithitsownsubdialects. ThedialectspokeninBeijingwasdesignatedtheofficiallanguageoftheRepublicofChinain 1911andwasretainedinthatfunctionbythecommunistregimewhenitcameintopowerin 1949.ItisusuallyreferredtoasPutonghua,orthecommonlanguage.Moreover,the pronunciationofChinesecharactersthatistaughtinschoolshasbeenstandardized,basedupon thepronunciationthatthesecharactershaveintheBeijingdialectarea.InadditiontoPutonghua, Pinyin,asystemfortranscribingChinesecharactersthatisbasedontheLatinalphabet,hasalso beenintroducedasadeviceforfacilitatingthelearningofstandardChinesecharacters.Children arefirsttaughttoreadinPinyinandthengraduallyintroducedtotheChinesecharacters themselves. AsidefromthedialectgroupsofHan,therearesome58otherminoritynationalitiesmakingup some6%ofthepopulationofthePRC(about56millionpeople).Themajorminoritylanguages are:Chuang,with7.8millionspeakers; ____________________ 4Theterm"Mandarin"isnolongerusedinthePRCasitsuseisassociatedwiththe prerevolutionaryera.Oftenitisreferredtoas"thecommonlanguage"orputonghua(seethe discussioninthetextfollowing). 22

Hui,with3.9million;Uighur,with3.9million;Miao,with2.6million;Yi,with3.3million; Tibetan,with2.8million;ManchuTungus,with2.4million;Mongol,with1.6million;Puyi, with1.3million;andKorean,with1.2million(Dreyer,1976). Autonomousnationaldistrictsexistinareaswithlargeconcentrationsofnationalminorities. Suchdistrictshavethesamedegreeofautonomyandpowerasthenationalrepublicsinthe SovietUnion.However,allgovernmentalinstitutionsinthosedistricts(e.g.,judicialcourts)are requiredtousethelocalminoritylanguage,andpublicdocumentsarewritteninbothPutonghua andthelocallanguage.Instructionisinthelocallanguageatboththeelementaryandmiddle schoollevels,andChinesefromtheHanspeakingareaswhoworkorliveinminoritylanguage areasarerequiredtolearnthelocallanguage(Mei,1984). LiketheSovietUnion,thePRChashadtocontendthroughoutitshistorywiththeproblemof reconcilingethnicandlinguisticpluralitywiththesocialandeconomicgoalsofacentralized communiststate.AlsoasintheSovietUnion,periodsofrelativeaccommodationtotheinterests ofminoritygroups(asindicatedforexample,bysupportforthesystemofautonomousnational districts)havealternatedwithmoreradicalattemptstoeliminatespecialconsiderationfor minoritygroupsandtoassimilatethemintotheHanspeakingmainstream.Thelattertendency wasespeciallyevidentduringtheCulturalRevolutionofthe1960s.Withthewidespread disruptionofChinesesocietyduringthisperiod,manyofthegovernmentsponsoredagencies andservicesforsupportingnonHanminoritygroupswereeffectivelydestroyedandhadtobe painstakinglyrebuiltduringthe1970s. SincetheCulturalRevolution,thePRCappearstohavereturnedtoamorebalancedpolicy towarditsminorities,supportingthemaintenanceofnonHanlanguagesandcultureswhileatthe sametimeencouragingthelearningofPutonghuabyallcitizensofthePRC.Thepresentviewis typifiedbythefollowingstatementissuedbytheCommissiononLanguageReformin1984: "BecauseofthevarietyoflanguagesinChina,everyoneisrequiredtospeakthecommon language.Thepolicyisnottoreplaceone'sowndialectbuttobecomebilingualinone'sown languageandthecommonlanguage[Putonghua]"(Mei,1984,p.78).TheuseofPutonghuais beingencouragedinallaspectsofpubliclife:governmentandlegislature,themilitary,mass media,schoolinstruction,andtheworkplace.Inaddition,schoolsarebeginningtocreate incentivesforlearningPutonghua,forexample,holdingcontestsandcompetitionswithinand betweenschools,throughtheuseofpeermonitoring,andsoon. AnyconclusionsabouttheroleofbilingualisminthelanguagepolicyofthePRCmustbe guardedbecauseofthepaucityofinformation.Moreover,theinformationthatisavailablehas comeprincipallyfromgroupsthat,havingtouredChinabyofficialinvitation(e.g.,Lehman, 1975;Mathias&Kennedy,1980;Mei,1984),havebasedmuchoftheirreportsoninformation providedby 23

governmentsources.Fromthisinformation,onegathersthattheofficialpolicyofthePRCisto maintainitsminoritylanguageswhileatthesametimeencouragingthelearningofPutonghuaas amediumofcommunicationwithothercitizensofthePRC.Itwouldbepremature,however,to drawanyfirmconclusionsabouttheactualimplementationandresultsofbilingualeducationin thePRCatthistime. SOMEPSYCHOLOGICALIMPLICATIONS Comparingthesefourcountries,then,oneseesverydifferentattitudesandapproachesto bilingualismandbilingualeducation.TheofficialpositionbothintheU.S.andintheSoviet Unionisthatbilingualismisprimarilyatransitionalstageintheprocessofassimilationintothe dominantculture.ThepositioninCanadaandthePRC,ontheotherhand,isthatbilingualismis anormal,everydayphenomenonandthattheaimofbilingualeducationshouldbetoenable citizenstofunctionequallywellintwolanguages. Theimplementationofbilingualeducationlikewiseappearstohavehadratherdifferentresultsat 5IntheU.S.,forexample,bilingualeducation leastfortheU.S.,theSovietUnion,andCanada. programshavesofarhadmixedsuccess(Cordes,1985;McLaughlin,1978;U.S.Departmentof Education,1984).Althoughseveralprogramshavebeenextremelysuccessful,withsubstantial gainsbeingshowninschoolperformance,othershaveshownlagsinperformance,whichactually increasewithtime. CriticsofbilingualeducationintheU.S.arguethatsuchvariedresultssupporttheconclusion thatbilingualeducationitselfisnotmoreeffectivethanothertypesofremedialprogramsand thatindividualcommunitiesshouldbegivenmorefreedomtousewhatevermethodtheyfeelis mosteffectiveinhelpingminoritylanguagechildren.Othersinsist,however,thattheseresults maybeduetoanumberoffactors,includingvariationsinthequalityofteaching,instudent background,andinthedegreeofcommunitysupport. InformationaboutbilingualeducationintheSovietUnionisfarlessextensivethanthatavailable fortheU.S.However,thereareatleastsomeindicationsthatSovietprogramsforteaching RussiantononRussianminoritychildrenmaybeencounteringsimilardifficulties.Kreindler (1982)hasnotedseveralpublishedreportsofahighincidenceofschoolfailuresamongthenon Russians. Canada,ontheotherhand,appearstohavehadenormoussuccessinimplementingbilingual educationsuccessfullyonaverylargescale.Thechildrenwho ____________________ 5WefeelthatinformationavailableaboutthestateofbilingualeducationinthePRCisso limited,weprefernottomakeanyfurthercommentsaboutitrelativetotheothercountries whichwearecomparing. 24

emergefromtheprogramsinCanadashowacademicperformanceatleasttheequivalentofthat shownbychildrenenrolledinmonolingualclasses(Lambert&Tucker,1972;McLaughlin,1978 ). Beyondtheirsocialorpoliticalsignificance,thesedifferentpoliciesundoubtedlyhaveenormous psychologicalimpact.Lambert(1977)hasdistinguishedbetweenadditiveandsubtractive processesinbilingualism.Withsubtractivebilingualism,asecondlanguagereplacesafirst language.Withadditivebilingualism,theindividualachievesandmaintainsabalanced 6 proficiencyinbothlanguages. Lambertsuggeststhatdifferentpsychologicaloutcomesareassociatedwitheachofthese processes.Childrenwhoundergosubtractivebilingualism(forexample,whenimmigrant childrenwereputintomonolingualEnglishlanguageclassesintheUnitedStatesuntiltheadvent ofbilingualeducation)areoftenpreventedfromusingwhateverlanguageskillstheyacquired withthefirstlanguagetolearnthesecondlanguage.Thesechildren,incomparisonwith monolingualchildren,mayshowdeficitsinperformancebecausetheirgenerallanguage processingskillshavenotbeenfullydevelopedineitherlanguage.Also,theexperienceofbeing betweentwoworlds,thatis,betweentwodifferentlanguagesandculturesand,moreover,having tochooseoneovertheother,mayalsobeasourceofemotionalconflictforthechildwho undergoesaprocessofsubtractivebilingualism. Ontheotherhand,childrenwhoexperienceadditivebilingualismusuallyacquirebothlanguages inanatmospherewhereinbothlanguagesareconsideredtobeequallysociallyrelevant.Thereis nopressuretogiveuponelanguagefortheother.Moreover,instructioninthesecondlanguage oftentakesadvantageofwhateverskillschildrenalreadypossessinthefirstlanguage.Those childrenwillpresumablydoaswellasmonolingualchildrenintheirgeneralscholastic performanceandmayevenderivesomesecondarycognitiveadvantagesfrombeingproficientin twolanguages(seePalij&Homel,thisvolume). UsingLambert'sdistinctions,wewouldclassifyCanadaasanationwhoselanguagepolicies promoteadditivebilingualism,whereastheU.S.andtheSovietUnionarenationswhosepolicies appeartopromotesubtractivebilingualism.Moreover,wefeelthatitisnomerecoincidencethat thepatternofsuccessofbilingualeducationorlackofit,ineachofthesecountriestendsto agreewiththepredictionsoutlinedbyLambert.Itremainstobeseenwhetherthosewhoare concernedwithevaluatingtheeducationaloutcomesofbilingualismwillconsidertheroleof socialcontext,evenatthelevelofgovernmentalpolicy,inthepsychologicaldevelopmentofthe bilingualchild. ____________________ 6Lamberthasalsosuggestedasimilardistinctionforthephenomenonofbiculturalism: additivebiculturalismwouldbetheprocessthatallowsapersontobecomeknowledgeable aboutandanactiveparticipantintwocultures,subtractivebiculturalismwouldbetheprocess bywhichknowledgeofandparticipationinonecultureisexchangedforknowledgeofand participationinanotherculture. 25

REFERENCES Beaujot,R.,&McQuillian,K.(1982).Growthanddualism:Thedemographicdevelopmentof Canadiansociety.Toronto:GagePress. BilingualEducationActof1974,20U.S.C.A.880bet.seq.(Supp.1975).InUnitedStates CommissiononCivilRights(1974).AbetterchancetoLearn:Bilingualbiculturaleducation (pp.185203).(Clearinghousepublication51.)Washington,DC:U.S.GovernmentPrinting Office. Bilinsky,Y.(1964).TheSecondSovietrepublic:theUkraineafterWorldWarII.New Brunswick,RutgersUniversityPress. Brunner,E.(1981).Sovietdemographictrendsandtheethniccompositionofdraftagemales 19801995(RandnoteN16541SantaMonica,CA:Rand. Comrie,B.(1981).ThelanguagesoftheSovietUnion.London:CambridgeUniversityPress. Cordes,C.(1985,November).StudiesdisputeBennet'sattackonbilingualism.APAMonitor,16, 6. Dreyer,J.T.(1976).China'sfortymillions.Cambridge,MA:HarvardUniversityPress. Fishman,J.A.(1980a)."Bilingualismandbiculturalismasindividualandassocietal phenomena".JournalofMultilingualandMulticulturalDevelopment,I,315. Fiske,E.B.(1985,September26)."Educationdepartmenttoalterbilingualefforts".TheNew YorkTimes,pp.A1,B9. Garcia,R.(1985,October7).Morenewfederalism[Lettertotheeditor].TheNewYorkTimes,p. A30. Glazer,N.,&Moynihan,D.P.(1970).Beyondthemeltingpot:TheNegroes,PuertoRicans, Jews,Italians,andIrishofNewYorkCity(2nded.).Cambridge,MA:MITPress. Greer,C.(1974).Dividedsociety:TheethnicexperienceinAmerica.NewYork:BasicBooks. Grosjean,F.(1982).Lifewithtwolanguages:Anintroductiontobilingualism.Cambridge,MA: HarvardUniversityPress. Howe,H.(1985,October7).Misguidedtamperingwithbilingualeducation[Lettertothe Editor].TheNewYorkTimes,p.A30. Isayev,M.I.(1977)NationallanguagesintheUSSR:Problemsandsolutions.Moscow: ProgressPublishers. Kaalt,J.(1977).Abriefintroductiontothecensuslanguagedata.InP.Lamy(Ed.),Language maintainenceandlanguageshiftinCanada:Newdimensionsintheuseofcensuslanguagedata

(pp.1621).Ottawa:UniversityofOttawaPress. Keefe,J.(1985,October24)."Analternativetobilingualism".TheNewYorkTimes,p.A27. Kreindler,I.(1982)."ThechangingstatusofRussianintheSovietUnion".International JournaloftheSociologyofLanguage,33,739. Lambert,W.E.(1977).Theeffectsofbilingualismontheindividual:cognitiveandsociocultural consequences.InP.A.Hornby(Ed.)Bilingualism:Psychological,social,andeducational implications(pp.1527).NewYork:AcademicPress. Lambert,W.E.,&Tucker,G.R.(1972).Bilingualeducationofchildren:TheSt.Lambert experiment.Rowley,MA:NewburyHouse. LauV.Nichols,No.726520(414U.S.at566,1974). Lehmann,W.P.(1975).LanguageandlinguisticsinthePeople'sRepublicofChina.Austin: UniversityofTexasPress. Lewis,E.G.(1972).MultilingualismintheSovietUnion.TheHague:Mouton. Luckyj,G.S.N.(1975).Socialistincontentnationalistinform.InG.S.N.Luckyj(Ed.), Discordantvoices:ThenonRussianSovietliteratures(pp.112).Oakville,Ont.:MosaicPress. Mathias,J.&Kennedy,T.L.(Eds.)(1980).Computers,languagereform,andlexicographyin China:AreportbytheCETAdelegation.Seattle,WA:WashingtonStateUniversityPress. McLaughlin,B.(1978).Secondlanguageacquisitioninchildhood.Hillsdale,NJ:Lawrence ErlbaumAssociates. 26

Mei,J.Y.(1984).ReadinginChina:ReportoftheU.S.readingstudyteamtothePeople's RepublicofChina.Washington,DC:NationalCommitteeonU.S.ChinaRelations. Moser,L.J.(1985).TheChinesemosiac:ThepeoplesandprovincesofChina.Boulder,CO: WestviewPress. NarodnoekhozjajstvoSSSR19221982[ThenationaleconomyoftheUSSR19221982]. (1982)Moscow:Financyistatistika. Novak,M.(1977).Theriseoftheunmeltableethnics:Politicsandcultureintheseventies.New York:Macmillan. Savchenko,F.(1970).Zaboronaukrajinstva1876[ThesuppressionoftheUkrainianactivities] (HarvardSeriesinUkrainianStudiesvol.14).Munich:WilhelmFinkVerlag. Simirenko,A.(1969).Thedevelopmentofsovietsocialscience.InA.Simirenko(Ed.),Social thoughtintheSovietUnion.Chicago:Quadrangle. Thernstrom,S.,Orlov,A.,&Handlin,O.(Eds.)(1980).HarvardencyclopaediaofAmerican ethnicminorities.Cambridge,MA:BelnapPressofHarvardUniversity. U.S.DepartmentofCommerce.BureauoftheCensus.(1983).1980census:Detailedpopulation characteristics;UnitedStatessummary,part1:sectionA:UnitedStates(PC80IDIAUS Summary).Washington,DC:U.S.GovernmentPrintingOffice. U.S.DepartmentofEducation(1984).Theconditionofbilingualeducationinthenation,1984. Rosslyn,VA:NationalClearinghouseforBilingualEducation. Whitaker,R.A.(1984).TheQuebeccauldron.InM.S.WhittingtonandG.Williams(Eds.), Canadianpoliticsinthe1980's(pp.3357).Toronto:Methuen. 27

[Thispageintentionallyleftblank.] 28

II LANGUAGEACQUISITIONANDPROCESSING 29

[Thispageintentionallyleftblank.] 30

3 TheSecondLanguageLearnerintheContextoftheStudyofLanguageAcquisition KenjiHakutaYaleUniversity Howcanthestudyofsecondlanguagelearnersilluminatethefundamentalissuesofdevelopment? Thatisthecentralconcernofthisconference,onewhosemanyfacetedresponseswillbeslowly revealedthroughthepaperspresented.Iwouldliketoaddressspecificallytheproblemofgrammar intheacquisitionofasecondlanguageaftertheprimarylanguagehasbeenestablished.The problemcanbeputintotheperspectivepresentedbyLilaGleitmaninthe10thAnniversaryspecial issueofthejournalCognition(Gleitman,1981).Essentially,Gleitmanarguesfortheinformative valueofthreedifferentkindsofvariationsininvestigatingthedifferentialrolesofmaturationand environmentindetermininglanguageacquisition.First,therearevariationsinthequalityofthe languagesampleavailabletothechild.Theseincludethetraditionalvariablesusedinthe investigationofmotherese.Second,therearevariationsintheinterpretiveinformationfromthe learner'sperspective(forexample,howdoesablindchildinterprettheverbsee?)Andthird,there arevariationsinthelearner'sendowment,specificallytheabilitytorepresentlanguage.Although everyonewouldhavetheirownpetvariationstoaddtothislist(myownbeingcrosslinguistic variations,moretobesaidonthislater),theframeworkisveryusefulindiscussingwherethestudy ofsecondlanguageacquisitionfitsinwithrespecttomajorissuesinlanguageacquisition. Withregardtothefirstpoint,thereareprobablymorevariationsinlinguisticenvironmentsfor secondlanguage(L2)acquisitionthanforfirstlanguage(L1)acquisition.MostL2studiesare concernedwith"naturalistic"situations,thatis,caseswherethelearnersarenotformallytutoredin thesecondlanguage.ThesestudiesbeartheclosestresemblancetotheL1inputsituation,although thesourceofinputcanvaryfromadultstopeers.Then,therearecaseswhere 31

childrenandadultslearnasecondlanguagethroughformallytutoredclasses.Suchvariationsinthe typesoflinguisticinputtoL2learnerscanprovidearichsourceofdataforassessingtheroleofthe input.ItisinterestingthatmostL2studieshavebeenconcernedwithnaturalisticacquisition, becausetutoredsettingsprovideactuallyanidealsituationinwhichinputdatacanberecorded accuratelyandexperimentallymanipulated. Gleitman'ssecondpoint,variationsininterpretiveinformation,canbewelladdressedbythestudy ofL2acquisition.AfacttooobvioustodwellonatanylengthisthatL2learnersaremore cognitivelyadvancedthantheL1learner,sothattheusualconfoundingsoflanguageandcognitive developmentcanbeseparatedout.Bylookingatthecorrelationsbetweenvariationsinthecourseof acquisitionofL1andL2learnersontheonehandandvariationsintheirdifferencesininterpretive capacitiesontheother,wecanhypothesizewhatfactsobservedinL1learnersare"artifacts"of cognitivedevelopment.Likewise,similaritiesbetweenL1andL2acquisitiondespitethevast cognitivedifferenceswouldbestrongevidenceforlanguagespecificprocesses.Leavingasidethe problemoflanguagetransferforthemoment,onemightevenarguethatL2learnerscanrevealmore aboutlinguisticbiasesinherentinlanguageacquisitionthanL1learners(GassArd,1980).This bringsustothethirdpointraisedbyGleitman,onvariationsoftheendowmentofthelearner. Perhaps,asGleitmanputsit,"theearly'conceptuallanguage'stageisnottraversedduringlearning bytheseolderchildrenfortheyarecapableofthematurerepresentationsoflinguisticdata"(p.111). InadditiontowhetherthefirsthypothesesbytheL2learnerarerepresentedlinguisticallyrather thanconceptually,thereismuchtobelearnedfromlanguagetransfer.Inthe1950sand1960s, duringtheheydayofcontrastivelinguistics,languagetransferwasseenasevidencefortheSRview oflanguageacquisition(seeHakutaCancino,1977,forahistoricalreview).However,morerecently, transferisseenasevidenceforthelearner'sworkinghypothesesconcerningthenatureofthetarget language.Bystudyingwhattypesoflinguisticrulestransferfromthenativelanguagetothesecond language,wecangarneraglimpseoftheentrenchmentofthenativelanguageinthechild.The existenceoftransferatteststothepsychologicalrealityofthelinguisticrulesinquestion.Another issueraisedwithregardtoendowmentinsecondlanguagelearnersisthatofthecriticalperiod. Doesthecapacitytorepresentlanguagedisappearwithage(Lenneberg,1967;Penfield&Roberts, 1959)?WecanlookatL2learnersvaryingintheageatwhichtheybeginacquisition. FirstIwouldliketostatetheconclusionsthatcanbedrawnfromtheliterature,andthenprovidethe relevantevidence.Generally,itisnottruethatasecondlanguagelearner,regardlessofhisherage oflearning,willperfectlymimicthedevelopmentalpatternsdisplayedbyachildlearningthenative language.Infact,moresimilaritiesaretobefoundbetweenanadultanda5yearoldsecond languagelearnerthanbetweena5yearoldandafirstlanguagelearner.However,therearealsoa largenumberofparallelsbetweenL1andL2 32

acquisitionthatcanbeobserved.Specifically,first,therearecertainstrikingfactsaboutL2 acquisitionthatcanbebestunderstoodwhencognitivedevelopmentalfactorsaretakeninto account.Thatistosay,someofthedifferencesbetweenL1andL2acquisitionareduetothefact thatsecondlanguagelearnersarecognitivelymorematurethantheirL1counterparts.Second, therearealsostrikingsimilaritiesbetweenL1andL2learners.Theyrevealvariationthatcanbe attributedtothebureaucraticstructureofthebeastthatalllanguagelearners,whetherL1orL2, mustmaster.Regardlessofthelearner'sendowment,certainstructuresaremoredifficultthan others.Thismaybeduetogeneralcognitivefactors,suchasprocessingconstraintsimposedby configurationalstructureormemory,oritmaybeduetomoreabstract,languagespecificfactors. Third,theextenttowhichahumanbecomesentrenchedinthenativelanguagecanbeseeninthe effectsoftransferfromthenativelanguagetothesecondlanguage.Thereisgoodevidencefor thenativelanguageofthelearnerbiasingL2acquisitionindifferentsortsofways.Andfinally, thereisgoodevidencetosuggestthatsometimeafterpubertyisaperiodwhenthecapacityto acquireasecondlanguagedeteriorates. Insupportoftheseconclusions,Idonotintendtoreviewalloftheavailableevidence.Rather, thediscussionisintendedtoexposethenewcomertosecondlanguageacquisitiontothekindsof datathatcanbeexpectedfromsuchresearch.Moreextensivereviewsoftheliteraturecanbe foundinMcLaughlin(1976),HakutaandCancino(1977),Hatch(1978),Schumannand Stenson(1975),OllerandRichards(1973).Thefirstsectionaddressesthequestionoftheeffects ofcognitivematurity.Thesecondsectionreviewstheeffectsoflinguisticstructurefromthe viewpointofsimilaritiesbetweenfirstandsecondlanguageacquisition.Thethirdsectionreviews evidencefortransferfromthenativelanguage.Thefourthsectionlooksatthequestionofage effects.Thefifthsectionproposesaframeworkforresearchfromtheviewpointoflanguage universals. EFFECTSOFCOGNITIVEMATURITY AwelldocumentedperiodinfirstlanguageacquisitionisthesocalledtwowordstageorStageI speechinBrown's(1973)outline.Itappearsthatarelativelysmallnumberofsemanticrelations characterizealargeproportionofthetwowordutterancesthatcanbefoundinchildrenduring thisperiodofdevelopment.Frequentlyfoundarerelationssuchasagentaction,agentobject, attributiveentity,andsoforth.Brownpointedtothecorrespondencebetweenthesemanticroles expressedduringthisperiodandthesensorimotorschemesoutlinedindependentlybyPiaget,the conclusionbeingthatearlyStageIspeechisconstrainedbythecognitivecapacitiesofthechild. Conspicuouslymissingaresemanticrelationssuchasifthenconditionals,sophisticated temporalandaspectualrelations,andlogicalconnectives.Theobviousneededtobedone. Lightbown(1977)lookedattheacquisitionofFrenchbytwo6yearoldchildrenwhose 33

nativelanguagewasEnglish.Shesubmittedthechildren'slanguageintheirinitialstagesoflearning tothesamekindsofanalysesconductedforL1learners.Essentially,Lightbownfoundthatthese childrenexpressedallkindsofsemanticrelationsfromtheverybeginning.Therewasnotthekind oforderlyprogressionfoundwithfirstlanguagelearners.Shefoundrelationsexpressedsuchas manner,intensifiers,andconjunctions.Inshort,L2children,evenwhenrelativelyyoung,donot seemtogothroughidenticalstagesofdevelopmentasL1children. Ithasbeenobserved(Hakuta,1975;Tiphine,personalcommunication)thatL2childrenuse sentencecoordinationfromquiteearlyonintheirdevelopment.InL1children,thisstructureis relativelylateinemerging(Brown,1973;Hakuta,deVilliers,&TagerFlusberg,1982;Tager Flusberg,deVilliers,&Hakuta,1982),especiallywhenusedinthecontextscorrespondingto logicalconnectives(Beilin,1976).AnillustrationoftheearlyuseofconjunctionsappearsinTable 3.1.ThedataarefromafiveyearoldJapanesegirl,namedUguisu,learningEnglish(Hakuta,1975 ).Table3.1isfrequencydistributionofvariousconjunctionsobservedinherspeechovertime.The monthlysamplesweremadeequivalentinlengthat200utteranceseach.Ascanbeseen,Uguisu usedconjunctionsfromquiteearlyon,inthecaseofandandbecause,fromtheveryfirstsample. Theusagewasinmostinstancesappropriatefromthetargetlanguagepointofview. Asidefromthestructureoftheirnativelanguage,secondlanguagelearnersmostlikelyknowcertain factsaboutthefunctionsoflanguage,suchasthatitisusedforconversations,thatconversations involveturntaking,andsoforth TABLE3.1 DistributionofCoordinatingConjunctionsinUguisu'sSpeechSamples.EachMonthly SamplingPeriodContains200Utterances. Month 1 * 2 3 4 5 6 7 8 9 10 and 3 8 5 5 20 4 5 7 7 6 but 1 4 4 1 6 2 5 4 because 2 3 4 11 10 5 8 12 2 so 8 4 2 4 4 3 if 11 3 1 5 3 2

Note:*representsthefirstmonthwhenUguisustartedpro ducinghighfrequencyofutterances,whichwas6monthsafter herinitialexposuretoEnglish. (Source:Hakuta,1975) 34

(KellerCohen,1979).Thisknowledgeabouttheglobalpropertiesoflanguage,alongwitha moredevelopedmemoryspanforrememberingwholesentences,hasbeenhypothesizedto accountforthelargenumberofprefabricatedpatternsobservedcommonlyinsecondlanguage learners(Hakuta,1974b;Huang,1971;WongFillmore,1976).Prefabricatedpatterns(or formulaicutterances,asWongFillmorecallsthem)arecharacterizedbylackofinternal structure.Itappearsthatsecondlanguagelearnersmemorizeentireutteranceswithoutknowledge ofunderlyingstructure.ThisisnotunheardofinL1acquisition(Clark,1974),butits preponderanceinL2acquisitionisstriking.Huang(1971)studieda5yearoldTaiwaneseboy, Paul,learningEnglish.HereportsanexcellentobservationofPaul'sfirstutteranceinEnglish: OnFebruary4,onlytwodaysafterthebeginningofPaul'snurseryschool experience,theinvestigatorheardhim(Paul)muttering:Getoutofhere.Ontheway homefromschoolthenextmorning,heaskedmeaboutthemeaningofthat utterance.Whentheinvestigator,insteadoftellinghimthemeaning,askedhimto relatewhathadhappened,hesaidthataboywhowantedtogetawayhadsaidit.The TaiwanesetranslationherenderedmeansDon'tbestayhere,whichisverycloseto Getoutofhereinmeaning.Anincidentinthenurseryschoolthenextdayproved Paul'scapacitynotonlytounderstandthisutterancebuttouseitappropriatelyas well. (PaulwasonatricyclewhileMicheleholdingontothehandlebar,keptonbotheringhim. Obviously,hewantedhertoleavehimalone.)Paul:Getoutofhere.(Michelewalkedaway, somewhatembarrassed.)(pp.1213)WongFillmore(1976),inherstudyoffiveSpanish speakingchildrenlearningEnglish,reportedthatoverhalfofthechildren'sutterancescontained prefabricatedforms.Shearguedthatthroughthegradualanalysisofsuchforms,laterlinguistic structuredeveloped:"Alloftheconstitutentsoftheformulabecomefreedfromtheoriginal construction,[and]whatthelearnerhasleftisanabstractstructureconsistingofapatternorrule bywhichhecanconstructlikeutterances"(p.645).Itisnotclear,however,whethersuchabstract structurecanemergethroughbruteforce.Thatisthetraditionalproblemassociatedwiththe emergenceofgrammar.Nevertheless,whatisclearisthatsecondlanguagelearnershavebeen observedtobeginwithwholesentences,andthatifthesameformisfollowedovertime,the emergenceofstructurecanbeobserved.AstrikingexampleofthiscanbeillustratedinUguisu's useofembedded"how"questions(Hakuta,1976).Duringthethirdmonthofobservation, Uguisumadethefollowingsetofutterances: Iknowhowtodoit. Iknowhowtodoreaditthis. 35

I know how to readit this. I know how to make. I know how to draw itcat. I know how to draw (it) butter fly. I know how to draw itboy.

Thesecanbecharacterizedbyaprefabricatedpattern,Iknowhowto+VP.This apparentlycorrectformchangedovertimeintoformssuchasthefollowing,whichwere observedinher15thmonth. FirstIgottawriteitandshowyouhowdoyouspell'Debra'. IknowhowdoyouspellVino. Weonlyknowhowdoyoumakeitlikethat. Iknowhowdoyouwritethis. Figure3.1plotstheproportionofformsusing"howto"overthetotalnumberofhow questions,showingthatthedeclineinperformanceisagradualone.Ihaveargued elsewherethatthischangeisinfactreflectiveofherotherusesofindirectwhquestions, whereformswerefirstusedwithsubjectauxinversion(Hakuta,1976).Atanyrate, prefabricatedpatternsarequitepredominantinearlysecond Fig.3.1.Proportionofcorrecthowembeddings(howto)overtotalhowembeddings. Biweeklysamplesarepaired.Source:Hakuta,1976. 36

languagespeech,butitisclearthatgraduallytheyareanalyzedintomoreflexibleform. Althoughprefabricatedpatternsmaybeagoodplacetostart,andmayprovidethelinguisticdata uponwhichthelanguageacquisitionmechanismmightwork,theproblemofwhattheanalytic processconsistsofisamystery,justasitisinfirstlanguageacquisition. Perhapsalongerprocessingspanwouldimplythatthesecondlanguagelearnerwouldnotshow effectsoflength(suchaswouldbereflectedinM.L.U.upperbounds).Oneindicationofthis canbefoundinUguisu'suseoftheEnglishpossessiveinflection's.Cazden(1968)reported differentialuseofthepossessiveinAdam,Eve,andSarah,dependingonwhethertheywereused innounnouncontexts(e.g.,Mom'spie)orinellipticcontexts(e.g.,Mom's).Forallthree children,therewasalargedifferenceinthepercentagetowhichthemorphemewassuppliedin obligatorycontext,withgreaterproportionbeingsuppliedinellipticcontextsthaninnounnoun. Uguisu,ontheotherhand,performedequallyinbothcontexts(seeHakuta,1976).Becausethe ellipticcontextrequiresonlyonenoun,whilethenounnouncontexttwo,itispossiblethatthis differencereflectsthelargerprocessingspaceavailabletosecondlanguagelearners.Inturn,this providesanexplanationfortheotherwiseperplexingdifferentialuseinLIchildren.First languagelearners,havingamorelimitedprocessingspan,mayomitthemorphemewhenthey havetoprocesstwonounsforthenonellipticform,butintheellipticformmayfinditpossible to"fitin"themorphemebecausethereisonlyonenountoprocess.Thephenomenonfoundin firstlanguagelearners,then,maybeasimpleperformancefactor. Finally,withregardtothemanifestationsofcognitivematurity,itisworthpointingtovocabulary developmentasanunderinvestigatedissue.Allindicationsarethatlexicaldevelopmentin secondlanguagelearnersisextremelyrapid,Gillis(1975),inherstudyof3sevenyearold JapanesechildrenlearningEnglish,reportedPPVT(PeabodyPictureVocabularyTest)increases inmentalscoreequivalentsof6monthstoIyearinaperiodofjustImonth.Snowand HoefnagelHohle(1978)reportthatageispositivelycorrelatedwithvocabularydevelopmentin AmericanchildrenlearningDutch,suggestingthatolderlearners,withmorecognitivecapacity, pickupvocabularyfaster.Whatisneededisastudyoftheorganizationofthelexiconin bilingualchildren,apromisingareaforfutureinvestigation. EFFECTSOFLINGUISTICSTRUCTURE:L1ANDL2SIMILARITIES Anactiveresearchtraditioninappliedlinguisticsisonecommonlycalled"erroranalysis"( Corder,1971;Nemser,1971;Oller&Richards,1973;Schumann&Stenson,1975;Selinker,1972; Svartvik,1973).Typically,erroranalysislooksatthekindsofsystematicdeviationsfromthe targetlanguagegrammarobserved 37

inlanguagelearners,andtheerrorsareclassifiedwithrespecttotheirhypothesizedsource. Usualcategoriesforclassificationaretransfer,simplification,andovergeneralization.Transfer errorsrefertothoseerrorswhosesourceisclearlyidentifiableasthenativelanguagegrammar. Forexample,anativespeakerofSpanishsaying"Henohavehappiness"isconsideredtohave constructedanEnglishutterancebasedonhisnativelanguage.Simplificationusuallyrefersto errorsofomission,particularlyinflectionsandauxiliaryverbs.Theutterance"Reaganalways sleep"isconsideredanexampleofomission,wherethethirdpersonsingularindicativemarkeris missing.Overgeneralizationerrorsaremoststriking,andusuallyinvolvethelearner"ironing out"irregularitiescommontolanguage.Forexample,"Cooneyrightedpoorly"shows overgeneralizationoftheregularpasttensemarkeredtoirregularinstances.Simplificationand overgeneralizationareerrorswellknowntothestudentoflanguageacquisition. Studiesemployingerroranalysistypicallyshowarelativelysmallnumberoftransfererrors,with simplificationbeingthemostfrequent(Cohen,1975;Dulay&Burt,1973;Duskova,1969; Politzer&Ramirez,1973;Selinker,Swain,&Dumas,1975).Forexample,DulayandBurt (1974a)lookedattheerrorsproducedby179SpanishspeakingchildrenlearningEnglish.Outof 513errorsthattheyconsidered,5%wereclassifiedasinterference,whereas87%wereeither simplificationorovergeneralizationerrors.However,theclassificatorysystemdifferssomewhat fromstudytostudy,anditisdifficulttoknowhowtointerprettheresults(Hakuta&Cancino, 1977).Itisnotasimplemattertoclassifyerrors.Japanesedoesnothavearticles,soshouldthe omissionoferrorsbyaJapanesechildlearningEnglishbeconsideredaninterferenceerrorora simplificationerror?Also,itisnotclearthatthedifferenttypesoferrorsarecomparable,since theiropportunitiesforoccurrenceareuncontrolledfor.Thesestudiesinerroranalysis,however, dosuggestcommonalitiesbetweenfirstandsecondlanguagelearnersevenwhenerrorproportion areleftaside.Similarkindsoferrorscanbefoundinbothkindoflearnersandacrosssecond languagelearnersofdifferentlanguagebackgrounds. Withregardtospecificstructures,thereissomeindicationofsimilarities.ErvinTripp(1974) reportsastudyinwhichshelookedatthecomprehensionofFrenchpassivesinAmerican childrenlearningFrench.Shefoundsystematicmisinterpretationofpassivessimilartothose reportedforFrenchL1childrenbySinclairdeZwart(1973).Interestinglyenough,thechildren whomisinterpretedthepassiveswereatthesametimecorrectlyinterpretingpassivesinEnglish. GassandArd(1980)reportastudyofEnglishrelativeclausecomprehensionbyadultsecond languagelearnersfromdifferentnativelanguagebackgrounds.Theirresults(thoughnottheir interpretationofthedata)aresimilartothoseobtainedforthesamestructuresinEnglishL1 children(forreviewsoftheEnglishL1literature,seedeVilliers,TagerFlusberg,Hakuta,& Cohen,1979;Hakuta,1981).Finally,D'AnglejanandTucker(1975)administeredtoadultL2 learnersEnglishcomplementizerstructuressimilartothoseusedbyCarolChomsky(1969)for olderL1children,andobtainedsimilarresults.Thetenta 38

tivegeneralizationemergingfromthesethreestudiesseemstobethatatleastwhencomprehension proceduresareemployedtoinvestigateparticularsyntacticstructures,L2learnersperformsimilarly toL1learners. EversinceBrown's(1973)reportontheorderofacquisitionof"grammaticalmorphemes" appeared,researchersinsecondlanguageacquisitionhaveconcernedthemselveswithwhetherthe sameordercanbeobservedinsecondlanguagelearners(Cancino,1976;Dulay&Burt,1973, 1974b;Gillis,1975;Hakuta,1974a,1976;Bailey,Madden,&Krashen,1974;LarsenFreeman,1976; Rosansky,1976).Theanswerisrelativelystraightforward:itisnot.However,therehasemergeda strikingsimilarityinorderofacquisitionacrosssecondlanguagelearners,regardlessoftheirnative language.Tobesure,thereareinfluencesofthenativelanguage(tobediscussedinthenext section),butthedifferencesseemtobeovershadowedbythesimilarities.Thisistakenasevidence thatthenativelanguageexertsminimalinfluenceontheorderofacquisitionofgrammatical morphemes(seeHakuta&Cancino,1977).Theinferencefromheretotheconclusionthatfirstand secondlanguageacquisitionsaresimilarisasomewhatdifficultone,althoughitisonecommonly made.Ontheonehand,thereisnoreasontoexpect,evenifthetwoprocesseswerethesame,that thelinguisticproductwouldbethesame.Ontheotherhand,withoutsomeexplanationforthe differences,oneisleftskeptical.Compoundingthisproblemisthecommonlyheldmisconception that"cognitionequalssemantics."Forexample,DulayandBurt(1974b),inexplainingthe differencefoundintheorderinfirstandsecondlanguagelearners,state:"Itseemsintuitivethat childrenwhoareacquiringtheirfirstlanguagehavetodealwithbothsemanticandsyntactic information.However,six,seven,andeightyearoldchildrenlearningasecondlanguageneednot strugglewithsemanticconceptstheyalreadyacquired,suchasconceptsofimmediatepast, possession,orprogressiveaction"(p.74). Theproblemwiththisreasoning,asSchlesinger(1974)pointsout,isthatcognitiondoesnotequal thesemanticsofalanguage.Iftheywerethesame,therewouldhavetobenodistinctionbetween thetwo.Thebestdemonstrationofthiscomplexrelationshipisthroughthefactthatthecognitive categoriesfromwhichlanguagesdrawarenotuniformacrosslanguage.Forexample,although manylanguagesobservethedistinctionbetweenalienableandinalienablepossessions,Englishdoes not.Genderisanothercognitivecategorythatisexpressedtowidelyvaryingdegreesindifferent languages.Althoughcognitivedevelopmentmaybeapacesetterforcognitivecategoriesavailableto thelearner,thesemanticsofeachparticularlanguage(oneaspectoftheformalcomplexityofthe language)isoftenspecifictothatlanguage(Slobin,1973).AsIarguelater,thesemantic distinctiondrawninEnglishbetweendefiniteandindefinitearticlespresentspreciselysucha stumblingblockforlearnersfromlanguagesthatdonotdrawuponthisdistinction. OneexplanationforthemorphemeorderinghasbeenprovidedbyLarsenFreeman(1976),who suggestsfrequency(althoughfrequencyitselfshouldbeexplained).LarsenFreemancorrelatedthe L2orderingswiththefrequencies 39

reportedbyBrownforthemothersofAdam,Eve,andSarah.AlthoughBrownfoundno correlationbetweenmaternalfrequencyandtheorderofacquisitionforthechildren,Larsen Freemanfoundrankordercorrelationsofroughly.80(dependingonthestudy).Shouldwe acceptthisconclusion,thatL2learnersaresensitivetofrequencywhereasLIlearnersarenot,at leastforclosedclassitems,whatarewetoconcludeaboutthesimilaritiesanddifferences? Gleitman(1981)suggeststhatL1childrencanbeinfluencedbydifferentialuseofclosedclass itemsinmaternalinput.PerhapsthiswouldforceareexaminationoftheissueoffrequencyinL1 acquisitionaswell,asa"modifiedfrequency"hypothesis. EFFECTSOFLINGUISTICSTRUCTURE:NATIVELANGUAGETRANSFER Justaslanguagecontactinsocietywasseenasareliableindicatorofthedynamicinteraction betweenculturesbythegreatsociolinguistWeinreich(1953),theinterplayofthetwolinguistic systemsintheindividualcanbeseenasreliableindicatorsofinterplaybetweenmental structures.Thiseffectisbestseeninlanguagetransfer.Whatbetterindicatoristhereforthe psychologicalrealityofalinguisticstructurethanthefactthatitcantransfertoanotherlanguage inthecourseofL2acquisition? Aconstantthorninthesideofthosewhowanttoargueforthesimilarityintheorderof acquisitionofgrammaticalmorphemesistheEnglisharticlesystem.Asmentionedabove, childrenlearningEnglishwithJapanese(Hakuta,1976)andKorean(Fathman,1975;Kang, 1982)astheirnativelanguage(neitherofwhichhasanarticlesystem)haddifficultyinlearning theEnglisharticlesystem.Thiscompares,forexample,toFrauenfelder's(1974)studyof EnglishspeakingchildreninaFrenchimmersionprograminCanada,inwhichhefoundthat althoughthechildrenmademanyerrorsingenderonarticles,theyneverconfusedthedefinite indefinitecontrast.Noticehowthisexemplifiesthedistinctionbetweensemanticdevelopment andcognitivedevelopment.ItcertainlycannotbethattheJapaneseandKoreanchildrenwere unabletoconceptualizethedifferencebetweendefiniteandindefinitereference.Rather,itwas thatthedistinctionwasnotmarkedinthesemanticsystemoftheirnativelanguage.Itappears thatthesemanticstructureofthenativelanguageguidestheformulationofspecifichypotheses aboutthetargetlanguage,notthecognitivesystem. Negationhasbeeninvestigatedbyanumberofresearchersinbothchildrenandadults(Adams, 1974;Butterworth,1972;Cazden,Cancino,Rosansky,&Schumann,1975;WongFillmore,1976 ).ThegeneralfindingseemstobethatnativespeakersofSpanishhaveacharacteristicstageof preverbalno,suchasCarolinanogotoplay.ThisstructureisseenasderivingfromtheSpanish structure,whereinfactamorphemeofverysimilarsoundastheEnglishformisused.Both childrenandadultsusesuchstructuresintheirearlystagesofacquir 40

ingEnglish.Recently,Herlinda.CancinoandIcompletedalargestudyof59adultspeakersof Spanish,allofwhomhadarrivedintheUnitedStatesafteradolescence.Oneofthestructureswe investigagedwasnegation.Inanimitationtask,weaskedsubjectstorepeatsentenceswherethe thirdpersonsingularnegativeauxiliarywasrequired(e.g.,"Themandoesnotlikecheese"). Responseswerescoredasfollows:1pointforno+Verbresponses;2pointsfornot+Verb responses;3pointsfordon't+Verbresponses;and4pointsfordoesn't+Verbresponses.They correspondedtotheroughorderofdevelopmentinthecasestudies.Figure3.2showsthe scatterplotofthemeanresponsescoreagainstanindependentmeasureofgrammaticalability basedonratingsandavocabularytest.Thesysternaticityintheirdevelopmentisremarkable, despitethelargerangeofvariationwithrespecttothesubjects'lengthofresidenceintheUnited Statesandeducationalstatus.Thesefindingsseemtosupporttheconclusionthatthefirst hypothesisthesecondlanguagelearnerentertainsisthat"thislanguageismineunlessproven otherwise." Fig.3.2.Relationshipbetweenmeannegationscoreonimitationtaskandglobalmeasureof EnglishfluencyinadultPuertoRicanslearningEnglish. 41

LestwemistakenlyconcludethatalllanguagetransferoccursduringtheinitialphasesofL2 acquisition,twospecificexamplesindicatingotherwisewillbementioned.Thefirstcaseinvolves Uguisu'suseofEnglishreflexives(reportedinHakuta,1976).Sometimebetweenthe5thand6th monthaftershestartedspeakingEnglish,UguisubeganusingreflexivesinutterancessuchasYou havetodoself,becauserememberIdoself?(meaningYouhavetodoitbyyourself,because rememberIdiditbymyself?)Althoughsuchuseofthereflexivehasnotbeenreported(tomy knowledge)forL1learners,itiscertainlysomethingthatmightbeexpected.Aroundthe10th month,shebeganplacingapronouninfrontofself,butinaddition,shebeganusingtheEnglish prepositionwith,resultinginutterancessuchasTheyhavetodoitwiththeirselfs,andIcanmake toastwithmyself.InJapanese,theinstrumentalprepositionisusedwithreflexivepronouns,and Uguisuapparentlycametoputthetwotogether,resultinginthetransfererror.Whatacleverchild. Inhernativelanguageknowledge,reflexivitywaspairedwiththeinstrumentalcase,presumablyat thesemanticlevel. TheothercaseinvolvesaSpanishspeakingfiveyearoldgirl,namedMarta(Cancino,1976). Martabeganusingthepossessiveincontextsdeterminedtorequire'susingtheSpanishderived formandwordorderde,asinbookdeMarta.Noticethattheorder,possessedpossessor,reverses theusualEnglishorder.Then,shereversedthenouns,butomittedthe's,asinMartabook.Now,her formswerejustlikeonewouldfindinanL1learner.Foursampleslater,shepickeduponthe Englishprepositionof,asinfrogofFreddie,andfortwosamplessheusedboththisformandthe earlierforminapparentlyfreevariation.Shortlyfollowingthisperiod,shegraduallybegan supplyingthe'sinobligatorycontexts.Essentially,thereweretwolevelsatwhichlanguagetransfer occurred,oneatagrosslevelofactuallyusingthenativelanguageform,thenamoresophisticated transferoncetherelevantEnglishformwasacquired. Anothererrorofpotentialinterestininferringtheorganizationofthenativelanguageoccurswhen thenativelanguageandthetargetlanguagedifferwithrespecttotheformclassrepresenting specificconcepts.OneexampleofthiscanbeseeninUguisu'suseofthewordmistakeinEnglish( Hakuta,1976).InJapanese,itisaverb,whereasinEnglish,itiscommonlyanoun.Uguisu,until thelastfewsamples,usedmistakeasaverb,resultinginutterancessuchasNotthere,Imustake; Don'tgivememorebecauseyou'remustaking;Ijustmustake,andIjustskipped.Sucherrorsgive anindicationthatUguisuwastransferringattheunitoftheformclass,andthatindividuallexical itemsinhernativelanguageweretaggedwiththeparticularprivilegesofoccurrencethatdefine membershipinformclasses(Maratsos&Chalkley,1980). Finally,itshouldbepointedoutthattransferdoesnotmanifestitselfonlyaserrors.Schachter (1974)showedthatinwrittenEnglishcompositionsbynativespeakersofdifferentlanguages,the nativelanguagemadeanimportantdifferenceinthefrequencywithwhichtheyusedrelativeclause structures.English 42

isa"rightbranching"language,thatis,therelativeclausecomestotherightoftheheadnoun, suchasin thebottle[thatfellonthefloor]. Thedirectionofbranching,leftorright,hasbeenconsideredanimportantvariableinthestudy oflanguagetypology(Lehmann,1973).ItturnsoutthatinSchachter'sstudy,subjectswhocame fromrightbranchinglanguages(Arabic,Spanish)usedmorerelativeclausesthanthosefrom leftbranchinglanguages(JapaneseandChinese).Itisimportantthattherewerenodifferences betweenthesetwogroupsintermsoferrorrates,thatis,thedegreetowhichtheycorrectlyused relativeclausestructurewhenitwasinfactused.Thus,thedatacanbeinterpretedasavoidance ofthestructure,ratherthaninability.InacomparisonofUguisuandMarta'sspeechsamples, Hakuta(1976)found,albeittentativelybecauseofthedifficultyincomparability,thatMarta, whosenativelanguageisSpanish,usedmorerelativeclausesthanUguisu,extendingSchachter's importantfindingtochildren.Thus,itisimportantinconsideringtransfertolookattheoverall patternofdevelopment. EFFECTSOFAGE Ifinallyturnnowtoadiscussionofwhetherthereisacriticalperiodforsecondlanguage acquisition.Itisassumedthatthisquestionbearsonthemoregeneralissueofconcern,thatof capacityforlanguage.Muchspeculationhasbeenmadeabouttheageatpuberty(roughly13to 14)asthepivotalpointbeyondwhichsecondlanguageacquisitionbecomesdifficult( Lenneberg,1967).Hence,itisnotsurprisingthatmoststudieshaveattemptedtospanthatage period. Williams(1974)lookedataveryspecificaspectofspeechperceptionininvestigatingthe problem.ShechosetheVoiceOnsetTime(VOT)parameterdifferentiating/pa/from/ba/.In Spanish,theboundaryfallsatthe0msec.VOT,whereasforEnglish,theboundaryisat+25 msec.Capitalizingonthiscontrast,WilliamstestedPuertoRicansubjectswhovariedalongtwo backgroundvariables:ageofarrivalintheUnitedStates(from8yearsoldto16yearsold)and lengthofexposuretoEnglish(from0yearsto31/2years).Sheusedbothdiscriminationand labelingparadigmstodeterminetheboundaryintheEnglish/pa/and/ba/.Williamsfoundmain effectsforbothlengthofexposureandforageofarrival,withcloserapproximationtothe English+25msec.boundaryassociatedwithlongerexposureandwithyoungerageofarrival. Shealsowasabletoobtainproductionofthesoundsfromthesamesubjects,andshowed throughspectographicanalysisaclosecorrespondencetotheperceptionresults. Oyama(1976)usedamoreglobalmeasureofrateddegreeofaccentednessinthespeechofadult ItalianimmigrantstotheUnitedStates.Hersubjectsvariedinageofarrivalfrom6yearsoldto 20yearsold,andforherlengthofexposure 43

variable,shetooktherangefarexceedingWilliam'srange:5yearsto18years.Oyamafoundthat ageofarrivalwasanimportantpredictorofratedpronunciation,butlengthofexposurehadno predictivepower.ThelatterresultdoesnotconflictwithWilliamsbecauseofthedifferencesin range.Oyama(1978)reportedtheresultsofasentencecomprehensiontaskwiththesame subjects,andreportedessentiallythesameresults. ArecentstudybyPatkowski(1980)lookedatsyntacticabilityinadultsasratedbyjudges. Subjectscamefromavarietyoflanguagebackgrounds.Ageofarrivalrangedfrom6yearsoldto 61yearsold,andalthoughlengthofstaywasnotcontrolledstrictly,allsubjectshadbeeninthe UnitedStatesforlongerthan5years.Patkowskireportsthatageofarrivalwascorrelated.74 withsyntacticrating.NeitherlengthofstaynoramountofformalinstructioninEnglishmadea difference.Thus,inallthreestudies,thereissubstantialsupportforanegativerelationship betweenageanddegreeofsuccessinsecondlanguageacquisition. Onestudy,inwhichtheageofsubjectsrangeddownfrom3yearsoldupto15yearsold,reports apositiverelationshipwithage(Snow&HoefnagelHohle,1978).SubjectswereAmerican childrenlearningDutch,andtheyweretestedonavarietyofmeasuresincludingpronunciation, auditorydiscrimination,andmorphology.Thefactthattheyfoundthatlearningincreasedwith ageisinteresting,especiallyinlightofourearlierdiscussionoftheadvantagesofhavingamore cognitivelymaturesystemtoworkwith.Perhapsupuntilsomewherearounda"critical"period, thereisapositiverelationship.Ontheotherhand,itisentirelypossiblethattheolderchildren wereperformingbetteronthetasksbecausetheyarebettertesttakers. Theevidencepreviouslysummarizedlooksfairlystronginsupportofanegativerelationshipof languagelearningabilitywithage,butthatisonlyabeginning.Itwillberecalledthattherewere manypiecesofevidencesuggestingsimilaritiesbetweenchildandadultsecondlanguage learning.Whataccountsforthesesimilarities?Withoutspecificationofwhatcapacityitisthat deteriorates,statementsregardingageeffectsareboundtobemisleadinganduninterpretable.In addition,therearemanysuccessfuladultsecondlanguagelearners.EveninPatkowski's distribution,whichwasfortunatelypublished,therewereoverlapsinthedistributions.The criticalperiodishardlylikelytobeevenanabsoluteboundary.Furtherlinguisticstudiesthat shedlightonthenatureoftheacquisitionprocesscanhelpsharpentheproblem. LANGUAGEUNIVERSALSASANINTEGRATEDPERSPECTIVE Howarewetounderstandthemazeofdatafromsecondlanguageacquisition,andits relationshiptofirstlanguageacquisition?Inanearlierpaper(Hakuta,1982),Iarguedforthe useofresearchinlanguageuniversalsandlanguagetypologyasaframework. 44

Anyparticularhumanlanguagecanbespecifiedbyitslocationwithinanndimensionalspace thatdefinesthelimitsofvariationofallhumanlanguages.Asapsycholinguist,Iunderstandthe studyoflanguagetypologyandlanguageuniversalstobeanattempttodeterminewhatthe relevantdimensionsare,andtodeterminehowmanymeaningfuldimensionsexist.Iassumeasa workinghypothesisthatthendimensionalspacebearssomerelevancetothefactsofhuman languagelearning.Perhapsthereissomeisomorphismbetweenthendimensionalspacedefined bylanguagetypologistsandthehypothesisspaceofthelanguagelearner. Onaconceptuallevel,thegoaloflanguagetypologistscanbeseenassimilartothefactor analyticpersonalitypsychologist(e.g.,Cattell,1965),withlanguagesbeingusedinsteadof peopleasrandomvariables.Unlikepsychologists,however,linguistsdonotgoaroundasking raterstoaratealargenumberoflanguagesalongvariouslinguisticdimensions.Inaddition,the typesofdata,orscalesofmeasurement,usedinlanguagetypologyaredifferent.Thelinguistic dimensionsaremostfrequentlyconsideredtobecategorical(Greenberg,1978),whereas personalitydimensionsareintervalscaledata,andassumedtobecontinuous.Butthese differencesconstitutedifferencesinstatisticaltreatment.Theworkingassumptionseemstobe thatonceenoughtypologicaldimensionshavebeenformulatedandinvestigatedwithrespecttoa largenumberoflanguages,thesevariablescanbecollapsedthroughstatisticaltechniquestoa smallernumberofunderlyingdimensions,theessenceofhumanlanguage.AsGreenberg recentlyputit,"Atheoreticalanalysisofbasictypologicalconceptshelpsustobroadenour conceptionofcrosslinguisticgeneralizations,whileitsapplicationprovidesauseful methodologyfordiscoveringsuchgeneralizationsatthelowerempiricallevelsandthus providingthematerialsforbroaderanddeeperconclusionsaboutthenatureofhumanlanguage" (p.58). Linguistshavetypicallykeptawayfromprovidinghypothesesastothereasonsfortheexistence ofcrosslinguisticgeneralizations(withnotableexceptionssuchasKuno,1974,andGivon,1979. Forexample,Downing(1978),informulatinguniversalcharacteristicsofrelativeclause structures,writes:"Intheirpresentformtheymayserveasasummaryofobservationsonthe natureofrelativeclausesacrosslanguages,withwhichthedataofadditionallanguagesmaybe compared.Assuchgeneralizationsarerefined,theyaffordanincreasinglysolidempiricalbasis fortheformulationofexplanatoryprinciplesinfunctionalandpsychologicalterms"(p.411). Alongsimilarlines,Steele(1978)formulatesconstraintstoaccountforwordordervariation, suchasthefollowing:"Avariationonbasicwordorderinwhichtheobjectprecedesandthe subjectfollowstheverbistobeavoided"(p.604).Althoughsuchconstraintsservetoexplainat oneleveltheobserveddataonwordordervariation,shewritesinafootnotethat"Iamnot offeringexplanationsfortheconstraints"(p.604,footnote15). Giventhistendency,thedevelopedproductoflanguagetypologicalresearchwillbeessentiallya setoffactorswithloadingsondifferentlinguisticvariables 45

(e.g.,directionofbranching,orderofverbandobject,etc.).Thesenfactorswillconstitutethen dimensionalspaceofhumanlanguages.Myunderstandingoflanguagetypologicalresearchinits presentstateisthatitisnotyetatthisstage,butitisperhapsnottooearlytodiscussatthe abstractlevelthepsychologicalquestion,"Inwhatwaysisthendimensionalspace psychologicallyreal,andhowcanweunderstandchangesintherelationshipunderdifferent psychologicalconditionsoflanguageacquisition?"Iusethetermpsychologicalcorrespondence torefertoanempiricalcorrespondencebetweenthelinguist'sdimensionsandpsychologicaldata obtainedfromstudiesoflanguagelearners.Fordiscussion'ssake,considerthefollowingvariables alongwhichlanguagesareknowntobedistributed: 1. POSITION(postposition/preposition) 2. BRANCHINGDIRECTION(leftbranching/rightbranching) 3. WORDORDERVARIABILITY(rigidwordorder/freewordorder) 4. DUMMYSUBJECT(hasnodummysubjects/hasdummysubjects) 5. OBJECTVERBORDER(verbobject/objectverb) 6. AGREEMENT(hasnosubjectverbagreement/hasagreement) 7. PASSIVIZATION(hasnopassives/haspassives) Table3.2presentsrawdatafor20hypotheticallanguages(fabricatedfrommyimaginationwitha littlehelpfromtheintuitionofseveralcolleaguesaboutsomereallanguages)withvalueson eachofthevariables.Avalueof"0"isenteredwherethelanguageexhibitspropertiesofthefirst levelofthevariable,and"1"isenteredwherethesecondlevelisexhibited.Therelationship betweenthevariablesacrosslanguagescanbeexpressedinacorrelationmatrix,whichappears inTable3.2.AcasualinspectionofTable3.3revealsthattherearemanyvariablesthatarewell correlated.Forexample,OBJECTVERBANDBRANCHINGNESSarecorrelated1.00,a perfectnegativerelationshiprevealingthatallOVlanguagesareleftbranching,andallVO languagesarerightbranching.AGREEMENTiscorrelated.70withDUMMYSUBJECT, indicatingthatlanguageswithsubjectverbagreementalsotendtohavedummysubjects.An underlyingstructureoftheintercorrelationsbetweenthevariablescanberevealedstrikingly throughfactoranalysis,theresultsofwhichappearinTable3.4.FactorIis"saturated"withthe variablesofbranchingness,objectverb,andposition.Thismaybeinterpretedasfollows: languagesthatareleftbranchingtendtobeobjectverbandhavepostpositions,whereasright branchinglanguagestendtobeverbobjectandhaveprepositions.Factor2issaturatedwiththe variablesdummysubject,agreement,andpassive.Theinterpretationisthatlanguageswith dummysubjectsalsotendtohavesubjectverbagreementandpassivization.WhenIfabricated thedataforTable3.1,Ihadinmindtwoclustersofvariablesthathavebeensuggestedinthe literature,onerelatedtothe 46

TABLE3.2 Distributionof20FabricatedLanguageswithRespecttoLanguageVariables. LanguageVariables _____________________________________________________ _________________________________ Word Dummy Object Position Branching Order Subject Verb Agreement Passive Language 1 0 1 1 0 0 0 0 2 1 1 1 0 0 0 0 3 0 1 1 1 0 1 1 4 1 1 1 1 0 1 0 5 0 1 0 0 0 0 0 6 1 1 0 0 0 0 1 7 0 1 0 1 0 1 1 8 1 1 0 1 0 1 1 9 0 0 1 0 1 0 0 10 0 0 1 0 1 0 0 11 0 0 1 1 1 1 1 12 1 1 1 0 0 1 1 13 0 0 1 0 1 0 0 14 1 1 0 1 0 1 1 15 1 1 0 1 0 1 1 16 1 1 1 1 0 1 1 17 0 0 1 0 1 0 0 18 1 1 1 0 0 1 1 19 0 0 1 0 1 0 1 20 0 0 0 1 1 0 1 Variablelabels: Position(0=postposition;1=preposition) Branching(0=leftbranching;1=rightbranching) WordOrder(0=rigidwordorder;1=freewordorder) DummySubject(0=hasnodummysubject;1=has dummysubject) ObjectVerb(0verbobjectorder;1=objectverborder) Agreement(0hasnosubjverbagrmnt;1=hassubjverb agrmnt) Passive(0=hasnopassives;1=haspassives) TABLE3.3 CorrelationMatrixofLanguageVariablesAcross20FabricatedLanguages.

Word Dummy Object Position Branch Order Position 1.00 0.66 0.17 Branching 0.66 1.00 0.31 Wordorder 0.17 0.31 1.00 Dummy 0.19 0.24 0.38 subject ObjectVerb 0.66 1.00 0.31 Agreement 0.50 0.52 0.10 Passive 0.32 0.25 0.38 47 Subj. Verb Agreement Passive 0.19 0.66 0.50 .32 0.24 1.00 0.52 .25 0.38 0.31 0.10 .38 1.00 0.24 0.24 1.00 0.70 0.52 0.53 0.25 0.70 0.52 1.00 0.61 .53 .25 .61 1.00

TABLE3.4 FactorAnalysisSolutionwithVarimaxRotationfor20FabricatedLanguages. Factor 1 Factor 2 Position 0.63 0.26 Branching 0.98 0.19 Wordorder 0.21 0.33 Dummysubject 0.07 0.84 Objectverb 0.98 0.19 Agreement 0.38 0.74 Passive 0.15 0.70 orderofelementsinsentences(e.g.,Greenberg,1963;Lehmann,1973)andtheotherrelatedto thesubjecttopictypologicaldimensionsuggestedbyLiandThompson(1976).Thefactor structuresinTable3.4reflectthesedimensions,althoughIshouldpointoutthat,forpurposesof thepresentpaper,theactualvariablesthatloadonthefactorsareirrelevant.What'simportantis simplythefactthatthisisthekindofwayinwhichtheultimateoutcomeofthecurrentthrustof languagetypologymightberepresented.Insubsequentdiscussionofthefactorstructureof languages,IsimplylabelthefactorsFactorAandB,andtheindividualvariablesthatloadonthe factorsVariables1,2,andsoforth,sothatourdiscussionisunclutteredbythetruthvalueof linguisticstatementsandconcentrateonthelogicofinquiry. INSEARCHOFPSYCHOLOGICALCORRESPONDENCE AsStephenJayGouldpointsoutinhiselegantlywrittenbookonintelligencetesting(Gould, 1981),wehumanconsumersofstatisticshaveaninherentbiastowardsreifyingfactorsderived throughfactoranalysis.Thisisahigherorderbiassimilartothebiasofinferringcausalityfrom correlation,againstwhichwearewarnedrepeatedlyinelementarystatisticsclasses.Language factorsarenomorethanstatementsaboutthedistributionoftheworld'slanguages.Weshouldbe waryofusingobservedlanguagefactorsasexplanationsforpsychologicaldata.Rather,the questionshouldbe"Whataretheprinciplesthatdeterminetheobservedfactorstructure?" Theobservationofthelanguagelearnercanconstrainthepsychologicalplausibilityofthen dimensionalspace.Wecanlookforthepreservationorfragmentationofthefactorstructurein thelanguagelearner,underdifferentcircumstances,i.e.,inL1andL2learners.Ifitcanbe observedinsomecircumstances,butnotothers,thenwemightbeabletoformulatehypotheses aboutitsgoverning 48

principles.Ifweconsiderthendimensionalspacedefinedbylanguagefactorstobeagood candidateasapsychologicallyrealhypothesisaboutthetargetlanguageonthepartofthelangage learner,weexpecttheretobesomecorrespondencebetweenthelanguagefactoranddataobtained fromlanguagelearners.Inthissection,Isketchoutsomeconsiderationsthatmustgointothesearch forpsychologicalcorrespondence. Thetaskforthelearnercanbedefinedasaprocessofdeterminationofthefactorscoreforthe particulartargetlanguage.Havingdeterminedthefactorscore,thelearnercanbeguidedinthe searchfortheparticularrealizationsoftheindividualvariablesthatgowiththefactor.Considerthe situationinTable3.5.AlanguagewithahighpositivescoreonFactorAwillhaveavalueof"1"on Variables13,andavalueof"0"onVariables4and5,asinLANGUAGEX.Alanguagewitha highnegativescoreonFactorAwillhavevaluesof"0"onVariables13,andvaluesof"1"on Variables4and5,asituationreflectedinLANGUAGEZ.Thetwohypotheticallanguagesare mirrorimagesofeachotherwithrespecttoFactorA.Althoughitwouldbehighlyinterestingifall languageswereofthetypesXandZabove,itwouldbedifficulttotestforpsychological correspondence,sincetherewouldbenovarianceacrosslanguages.However,suchasituationis unlikelyandiscertainlyinconsistentwithcurrentknowledgeaboutcrosslinguisticvariation.Then, variationsacrosslanguageswithrespecttotheirlanguagefactorscores(i.e.,theextenttowhichthey reflecttheidealfactorstructure)canbeusedtotestthepsychologicalcoherenceofthefactor.Take forexampleLANGUAGEYinTable3.5,convenientlycreatedforourpurposes.Thevaluesonthe variablesmostlyreflectahighpositiveloadingonFactorA,withtheexceptionofVariable3.The structureofVariable3forLANGUAGEYinfactmatchesthatforLANGUAGEZ,whichisthe ideallanguagewithnegativeloadingonFactorA.Thereareseveralpredictionsthatcanbemade, andempiricallytested,givensuchasituation.OnewouldexpectthatthelearnerofLANGUAGEY wouldhavefewercuesthanthelearnerofLANGUAGEX,duetothemismatchonVariable3.Ifthe determinationofthefactorscoreisapsychologicallyrealprocess,thenonecanpredictdifferences in TABLE3.5 DistributionofValuesonVariablesforThreeHypotheticalLanguages. VAR1 LoadingonFactor A LanguageX LanguageY LanguageZ + VAR2 + VAR3 + VAR4 VAR5 VAR6 0

________________________________________________________ 1 1 1 0 0 1 1 1 0 0 0 1 0 0 0 1 1 1 49

theeaseofacquisitionofstructuresthatreflectvariableswiththesamevaluesforboth languages,assumingthatothersourcesofdifferences,suchasfrequency,canbecontrolled. Thus,forVariables1,2,4,and5,thelearnerofLANGUAGEXisatanadvantageoverthe learnerofLANGUAGEY.InTable3.5,IhavealsoinsertedVariable6,whichhasnoloadingon FactorA.Thismightbeconsideredacontrolvariable,forwhichnodifferencewouldbe predictedbetweenthetwolanguages. IfwehadLANGUAGEZforcomparison,wecanmakefurtherpredictions,sincethevalueson Variable3aresimilarforbothLANGUAGESYandZ.SincethevalueforLANGUAGEZis consonantwiththefactorstructure,whileitisnotforLANGUAGEY,wewouldpredictthatthe structureforVariable3wouldbeeasierforthelearnerofLANGUAGEZthanforthelearnerof LANGUAGEY.Furthermore,wecanmakepredictionsaboutthefrequencyandkindsoferrors thatmightbeexpectedinthecourseoflearning.LearnersofLANGUAGEXwillbelikelyto makeerrorsonstructuresreflectingVariable3thatdeviatetowardthevalueof"1."Thiscanbe comparedtothelikelihoodofsucherrorsforlearnersofLANGUAGEZ. Whetherthepsychologicalcorrespondencecanbedeterminedornotisanempiricalquestion. Ideally,oneshouldbeabletoiteratetheaboveprocessacrosseachofthevariables,finding strategicallylocatedlanguages.Ifwefindthatcertainvariablesconsistentlydonotaffectthe acquisitionofitsrelatedvariables(i.e.,variableswithwhichitisrelatedthroughthefactor structure),wecanweedthemoutfromourmappingofpsychologicalcorrespondence.Theend resultwouldbeapsychologicallyrealhypothesisspaceoflanguagelearners,whichcanbeused inthefurther,andnecessary,investigationsintothenatureofthetaskspecificityandspecies specificityoflanguage. Inthediscussionabove,Ihavesimplisticallyreducedthestudyoflanguagetypologyand universalstofactoranalysis,andglossedovermanyofthetechnicaldifficultiesthatthe researcherwouldencounterinsearchingforpsychologicalcorrespondenceforlanguagefactors.I undertookthisexercisebecauseIwantedtoemphasizetheviewpointtowardfirstandsecond languageacquisitionthatisimplicitinanapproachthatincorporateslanguagetypologyand universals,aviewpointthatIbelieveatpresenttobepotentiallythemostproductive.Insum,in thisapproach,wedonottreateachlinguisticvariableasanisolatedentity,butratherasoneofa constellationofvariablesrelatedtoeachother,concretizedfortheresearcherthroughfactor structures. MycurrentanalogycanbefoundinarecentarticleintheAmericanScientistthatreviewedsome recentresearchindevelopmentalbiology(Tickle,1981).Inparticular,thearticlewasconcerned withsimilaritiesanddifferencesbetweenthegrowthoflimbsduringontogenyontheonehand, andtheregenerationofseveredlimbs(incertainspecies)ontheother.Essentially,theproblems facedbythecellsinthesetwoprocessesaresimilar:howdoesagrowingcellknowwhatpartof thelimbitisultimatelytobecome?Tomakealongstoryshort,there 50

aremarkeddifferencesbetweenthetwo,whichcanbecharacterizedbythedegreetowhich developingcellsaresensitiveto,andinteractwith,positionalspecificationsofneighboringcells. Indevelopment,thepositionalspecificationsaredevelopedincellgeneration,whereasinlimb regeneration,thepositionalspecificationsofthenewgrowthinteractswiththealready establishedpositionalspecificationsofitsneighbors.Thisisaninterestingstatementaboutthe relationshipbetweenthetwoprocessesthatcouldnothavebeenpossiblewithoutasystemfor specifyingposition(itturnsoutthatpositioncanbespecifiedbythreedimensions:anterior posterior,dorsalventral,andproximaldistal,andsomepromisingmechanismsforhowthis informationissignaledhavebeenproposed). Therelationshipbetweenfirstandsecondlanguageacquisitionandlanguagetypologymightbe regardedinasimilarway.Thendimensionalspacehypothesizedbylanguagetypologists,whose psychologicalcorrespondenceisverified,canbecomeatoolsimilartothepositional specificationofthedevelopmentalbiologistconcernedwithlimbgrowth.Itdefinestheproblem, andtheproblemforthelanguageacquisitionresearcherthenbecomestoobserveandexplainthe rolethatthisndimensionalspacemightplayinthedifferentconditionsunderwhichlanguageis learned.Althoughwearestilluncertainastothenatureofthisndimensionalspace,Ibelieve thatitisnottooearlytobeginspeculatingandformulatingourresearchquestionswithrespectto itsmanifestationsunderdifferentpsychologicalcontexts. CONCLUSIONS Conclusionsneednotbestatedsincetheywerelistedintheintroductoryremarks.Ionlyhope thattheevidencepresentedwastantalizingtothenewcomertosecondlanguageacquisition. Therehasalwaysbeenakindofunstatedsnobberyamongpsychologistsandlinguiststoward appliedlinguisticsanditsinterestintheappliedproblemofsecondlanguageacquisition.Some ofitmaybejustified.Appliedlinguistsarenottrainedinconductingsoundresearchfroman empiricalperspective.However,muchofthedenigrationofstudyingsecondlanguagelearners mostlikelycomesfromlackofknowledgeandfromartificialdean'sboundariesintheacademic world.Onceweadmitthatsecondlanguageacquisitioncanshedlightonthenatureofour capacitytolearnlanguage,thedoorswingswideopenforaddingnumerousvariationsthatcan shedlightonthebeast.Ihavearguedthatthestudyoflanguagetypologyanduniversalscanbea workingframeworkaroundwhichLIandL2acquisitioncanbeunified.Theyare,afterall,the sameproblem.Andgiventhelackofprogresswehavemadeinthepastfewyearstoward understandingourabilitytolearnlanguage,whocanturnawaysuchaninterestingsourceof variationontheproblem?JohnMacnamara(1976)wasrightwhenhesaid,"Whenaninfant,a tenyearoldchild,andanadultlearnRussian,themoststrikingoutcomeisRussian"(p.175). 51

ACKNOWLEDGMENT PreparationofthispaperwassupportedinpartbyGrantG810123fromtheNationalInstituteof Education. REFERENCES Adams,M.(1974).Secondlanguageacquisitioninchildren:Astudyinexperimentalmethods: Observationsofspontaneousspeechandcontrolledproductiontests.Unpublishedmaster's thesis,EnglishDepartment,UniversityofCaliforniaatLosAngeles Bailey,N.,MaddenC.,&Krashen,S.(1974).Istherea"naturalsequence"inadultsecond languagelearning?LanguageLearning,21,235243 Beilin,H.(1976).TheCognitiveBasisofLanguageDevelopment.NewYork:AcademicPress Brown,R.(1973).AFirstLanguage:TheEarlyStages.Cambridge,MA:HarvardUniversity Press Butterworth,G.(1972).ASpanishspeakingadolescent'sacquisitionofEnglishsyntax. Unpublishedmaster'sthesis,EnglishDepartment,UniversityofCaliforniaatLosAngeles Cancino,H.(1976).GrammaticalmorphemesinsecondlanguageacquisitionMarta. Unpublishedqualifyingpaper,HarvardGraduateSchoolofEducation Cattell,R.(1965).TheScientificAnalysisofPersonality.Chicago:Aldine Cazden,C.B.(1968).Theacquisitionofnounandverbinflections.ChildDevelopment,39, 433448 Cazden,C.,Cancino,H.,Rosansky,E.&Schumann,J.(1975).Secondlanguageacquisition sequencesinchildren,adolescents,andadults.Cambridge,MA:HarvardGraduateSchoolof Education.(ERICDocumentReproductionServiceNo.ED123873.) Chomsky,C.(1969).Theacquisitionofsyntaxinchildrenfrom5to10.Cambridge,MA:M.I. T.Press Clark,R.(1974).Performingwithoutcompetence.JournalofChildLanguage,I,110 Cohen,A.D.,(1975).Asociolinguisticapproachtobilingualeducation:Experimentsinthe AmericanSouthwest.Rowley,MA:NewburyHouse Corder,S.P.(1971).Idiosyncraticdialectsanderroranalysis.InternationalReviewofApplied Linguistics,9,147160 d'Anglejan,A.,&Tucker,G.R.(1975).TheacquisitionofcomplexEnglishstructuresbyadult learners.LanguageLearning,25,281196 deJ.Villiers,TagerFlusberg,H.,Hakuta,K.,&Cohen,M.(1979).Children'scomprehensionof

relativeclauses.JournalofPsycholinguisticResearch,8,499518 Downing,B.(1978).Someuniversalsofrelativeclausestructure.InJ.H.Greenberg(Ed.), Universalsofhumanlanguage:Vol.4,Syntax.Stanford,CA:StanfordUniversityPress Dulay,H.,&Burt,M.(1973).Shouldweteachchildrensyntax?LanguageLearning,23,245258 Dulay,H.,&Burt,M.(1974a).Naturalsequencesinchildsecondlanguageacquisition. LanguageLearning,24,3753 Dulay,H.,&Burt,M.(1974b).Anewperspectiveonthecreativeconstructionprocessinchild secondlanguageacquisition.WorkingPapersonBilingualism,4,7198 Duskova,L.(1969).Onsourcesoferrorsinforeighlanguages.InternationalReviewofApplied Linguistics,7,1136 ErvinTripp,S.(1974).Issecondlanguagelearninglikethefirst'?TESOLQuarterly,8,111127 Fathman,A.(1975).Languagebackground,age,andtheorderofEnglishstructures.Paper presentedattheTESOLConvention,LosAngeles. 52

Frauenfelder,U.(1974).TheacquisitionofFrenchgenderinTorontoFrenchImmersionschool children.Unpublishedseniorhonorsthesis,UniversityofWashington Gass,S.,&Ard,J.(1980).L2data:Theirrelevanceforlanguageuniversals.Paperpresentedatthe TESOL(TeachersofEnglishtoSpeakersofOtherLanguages)Convention,SanFrancisco Gillis,M.(1975).TheacquisitiontheEnglishverbalsystembytwoJapanesechildreninanatural setting.Unpublishedmaster'sthesis,McGillUniversity Givon,T.(1979).Onunderstandinggrammar.NewYork:AcademicPress Gleitman,L.R.(1981).Maturationaldeterminantsoflanguagegrowth.Cognition,10,103114 Gould,S.J.(1981).Themismeasureofman.NewYork:Norton Greenberg,J.H.(1963).Someuniversalsofgrammarwithparticularreferencetotheorderof meaningfulelements.InJ.J.Greenberg(Ed.),Universalsoflanguage.Cambridge,MA:M.I.T. Press Greenberg,J.H.(1978).Typologyandcrosslinguisticgeneralizations.InJ.H.Greenberg(Ed.), Universalsofhumanlanguage:Vol.1,MethodandTheory,(pp.3360).Stanford,CA:Stanford UniversityPress Hakuta,K.(1974a).Apreliminaryreportonthedevelopmentofgrammaticalmorphemesina JapanesegirllearningEnglishasasecondlanguage.WorkingPapersonBilingualism,3,1843 Hakuta,K.(1974b).Prefabricatedpatternsandtheemergenceofstructureinsecondlanguage acquisition.LanguageLearning,24,287297 Hakuta,K.(1975).Becomingbilingualatagefive:ThestoryofUguisu.Unpublishedseniorhonors thesis,DepartmentofPsychologyandSocialRelations,HarvardUniversity Hakuta,K.(1976).AcasestudyofaJapanesechildlearningEnglish.LanguageLearning,26,321 351 Hakuta,K.(1981).Grammaticaldescriptionversusconfigurationalarrangementinlanguage acquisition:ThecaseofrelativeclausesinJapanese.Cognition,9,197236 Hakuta,K.(1982).Inwhatwaysarelanguageuniversalspsychologicalyreal?Paperpresentedat theconferenceonLanguageUniversalsandSecondLanguageAcquisition,UniversityofSouthern California Hakuta,K.(inpress).Theoreticalissuesandfuturedirectionsinsecondlanguageacquisition research,withspecialreferencetoAsianAmericans.InM.ChuChang(Ed.),Comparativeresearch inbilingualeducation:AsianPacificAmericanperspectives.(pp.3155)NewYork:Teachers CollegePress Hakuta,K.,&Cancino,H.(1977).Trendsinsecondlanguageacquisitionresearch.Harvard EducationalReview,47,294316 Hakuta,K.,deJ.Villiers,&TagerFlusberg,H.(1982).SentencecoordinationinJapaneseand English.JournalofChildLanguage,9,193207 Hatch,E.M.(Ed.)(1978).Secondlanguageacquisition:Abookofreadings.Rowley,MA: NewburyHouse Huang,J.(1971).AChinesechild'sacquisitionofEnglishsyntax.Unpublishedmaster'sthesis, UniversityofCaliforniaatLosAngeles Kang,H.(1982).Thecontextsinwhicherrorsofarticleusageoccurandtheirimplicationfor nativelanguagetransference.Unpublishedsenioressay,DepartmentofPsychology,YaleUniversity

KellerCohen,D.(1979).Systematicityandvariationinthenonnativechild'sacquisitionof conversationalskills.LanguageLearning,29,2744 Kuno,S.(1974).Thepositionofrelativeclausesandconjunction.LinguisticInquiry,5,117136 LarsenFreeman,D.(1976).Anexplanationforthemorphemeacquisitionorderofsecondlanguage learners.LanguageLearning,26,125134 Lehmann,W.(1973).Astructuralprincipleoflanguageanditsimplications.Language,49,4766. 53

Lenneberg,E.H.(1967).Biologicalfoundationsoflanguage.NewYork:Wiley Li,C.,&Thompson,S.(1976).Subjectandtopic:Anewtypologyoflanguage.InC.Li(Ed.), Subjectandtopic.NewYork:AcademicPress Lightbown,P.(1977).FrenchL2learners:Whatthey'retalkingabout.Paperpresentedatthefirst LosAngelesSecondLanguageResearchForum,UCLA Macnamara,J.(1976).Comparisonbetweenfirstandsecondlanguagelearning.DieNeueren Sprachen,2,175188 Maratsos,M.P.,&Chalkley,M.A.(1980).Theinternallanguageofchildren'ssyntax:The ontogenesisandrepresentationofsyntacticcategories.InK.E.Nelson(Ed.),Childlanguage(Vol. 2,pp.127214).NewYork:GardnerPress McLaughlin,B.(1976).Secondlanguageacquisitioninchildhood.Hillsdale,NJ:Lawrence ErlbaumAssociates Nemser,W.(1977).Approximativesystemsofforeignlanguagelearners.InternationalReviewof AppliedLinguistics,9,115123 Oller,J.,&Richards,J.(Eds.).(1973).Focusonthelearner:Pragmaticperspectivesforthe languageteacher.Rowley,MA:NewburyHouse Oyama,S.(1976).Asensitiveperiodfortheacquisitionofanonnativephonologicalsystem. JournalofPsycholinguisticResearch,5,261185 Oyama,S.(1978).Thesensitiveperiodandcomprehensionofspeech.WorkingPaperson Bilingualism,16,117 Patkowski,M.(1980).Thesensitiveperiodfortheacquisitionofsyntaxinasecondlanguage. LanguageLearning,3,449472 Penfield,W.,&Roberts,L.(1959).Speechandbrainmechanisms.Princeton:PrincetonUniversity Press Politzer,R.,&RamirezA.(1973).AnerroranalysisofthespokenEnglishofMexicanAmerican pupilsinabilingualschoolandamonolingualschool.LanguageLearning,23,3961 Rosansky,E.(1976).Secondlanguageacquisitionresearch:Aquestionofmethods.Doctoral dissertation,HarvardGraduateSchoolofEducation Schachter,J.(1974).Anerrorinerroranalysis.LanguageLearning,24,205214 Schlesinger,I.M.(1974).Relationalconceptsunderlyinglanguage.InR.L.Schiefelbusch&L.L. Lloyd(Eds.),Languageperspectives:Acquisition,retardationandintervention(pp.129151). Baltimore:UniversityParkPress Schumann,J.,&Stenson,N.(Eds.).(1975).Newfrontiersinsecondlanguagelearning.Rowley, MA:NewburyHouse Selinker,L.(1972).Interlanguage.InternationalReviewofAppliedLinguistics,10,219231 Selinker,L.,Swain,M.,&Dumas,G.(1975).Theinterlanguagehypothesisextendedtochildren. LanguageLearning,25,139152 SinclairdeZwart,H.(1973).Languageacquisitionandcognitivedevelopment.InT.Moore(Ed.), Cognitivedevelopmentandtheacquisitionoflanguage(pp.925).NewYork:AcademicPress Slobin,D.I.(1973).Cognitiveprerequisitesforthedevelopmentofgrammar.InC.A.Ferguson& D.I.Slobin(Eds.),Studiesofchildlanguagedevelopment.NewYork:Holt,Rinehart&Winston

Snow,C.,&HoefnagelHohle,M.(1978).Thecriticalperiodforlanguageacquisition:Evidence fromsecondlanguagelearning.ChildDevelopment,49,11141128 Steele,S.(1978).Wordordervariation:Atypologicalstudy.InJ.H.Greensbert(Ed.),Universals ofhumanlanguage:Vol.4,Syntax(pp.585623).Stanford,CA:StanfordUniversityPress Svartvik,J.(1973).Errata:Papersinerroranalysis.Lund,Sweden;Gleerup TagerFlusberg,H.,deJ.Villiers,&Hakuta,K.(1982).Thedevelopmentofsentencecoordination. InS.A.Kuczaj(Ed.),Languagedevelopment:Vol.1,Syntaxandsemantics(pp.201243).Hillsdale, NJ:LawrenceErlbaumAssociates. 54

Tickle,C.(1981).Limbregeneration.AmericanScientist,69,634646 Weinreich,U.(1953).Languagesincontact.NewYork:LinguisticCircleofNewYork Williams,L.(1974).Speechperceptionandproductionasafunctionofexposuretoasecond language.Unpublisheddoctoraldissertation,HarvardUniversity WongFillmore,L.(1976).Thesecondtimearound:Cognitiveandsocialstrategiesinsecond languageacquisition.Unpublisheddoctoraldissertation,StanfordUniversity. 55 [Thispageintentionallyleftblank.] 56

4 Bilingualism,LanguageProfiency,andMetalinguisticDevelopment JimCumminsOntarioInstituteforStudiesinEducation By"metalinguisticdevelopment"Irefertoboththedevelopmentofchildren'sawarenessofcertain propertiesoflanguageandtheirabilitytoanalyzelinguisticinput,i.e.,tomakethelanguageforms theobjectsoffocalattentionandtolookatlanguageratherthanthroughittotheintendedmeaning. Earlyinvestigatorsofchildren'slanguagedevelopmenthavespeculatedthataccesstotwolanguages inearlychildhoodmightpromoteanawarenessoflinguisticoperationsandamoreanalytic orientationtolinguisticinput.Vygotsky(1935,1962),forexample,arguedthatbeingableto expressthesamethoughtindifferentlanguagesenablesthechildto"seehislanguageasone particularsystemamongmany,toviewitsphenomenaundermoregeneralcategories,andthisleads toawarenessofhislinguisticoperations"(1962,p.110).Inanearlierworkdirectlyconcernedwith multilingualisminchildren,Vygotsky(1935)suggestedthatwhentheapplicationofsound pedagogicalprinciplesensuredthateachlanguagehadanindependentsphereofinfluence, bilingualismcouldorientthechildtowardmoreabstractthoughtprocesses"fromtheprisonof concretelanguageformsandphenomena"(p.14).Leopold(1949)andImedadze(1960)haveboth arguedonthebasisofobservationalstudiesofchildren'ssimultaneousacquisitionoftwolanguages thatbilingualismcanacceleratetheseparationofnameandobjectandcanfocusthechild'sattention oncertainaspectsoflanguage.Inmorerecentyears,LambertandTucker(1972)suggestedthatthe experimentalgroupintheSt.Lambertbilingualeducationprojecthadlearnedtoengageinaform ofcontrastivelin 57

guisticsbycomparingsimilaritiesanddifferencesintheirtwolanguages. Thus,thehypothesisthataccesstotwolanguagesinearlychildhoodshouldaffectthe developmentofchildren'smetalinguisticskillsappearsplausiblebothtotheoristsofcognitive developmentandtothosewhohavemadedetailedsystematicobservationsofbilingualchildren's languageacquisition.Beforeexaminingtheresearchstudiesthathaveformallycompared bilingualandunilingualchildrenonmetalinguistictasks,itisusefultoplacethetwoconstructs weareconsidering,bilingualismandmetalinguisticdevelopment,intoabroadertheoretical framework.Manyofthelimitationsoftheresearchstudiesandthedifficultyofinterpretingthe resultsderivefromthefactthattheyhavebeencarriedoutwithaverycrudeconceptionofthe natureofbilingualismandinatheoreticalvacuumasfarasmetalinguisticdevelopmentis concerned. THECONSTRUCTOFBILINGUALISM Anextremelylargenumberoftypesofbilingualismandbilinguallearningsituationshavebeen suggested(seeBaetensBeardsmore,1981,forareview)butforpresentpurposes,onlytwo refinementsthatappeardirectlyrelatedtointerpretingresearchontheeffectsofbilingualismare considered.Theserefinementsarecloselyrelatedtooneanotherthefirstdealingwiththe additivesubtractivenatureofthebilingualismandthesecondwiththelevelsofbilingual proficiencyattained.Bothsetsofconstructsderivefromattemptstoreconcileapparently contradictoryfindingsrelatingtotheeffectsofbilingualismoncognition,specifically,the findingfrommanyearlystudies(19201960)thatbilingualsperformedpoorlyonverbal academictaskswhereasmorerecentresearchhasreportedcognitiveandlinguisticadvantagesfor bilinguals(seeCummins,1978a,foradetailedreview).Inaddition,adistinctionbetweenshort termandlongtermeffectsofbilingualexperiencesismade. AdditiveandSubtractiveBilingualism Lambert(1975)haspointedoutthatalargemajorityofearlystudieswerecarriedoutwith immigrantorminoritylanguagechildrenwhosefirstlanguage(L1)wasgraduallybeingreplaced byamoredominantandprestigioussecondlanguage(L2).Hetermstheresultingformof bilingualismsubtractiveinthatbilingualchildren'sproficiencyintheirtwolanguagesatany pointintimeislikelytoreflectsomestageinthesubtractionofL1anditsreplacementbyL2. Lambertcontraststhesubtractivebilingualismofmanyminoritylanguage 58

childrenwiththeadditivebilingualismgenerallyachievedbychildrenwhoseLIisdominantand prestigiousandinnodangerofreplacementbyL2.Thisisthesituationofanglophonechildren inFrenchimmersionprograms.Thebilingualismofthesechildrenistermedadditivebecause anothersociallyrelevantlanguageisbeingaddedtothebilingual'srepertoryofskillsatnocost toproficiencyinL1.Themajorityofstudiesreportingcognitiveadvantagesassociatedwith bilingualismhavebeencarriedoutincontextswherechildrenhaveattainedanadditiveformof bilingualism,thatis,relativelyhighlevelsofproficiencyinbothlanguages. Thisanalysissuggeststhatthelevelofproficiencybilingualchildrenachieveintheirtwo languagesmaybeanimportantfactorindeterminingthecognitiveeffectsofbilingualism.This ideaiselaboratedinthe"threshold"hypothesis(Cummins,1976). TheThresholdHypothesis Thethresholdhypothesisproposesthattheremaybethresholdlevelsoflinguisticproficiency thatbilingualchildrenmustattaininordertoavoidcognitivedisadvantagesandtoallowthe potentiallybeneficialaspectsofbecoming FIGURE4.1.CognitiveEffectsofDifferentTypesofBilingualism(adaptedfromToukomaa& SkutnaabKangas,1977). 59

bilingualtoinfluencecognitivegrowth.Thishypothesisassumesthatthoseaspectsof bilingualismthatmightpositivelyinfluencecognitivegrowthareunlikelytoexertasignificant longtermeffectuntilthechildhasattainedacertainminimumorthresholdlevelofproficiency inbothlanguages.Conversely,ifbilingualchildrenattainonlyaverylowlevelofproficiencyin L2orL1,therangeofpotentialinteractionwiththeenvironmentthroughthatlanguageislikely tobelimited. Theformofthethresholdhypothesisthatseemstobemostconsistentwiththeavailabledatais thatthereisnotone,buttwo,thresholds(Cummins,1976;Toukomaa&SkutnabbKangas,1977 ).Theattainmentofalowerthresholdlevelofbilingualproficiencywouldbesufficienttoavoid anynegativecognitiveeffects,buttheattainmentofasecond,higherlevelofbilingual proficiencymaybenecessarytoleadtolongtermcognitivebenefits.Essentially,thehypothesis ofalowerthresholdlevelofbilingualproficiencyproposesthatbilingualchildren'sproficiency inalanguagecanbesufficientlyweaktoimpairthequalityoftheirinteractionwiththeir educationalenvironmentthroughthatlanguage.Thus,theearlyfindingsofnegativecognitive effectsassociatedwithbilingualismwouldbeexplainedbythefactthattheminoritylanguage childreninthesestudiesoftenfailedtodevelopasufficientlyhighlevelofproficiencyinthe schoollanguagetobenefitfullyfromtheireducationalexperience. Therapidcognitiveandacademicprogressofbilingualsreportedinrecentstudiesisexplained bythefactthatthesechildrenhadattainedtheupperthresholdlevelofbilingualproficiency,that is,highlevelsofbothL1andL2skills.Thethresholdhypothesiswouldpredictthatneither positivenornegativecognitiveeffectswouldresultfromdominantbilingualismwherechildren developnativelikeproficiencyintheirdominantlanguagebutachieveonlyintermediatelevels ofproficiencyintheirweakerlanguage. Notehere,however,thattherelevanceofthishypothesisforinterpretingthedataregardingthe effectsofbilingualismonmetalinguisticdevelopmentrepresentsonlyarelativelycrude refinementtotheconstructofbilingualismbecausethenatureofthelanguageproficiency involvedinthethresholdsisleftlargelyundefined.Theissueofwhatconstituteslanguage proficiencyandmetalinguisticproficiencyisconsideredinalatersection. ShortTermandLongTermEffectsofBilingualism Thethresholdhypothesisisprimarilyrelevantonlytolongtermeffectsofbilingualism,i.e., effectsthatresultfromactuallyfunctioningbilinguallyoveraprolongedperiodoftime. ProficiencyinL1andorL2becomesimportantforcognitivedevelopmentonlywhenitis necessarytouseL1andorL2tointeractwiththeenvironmentoveraprolongedperiod.This typeofprolonged 60

bilingualinteractionwillresultinrelativelyfixedpatternsoforientationorproficiency. However,itispossibletodistinguishtheselongtermeffectsofbilingualfunctioningfromshort termeffectsthatmayresultfromthesecondlanguagelearningexperienceitself.Thedefining characteristicofashorttermeffectisthatitresultsfromtheprocessofacquiringproficiencyinL2 anditdoesnoteffectafundamentalorpermanentchangeinanindividual'scognitivefunctioningor development.Whetherornotshorttermeffectswillexertamorepermanentlongterminfluence maydependonthetypeofbilingualproficiencyattained.Thedistinctionbetweenshorttermand longtermeffectsisillustratedwithreferencetotheresearchfindingstobeconsideredinalater section. Insummary,Ihavesuggestedsomerefinementstotheconstructofbilingualismthatmaybehelpful ininterpretingtheresearchfindings.Theserefinementscomefromexaminingpatternsof associationsthatemergeinstudiesrelatingbilingualismandcognition.Whatissuggestedbythese hypothesesisthatthelanguageacquisitioncontextwilleffectthetypeofbilingualismattained, whichinturnwillstronglyinfluencethelongtermconsequencesofthatbilingualismforoverall cognitiveandlinguisticfunctioning.However,moretransienteffectsoncognitivefunctioningmay resultfromtheactualsecondlanguageacquisitionprocess. Obviouslyinanydiscussionoftheeffectsofbilingualproficiencythereissomeimplicitnotionof languageproficiency.Thetheoreticalframeworkpresentednextisintendedtopermittheconstructs oflanguageproficiencyandmetalinguisticskillstoberelatedtooverallcognitiveandacademic functioning. THECONSTRUCTSOFLINGUISTICANDMETALINGUISTICPROFICIENCY Muchoftheconfusionsurroundingthemeasurementofbilingualproficiencyandtheeffectsof bilingualeducationonchildren'sdevelopment(e.g.,Baker&deKantor,1981)derivesfroma failuretoadequatelyconceptualizethenatureoflanguageproficiencyanditscrosslingual dimensions(Cummins,1981).Similarly,theconstructofmetalinguisticdevelopmenthasnotbeen integratedintoanoverallframeworkofcognitiveandacademicfunctioning.Theframework outlinedhasbeenappliedtoavarietyofissues(e.g.,secondlanguageacquisition,measurementof languageproficiencyandcommunicativecompetence,languagepedagogy,learningdisabilityand psychologicalassessment,andthenatureofbilingualproficiency)(Cummins,1983). TheframeworkpresentedinFig.4.2proposesthatlanguageproficiencycanbeconceptualized alongtwocontinua.Firstisacontinuumrelatingtotherangeofcontextualsupportavailablefor expressingorreceivingmeaning.Theex 61

FIGURE4.2RangeofContextualSupportandDegreeofCongnitiveInvolvementin CommunicativeActivites.

tremesofthiscontinuumaredescribedintermsofcontextembeddedversuscontextreduced communication.Theyaredistinguishedbythefactthatincontextembeddedcommunicationthe participantscanactivelynegotiatemeaning(e.g.,byprovidingfeedbackthatthemessagehasnot beenunderstood)andthelanguageissupportedbyawiderangeofmeaningfulparalinguisticand situationalcues;contextreducedcommunication,ontheotherhand,reliesprimarily(oratthe extremeofthecontinuum,exclusively)onlinguisticcuestomeaningandmayinsomecasesinvolve suspendingknowledgeofthe"real"worldinordertointerpret(ormanipulate)thelogicofthe communicationappropriately. Ingeneral,contextembeddedcommunicationderivesfrominterpersonalinvolvementinashared realitythatobviatestheneedforexplicitlinguisticelaborationofthemessage.Contextreduced communication,ontheotherhand,derivesfromthefactthatthissharedrealitycannotbeassumed, andthuslinguisticmessagesmustbeelaboratedpreciselyandexplicitlysothattheriskof misinterpretationisminimized.Itisimportanttoemphasizethatthisisacontinuumandnota dichotomy.Thus,examplesofcommunicativebehaviorsgoingfromlefttorightalongthe continuummightbeengaginginadiscussion,writingalettertoaclosefriend,writing(orreading) anacademicarticle.Clearly,contextembeddedcommunicationismoretypicaloftheeveryday worldoutsidethe 62

classroom,whereasmanyofthelinguisticdemandsoftheclassroomreflectcommunicationthatis closertothecontextreducedendofthecontinuum. Theverticalcontinuumisintendedtoaddressthedevelopmentalaspectsofcommunicative proficiencyintermsofthedegreeofactivecognitiveinvolvementinthetaskofactivity.Cognitive involvementcanbeconceptualizedintermsoftheamountofinformationthatmustbeprocessed simultaneouslyorinclosesuccessionbytheindividualinordertocarryouttheactivity. Thiscontinuumincorporatesadevelopmentalperspectiveinthatsomelanguagesubskills(e.g.,L1 phonologyandsyntax)quicklyreachplateaulevelsatwhichtherearenolongersignificant differencesinmasterybetweenindividuals(atleaseincontextembeddedsituations).Other subskillscontinuetodevelopthroughouttheschoolyearsandbeyond,dependingontheindividual's communicativeneedsinparticularculturalandinstitutionalmilieu. Theupperpartsoftheverticalcontinuumconsistofcommunicativetasksandactivitiesinwhichthe linguistictoolshavebecomelargelyautomatized(mastered)andthusrequirelittleactivecognitive involvementforappropriateperformance.Atthelowerendofthecontinuumaretasksandactivities inwhichthecommunicativetoolshavenotbecomeautomatizedandthereforerequireactive cognitiveinvolvement.Persuadinganotherindividualthatyourpointofviewratherthanhersorhis iscorrect,orwritinganessayonacomplexthemeareexamplesofsuchactivities.Inthese situations,itisnecessarytostretchone'slinguisticresourcestothelimitinordertoachieveone's communicativegoals.Cognitiveinvolvement,inthesenseofamountofinformationprocessing,can bejustasintenseincontextembeddedasincontextreducedactivities. Asmasteryisdeveloped,specificlinguistictasksandskillstravelfromthebottomtowardthetopof theverticalcontinuum.Theretendstobeahighlevelofcognitiveinvolvementintaskoractivity performanceuntilmasteryhasbeenachievedor,alternately,untilaplateaulevelatlessthanmastery levelshasbeenreached(e.g.,L2pronunciationinmanyadultimmigrants,"fossilization"ofcertain grammaticalfeaturesamongFrenchimmersionstudents,etc.).Thus,acquiringthephonologyand syntaxofL1requiresconsiderablecognitiveinvolvementforthetwoandthreeyearoldchild,and thereforethesetaskswouldbeplacedinquadrantB(contextembedded,cognitivelydemanding). However,asmasteryoftheseskillsdevelops,tasksinvolvingthemwouldmovefromquadrantBto quadrantAbecauseperformancebecomesincreasinglyautomatizedandcognitivelyundemanding. Inasecondlanguagecontextthesametypeofdevelopmentalprogressionoccurs. TheframeworkisrelatedtotheoreticalconstructselaboratedbyDonaldson,1978(embedded disembeddedthought),Olson,1977(utterancetext),Bruner(communicativeanalyticcompetence), BereiterandScardamelia,1982(conversationcomposition).Inaddition,depthofcognitive processingisaconstructemployedbymanypsychologistsandappliedlinguists(e.g.,Oller,1980). 63

Wherethepresentframeworkdiffersfromothersisinpositingtwoindependentcontinuaandin applyingthemtointerpreting(andsynthesizing)datafromawidevarietyofresearchareas. Bydefinitionmetalinguisticskillsareclearlyattheextremecontextreducedendofthe horizontalcontinuum.However,thelocationofanyparticularmetalinguistictaskonthevertical continuumwilldependonavarietyoffactors,primarilydegreeofdifficultyoftaskinrelationto child'scognitivedevelopmentallevelanddegreeofpracticeorexposuretosimilartasksor situations.Thus,intermsofthedefinitionofmetalinguisticdevelopment(providedatthe beginningofthechapter)itisclearthattaskscanrequiremetalinguisticskills(insofarasthey requireindividualstolookatlanguageratherthanthroughit)atverydifferentcognitivelevels. Boththesedimensions(cognitivelevelanddegreeofcontextreduction)mustbekeptinmind whenconsideringtherelationshipsamongmetalinguistictasksorbetweenmetalinguisticand cognitiveacademictasks.BasedonthelocationinquadrantDofmetalinguistictasksonwhich degreeofmasteryvariesamongstudents,wewouldexpectstrongrelationshipsbetweensuch tasksandotherquadrantDtasks(e.g.,readingproficiency,verbalintellectualskills,etc.).There isconsiderableevidenceforsuchrelationships(seeRyan&Ledger,1978). BilingualismandMetalinguisticSkills:ResearchEvidence Thereisaconsiderablebodyofresearch(seeCummins,1981)whichshowsthatcognitive academic(quadrantD)skillstransferacrosslanguagesgivensufficientexposureandmotivation toacquireboth.Thus,inFrenchimmersionprograms,Englishbackgroundstudentswhohave notbeenexposedtoEnglishlanguageinstructionpriortoGrade2or3catchuptoequivalent studentsinallEnglishprogramswithinarelativelyshorttimeoftheintroductionofformal Englishlanguagearts.Thistransferofunderlyingcognitiveacademicproficienciesoccurseven acrosslanguagesthathavenosurfacesimilarities,forexample,JapaneseandEnglish(Cummins etal.,1984). Animplicationofthisgeneralfindingisthatmetalinguisticskillswillalsotransferacross languagesand,infact,thepresenceanduseoftwocodesmaypromptgreatermonitoringand inspectionofeachsuchthatmetalinguisticawarenessisenhanced.Manyoftheresearchfindings suggestthatbilingualismmayexertsomesubtleeffectsonchildren'sorientationtolinguisticand perceptualstructures. Oneofthefirststudiestoreportdatarelatingtobilingualismandmetalinguisticdevelopmentis thatconductedbyFeldmanandShen(1971)whofoundthatfiveyearoldHeadStartbilingual childrenweresuperiortounilingualsin 64

theirabilitytoswitchnamesandintheiruseofcommonnamesandnonsensenamesinrelational statements.Bilingualandunilingualgroups,however,werenotmatchedonIQorothercognitive measuresandthustheresultsmustbeconsideredtentative. IancoWorrall(1972),inastudyconductedinSouthAfrica,reportedthatbilingualchildren, broughtupinaoneperson,onelanguagehomeenvironment,weresignificantlymoresensitivethan unilingualchildrenmatchedonIQ,tosemanticrelationsbetweenwordsandwerealsomore advancedinrealizingthearbitraryassignmentofnamestoreferents.Unilingualchildrenweremore likelytointerpretsimilaritybetweenwordsintermsofanacousticratherthanasemanticdimension (e.g.,capcanratherthancaphat)andfeltthatthenamesofobjectscouldnotbeinterchanged. ThefactthatinlancoWorrall'sstudybilingualsagreedmoreoftenthatnamesofobjectscouldbe interchangediscapableofanotherinterpretation.Therewerethreepairsofnames:dog,cow;chair, jam;book,water.Childrenwerefirstasked,"couldyoucalladog'cow'andacow'dog'?"High schoolchildrentendedtoanswernotothisquestionbecausetheyfeltthereweresocialand linguisticconventionsregardingnamesthatcouldnotbebroken,thereforeasecondquestionwas added.Thisquestionwas,"supposeyouweremakingupnamesforthings,couldyouthencallacow 'dog'andadog'cow'?"Highschoolchildreninvariablyagreedthat,inprinciple,thiscouldbedone. IancoWorrall'sassertionthatbilingualsweremoreawareofthearbitrarynatureofwordreferent relationshipsisbasedonthefactthatatboththe46and79agelevelsasignificantlyhigher proportionofbilingualsfellintothis"noyes"category.However,thereisevidence(seeCummins, 1978a,forareview)thatbilingualchildrenaremoresensitivetofeedbackcues.Thechangeinthe formofthequestionmayhaveprovidedcuestochildrenthatwouldcausethemtochangetheir responsefromnotoyesandbilingualchildrenmayhavebeenmoresensitivetothesecues.Thus, becausechildrenwerenotrequiredtojustifytheirresponses,IancoWorrall'sfindingsare inconclusiveastowhetherthebilingualchildreninhersamplewereinfactmoreawareofthe arbitrarynatureofwordreferentrelationshipsorwhethertheywerejustmoresensitivetofeedback cues. Examplesofbothshorttermandlongtermeffectsofbilingualismaswellasthepossible consequencesoflevelofbilingualproficiencycomefromastudybyCumminsandMulcahy(1978) inthecontextoftheUkrainianEnglishbilingualprogram(50%Ukrainian,K6)inEdmonton, Alberta.Threegroupsofchildrenwerecomparedonavarietyoflanguageprocessingtasks.One groupofchildrenintheprogramspokeextensiveUkrainianathomeandwerejudgedbytheir teacherstoberelativelyfluentinUkrainian(BilingualGroup).Asecondgroupspokelittle UkrainianathomeandwerejudgedbyteacherstohaverelativelylittlefluencyinUkrainian(L2 Learners).Thethirdgroupwasaunilingualgroupattendingregularclassesinthesameschoolsas thebilingualprogramstudents. 65

AllthreegroupswerematchedonnonverbalIQandSESandwerealsoequivalentinEnglish readingskills.AmongthemeasuresadministeredweretheSemanticPhoneticPreferenceTest,a WordAssociationTest,andanAmbiguitiesTest. TheSemanticPhoneticPreferenceTestwasdevelopedbyPurbhooandShapson(1975)onthe basisofasimilartaskusedbyIancoWorrall(1972).Childrenwereasked16questionsofthe type"WhichwordismorelikeARM,isitHAND,orisitART?"Semanticandphoneticchoices occupiedeachposition50%ofthetime.Childrenwereclassifiedassemanticallyorientedifthey madeatleast12semanticchoices.TheWordAssociationTestconsistedof20stimuluswords readtoeachsubjectindividually.Subjectswererequiredtogiveoneassociationforeachstimulus wordandtheirresponsesweretaperecordedandlatertimed.TheAmbiguitiestestwastaken fromKessel(1970)andassessedchildren'ssensitivitytolexical,surfacestructureand underlyingstructureambiguities.Thetestconsistedoffouritemsofeachtyperandomlyordered. Childrenwereshownfourlinedrawings,twoofwhichdepicteddifferentmeaningsofasentence thatwasreadaloudbytheexperimenter.Thefollowingareexamplesofeachtypeofambiguous sentence:L"Theboypickedupthebat"(baseballbat,flyingbat);SS"Hetoldherbaby stories."US"Shehitthemanwithglasses."Childrenweregivenascoreof2iftheygavethe twomeaningsandcorrectjustificationsspontaneously,1iftheygavethetwomeaningsand justificationsafterprobing,and0iftheychosethewrongpicturesoronlyonecorrectpictureor iftheirjustificationsshowedtheyhadnotgraspedthetwomeanings. TheresultsoftheSemanticPhonetictestarenotconsistentwithIancoWorrall'sfindingthat bilingualchildrenweremoresemanticallyorientedthanunilingualchildren.Infact,attheGrade 1level,theL2learnersweresignificantlymoreacoustically(andlesssemantically)orientedthan theunilingualgroup.ThisdifferencedisappearsbyGrade3andisprobablyduetothenecessity to"trainone'sear"andpayattentiontophoneticsimilaritiesanddissimilaritiesintheinitial stagesofthebilingualprogram.Itwouldthusbeclassifiedasashorttermeffectofbilingual experiences. ThepatternofresultsontheAmbiguitiestestprovidesdirectsupportforthethresholdhypothesis inthatfluentbilingualsperformedbetterthanL2learnerswho,inturn,performedbetterthan unilinguals. ConsiderationoftheSemanticPhoneticPreferenceresults,inrelationtotheAmbiguitiesresults, suggestsageneralhypothesisregardingthedevelopmentofananalyticorientationtolanguage amongbilingualchildren.Realizationofsentenceambiguitiesinvolvesanalysisofbothsemantic andsyntacticaspectsofthesentenceandthusthisresultcouldbeinterpretedasevidenceof greatersemanticorientationamongbilinguals.However,thisorientationdoesnotemergeonthe SemanticPhoneticPreferencetaskbecauseofthenecessityforaforcedchoicebetweensemantic andphoneticsimilaritiesbetweenwords.An 66

equallystrongcasecanbemadethatbilingualismshouldleadtogreatersensitivitytophonetic patternsandseveralstudiessupportthisgeneralhypothesis(see,e.g.,Cohen,Tucker,&Lambert, 1967).ThegreaterphoneticorientationoftheGrade1L2learnerssuggeststhatintheinitial stagesofexposuretothesecondlanguage,phoneticfeaturesofwordsaremoresalientthan semanticfeatures.Infact,phoneticdiscriminationisprobablyanecessaryconditionforsemantic understanding.Withincreasingcompetenceinthesecondlanguage,phoneticdiscrimination becomeslessproblematicandsemanticandsyntacticanalysismorenecessarytounderstandthe linguisticinput. BecauseIancoWorrall'ssamplehadbeenexposedtobothlanguagesinthehomesinceinfancy, phoneticfeaturesoftheirlanguageswerenolongerproblematic,whereasthesamemaynothave beentrueformanyofthefluentbilingualandcertainlynotforthenonfluentGradeIbilingual childrenintheUkrainianstudy. IncontrasttotheSemanticPhonetictestresults,thedifferencesbetweentheBilingualandother groupsinTable4.2representlongtermeffectssincethebilingualchildrenhavebeenfunctioning bilinguallyforaconsiderableperiodoftime.Thefindingthatthebilingualsatbothgradelevels takelongertorespondonthewordassociationtaskagreeswithBenZeev's(1977a)findingand suggeststhatbilingualsmayprocessthesemanticinformationmoredeeplythantheothergroups. Itisalsopossiblethatgenerationofaresponseismoredifficultforthebilingualsasaresultof linguisticinterference(Torrance,Gowan,Wu,&Aliotti,1970). lancoWorrall'sfindingthatbilingualsweremoreawareofthearbitrarynatureoftheword referentrelationshipswasfurtherinvestigatedinstudiesconductedwithIrishEnglishbilingual children(Cummins,1978a)andwiththeUkrainianEnglishbilingualgroup(Cummins& Mulcahy,1978).UnlikeIancoWorrall'sprocedure,childrenwererequiredtojustifytheir responses,andjustificationsratherthantheactualresponseswerescoredcorrectorincorrect.In theIrishstudythereweresignificantdifferencesbetweenbilingualandunilingualgroupsinfavor ofthebilingualsatbothGrades3and6,buttheUkrainianstudyfoundnogroupdifferences eitheratGrade1orGrade3.Theequivocalnatureofthefindingsmaybeareflectionofthe relativecrudenessofthemeasurementinstruments.Thephenomenonofmetalinguistic awarenessisstillinadequatelyunderstoodandtheliteratureisdevoidofinstrumentswhose constructvalidityhasbeendemonstrated. TheIrishstudyalsoreportedthatGrades3and6bilingualchildrenwerebetterablethan unilingualchildrenmatchedforIQ,SES,andagetoevaluatenonempiricalcontradictory statements;e.g.,"Thepokerchip(hidden)inmyhandisblueanditisnotblueTrue,Falseor Can'tTell?"Childrenwererequiredtojustifytheirresponsesand,asintheArbitrarinessof Languagetask,justificationsratherthanactualresponseswerescoredcorrectorincorrect.These 67

findingsareconsistentwiththefindingsoftheUkrainianstudyandsuggestthatbilingualismcan promoteananalyticorientationtolinguisticinput. TheresultsoftwostudiesconductedbyBenZeev(1977a,1977b)arealsoconsistentwiththis hypothesis.BenZeevarguesthatinordertoovercomeinterlingualinterference,bilinguals developstrategiesoflinguisticprocessingthatcanpromotecognitivegrowth.Sheproposesfour differentmechanismsbymeansofwhichthebilingualchildattemptstoresolvetheinterference betweenhislanguages.Thesemechanismsare(a)analysisoflanguage;(b)sensitivityto feedbackcues;(c)maximizationofstructuraldifferencesbetweenlanguages;and(d) neutralizationofstructurewithinalanguage.ThefirsttwomechanismsaresupportedbyBen Zeev'sstudiesofmiddleclassHebrewEnglishbilingualsandlowerclassSpanishEnglish bilinguals.InbothstudiesIQdifferencesbetweenbilingualandunilingualgroupswere controlled. BenZeevusedasymbolsubstitutiontasktoinvestigatechildren'sabilitytoplaywithwords.For example,childrenwereasked,"Howdowesay'Theyaregoodchildren'?"substituting "Spaghetti"for"They."TheHebrewEnglishbilingualsweresignificantlysuperioronthistask thantheirunilingualcontrols.TherewerenodifferencesbetweentheSpanishEnglishbilinguals andtheircontrolsonthistask.However,BenZeevreportsthattheSpanishEnglishbilinguals madesignificantlyfewererrorsofaglobal,primitivetype,i.e.,simplyutteringthesubstitute wordinplaceoftheentiresentence.Thistypeoferror,sheargues,isindicativeofinabilityto treatthesentenceanalytically.Thisinterpretationshouldbetreatedcautiouslybecausethe symbolsubstitutionissubjecttothesamelimitationsasmanyofthemetalinguistictasksusedin otherstudies,i.e.,uncertainconstructvalidity. AlthoughBenZeevreportsthatbilingualswerebetterabletotreatsentencestructure analytically,inbothstudiestheyhadsignificantlylowervocabularyscoresandtheSpanish Englishbilingualsmadesignificantlymoregrammaticalmistakesonastorytellingtask.Doyle, Champagne,andSegalowitz(1977)havealsoreportedlowerlevelsofvocabularyamong preschoolbilingualchildreninMontreal.BenZeevsuggeststhattherelativelackofexperience witheachlanguageprobablyhassomelimitingeffectonbothvocabularyandknowledgeof standardgrammaticalrules. BenZeevinterpretsfindingsofbilingualunilingualdifferencesonseveralaspectsofnonverbal tasksasevidenceforthegeneralizationofbilinguals'analyticstrategytowardlanguagestoother kindsofstructures.TheresponsesoftheSpanishEnglishbilingualswerecharacterizedby attentiontostructureonclassification,matrixtranspositionandpicturecompletiontasks.Inthe HebrewEnglishstudythesamestrategywasevidentonthematrixtranspositiontask.Onmany ofthetasksintheBenZeevstudies,theoverallperformanceofbilingualandunilingualgroups didnotdifferbuttheresponsestrategiesofthebilingualswerecharacterizedbyattentionto structureandreadinesstoreorganizecognitiveschemata. 68

Superiorperformancebybilingualsontasksinvolvinganabilitytorestructurelinguisticand perceptualschematahasalsobeenreportedbyBalkan(1970).InastudyconductedinSwitzerland, Balkanmatchedbalancedbilingualsandunilingualsonnonverbalintelligenceandfoundthatthe bilingualgroupperformedsignificantlybetterontwovariablesthatheclaimsmeasurecognitive flexibility.OneofthesetestswassimilartotheEmbeddedFiguresTest,andinvolvedanabilityto restructureaperceptualsituation(FiguresCaches).Theothertestrequiredasensitivitytothe differentmeaningsofwords(Histoires).Balkanalsodividedhisbilingualgroupintoearly(those wholearnedtheirsecondlanguagebeforetheageoffour)andlate(thosewholearnedtheirsecond languagebetweenfourandeight)bilingualsandfoundthatthesuperiorityoftheearlybilinguals overtheirmatchedunilingualcounterpartsonthesetestswasmuchmorepronouncedthanthe superiorityofthelatebilinguals.Balkansuggeststhatthehabitofswitchingfromonelanguageto anotherleadstoagreaterdegreeofcognitiveflexibilityinbilingualchildren.Itisworthnotingthat Bruck,Lambert,andTucker(1978)foundlargedifferencesbetweenexperimentalandcontrol groupsintheSt.LambertprojectattheGrade6levelontheEmbeddedFiguresTest. Starck,Genesee,Lambert,andSeitz(1977)testedthehypothesisthatchildrenwhohavehada linguisticallyenrichedschoolingmaydevelopamorelateralizedverbalprocessingsystemthan comparablechildrenwithoutthatformoflinguisticenrichment.Starcketal.comparednative speakersofEnglishfromKindergarten,andGradesIand2attendingaHebrewFrenchimmersion schoolwithunilingualEnglishspeakerswhoseinstructionwastotallyinEnglishonadichotic listeningmeasureofearasymmetry.GroupswerematchedonIQandSES.Inthefirststudy,the trilingualgroupdemonstratedarightearadvantageonbothaccuracyandorderofrecallwhereas theunilingualgroupdidnot.Thesecondstudyfailedtoreplicatethefirstinthatthetrilingualsdid notshowagreaterrightearadvantagethantheunilinguals.However,theirrecallwassignificantly moreaccuratethanthatoftheunilinguals.Starcketal.concludethat"biormultilingual experiencesdoseemtohelpestablishmorereliableearasymmetryeffects,whicharethoughtto reflectenhancedcerebralasymmetry"(pp.5354).Therelationshipbetweenthisfindingand findingsofgreaterabilitytoanalyzestructureamongbilingualsisthatrightearadvantageon dichoticlisteningtasksreflectsgreaterdevelopmentofthemoreanalyticlefthemispherefunctions incomparisontorighthemispherefunctions. Severalotherstudiesaddtotheevidencethatbilingualismpromotessomesubtleformsof metalinguisticawareness.DiazandHakuta(1981)comparedagroupofSpanishEnglishbalanced bilingualchildrentoagroupofSpanishspeakingchildrenwhowerejustbeginningtolearnEnglish asasecondlanguageatschool;thecomparisongroupcouldbeconsidered,therefore,relatively monolingualchildrenwhowereatbeginningstagesofsecondlanguagelearning.Thetwogroupsof childrenwereequivalentintheirSpanishability,livedinthesame 69

neighborhoods,andattendedthesamekindergartenandfirstgradebilingualclasses. ThemetalinguisticawarenesstasksconsistedofeightungrammaticalSpanishsentencesandeight SpanishsentenceswithoneEnglishwordineach(e.g.,LateacherestaenlaclaseorE1doges grande);severalcorrectSpanishsentenceswereintermixedwithineachsetofwrongsentences.For thefirstsetofsentences,childrenwereaskedtogiveacorrectorgrammaticalversionofthe sentencespresented.Theresultsshowednodifferencesbetweenthetwogroupsofchildrenintheir abilitytodetectgrammaticalerrorsintheirnativelanguage.However,balancedbilingualsshoweda greaterabilitytomakegrammaticalcorrectionsandtodetectconfusionsbetweentheirtwo languages.DiazandHakutaconcludethatthebalancedbilingualchildrenshowedanawarenessof theindependenceandappropriateseparateusageoftheirtwolanguages. Mohanty(1982)carriedouttwostudiesinIndiathatsupportthenotionthatbilingualismconfers metalinguistic(andgeneralcognitive)advantagesonchildren.Thefirststudyexaminedthe metalinguisticdevelopmentof300Grades6,8,and10bilingualandunilingualchildrenfromurban andtribalculturesmatchedforSES.Asignificantmaineffectwasfoundforbilingualismandan interactionbetweenlanguageandculturewasalsofoundindicatingthatbilingualismhada somewhatmorepositiveeffectonmetalinguisticskillsinurbanculturethanamongtribals. Thesecondstudywascarriedoutamong180Grades6,8,and10tribals,someofwhomwere bilingualinKuiandOriyaandothersunilingualinOriya.Detailedsocioculturalbackground analysisrevealedthatculturally,racially,andintermsofsocioeconomicdifferencestheKondsare verysimilarregardlessoflanguagedifferences.Mohanty(1982)foundthat"incomparisontothe unilingualcontrolstheKuiOriyabilingualKondchildrenperformbetterintestsofcognitive informationprocessingability,haveamoreobjectiveandanalyticorientationtolanguage,and demonstratealeveloflinguisticdevelopmentintheirsecondlanguageatparwiththedevelopment oftheonlylanguageoftheunilinguals"(p.7). Insummary,thestudiesdiscussedinthissectiontendtosupportthehypothesisthatbilingualism promotesananalyticorientationtobothlinguisticandperceptualstructures.However,theevidence thatbilingualsaremoresemanticallyorientedthanunilingualsandhaveagreaterawarenessof certainpropertiesoflanguageisequivocal.Therewasalsoevidenceintwostudiesthatbilinguals hadlongerresponselatenciesonawordassociationtaskthanunilinguals.Thismightbedueto eitherlinguisticinterferenceinresponsegenerationordeepersemanticprocessing(orbothofthese hypotheses). Amajordifficultyininterpretingthesestudies(apartfromtheusualbilingualunilingualcontrol problems)isthatthemeasuresusedtoassessmetalinguisticskillsusuallyonlyhavefacevalidity. Wherecorrelationsbetweentasksare 70

reported(e.g.,Cummins,1978b)theytendtobelow,thusraisingnotonlytheempiricalvalidity questionbutalsothetheoreticalquestionofwhatthedimensionsareoftheconstructof metalinguisticawarenessorskillandwhatdevelopmentalstagesitgoesthrough.Somepromising workdirectedtowardansweringthesequestionshasrecentlystarted(Bialystok&Ryan,1985). Untilthetheoreticalissuesrelatingtotheconstructofmetalinguisticdevelopmentaretackled,little furtherempiricalprogressislikelytobemadeinelucidatingitsrelationshiptobilingualism. REFERENCES BaetensH.Beardsmore(1981).Bilingualism.Clevedon,Avon:Tieto. Baker,K.A.,&deA.AKanter.(1981).Effectivenessofbilingualeducation:Areviewofthe literature.Finaldraftreport,OfficeofTechnicalandAnalyticSystems,OfficeofPlanningand Budget,U.S.Dept.ofEducation. Balkan,L.(1970).Leseffetsdubilinguismefrancaisanglaissurlesaptitudesintellectuelles. Bruxelles:Aimav. BenZeev,S.(1977a)."TheeffectofSpanishEnglishbilingualisminchildrenfromlessprivileged neighbourhoodsoncognitivedevelopmentandcognitivestrategy".WorkingPapersonBilingualism, 14,83122. BenZeev,S.(1977b)."Theinfluenceofbilingualismoncognitivedevelopmentandcognitive strategy".ChildDevelopment,48,10091018. Bereiter,C.,&Scardamelia,M.(1982)."Fromconversationtocomposition:theroleofinstruction inadevelopmentalprocess".InR.Glasser(Ed.)Advancesininstructionalpsychology,Volume2. HillsdaleNJ:LawrenceErlbaumAssociates. Bialystok,E.&Ryan,E.B.(1985)."Ametacognitiveframeworkforthedevelopmentoffirstand secondlanguageskills".InD.L.ForrestPressley,G.E.Mackinnon,&T.G.Wailes(Eds.).,Meta cognition,cognitionandhumanperformance.NewYork:AcademicPress. Bruck,M.,Lambert,W.E.,&Tucker,G.R.(1976)."Cognitiveandattitudinalconsequencesof bilingualschooling:TheSt.Lambertprojectthroughgradesix".InternationalJournalof Psycholinguistics,6,1333. Cohen,S.P.,Tucker,G.R.,&Lambert,W.E.(1967)."Thecomparativeskillsofmonolingualsand bilingualsinperceivingphonemesequences".LanguageandSpeech,10,159168. Cummins,J.(1976)."Theinfluenceofbilingualismoncognitivegrowth:Asynthesisofresearch findingsandexplanatoryhypotheses".WorkingPapersonBilingualism,No.9,143. Cummins,J.(1978a)."Thecognitivedevelopmentofchildreninimmersionprograms".The CanadianModernLanguageReview,34,855883.(a) Cummins,J.(1978b)."Languageandchildren'sabilitytoevaluatecontradictionsandtautologies:A critiqueofOshersonandMarkman'sfindings".ChildDevelopment,49,895897. Cummins,J.(1981).Theroleofprimarylanguagedevelopmentinpromotingeducationalsuccess forlanguageminoritystudents.InCaliforniaStateDept.ofEducation,Schoolingandlanguage minoritystudents:Atheoreticalframework.LosAngeles:NationalDisseminationandAssessment Center. Cummins,J.(1983)."Languageproficiencyandacademicachievement".InJ.W.OilerJr.(Ed.), Currentissuesinlanguagetestingresearch.Rowley,MA:NewburyHouse. 71

Cummins,J.,&Mulcahy,R.(1978)."OrientationtolanguageinUkrainianEnglishbilingual children".ChildDevelopment,49,12391242. Cummins,J.,Swain,M.,Nakajima,K.,Hanuscombe,J.,Green,D.,&Tran,C.(1984)."Linguistic interdependenceamongJapaneseandVietnameseimmigrantstudents".InC.Rivera(Ed.), Communicativecompetenceapproachestolanguageproficiencyassessment:researchand application..Clevedon,England:Multilingualmatters. Diaz,R.M.,&Hakuta,K.(1981,April)."Bilingualismandcognitivedevelopment:Acomparison ofbalancedandnonbalancedbilinguals".PaperpresentedatthemeetingsoftheSocietyfor ResearchinChildDevelopment,Boston,MA. Donaldson,M.(1978).Children'sminds.Glasgow:Collins Doyle,A.B.,Champagne,M.,&Segalowitz,N.(1977)."Someissuesintheassessmentof linguisticconsequencesofearlybilingualism".WorkingPapersonBilingualism,No.14,2130. Feldman,C.,&Shen,M.(1971)."Somelanguagerelatedcognitiveadvantagesofbilingual fiveyearolds".JournalofGeneticPsychology,118,234235. IancoWorrall,A.(1972)."Bilingualismandcognitivedevelopment".ChildDevelopment,43, 13901400. Imedadze,N.V.(1960)."Kpsikhologichoskoyprioroderannegodvuyazyehiya(onthe psychologicalnatureofearlybilingualism)".VoprosyPsikhologii,6,6068. Kessel,F.(1970)."Theroleofsyntaxinchildren'scomprehensionfromagessixtotwelve". MonographsoftheSocietyforResearchinChildDevelopment,35, Leopold,W.F.(1949).Speechdevelopmentofabilingualchild.(Vol.3).Evanston:Northwestern UniversityPress. Lambert,W.E.(1975)."Cultureandlanguageasfactorsinlearningandeducation".InA.Wolfgang (Ed.),Educationofimmigrantstudents.Toronto:OntarioInstituteforStudiesinEducation.(Ed.), Educationofimmigrantstudents.Toronto:OntarioInstituteforStudiesinEducation. Lambert,W.E.,&Tucker,G.R.(1972).Bilingualeducationofchildren:TheSt.Lambert experiment.Rowley,MA:NewburyHouse. Mohanty,A.K.(1982).BilingualismamongKondTribalsinOrissa(India):Consequences,issues andimplications.Unpublishedmanuscript,UktalUniversity. Oller,J.W.,Jr.(1980)."Alanguagefactordeeperthanspeech:Moredataandtheoryforbilingual assessment".InJ.Alatis(Ed.),31stAnnualGeorgetownUniversityRoundTableonLanguagesand Linguistics.Washington,DC:GeorgetownUniversityPress. Olson,D.R.(1977)."Fromutterancetotext".Thebiasoflanguageinspeechandwriting.Harvard EducationalReview,47,257281. Purbhoo,M.,&Shapson,S.(1975).TransitionfromItalian.Toronto:TorontoBoardofEducation. Ryan,E.B.,&Ledger,(1979)."Differencesinsyntacticskillsbetweengoodandpoorreadersin thefirstgrade".PaperpresentedatthemeetingoftheMidwesternPsychologicalAssociation, Chicago. Starck,R.,Genesee,F.,Lambert,W.E.,&Seitz,M.(1977)."Multiplelanguageexperienceandthe developmentofcerebraldominance".InS.J.Segalowitz&F.A.Gruber(Eds.),Language developmentandneurologicaltheory,(pp.).NewYork:AcademicPress. Torrance,E.P.,Gowan,J.C.,Wu,J.M.,&Aliotti,N.C.(1970).Creativefunctioningof

monolingualandbilingualchildreninSingapore.JournalofEducationalPsychology,61,7275. Toukomaa,P.,&SkutnabbKangas,T.(1977).Theintensiveteachingofthemothertongueto migrantchildrenofpreschoolageandchildreninthelowerlevelofcomprehensiveschool. Helsinki:TheFinnishNationalCommissionforUNESCO. Vygotsky,L.S.(1935).Multilingualisminchildren.TranslatedbyMetroGulutsanandIreneArki, CentreforEastEuropean&SovietStudies,theUniversityofAlberta,mimeo,1975.Theessay 72

appearsinacollectionofessayswrittenbyVygotskyandeditedbyL.V.Zankov,Zh.1.Shif,& D.B.E1konin,Umstvennoerazvitiedeteivprotesesseobucheniia,spornikstatei(Mental developmentofchildrenintheprocessofeducation,acollectionofessays).Moscow& Leningrade:StatePedagogicalPublishingHouse. Vygotsky,L.S.(1962).Thoughtandlanguage.Cambridge,MA:MIT.Press. 73

[Thispageintentionallyleftblank.] 74

5 TheImpactofLanguageDifferencesonLanguageProcessing:AnExamplefromChinese EnglishBilingualism DorisAaronsonStevenFerresNewYorkUniversity InthischapterwecomparetheprocessingofEnglishbyChineseEnglishbilingualswiththatof monolingualEnglishspeakers.WehaveselectedChineseEnglishbilinguals(whomweoftencall "Chinese")forcomparisontomonolingualEnglishspeakers(whomweoftencall"Americans")for threeinterrelatedreasons.(a)First,theChineseandEnglishlanguagesdifferonimportantlinguistic dimensions,someofwhichquestiontheoriesaboutlinguisticuniversals.Incontrast,mostofthe pastbilingualresearchhasfocusedoneitherSpanish/EnglishorFrench/Englishspeakers.Focusing onanonIndoEuropeanlanguagemayrevealdifferencesthatdonotemergewhencomparingtwo membersofthesamelanguagefamily. (b)Second,mostofthepastbilingualresearchwithintheframeworkofcognitivepsychologyisofa generalnature,dealingwiththepossibilityofdualsemanticcoding,ofextraprocessingtimefor mentaltranslation,orofperformancedecrementsduetocapacitylimitationswhenanytwo languagesmustshareattentionormemoryspace(Hatch,1978;Spolsky&Cooper,1978).Instead wefocushereonlanguagespecificeffects,i.e.,onperformancedifferencesattributabletoparticular semanticandsyntacticdifferencesbetweenChineseandEnglish. (c)Fromanappliedstandpoint,thedeterminationofspecificlinguisticdifferencesthathave psychologicalcorrelateshasimplicationsforunderstandingsecondlanguageacquisitionand performanceoflanguagebasedcognitivetasksbybilinguals.Thereareprobablyover2million Chineseimmigrantsorimmediatedescendantsofthoseimmigrantsnowlivingincountries speakingIndoEuropeanlanguages(Cheng,1978).Relativelylittlehasbeendonetodevelophigh qualitybilingualeducationinthosecountries,ortodealwithotherso 75

cioculturaldifferencesthatmaysteminpartfromlinguisticdifferences.Amoredetailed understandingofthosedifferencesisanecessaryprecursortodevelopingeffectiveprograms. THESTRUCTUREANDMEANINGOFLEXICALCATEGORIESINENGLISH Thischapterdealslargelywithpropertiesofwordsbecausetheyareanimportantlinguisticand cognitiveunitinEnglish.Linguistically,awordcanbedefinedasa"minimumfreeform"whichis theunionofaparticularcomplexofsounds(phonology)withaparticularmeaning(semantics) capableofaparticulargrammaticalemployment(syntax)(Lyons,1979).Cognitively,wordsarean importantunitintheperception,comprehension,andmemoryoftext(Clark&Clark,1977;Miller &JohnsonLaird,1976).Thischapterfocusesonthelexicalcategoryorformclassofwordsas definedindictionaries(e.g.,noun,verb),becausethatisanorganizationallevelwhichisrelatively wordspecific.However,wewillseelaterthatChineseraisesdifficultiesforsuchdefinitionsof wordsandoflexicalcategories,whicharebasedonIndoEuropeanlanguages. Traditionally,thelexicalcategorieshavebeendividedintocontentwordsandfunctionwords.In additionweconsiderthepsychologicalattributesof"meaning"and"structure"thatcutacrossthe contentandfunctionwords,asallwordscontributebothmeaningandstructuretoasentence.A person'scodingofmeaningandstructurefromthevariouslexicalcategoriesmaybebasedonboth "linguistic"and"verbal"information.Forexample,thecognitiveprocessingofstructural informationfromwordsmaybebasedonpropertiessuchastheirsyntactic(functional)rolewithin thesurfacephrasestructure,theirformclass,theirlengthinlettersorsyllables(e.g.,functionwords areoftenshorterthancontentwords),ortheirrhythmicvaluewithinthetextbasedoncodingtime, stress,andintonationpatterns.Processingmeaninginformationfromwordsmayincludeprocessing theirsemanticattributes,theirreferentialrelationtoothertextcomponentsandtoworldlyentities, theirpropositionalinvolvement,theirwordfrequency(wordswithmoreselectiveorspecific meaningoftenhavelowerfrequencyinthevocabulary),ortheirstressvaluewithinacovertacoustic representation(wordsthatconveymoremeaningwithinacontextoftencarryaheavierstressvalue). ThroughoutmuchofthischapterweareconcernedwithdifferencesinthewaysChineseEnglish bilingualsandmonolingualEnglishspeakersprocessthemeaningandstructureofwords,andof sentencesasawhole.ThenextsectionsuggestssomespecificdifferencesbetweentheChineseand Englishlanguagesthatshouldleadtodifferencesinprocessingmeaningandstructurebetweenthe twolanguagegroups. 76

ANOVERVIEWOFCHINESEENGLISHDIFFERENCES AlthoughmostoftheinformationinthispaperisgeneralenoughtoapplytoallChinesedialects, specificexamplesareprimarilytakenfromwhatistermedModernStandardChinese(forthemost part,theBiejingdialectofMandarin).Since1966,theChinesegovernmenthasencouragedthe developmentofMSCasasingleunifyinglanguagethroughoutthemainlandand70%ofthepeople nowuseit(Fincher,1978;Li&Thompson,1981;Mathias&Kennedy,1980).MSCisalsoan importantlanguageinTaiwan.AlthoughmanyChineseAmericansspeakCantonese,thewritten formsofthevariousChinesedialectsareverysimilar. Asisgenerallywellknown,theminimumfreeforminChineseisthesinglecharacterideogramor pictogram,asopposedtotheword(Lyons,1979;Sapir,1921;Venezky,1984).Incontrasttowords, therearenoregularorsystematicrelationshipsbetweenChineseorthographyandphonology( Henne,Rongen&Hansen,1977).Further,theacousticrepresentationofagivencharactermayvary withcontext.AsChinesedictionariesareoftenorganizedintermsofnumberofstrokesinawritten character(orintherootofafamilyofcharacters),apersonwhowantstolearnhowtowritea characterheheardwillhavedifficultylocatingitinthedictionary. TheChinesecharacterisamonosyllabicmorpheme.Sentencesconsistofmorphemicstrings.In mostcasesasinglemorphemeisaconceptualandsemantic"unit."However,Chinesehasthree waystoformmultimorphemicunits(Henneetal.,1977;Li&Thompson,1981).First,compounds consistoftwoadjacentcharacters(e.g.,fengche=windvehicle=windmill;rexin=hotheart= enthusiastic).Second,reduplicationinvolvesrepeatingacharactersuchthatthedoubleformsanew semanticand/orsyntacticentity(e.g.,xie=rest;xiexie=restalittle,i.e.,nap;man=slow;man man=slowly).Finally,Chinesehassomeaffixes(althoughtheyarefewcomparedtoIndoEuropean languages)thatcanindicategrammaticalorsemanticattributes.Themultimorphemicentities appeartovaryalongacontinuumfromstronglybound,tolooselyconcatenated,tofreebut linguisticallyrelatedpairs.Occasionallymorethantwocharactersmaybe"combined."Asthereis oftennoclearagreementonthestrengthofapairedrelationship,anopenlinguisticquestionregards thenatureof"words"inChinese(Halliday,1956;Henneetal.,1977;Simon,1937).Whenwritten, thereisnospacingdifferencebetweentwomonomorphemicandonecompoundmorphemicentity. Theratioofmorphemestowords(includingcompounds)isrespectively1.2inChinese(basedon Chen,1982,andLiu,Chaung,&Wang,1975),and1.7inEnglish(Lyons,1979). InthischapterwearemakingthepsychologicalassumptionthatChinesemonomorphemic charactersareimportantperceptualandcognitiveunitsthatmayfunctionascodingunitsinways somewhatanalogoustothoseofEnglish 77

words.(Ideogramsandwordsaretheminimumfreeformsintheirrespectivelanguages.)However, thedifferencesbetweenChinesecharactersandEnglishwordsmayleadChineseEnglishbilinguals toperceiveEnglishwordsdifferentlythanmonolingualEnglishspeakersdo.Throughoutthis chapter,welooselyusetheterm"Chineseword"torefertoeitherasinglemorphemeora combination,withtheunderstandingthatthesinglecharacteristhebasicunit. NotonlythenatureofChinesewords,butalsothenatureoflexicalcategoriesisquestioned( Halliday,1956;Simon,1937).Sinologistshavedividedmorphemesinto"full"wordshaving semanticcontentand"empty"words,whicharegrammaticalmarkersdevoidofsubstance(e.g.,a morphemetoindicatethatpriorcharactersmodifyasubsequentcharacter;amorphemetoindicate thatthepreviousstringisaquestionorawarning).Recentlinguisticresearchhasfurtherdividedthe Chinese"full"wordsintotwofamilies,"verbals"and"nominals,"whichtogethercontainalmost everythingcharacterizedasacontentwordinEnglish,andalsosomeofthefunctionwords. ThesedifferencesbetweenlanguagesarefurtherincreasedbythefactthattheexactsameChinese charactercanrepresentmanydifferentlexicalcategories(e.g.,thesamecharactercanmeanthe nounhappiness,theadjectivehappy,theadverbhappily,orthestativeverbishappy).Thissituation arisesinpartbecauseChinesewordsgenerallyhavenogrammaticalinflectionstoindicatecase, number,gender,tense,ordegree(Henneetal.,1977;Li&Thompson,1981).Further,even sententialpositionisnotafullyreliablecuetoaword'sfunctionalrole(e.g.,subject,object)ina sentence.ThisispartlybecauseChineseisa"topicprominent"language(Li&Thompson,1981). Thatis,thesemantictopic(ratherthanthesyntacticsubject)ofasentencegenerallycomesfirst (e.g.,Thistree,[its]leavesareverybig.Yesterday,[I]readfortwohours). Intheaboveexamples,someofthefunctionwordsareinparentheses.Inillustrativeexamples throughoutthischapter,wordsinparenthesesaretobeunderstoodasinstanceswhereChinese wouldomitthewordentirely.IncontrasttoEnglish,mostChinese"function"wordsareoptional andaresuppressedwhenthatwouldnotleadtoambiguity. Insum,weseemajordifferencesbetweenChineseandEnglishthatcouldaffectthecognitiveand linguisticprocessingofEnglishwordsbybilinguals.ThenatureofChineseandEnglishwordsand lexicalcategoriesappeartodiffer.Asconsideredlaterinthispaper,Chinesefullandemptywords donotsystematicallymapintoEnglishcontentandfunctionwords.AgivenChinesesubstantivecan representmanydifferentEnglishlexicalcategories(includingfunctionwords),andemptywords(as wellassomesubstantives)aresuppressedwheneverpossible.Asweseeinthenextsection, processingthe"structure"and"meaning"ofwordsinEnglishsentencesdiffersbetweenbilinguals andmonolinguals.Wearguethatthosedifferencesbetweenlanguagepopulationsinpsychological performancetasksareconsistentwiththespecificlinguisticdifferencesbetweenChineseand English. 78

LINGUISTICPERFORMANCEBYCHINESEENGLISHBILINGUALS ThespecificsyntacticandsemanticdifferencesbetweenChineseandEnglishmayleadto differencesintheprocessingofEnglishwordsinvariouslinguisticperformancetasks.Inthis section,weconsiderseveraldifferenttypesofdatatoprovideevidenceonperformanceeffects. First,wepresentsomecasestudydataontheEnglishwritingofChineseEnglishbilinguals. Next,wereviewsomeresearchbyA.Bloom(1981)showingthatlinguisticdifferencesbetween ChineseandEnglishindealingwithabstractionleadtocognitiveperformancedifferences betweenthetwopopulations.Finally,wepresentdatafromasentenceperceptionexperimentthat comparesChineseEnglishbilingualswithmonolingualEnglishspeakers.Thesedataare consideredinsomedetail,because(a)theyillustratemanyofthelanguagespecificdifferences mentionedintheoverview,and(b)theyhaveimplicationsforreadingdifferencesbetweenthe twolanguagepopulations.Wesuggestthatthelinguisticperformancedatatobeconsidered supporttwogeneralhypothesesaboutChineseEnglishbilinguals. B1.TheComparisonHypothesis.BilingualsperceiveEnglishinrelationtotheirpastcombined linguisticexperienceinbothlanguages. B2.TheCognitionHypothesis.Languagespecificdifferencescanleadtoassociatedcognitive performancedifferencesbetweenbilingualsandmonolingualsinprocessingthemeaningand structureofsentences. WritingPerformance:ChineseAmericanUniversityStudents WeexaminedthewrittencourseworkoftwoChineseAmericanNYUstudentsforwriting problems(botherrorsandnontraditionalstyle).Itappearsthatalmostallofthesewouldbe consistentwiththeChinese/Englishdifferencesindicatedabove.JCis21yearsold,wasbornin HongKong,andimmigratedtotheU.S.atage5.SheattendedtheNewYorkCitypublicschool systemandshealsoattendedaChineselanguageschoolfor8years.Sheisasenior psychobiology(premed)major,probablyinthetop10%ofherclassintermsofgrades,andis doingindependentresearchontheroleofthelateralhypothalamusinfeeding.Thedatabasefor ourevaluationwasher10pagetypedreportonthatresearch(gradedA).JC'sratherdiscrepant SATscores,620math,420verbal,alongwithherhighGPA(about3.5)suggestspecific languageproblems. Thesecondstudent,JY,is21yearsoldandwasbornintheU.S.LikeJC,Chinese(Cantonese dialect)istheprimaryhomelanguage.SheattendedtheNewYorkCitypublicschools,aChinese languageschoolfor10years,andtookMandarincoursesatNYU.JYisajuniorandinaboutthe top30%ofherclass.HerGPAis3.0,andshehasverbalandquantitativeSATscoresof510and 520respectively.Thedatabaseforourevaluationinhercaseisa4pagetyped 79

researchproposal(gradedA)designedtotestthehypothesisthatcertaintypesofeducationalTV programscanincreasethevocabularyofpreschoolers. Table5.1presentsdataonthetypesoflinguisticproblemsexhibitedbythetwobilingualstudents. Inreality,agivenproblemcannotbeclassifiedasexclusivelyaffectingstructure/syntaxversus meaning/semanticsorcontentversusfunctionwords,asneitherthecategoriesnortheproblemsare dichotomous.However,thepredominantfocus(inouropinion)isindicatedinthetable,alongwith thefrequencywithwhichtheproblemsoccurred.Thesedatasuggestthatmoreoftheproblems centeroncontentthanfunctionwords,andmoreconcernstructurethanmeaning. 1.ContentWords.ThemajorproblemsforbothstudentsmayberelatedtothefactthatEnglish contentwordsareofteninflectedtoindicatesyntacticinformation,whereasChinesewordsarenot. Thislanguagedifferenceappearstounderliethreetypesofwritingerrors.Themostfrequent problemappearstobelackofagreementbetweenwordsinnumber(i.e.,singular/plural)for noun/verb,noun/noun,andnoun/pronounrelationships.Asecondcontentwordproblemforboth studentsinvolvedtheuseofincorrectverbforms,includingtense,aspect,andmodalinformation. Forexample,lexicalverbshadincorrectinflectionsandwereaccompaniedbyincorrectauxiliary verbs.Athirdproblemforbothstudentsinvolvedincorrectinflectionsonadjectivalandadverbial modifiersthatyieldedthewronglexicalcategory(e.g.,randomvs.randomly)orthewrongdegree (e.g.,largervs.largest). Atthispoint,letusgiveawordofcautionregardingoverinterpretingthedata.Theaboveexamples frombothstudentsareindeedconsistentwiththeinflectiondifferencesbetweenChineseand English.ButmostteachershavealsoseensucherrorsonthepartofmonolingualEnglishspeaking students.However,thelargenumberoflinguisticerrorsinconjunctionwiththeotherwisehigh qualityofthesestudents'academicperformanceisstriking,andiscertainlysuggestiveofproblems specifictoChinese/Englishdifferences. 2.FunctionWords.BothstudentshaddifficultieswithEnglishfunctionwordsthatappeartobe consistentwithsomeoftheChinese/Englishdifferencesmentionedintheoverview.First,both studentsusedfunctionwordsthatdidn'thavepreciselythecorrectmeaning.JCfrequentlyselected prepositionsthatwerenotquiteappropriatewithinhersentencecontext,asillustratedinTable5.1. Shealsoinventedverbprepositioncompounds,withnonominalobjectsfortheprepositions(e.g., codedfor).HeruseofthesecompoundsoftengavethefeelingofChinesecoverbs(discussedlater). JYused"that"asauniversalrelativepronoun,whenwhwordswouldhavebeensemanticallymore appropriate(seethediscussionbelowoftheChinese"universal"subordinatingmorpheme"de"). Asdiscussedbelow,manyChinesefunctionwords,includingprepositionsandrelativepronouns, canbeusedinabroaderrangeofsemanticcontextsthanthe 80

TABLE5.1 WritingErrorsofChineseEnglishBilinguals TypeofError Frequency A. StudentJC Contentcategories Singular/pl 16 ural Verbform 5 Examplesa

Wordorder

Adjective inflection

biteswas; stimulationwere Thelatency increasedwhile theduration ofeatingis shortened theamount(of food)eatenper biteof food Theelectrode designatedas more(most) effective

Function categories Preposition

Classifier Article

5 4

Thelatencywas codedforinthe follow ingmanner convertedinto timeintheunit of seconds If(the)mouthis incontactwith (the) food

B. StudentJY Content categories Verbform

Singular/pl

Thechildren willbetested theyhad (will);Iwould want...sothat Ican (could) TVisoneofthe

TypeofError Frequency ural

Lexical category Function categories That/who

Examplesa formofmedia; the experiment were sampletobe selected random(ly)

Missing function word C. Szechuan menu Content categories Singular/pl ural Lexical category

thechildrenthat played;theones that watched (On)thefirst day;Ithelps (them)to improve

25 11

Semantic errors

11

Verbform Wordorder Function categories Preposition

6 4

Coldnoodle withsesame sauce Tenderly chickenbreast; Choicedjumbo shrimps;a natureflavor Lotusstems keepturnover averagelyin thewok Deepfriedand sprinklewith scallions Watermouthing gingersauce

Article

Shrimpsare combinedto match(with) broccoli Weofferthe(a)

TypeofError Frequency

Examplesa reasonableprice

aErrorsareunderlined.Wordsthatshould

havebeenincludedareinparentheses. 81

Englishfunctors.Inaddition,bothstudentsentirelyomittedfunctionwords(articles, prepositions,pronouns)toyieldgrammaticallyquestionableorerroneousconstructions.Inthe equivalentChinesesentences,thesefunctionwordswouldbeoptionalandunnecessary. Finally,bothstudentsusedEnglishconstructions(involvingbothcontentandfunctionwords) thatwerestylisticallyunusual,butappropriateforChinesesentences.Forexample,Chinese permitsnonstandardwordorderforthepurposesofemphasistoagreaterextentthanEnglish 1incontextsthatwere does.Inaddition,JCusedprepositionalphrasesas"classifiers" stylisticallyunnecessaryorunusualforEnglish. Insum,almostallofthewritingdifficultiesexhibitedbyJCandJYappeartobeconsistentwith linguisticdifferencesbetweenChineseandEnglish,andmaybeattributableinparttotheir bilingualabilities,supportingbothhypothesesB1andB2.Hakuta'schapterinthisvolume providesevidenceforsimilartrendsinJapaneseEnglishbilinguals.Hesuggestscausal mechanismsofinterlanguagetransferandlinguisticsimplification. WritingPerformance:CommercialBusiness AsalastexampleofwrittenperformancebyChineseAmericanbilinguals,weexaminedthetwo pagemenuofaneighborhoodSzechuanrestaurant.Thecontentandfunctionworderrorswere similartothoseofthestudents,butoccurredmuchmorefrequently.Themajortypesoferrors, alongwithspecificexamples,areillustratedinTable5.1c. 1.ContentWords.Aswiththestudents,themostfrequenterrorwasaconfusionofsingularand pluraldesignations,mostlyfornouns,asChinesedoesnotdiscriminatebetweentheseforms syntactically.Second,almostallcontentwordlexicalcategorieswereincorrectlysubstitutedfor eachotheratonepointoranotheronthemenu,consistentwiththelackofinherentspecific markers(e.g.,inflections)inChinesewords.Analogously,Hakuta(thisvolume)mentionsa bilingualchildwhoregularlyusedthenoun"mistake"asaverb,consistentwithitsusagein Japanese.Third,themenuhadafairlyhighnumberofsemanticerrors,notshownbythe students.But,consistentwiththestudents,frequenterrorsinvolvedincorrectverbforms.These weremostlytenseinflectionsinaccordwiththelackofsuchinflectionsinChinese. 2.FunctionWords.Themostfrequentfunctionworderrorsinvolvedprepositions,including missing,extra,andsemanticallyincorrectwords.Also,errors ____________________ 1Chineseclassifiersaresomewherebetweenarticles,quantifiersandadjectives.Examplesof theiranaloguesinEnglishwouldbe:bunchofbananas,sheetofpaper,heardofcattle,packof wolves,squadofsoldiers. 82

witharticlesincludedincorrectsubstitutionsof"the"for"a"andviceversa,aswellasmissingand extraarticles.TheseerrorsareconsistentwiththelackofarticlesinChinese,withthelackofsome oftheusualIndoEuropeanprepositionsinChinese,andwiththegeneralsuppressionofChinese prepositionsinsentenceswheretheirabsencewouldnotcauseambiguitywithinthetotalcontext. Inthecaseofarestaurantmenu,thelinguisticerrorsmayindeedhaveapositiveimpactonbusiness. TheysuggestthatwewillbeprovidedwithauthenticChinesecuisine,asadvertised,ratherthanan Americanimitation.However,inotherbusinesscontexts,thesetypesoferrorscouldleadto financiallosses. AbstractThinkinginChineseEnglishBilinguals InaccordwithotherlinguistsandwithanecdotalcommentsbyChineseEnglishbilinguals,Bloom( 1981)maintainsthatChinesedoesnothavesomelinguisticmarkers,inherentinwordsofWestern languages,thatareusedtoexpressabstractideas.Hehypothesizesthattheabsenceofsuchmarkers leads(a)toarelativeinabilitytodocertaintypesofabstractthinkingand(b)todifficultiesin communicatingwithWesterners.Aseriesofexperimentswithmonolingualandbilingualspeakers providesevidencetosupporthispoints. 1.Counterfactuals.Englishhasspecificstructures,usingsubjunctiveverbformsplusfunction words,thatlinguisticallymarkthecounterfactualrealm.Someexamplesare:"IfIwerehe,Iwould respondtotheletter";"IfJohnhadcome,hecouldhaveseenMary."Insteadofcounterfactuals,the Chineseusedescriptiveorimplicationalstatementsthatconveyaconcreteorfactualsense,rather thananabstracttheoreticalorhypotheticalsense.TheEnglishcounterfactual"Ifthelecturehad endedearlier,Billwouldhavehadtimetopreparefortheexam"wouldmorelikelybeexpressedin Chinese(orinEnglishbybilinguals)as"Thelectureendedtoolate,soBilldidnothavetimeto preparefortheexam."Thus,ahypotheticalstatementwouldbeturnedintoafactualstatement. Bloomhypothesizedthatthelackofalinguisticallymarkedornaturalcounterfactualstructure wouldleadtodifficultiesinthecognitiveprocessingofsuchinformation.Totestthathypothesis,he developedChineseandEnglishstoriesincludingcounterfactualinformation.Thelinguisticstructure ofthestimuluswasmodifiedtoenablesomewhatsimilarexpressionsinbothlanguages:"Xwasnot thecase,butifXhadbeenthecase,thenYwouldhavebeenthecase."Basedoncomprehension questions,97%ofthemonolingualAmericansunderstoodtheEnglishstories,butonly29%ofthe ChineseunderstoodtheChinesestories.Further,ChinesecomprehensionofChinesestories correlatedsignificantlywiththeirEnglishability.Finally,agroupofChinesewhousedEnglish everydayinbusinesswerefirsttestedonChinesestories,andamonthlateronmatchedEnglish stories.Theircomprehensionwas6%ontheChineseand94%onthe 83

Englishstories.Thisstrikingdifferenceinthesamesubjectssuggeststousthattherearelanguage specificdifferencesinverbalprocessingstrategies,ratherthandifferencesingeneralthought processesbetweenthetwopopulations. WeshouldnotethatAu(1983,1984)failedtoreplicatesomeofBloom'sresultsandquestionssome ofhismethodology.AlthoughshefounddifferencesbetweenEnglishandChinesepeoplein representingandproducingcounterfactualslinguistically(ineitherlanguage),shedidnotfind failuresinChinesestudents'abilitiestounderstandcounterfactualreasoning.However,accordingto Bloom(1984)differencesbetweenthestudiesmightbeattributedtodifferencesintheabstractness ofthetextmaterials(Bloom'sweremoreabstract),andintheEnglishabilitiesofthesubjects (Bloom'sknewlessEnglish). Bloomgoesontocitenumerousexamplesofcrossculturalconversationswherecounterfactual statementscausedmiscommunications,confusions,andtheperceptionthatAmericanspeakerswere outtomisleadtheChineselisteners,ortoimposeWesternideasonthem.Onesociallyand politicallyembarrassingsituationinvolvedanAmericanjudgeinacustomscaseagainstaChinese man.EvenwithanexperiencedChinesetranslatorpresent,themancouldnotunderstandthatthe judgewasraisinghypotheticalsituationstotestwhetherheunderstoodthelaws,ratherthanmaking realisticthreats.InacontentanalysisofaTaiwannewspaper,in3weeksofissuesBloomfoundonly oneexampleofwhatmightbecalledcounterfactualargumentinthetranslationofaspeechby HenryKissinger!Bloom'sworkisconsistentwithourownnotionsintwoimportantways.(1) EnglishisinterpretedbyChineseEnglishbilingualsinrelationtotheircombinedpastexperiencein bothlanguages.Chinesebilinguals,quitefluentinEnglish,stillperceiveEnglishsentences differentlythanmonolingualEnglishspeakersdo(hypothesisB1).(2)Linguisticdifferencesin syntaxandsemanticsleadtocognitiveperformancedifferencesinprocessingstructureandmeaning byChineseEnglishbilinguals(hypothesisB2). 2.Definite"Generic"Articles.Englishusesthearticle"the"intwodifferentways.First,itisused tointroduceananaphoricreferenttoaparticularnounintherecentsharedexperienceofthespeaker andthelistener(e.g.,Abuffalowasdrinkingfreshrainwater.Thebuffalostoppedsuddenlywhena coyoteapproached).Second,itisusedtorefertoagenericnominalconcept,withnoparticular instanceathand(e.g.,Thebuffaloisbecomingextinct).AlthoughChinesedoesnothavedefinite articles,thedemonstrativeadjectives"this,that,these,those"aresometimesusedtofulfillthefirst usage.However,thereisnolinguisticdevicetocommunicatethesecondnotionof"genericthe." Bloommaintainsthatthelackofsuchfunctionwordsisanotherexampleofthelackofaverbal paradigmtoexpressabstractideas.Evenwhenexplicitlytoldthattheexampleaboverefersto"a conceptual"buffalo,ChineseEnglishbilingualswillnotacceptthenotion.Theywillrespondwith statementslike 84

"Whatdoyoumeanby'conceptual'buffalo?Eitheryouaretalkingaboutasinglebuffaloorabout allbuffaloes.Whatelseisthere?"(RememberthatChinesehasnoreliablesyntacticinflectionsfor singular/plural.)InanexperimentinwhichChineseEnglishbilingualsweregivengeneric explanationsofChinesesentences,only37%werewillingtorespondyesthatthesentencescould refertoconceptualentities,andmostofthose37%hadhadextensiveexperiencewithEnglish. 3.EntificationofProperties,Actions,andConditions.Chinesedoesnothavegeneralaffixes(or otherlinguisticdevices)toconvertconcretecontentwordsintoabstractentities(e.g.,red,redness;to accept,theacceptanceof;capital,capitalism;modern,modernize).NotonlycanEnglishentify individualcontentwords,butsyntactictransformationsexisttoconvertentiresentencestoabstract nominals(e.g.,"Interestratesrose";"Theriseofinterestrates").Bloomnotes(p.40)that"whenthe Englishspeakershiftsfrom'ThatmeasurewillbeapprovedbyCongress'to'Theapprovalofthat measurebyCongress,'henotonlyconvertsacompletedsentenceintoanounphrase,butona semanticlevelhemovesfromthedescriptionofaneventthathashappened,ishappening,orwill happentoanabstractionoftheideaortheeventasapurelytheoreticalnotion,"(e.g.,"Theapproval ofthatmeasurebyCongresswilldependonthesubcommittee'sreport.") Insomecases,Chinesecanexpressentificationintermsofsubordination,withthemarkerde. However,theuseof"de"requiresthatthemodifiersprecedethenominal,yieldingchainsofleft embeddedstructureswhichrapidlyexceedthelistener'smemoryspan.Forexample,English speakershadnotroubleunderstandingsentenceslike"wewillputofftonextweekdiscussionofthe furtherimplicationsofthenewmethodforcalculatingtherelationshipsbetweentherateof economicdevelopmentandtheindividualstandardofliving."Butonly58%ofTaiwanuniversity studentsknewwhatwouldbediscussedwhengivenanappropriateChinesetranslationwith"de" yieldingsomethinglike"Wewillputoffdiscussinguntilnextweekcalculateeconomic development'de'rateandindividualstandardofliving'de'relationship'de'newmethod'de'further implications." Inarelatedexperiment,ChineseandAmericansweregivenseveralsimpleexamplesofentification transformationsintheirnativelanguages(e.g.,JohnandMarygotmarriedThemarriageofJohn andMary;ThisthingisimportantTheimportanceofthisthing)andwereaskedtogeneralizeto similarlyconstructednewstrings.Bloomfoundthat87%oftheAmericans,butonly11%ofthe Chinese,coulddothetaskconsistentlyandaccurately.Bloomsuggeststhatthesyntacticstructurein Englishmakestheabstractsemanticrelationshipbetweeneventspredominateovertheindividual conditionsoreventsbeingrelated,andalsomakesittakeonarealityofitsown.Incontrast,Chinese sentences 85

callattentiontothetwoindividualconditions,andtheninaddition,stipulatetheirrelationship. Whensubjectsweregivencomplexstoriesintheirnativelanguageincludingsuchentification, responseaccuracyforcomprehensionquestionswas59%and79%respectivelyforChineseand Americans. Bloommaintainsthatdifficultiesinusingentification,thecounterfactual,andthegeneric"the" illustratethatChinesedonothavecognitivestrategiesformanytypesofabstractions.Hesuggests thatthelackoflinguisticstructurescanoftenleadtochildrendevelopingwithouttheassociated cognitiveschema,andtoadultslackingtheassociatedperformanceability.However,Bloomhasnot providedevidenceontheextenttowhichthedifficultiesarerelatedtolanguageprocessingas opposedtomoregeneralthoughtprocesses. ACOMPARISONOFSENTENCEPROCESSINGINBILINGUALSAND MONOLINGUALS Inthestudyreportedhere,weexaminethedifferencesbetweenmonolingualsandbilingualswho evaluatethecontributionsofcontentwords(nouns,verbs,adjectives,adverbs)andfunctionwords (prepositions,conjunctions,articles,pronouns)tothestructureandmeaningofEnglishsentences. Lexicalcategoriescanbeconceptualizedbasedontheirrelativecontributiontostructureand meaning(Aaronson&Ferres,1983a,1983b,1984a).Theprimaryimportanceoffunctionwordsis insignalingstructureandthatofcontentwordsisinconveyingmeaning.However,withinboth majordivisions,thesubcategoriesvaryintheiremphasisonstructureormeaning.Forexample, withinthecontentwords,verbsaremoreimportantforthestructuralorganizationofthetext, whereasnouns(andtheiradjectives)aremorecriticalforitsmeaning.Withinthefunctionwords, prepositionsandconjunctionsaremoreimportantforstructure,whereaspronounsandarticlesare moreimportantformeaning. Empiricalevidenceshowsthatpeoplecanusethe"structureoriented"lexicalcategoriesascuesto initiatephraseunitcodinginasentence.Theverbisscoredasthemostimportantsentential divisioninclusteranalysesofrecalldata(Levelt,1970)andinsubjectiveparsingdata(Bond& Gray,1973).Conjunctionsarecodeddifferentlyincontextswheretheyconjoinsentencesas opposedtophrases(Jeremy,1978).ClarkandClark(1977)htprepositionsasstructuralcuesin theirlistofsyntacticcodingstrategies:"Wheneveryoufindapreposition,beginanewprepositional phrase"(p.59);"Afteridentifyingapreposition,lookforanounphrasewhichclosesoutthe prepositionalphrase"(p.62).Whentheperformancetaskrequiresextensivememory,silentreading showsselectiveincreasesbothineyefixations(Gibson&Levin,1975)andinkeypressRTsatthe structurallyorientedcategoriesthatinitiatephrases(Aaronson&Ferres,1983a,1983b;Aaronson &Scarborough,1976,1977). 86

Incontrast,empiricalevidencesuggeststhatpeoplerelymoreonthe"meaningoriented"lexical categorieswhentheperformancetaskemphasizescomprehension.Whennounsaredeleted, sentencecomprehensibilitydropsmarkedly(Healy&Miller,1971).Theuseofdeterminers suchas"a"and"the"influenceswhetherornotpeopleintegratethesemanticinformationfrom onesentencetothenext(deVilliers,1974).Anaphoricpronounsplayanimportantrolein aidingthesemanticintegrationofsentencecomponents(Caramazza,Grober,&Garvey,1977). Adjectivessetupexpectanciesthatfacilitatetheprocessingofrelatednouns(Foss,Cirilo,& Blank,1979). Toprovidedetailedinformationaboutthedifferencesbetweenbilingualsandmonolingualsin codinglexicalcategoriesinEnglishsentences,wehavetakentwoapproaches.First,weasked thesegroupsofpeopletoindicatethecontributionsofwordsfromthevariouscategoriestothe meaningandstructureoftheirsentencesoneachoftwo5pointratingscales.Second,we reviewedasubstantialamountofliteratureontheChineselanguageanditsusage,inorderto makecomparisonstoEnglishusage.Thus,wefirstpresenttheratingstudy,whichthenservesas anorganizationalframeworkforatheoreticalanalysisofthepastliterature. StimulusSentences METHODSFORTHERATINGSTUDY

The90Englishstimulussentencesrangedinlengthfrom9to19words,andhadameanof14.4 words.Thesentenceswerevariedwidelyinsyntacticstructure,includingclausesandphrases thatvariedinnumber,length,location,andsubjectratedcomplexity.Thesemanticcontentwas notconstrainedbyanyspecificrulesandwasobtainedprimarilyfromnewspapersand magazines.ExamplesentencesareinTable5.2.Table5.3containsthesamplesizesforwords ratedineachlexicalcategory. TABLE5.2 SampleStimulusSentences 1. 2. 3. 4. 5. OnlythelowereconomicsectionofWashingtonwasburnedbythemilitants. InameetingwiththecabinetthePresidentmadeknownhisfeelingsabout foreignpolicy. Becauseofthedoctor'snewcurethepatientwasabletocontinuehiswork. Aftermanylonghoursofdebatethehousingbillwasapprovedbythe legislature. WithinallofthegovernmentonlytheofficeofthePresidentissuedastatement aboutthearmstreaty. 87

TABLE5.3 SamplingofLexicalCategories NforAll Lexicalcategory Stimuli Contentwords 662 Verbset 172 Lexicalverbs 89 Auxiliaryverbs 60 Adverbsforverbs 23 Nounset 490 Nouns 302 Adjectives 167 Adverbsfor 21 adjec. Subjects SixteenNYUundergraduatesprovided5pointjudgmentsofstructureandmeaningforsetsofthree underlinedwordsineachsentence,generallyonecontentandtwofunctionwordsasindicatedin Table5.2.AllsubjectsspokeEnglishfrombirth.ThebilingualsalsospokeChinesefrombirth,as bothEnglishandChinesewerespokenintheirhomes.Mostoftheirparentswereimmigrantsfrom mainlandChinaorTaiwanandarecurrentlymiddleclassfamilieslivinginNewYorkCity.Mostof thesestudentsattendedNewYorkpublicschools.ManyalsoattendedChineselanguageschoolfor severalyearsandsomeareenrolledinNYUChinesecourses. Asthesubjectsinthisstudywere1820yearsofage,theyareatthefarendofalanguageand cognitivedevelopmentscale.Continuedsyntacticandsemanticdevelopmentoccursforstudents havingEnglishaseitherafirstorsecondlanguage,accordingtoteachersfortherequiredNYU WritingWorkshopCourse.Ourdataprovideevidencethatdifferencesinchildhoodlanguage backgroundsystematicallyinfluencesentenceprocessingintheseyoungadults. RatingProcedures Thesubjectsratedeachunderlinedwordseparatelyforitscontribution(1)tothestructureand(2)to themeaningofitssentence.Theratinginstructionsweredesignedtobeunrestrictiveand theoreticallyneutral,asillustratedbythefollowingexcerpt:"Wordsinsentencesplaytworolesin helpingyouunderstandthesentence.Theycontributetothemeaningandtheyprovidestructural informationtohelpyouorganizethecontextintoideas.Mostwordscontributetobothmeaningand structuralaspects.Butanyparticularwordmightcontributemoreor 88 NforS,M Ratings 93 45 15 15 15 48 15 18 15 Lexicalcategory Functionwords Organizationalset Conjuctions Prepositions Definitenessset A,an The Absolutepronouns Relativepronouns NforAll Stimuli 300 107 17 90 193 33 109 36 15 NforS,M Ratings 161 61 17 44 100 17 36 33 14

lesstoeithertypeofinformationthanotherwordsdo.Howmuchaparticularwordcontributesto structureortomeaningdependsonmanythings,including:(a)theparticularworditself;(b)its localandtotalsentencecontext;aswellas(c)theknowledgeororientationofthereader.We wouldlikeyoutoratesomeofthewordsin90sentencesaccordingtohowmuchyouthinkeach particularwordcontributestostructureandtomeaning.Youwillhaveafivepointscalefor indicatingeachoftheseattributesseparatelyforeachwordinitsparticularcontext.Thereareno rightorwronganswers." Infurtherinstructions,subjectswereaskedtouseasaframeworkfortheirstructureandmeaning ratingsthecontributionsofwordsencounteredintheirgeneralpastlanguageexperience.Their ratingswerebasedonthefollowingscale:(1)verylowcontribution;(2)moderatelylow contribution;(3)mediumcontribution;(4)moderatelyhighcontribution;and(5)veryhigh contribution.Subjectsweregivenexamplesentencesandwereencouragedtoaskquestionsifthe writteninstructionswerenotcleartothem. MeaningRatings RESULTSOFTHERATINGSTUDY

Threetrendsinthemeaningratingsareofinterest.First,Fig.5.1showsthattheratingsare consistentlyhigherforthebilingualsthanthemonolinguals(ttest:p<.05)foreverylexical category.The"Chinese"mayperceiveEnglishwordstobesemanticallyricherthandothe "Americans."Second,Fig.5.1showsthatboththebilingualsandmonolingualsgivehigher meaningratingstothecontentwordsthantothefunctionwords(ttest:p<.05).Thiswouldbe expectedbasedonthedifferentialcontributionsofcontentandfunctionwordstothesemantics ofEnglishsentences. Third,forthemeaningratings,thereisaninteractionbetweenlanguagebackgroundandlexical category.Theratingdifferencesbetweenthetwosubjectgroupsaregreaterforfunctionthan contentwords(ANOVA:p<.05).ThisisobservedinFig.5.1,andmoreexplicitlyinFigs.5.2a 5.2d.ThefourpanelsofFig.5.2comparepairsofcontentandfunctioncategoriesthathave relatedfunctionalrolesinEnglishsentencesbutthatdifferinsemanticvalue.Fig.5.2acompares nouns(e.g.,friends,engine)with"absolute"pronouns(e.g.,she,it,his,their)thatcansubstitute fornouns.(Relativepronounswerenotusedinthispanelbecausetheyhaveasubordinateaswell asanominalfeature.)Fig.5.2bcompareslexicalverbs(e.g.,hunting,run)withauxiliaryverbs (e.g.,had,can)thatprovideaspectormodalinformation.Fig.5.2ccomparesadjectivesand indefinitearticles,whichbothmodifynouns.Fig.5.2dcomparesadverbialmodifiers(e.g., rapidly)withprepositionsintroducingadverbialmodifierphrases(e.g.,swamwithgreatspeed). Allfourpanelsshowlargermeaning 89

FIGURE5.1MeaningratingsofmonolingualEnglishspeakersandbilingualChineseAmericans foreightlexicalcategories.Abbreviations:AM=American,CH=Chinese,AJ=adjectives,AA =adverbsthatmodifyadjectives;ADV=adverbsthatmodifyverbs;PRO=pronouns;ART= articles;CONJ=conjunctions;PREP=prepositions;DEF=definiteness;ORG=organization differencesbetweenChineseandAmericansforfunctionwordsthanfortheirrelatedcontent wordcategories,illustratingtheaboveinteraction. StructureRatings Twotrendsinthestructureratingsareofinterest.First,Fig.5.3showsthattheratingsare generallyhigherforthebilingualsthanthemonolinguals(ttest:p<.05).Thereversalforverbsis importantandisconsideredlater.ThesubjecteffectsinFig.5.3areconsistentwiththoseinFig. 5.1,withChineseratingsexceedingthoseofAmericansforbothstructureandmeaning. Figures5.2and5.3suggestaninteractionforstructureratingsbetweenlanguagebackgroundand lexicalcategorythatisoppositetotheinteractionobservedformeaningratings.Ingeneral, structuredifferencesbetweenthetwosubjectgroupsaregreaterforcontentthanfunctionwords. Thisisobservedfor 90

FIGURE5.2Meaning(AD)andstructure(EH)ratingsforrelatedpairsofcontentandfunction wordsformonolinguals(AM)andbilinguals(CH).Abbreviations:LEX=lexicalverbs;AUX= auxiliaryverbs;AVPREP=adverbialprepositions(introducingphrasesthatmodifyverbs);for otherabbreviations,seeFig.5.1. 91

FIGURE5.3StructureratingsofmonolingualEnglishspeakers(AM)andbilingualChinese Americans(CH)foreightlexicalcategories.Abbreviations:seeFig.5.1. thefunctionallyrelatedpairsoflexicalcategoriesinFigs.5.2e,g,andh.(Notethereversalof subjecteffectsforverbsinFig.5.2f.) ComparisonofStructureandMeaning Incomparingstructureandmeaningratings,thedatashowtwotypesofeffectsthatareof interest:(a)alexicalcategoryeffectand(b)asubjectgroupeffect. Figures5.4aand5.4bshowthattheratingsarehigherformeaningthanstructureforcontent words,butthatthereverseholdsforfunctionwords.Thisinteractionbetweenlexicalcategories andratingtypesholdsforbothmonolingualandbilingualsubjects(Ftest:p<.05).Thedatain Fig.5.5aretheproportionsofthetotalmeaning(M)plusstructure(S)valuesthatareassignedto structureforeachlexicalcategory,i.e.,S/(S+M).AgraphofM/(S+M)wouldobviouslyshow complementaryresults.TheFig.5.5dataareconsistentwiththoseinFig.5.4.Namely,both subjectgroupsshowhighervaluesofS/(S+M)forfunctionthancontentwords.Thus,higher valuesofM/(S+M)wouldoccurforcontentthanfunctionwords.Onewouldexpectthesetrends basedontherelativecontributionsoffunctionandcontentwordstothesyntaxandsemanticsof Englishsentences. 92

FIGURE5.4Acomparisonofstructure(S)andmeaning(M)ratingsformonolinguals(top graph)andbilinguals(bottomgraph).Abbreviations:S=structureratings,M=meaningratings; forotherabbreviations,seeFig.5.1. 93

FIGURE5.5Theproportionofstructureplusmeaning(S$PLM)assignedtostructure(S)for monolinguals(AM)andbilinguals(CH).Abbreviations:seeFig.5.1. Figure5.5alsoshowsaninteractionbetweensubjectgroupsandlexicalategories(Ftest:p<.05). (Again,verbsareanimportantexception.)Forfunctionwords,agreaterproportionofthetotal ratingvalueisdevotedtomeaningbyChinesethanbyAmericans.Incontrast,forcontentwords, agreaterproportionofthetotalratingvalueisdevotedtostructurebyChinesethanby Americans.Thisinteractionisconsistentwiththeinteractionbetweensubjectgroupsandlexical categoriesobservedinFig.5.2. ATHEORYOFSENTENCEPERCEPTIONINCHINESEENGLISHBILINGUALS Thissectionfocusesonreasonsforthedifferencesintheratingsbetweenthetwolanguage groups.Theratingdifferencesareconsideredinthecontextofsemanticandsyntacticdifferences betweentheChineseandEnglishlanguages,asdiscussedinthelinguisticsliterature.In particular,wefocusonanexpansionoftheComparisonHypothesis:(a)Inratinganyparticular Englishword'slinguisticcontributiontoitssentence,monolingualAmericanscompareitsroleto thatofmanyotherEnglishwordsinmanyotherEnglishsentencesthattheyhaveexperiencedin thepast.(b)Incontrast,thebilingualChinesecompareitsrole,inpart, 94

FIGURE5.6SchematicillustrationoftheComparisonHypothesis. tothatofChinese"words"inChinesesentencesthattheyhaveexperienced. Inessencethisisan"adaptationlevel"theory,asillustratedinFig.5.6.Thecollegeaged subjectshaveabout20yearsoflanguageexperience,andarepresumedtocompareanyparticular stimulustosomekindofaveragedbaseline,perhapsweightedbyrecency,aswellasfrequencyof languageexperience.Thus,AmericanandChineseratingsofthesameEnglishstimulishould differtotheextentthattheadaptationlevelsofthesetwosubjectgroupsdiffer.IftheEnglish stimuliareperceivedtocarrymore(orless)linguisticinformationthanbroadlyanalogous Chinese"words"comprisingpartofabilingual'sadaptationlevel,thenthosestimulishouldbe assignedcorrespondinglyhigher(orlower)ratings.InFig.5.6,theseperceptionsareindicatedby thedistancesbetweenwordsandtherelevantadaptationlevelforeachgroup. Wearguethat,ingeneral,Chinese"words"carrylesslinguisticinformationthananalogous Englishwords,producingalower(combined)adaptationlevelforthebilingualthanthe monolingualsubjects.ThiswouldyieldhigherratingsforEnglishwordsbytheChinesethanthe Americans,incontrastorcomparisontotheirparticularadaptionlevel.Table5.4containsan outlineoflinguisticdifferencesbetweenEnglishandChinesethatcouldpossiblyleadtodifferent perceptions,andthustodifferentratings,betweenthetwolanguagepopulations.Theselinguistic factorsarediscussedindetailbelow.Asafairamountofinformationiscovered,thesummaryin Table5.4mightbeausefulorganizationalaidtothereader. Thenextsectionsareorganizedasfollows.Firstweconsiderthemeaningratings,second,the structureratings,withaseparatepartonverbs,andthird,acomparisonofmeaningandstructure. Eachmainsectioncontainssubsectionsoncontentandfunctionwords. 95

TABLE5.4 LinguisticDifferencesBetweenEnglishandChineseWhichCouldLeadtotheRatingDifferences BetweenBilingualsandMonolinguals Meaningratings:bilinguals Forcontentandfunctionwords,ontheaverage: 1.EnglishwordshavemoremeaningperunitthanChinese. 2.EnglishwordsarelesscontextuallydependentthanChinese. 3.EnglishwordsaremoreabstractthanChinese. Structureratings:bilinguals Forcontentwords,ontheaverage: 1.Englishwordshavelessoverlapamonggrammaticalcategoriesthan Chinese. 2.EnglishwordshavegrammaticalinflectionsmoreoftenthanChinese. Forfunctionwords,ontheaverage: 1.EnglishwordsservemultiplesyntacticfunctionslessoftenthanChinese. 2.Englishwordsareoptional(vs.obligatory)lessoftenthanChinese. MEANINGRATINGS ContentWords Fig.5.1showsthatbilingualsgivehighermeaningratingsthanmonolingualsforallfour categoriesofcontentwords.LetuscompareChineseandEnglishtogainsomeinsightonwhy thesetrendsmighthaveoccurred. AssuggestedearlierapsychologicalassumptionunderlyingtheComparisonHypothesisisthat theperceivedcomparisonunitforthewrittenEnglishwordisgenerallythesingleChinese characterormorpheme,orperhapsarootmorphemeplusanoccasionalsecondcharacter( Kratochvil,1968).Thisisreasonable,because(a)formalunitboundaries(i.e.,spaces)inwritten Chinesearebetweenindividualcharacters(Henneetal.,1977),(b)pausesarepermissible betweenspokencharacters(Brandt,1943),and(c)Chinesemorphemesaretheminimalfree standingunitofmeaning(Halliday,1956).Giventheabove"morpheme/word"assumption, threedifferencesbetweenEnglishandChinese"contentwords"wouldsupporttheComparison Hypothesis. 1.AmountofMeaningperUnit.Englishcontentwordsinnewspapers(e.g.,theSouthampton Press)andmagazines(e.g.,TheReadersDigest)averageabout2syllables.Thecontentwords ratedinthepresentstudyaveraged2.1syllablesand1.8morphemes.BecauseChinese"content words"averageclosertoonemorpheme(Chen,1982;Karlgren,1962;Liuetal.,1975),the presentstimulishouldbeperceivedashavingmoremeaningthanabilingualadaptationlevel. Evenifthebilingualsubjectshadanadaptationlevelthatwasaweightedaverage 96

ofthetwolanguages(perhapswithEnglishweightedmoreheavily),pureEnglishwordsshouldstill beperceivedasconveyingmoremeaningthanthebilingualbaseline. Thefollowingisjustoneillustrationoftheabovelanguagedifference.AsChinese"words" generallyconsistofasinglerootmorpheme,withnoinflectionstoindicatecase,tense,andsoforth, whatlinguiststerm"grammaticalmeaning"isoftennotinherentinindividualcontentwords.For example,adjectivescontainnofeaturestodistinguishdegreesofcomparisonorgradationsof intensityintheirmeaning(e.g.,great,greater,greatest;hot,hotter,hottest).Instead,anentirephrase mustbeused,e.g.,"greatcompared(with)me,"or"greatsurpass(ing)me."Thus,theamountof meaninggenerallyassociatedwithasingleEnglishcontentwordisoftenspreadoverseveral Chinese"words,"resultinginlessmeaningperword(Forrest,1973). 2.ContextDependencyoftheUnit.Englishwordsgenerally,butnotalways,haveafairlywell definedmeaningassociatedwithaparticularorthographicrepresentation.Incontrast,themeaning ofaChinesecharacterisoftenlesspreciseandhighlyvariable.Itmayhaveseveralmeanings (homographs)thatarecontextdependent,andthatarenotcompletelydisambiguatedbycontext. Someexamplesofhomographsin"standard"Chinese(Mandarin)arethefollowing:yang=ocean, foreign;ken=theheel,tofollow,with,and;ts'ai=tocutcloth,todecide;ho=toshut,harmonious; ching=clean,only;keng=more,toalter,anightwatch(Brandt,1943).Farmorefrequentthan suchdistinctlydifferentmeaningsaresetsofrelatedmeaningsthatcorrespondtoasinglecharacter. Someexamplesarethefollowing:tien=lightning,electricity;chu=host,master,ruler;hsia=blind, reckless,heedless;t'ung=together,with,alike,identical;p'ai=hit,clap,pat;wen=hear,smell, news;shu=anumber,some,several,tocount,tocalculate;ming=aname,fame,reputation( Brandt,1943). Twokindsofempiricaldatasupporttheaboveillustrations,basedon(a)typesand(b)tokensinthe language.First,Karlgren(1962)reportsanaverageof10differentmeaningsperwordinasmall Chinesedictionaryof4,200words,withsomewordshaving3070differentmeanings.Second, 2,460differentcharactersaccountfor99%ofthetotalcorpusofaImillionwordChinesefrequency count.Ifweconsiderwords,ratherthancharacters,includingthemultimorphemiccompounds, about15,000wordsaccountfor99%ofthetext(Chen,1982;Liuetal.,1975).Incomparison, 40,000wordsaccountfor99%oftheImillionwordKueraFrancis(1967)Englishword frequencydictionary.Ifweassumethatthesetwoverylargecorpusescontainaboutthesame amountofmeaning,weseethatChineseexpressesthatmeaninginfarfewerunits,andthusfar moremeaningsperunitthanEnglish.Thus,theextenttowhichcontextdeterminestheexact meaningwillbemuchgreaterinChinesethaninEnglish. 97

BecauseChinesecontentwordshavemultipleshadesofmeaning,aswellasmultiplemeanings,and becausemanycontentwordscanserveasnoun,verb,adjective,oradverb(withoutorthographic modifications)dependingonthecontext,anational"languagegame"hasdevelopedthatmakesuse oftheseattributes(Herdan,1964).Thegameisakindofsentencelevelanagrams,andis"played" ratherseriouslybybothpoetsandmathematicians.Forexample,poemsmightbewritteninmatrix formatwitheachrowandcolumnhavingfoursentences,eachcomposedoffourwords.Thewords areselectedsothatvariouspermutationsoftherowsand/orcolumnswillresultinanewand meaningfulpoem.Forthepoet,the"game"istowriteinthisformatwithhighliteraryquality.For themathematician,thegamegoesbeyondasimpletwodimensionalmatrix,andinvolves topologicalstructureswithawidevarietyofmathematicalconstraints. Inlightoftheabove,ifthemeaningsofindividualEnglishwordsaremoreprecise,welldefined, reliable,orunambiguousthanthoseofChinesewords,bothinisolationandwithinasentential context,thenoneshouldratethoseEnglishwordshigherinrelationtoabilingualadaptationlevel. However,thereisnoreasonthatthestimuluswordsshoulddifferinamountorprecisionofmeaning fromwhattheAmericansubjectsareusedtoreadingintheirdailylives. 3.QualityofMeaning:AbstractnessandFrequencyofUnit.Pastresearchhasshownthat subjectsgivehigherratingsof"meaningfulness"and"importance,"andthattheyspendmoretime whenreadingwords(a)thataremoreabstract(asopposedtoconcrete),and(b)thatarelower frequencyorrarerinthevocabulary(Aaronson&Ferres,1984a;Ferres,1981).Bothofthese attributesshouldberelatedtotheratingdifferencesbetweenChineseandAmericansubjectsas follows. Regardingabstractness,itiswidelybelievedthat,ontheaverage,Chinese"words"arenotas abstractasthewordsofEnglishandotherWesternlanguages.InpartthisisbecauseChinesemakes littleuseofaffixestobuildabstractwords(e.g.,poly,super,anti;tion,ment,ist),asisdonein manyotherlanguages(Forrest,1973).Inaddition,thelackofabstractionmaysteminpartfromthe earlydevelopmentofChineseorthographyas"pictograms."Today,however,lessthan10%ofthe charactersmightbeconsideredtoembodysomevisualrepresentationoftheirsemanticcontent( Karlgren,1962;Venezky,1984). ToexpressabstractconceptsinmodernstandardChinese,onecanuseacombinationofcharacters that,inandofthemselves,havemoreconcretemeanings.Forexample,theEnglishword"truth" correspondstotheChinesestringmeaning"saytruewords;""sincerity"correspondsto"trueheart" (Forrest,1973);"emancipate"correspondsto"turnoverone'sbody,"and"youthful"correspondsto "yearsarelight"(Henneetal.,1977).Someabstractconceptsareformedbycompoundingor concatenatingtwomoreconcretemorphemepairs.Examplesofthisare:dawn(tianliang)=day+ brighten;adult(daren)= 98

big+person;algebra(daishu)=tosubstitute+numbers;traffic(chema)=vehicles+horses; language(yuyan)=words+speech;landscape(shanshui)=mountains+waters;length(chang duan)=long+short;analysis(fenxi)=divide+separate;tailor(caifeng)=cut+sew;start (dongshen)=move+body(Henneetal.,1977). Finally,therearemanycaseswhereChineselacksthewordforanabstractclassname(or superordinate),butdoeshaveawidevarietyofwordsforthesimpleconcreteinstancesofthe class.Forexample,Chinesehasnogeneralwordfor"orange,"butdoeshavewordsfor"thin skinnedorange,""orangewiththicklooseskin,"andforakindofsmalltangerine.Thereisno generalwordfor"carry,"buttherearewordsfor"carryonapole,""carryontheshoulders," "carrybetweentwomen,"or"carryinthearms"(Forrest,1973). Thus,relativetoChinese,manyEnglishwordswouldattainahighmeaningratingeitherbecause theyareconceptuallymoreabstractthantheChineseequivalent,orbecauseoneEnglishwordis worthtwoChinesecharacters.Asmallpercentageofpictogramsstillhavesomeresemblancetoa concreteimage(Kratochvil,1968),whereasEnglishlettersandwordshaveabsolutelyno correspondencetotheirmeaningortoanyrealworldimage. Regardingfrequency,theequivalentsofrarewordsintheChinesevocabularyareoftenexpressed ascombinationsofmorefrequentlyoccurringmorphemes.Aneducatedadultmayhavea vocabularyofover50,000concepts,butthesemayberepresentedbycombinationsoffarfewer monomorpherniccharacters.Chineseprintinghousesmayhaveabout6,000charactersandthe Biblewasprintedwithonly2,000characters(Brandt,1943).Frequentwordsmayberatedas lessmeaningfulthanrarewords(a)inpartbecausetheirhighfrequencyofoccurrencegains themmorelinksorassociationstootherwords,andthusmorecontextualredundancy,and(b)in partduetotheirlackofnovelty. Asnotedearlier,someChineseconceptsareformedasreduplicationsofindividualcharacters, thusyieldinganevenhigherfrequencyofoccurrenceforthosecharacters.Forexample, reduplicationisdonetoindicateadistributivemeaning:renren=everybody(ren=person). Reduplicationisalsodonetoindicateanintensificationofquality:Haohaoxuexi=studyreally well!(hao=well).Finally,reduplicationofmodifiersisusedforitemization(asisdonein English)yijuyijudejieshi=explainsentencebysentence;yiciyicidelai=cometimeandagain; yigeyigedenaqilai=pickuponebyone.Thus,Chineseformsmanyverbalconceptsfromfar fewer,butmorefrequentlyused,singlecharacters.Englishalsodoesthisonasyllablelevel(e.g., carefree,mailman,sometimes),butitappearsthatChineseusescombinationsmoreoften. Insum,thelinguisticsliteratureprovidesthreereasonswhyEnglishcontentwordsshouldbe perceivedascontainingmoremeaningthanChinesemorphemes.Ontheaverage,(a)English wordscontainmoremeaningperunit,(b)theyhavemeaningsthatarelesscontextually dependent,and(c)thewordsaresometimesmoreabstractandoccurlessfrequently.These linguisticdifferences 99

shouldyieldhighermeaningratingsbythebilingualsthanthemonolingualsubjects,supportingthe ComparisonHypothesis.Note,however,wearenotsuggestingthatChinesesentencesorparagraphs carrylessmeaningthanEnglishonesdo.LinguisticinformationinherentinindividualEnglish wordsmaybecarriedathigherlevelsinChinese.Chineseappearstobeamorecontextdependent languagethanEnglish. FunctionWords AsgraphedinFig.5.1,functionwords,also,aregivenhighermeaningratingsbybilingualthan monolingualsubjects.Explanationsanalogoustothosesuggestedforthecontentwordeffects appeartoholdforthefunctionwords. 1.AmountofMeaningperUnit.TheequivalentsofEnglishfunctionwordssometimesdon'texist inChinese.Further,toservesomeofthepurposesofEnglishfunctionwords,Chinesehassyntactic 2classifiedbyChineselinguistsas"empty"or"vacuum"words(i.e.,devoidofsubstantive markers meaning)andparticles,whichmayhavelessmeaningthantheapproximateEnglishequivalent. ThefirstillustrationoftheaboveEnglishChinesedifferenceisthatChinesereallyhasnoexact analoguesforEnglisharticles(Forrest,1973;Henneetal.,1977).Wordsfunctioninginsentences asnominalsarenotaccompaniedby"a"or"the."Thedefiniteorindefinitestatusofanounis generallyinferred,basedoncontext.However,ifitisdesired,forclarityoremphasis,thecardinal number"one"(yiben)andthedemonstrativeadjective"this,that"(i.e.,endboundmorphemes "zhei"and"nei")aresometimesusedrespectivelytofunctionasindefiniteanddefinitearticles. Thus,relativetothegeneralabsenceofarticlesinChinese,theEnglisharticlesshouldbeperceived tohaveafairamountofmeaningbythebilingualsubjects. Second,ChinesehasfewerpronounsthanEnglish,becauseseparateformsgenerallydonotexistor arenotusedtodiscriminateamongthefinegrainmeaningattributesofgenderand,tosomeextent, case.Forexample,he,him,she,her,anditarealltranslatedas"ta";I,me,andmyarealltranslated as"wo"(Henneetal.,1977).Thus,manyChinesepronounsserveasgeneralnounsubstitutesand havelessspecificmeaningthanEnglishpronouns.Further,eventhesemanticallygeneralpronouns aresuppressedwheneverpossibleasindicatedbelowbyparentheses.Forexample,itwouldbe semanticallyappropriateandsyntacticallycorrectforonepersontoansweraquestion"(I)don't know,"andforanothertorepeat"Hesays(he)doesn'tknow."Inneithercaseisthesubject ____________________ 2Englishhasnowordsservingaspuresyntacticmarkers.ForreadersfamiliarwithHebrew,the wordpronounced"et,"whichisobligatorybeforeadefinitedirectobject,providesanexampleof asyntacticmarker. 100

oftheverb"know"expressed(thepronounsaresuppressedandtheverbisnotinflected),andthe communicationisperfectlyintelligiblewithoutsuchexplicitsemanticinformation(Forrest,1973). AlthoughChinesedoeshavesomerealprepositions(andtheirassociatedlocativeadverbs)which indicaterelationshipsamongcontentwords,theyareoftennotused,astherelationshipisdeduced fromcontext.Further,ChineseequivalentsofmanyIndoEuropeanprepositionsaresuppliedby nounsandverbsplacedbefore,after,orcompoundedwithothersubstantives.Forexample,"inthe sea"wouldoccurinChineseasthenounpair"(the)midst(ofthe)sea."Insteadofsaying"eatswith chopsticks,"onewouldusetheverbpair"eatsusingchopsticks."Otherverblikesubstitutesfor prepositionsare"enter"for"into,"and"arriveat"for"until"(Forrest,1973).Thus,semantic attributescarriedbyEnglishfunctionwordsarecarried,inpart,byChinesecontentwords. PrepositionsprovideyetanotherlexicalcategorywherebilingualsubjectswouldcompareEnglish wordstoageneralpaucityofthegivenformclassinChinese. ConjunctionsalsoappeartoplayaweakersemanticroleinChinesethaninEnglish(Henneetal., 1977).Thecoordinateconjunctions"and"and"but"areoftenomittedfromcontext.Their functionalrolemightbeindicatedbypunctuation(orbypausesinspeech).The"universal" subordinatemarker"de"indicatesthatwords,phrases,orclauselikestringsprecedingitare modifiers.Otherthanitsgrammaticalfunction,"de"carrieslittlelexicalmeaning.Dependingon context,itsubstitutesforthat,which,who,whose,towhom,how,what,of,andthepossessivecase fornouns.Further,itisoftenomitted,yieldingunmarkedsubordination,wherethemodifyingword orstringimmediatelyprecedesthesubstantive. ThefrequentomissionoffunctionwordsinChineseresultsinsentenceswhosemeaningmustbe inferredfromnonlinguisticaswellaslinguisticcontext.Forexample,thecharacterstring"Jibuchi le"(literally,chicken(s)noteat[perfective])isinterpretedtomeaneither"Thechickensarenot eatinganymore"or"Asforchicken,Iamnotgoingtoeatanymore,"dependingonwhetherthe communicationcontextisthatoffarmsorrestaurants(Li,1971).Likewise,omissionoffunction wordsresultsinthesamecharacterstringmeaning"Iamgoingtocuthishair"or"Iamgoingtoget ahaircut,"dependingonwhetherthespeakerisabarberoracustomer(Forrest,1973).Such omissionsoffunctionwordswouldseeminglyimplythattheyaddlittleornoadditionalmeaningto apresumablyoverredundantcommunicationcontext.Hence,theyareaptlyclassifiedas"empty" words(Forrest,1973;Kratochvil,1968). Insum,althoughChinesedoeshavefunctionwords,(a)theyprimarilyserveassyntacticmarkers withlittlelexicalmeaning,(b)thesemanticsthattheydocarryisgeneralinnature,and(c)muchof thetimetheyareomitted,leavingunmarkedrelationshipstobeinferredfromcontext.Thus,by comparison,bilingualsshouldviewEnglishfunctionwordsashavingmoremeaningthantheir monolingualcohorts. 101

2.ContextDependencyoftheUnit.Aswithcontentwords,manyoftheChinesefunctionwords havemultiplemeanings,aswellasmultiplesyntacticfunctions,sothattheirexactsenseisheavily contextdependent.Forexample,dependingonthecontext,thesamecharacter"gen"canmean "with"(i.e.,instrumentalcase)"togetherwith,""and,""using,""following,"andincomparative constructions"as"or"from"(thesamelengthas;differentfrom)(Henneetal.,1977;Tewksbury, 1948).Asthecharacter"gen"(a)isoftenusedincombinationwithverbs,(b)oftenservesin 3inWesternlanguages,and(c)often linguisticunitsthatwouldbeadverbialprepositionalphrases hasverblikesemantics(e.g.,"using"and"following"arefeaturesofgen),itistermeda"coverb"by sinologists. Somecharactersservingprepositionalfunctionsarecontextdependentinadifferentway.The relative"place"and"time"wordsareoftennotfreestandingfunctionwords,butarerelegatedto beingsuffixesof,orcompoundedwithnouns.Thus,sinologistsdonotclassifythemasaseparate formclass,buttermthecombinedunitaplaceortimenominal,i.e.,acontent"word."Forexample, thestartboundmorpheme"li"meaning"in,on,during,insideof"occurswithnounsasfollows: jiangli(intheriver),wocheli(insidethesleepingcar),chuntianli(duringthespring)(Henneetal., 1977).Themorpheme"t'ou"formsadverbsorprepositionsoflocalitysuchaslit'ou(inside), shangt'ou(onthetop,i.e.,topside),wait'ou(outside),hsiat'ou(below,i.e.,underside).Itshould againbenotedthatsuchmorphemesmayserveratherdifferentlinguisticfunctionsinother contexts.Forexample,thesamecharactert'oualsoservesasatotallymeaninglesssyntacticmarker toindicatethatanothercharacterisbeingusedasanoun(e.g.,jiht'ou,thesun;shiht'ou,astone; chiht'ou,afinger)(Brandt,1943).Insum,manyChinesefunctioncharactershavegeneralor multiplemeaningsthataredeterminedinpartbythecontextratherthanbeinginherentinthelexical itemitself,orthatarenominalsuffixesratherthanbeingindependentfunctionwords. 3.QualityofMeaning:AbstractnessandFrequency.ThecomparisonofChineseandEnglish "functionwords"intermsofqualityofmeaningisnotasstraightforwardasitwasforcontent words.First,considertheperceivedabstractnessofChinesefunctionwords.Ontheonehand,many ofthem(e.g.,thesubordinatede,andthetimeandplace"prepositions")maybeperceivedasmore concretethantheEnglishequivalents.Thisisbecausetheyareoften(oralways)boundmorphemes thatoccurincombinationwithnominals(andothercontentwords),andsuchcontentwordsare generallyfarmoreconcretethanfunctionwords.Ontheotherhand,thefactthatmanyChinese functionwords ____________________ 3Indeedtheadverbialprepositionalphrases"paintswithabrush"and"walkswithhisfriends"can becloselyparaphrasedusingverbals:"paintsusingabrush"and"walksaccompanyinghis friends." 102

arealmostpuresyntacticmarkers,devoidofsemanticcontent,couldmeanthattheyareperceivedto bemoreabstractthantheirEnglishanalogue. Ifoneconsidersperceivedwordfrequency,acomparisonofEnglishandChinesefunctionwords mightalsoyieldeitherordering.Ontheonehand,thefact(a)thatmanyoftheChinesefunction charactersservemultiplepurposesand(b)thatmanyofthematoncecorrespondtoseveralEnglish words,shouldcauseChinesefunctorstohavehigherperceivedwordfrequencythanEnglish functors.Alternatively,thefactthatmanyChinesefunctorsareoftenomitted,leavingtheirroleto beinferredfromcontext,mayyieldlowerperceivedfrequencyvalues. Insum,althoughtheevaluationsofabstractnessandfrequencyattributesareequivocal,theabove evaluationsof(a)meaningperunitand(b)contextdependencyprovideanaccountingthatis consistentwiththemeaningratingsforfunctionwordsinFig.5.1.IncomparisontoEnglish, Chinesefunctorsappeartohavelessmeaningandtobemoreheavilycontextdependent.Thus,in linewiththeComparisonHypothesis,thisshouldleadtohighermeaningratingsforEnglish functorsbythebilingualthanthemonolingualsubjects. STRUCTURERATINGS ContentWords Figure5.3showshigherstructureratingsforbilingualsthanmonolingualsforthreeofthefour classesofcontentwords.Below,weconsidersomepossiblereasonsforthedominanttrend.The exceptioncategory,verbs,isdiscussedlater.TherearetwoimportantdifferencesbetweenChinese andEnglishthatcouldyieldaperceptionoflessstructureorsyntax,andthusaloweradaptation level,forChinesecontentwords:Chinesehasextensiveoverlapamonggrammaticalcategories,and alackofgrammaticalinflectionstodesignatecategories. 1.OverlapAmongGrammaticalCategories.InEnglish(andotherIndoEuropeanlanguages)a reliablepropertyofmostwordsistheirgrammaticalorlexicalcategory.Althoughasmall percentageofthewordsmaybelongtomorethanonecategory(e.g.,toplant,theplant),thisisso unusualthatschoolchildrenaretaughttolabelandtolearntheparticular"partsofspeech" associatedwitheachword(e.g.,"Nounsarethenamesofpersons,placesandthings").Byandlarge, linguistsagreeonthemajorlexicalcategoriesandonwhichwordsshouldbeclassifiedintowhich categories,anddictionariesindicatethosecategories.Thecategorystructurehassomuch commonalityacrossmajorfamiliesoflanguagesthatlinguistshavehypothesizedittobealinguistic universal(Halliday,1956;Kratochovil,1968). Insharpcontrast,classicalsinologistsdidnotclassifyideogramsintolexicalcategories.Asnoted earlier,onlyadistinctionbetween"fullwords"and"emp 103

tywords"wasmade.Therecentsubdivisionofthesubstantivesintonominalsandverbalsis basedmoreonthebehaviororfunctionofwordsintheirsentencecontextsthanonanyinherent propertiesoftheworditself.Tothisday,thereislittleagreementonsuchmatterswithinnational Chineselinguisticcommitteesthatarerevising,simplifying,andstandardizingthelanguage( Mathias&Kennedy,1980).Chinesechildrenarenottaughtlexicalcategoriesanddonotmake syntactictreediagramsofsentences(accordingtorecentimmigrantstudentsatNYU).Western linguistshavewrittenpapersillustratinghowtheIndoEuropeangrammaticalcategoriescanalso provideaconceptualframeworkforChinese(Halliday,1956;Simon,1937).Theiranalysesare notbasedoninherentlexicalattributes,butratheronthefunctioningofcharacterswithinstrings. However,evenafunctionalanalysisisoftencontroversial,becausesomanywordscanserve multiplefunctions.Forexample,thestring"babaxiedehao"hastwotranslationsintoEnglish( Kratochvil,1968).(1)"Fatherwriteswell,"wherebaba(father)isanoun,xie(towrite)isa verb,hao(well)isanadverb,andthemarkerdedenotestherelationshipbetweentheverband theadverb.(2)"Whatfatherwritesisgood,"wherebabaxie(fatherwritesorfather'swritings)is asubstantive,themarkerdedenotesthatthewordgrouporcharactercompoundisusedinstead ofanoun,andhaoisanadjective.Kratochvilmakesthepoints(a)thatthesemanticsofthe stringarenotsufficienttodeterminethegrammaticalcategories,(b)thatthefunctioningof characterswithinthestringasawholearenotsufficient,and(c)thatexampleslikethisquestion somesupposeduniversalsofbothsyntaxandsemantics.Infurthersupportofthenotionthat Chinese"words"donothaveinherentgrammaticalcategoriesaretheexamplesinHerdan's (1964)mathematical"languagegames."Inonesetofsententialanagrams,hestatesthatthere are108waysofselectingsetsofthreecharactersfromninetomakecoherentsentencesthat satisfyseveretopologicalconstraintsspecifiedinoneofPascal'stheorems.(Theseconstraintsare thetypethatcrosswordpuzzlesplaceonrowsandcolumns,butatthesyntacticandsemantic levels.)Asmentionedearlier,thiscanbedoneinChinesebecausealmostanysubstantive charactercanoccupyalmostany"contentword"category.Fromhisexample,herearethree charactersthatcaneachservemultiplegrammaticalcategories. sheng:life(noun);living,growing(adjective);tolive,produce,grow(verb) hsi:happiness(noun);happy(adjective);torejoice(verb);happily(adverb) chang:length,seniority(noun);long,elderly(adjective);prolong,grow(verb) InEnglishitisalsopossibletocreatesyntacticallyambiguousconstructionsbasedonmultiple grammaticalcategoriesforsomewords,buttheyaretheexceptionratherthantherule.Indeed, psychologistshavedevotedalotofdiscussiontostructureshavingthetwoparsings"theyare (eatingapples)"and"they(areeating)apples."Thisambiguityispossiblebecausethesameing 104

inflectionboth(a)denotesthepresentprogressiveverbformand(b)convertssomeverbclasses toadjectivalform.Therearefew(ifany)othertypesofsyntacticambiguitydiscussedinthe literatureonEnglish.Incontrast,asindicatedabove,bothChinesepoetryandwordgamesare basedontheveryhighfrequencyofwordswithmultiplesyntacticroles.StandardChinese grammarbookshaveentirechaptersdevotedtosuchsyntacticallyambiguousstructures.For example,the"complexstativeconstruction"isillustratedbythesentence"womenchidehen kaixin,"literallymeaning"weeat(marker)veryhappy."Dependingonthecontext,Chinese wouldinterpretthisaseither"Weateveryhappily"(happily=adverb)or"weatetothepointof beingveryhappy"(happy=adjective)(Li&Thompson,1981).Syntacticambiguitybasedon multicategorywordsisnotuniquetoChinese,butitiscertainlymoreprevalentthaninEnglish. AlthoughthereisenormoussyntacticflexibilitywithinChinese,thereareseverallinguistic devicesthatareusedtogethertoconveystructureforasentence.Thus,inaccordwithHalliday's andSimon'sviewsthatChinesedoeshaveatleastfunctionalgrammaticalcategorieswithina sentence,onefindsthefollowingtypesofpartiallyreliablecuestosyntax:(a)thesemanticsof thewordsand(b)ofthestring,(c)grammaticalmarkersandparticles(whichareoftenoptional), (d)syntacticaffixes(againoptional),(e)positionsofwordsrelativetoeachother(Chineseoften hassubjectverbobjectstringswithmodifiersgenerallyprecedingtheirmajorsubstantives),(f) extracontentwordstoindicateconjugationanddeclensioninformation(e.g.,yesterdayand tomorrowindicatepastandfuturetenseforverbals),(g)tone/stresspatternsforindividualwords, and(h)suprasegmentalintonationcontours. Grantedtheabovesyntacticcues,theComparisonHypothesisconcernsthecontributionto sentencestructurefromindividualEnglishwords.Totheextentthatstructuralinformationin Chineseisembodiedheavilyinthesentenceasawhole,ratherthaninherentintheindividual words,awordbasedadaptationlevelshouldbelowerforthebilingualthanthemonolingual subjects. 2.PaucityofGrammaticalInflections.Animportantpartofaword'ssyntacticcontributiontothe sentenceliesinitsgrammaticalinflections.InEnglishtheinflectionssignaltheword'slexical category(e.g.,noun),functionalrole(e.g.,subject),andgrammaticalrelationstootherwordsin thesentence.KratochvilmakesanimportantpointthataffixesinEnglishhaveakindofbinary significance,withtheabsenceoftheaffixbeingassignificantasitspresence.Thus,ifanEnglish nouncanoccurwithapluralsuffix(e.g.,students),itsoccurrenceinthesingularimmediately providesgrammaticalinformationaboutnumber.ThatisnotthecaseforChinese.Occasionally therearegrammaticalaffixes(e.g.,men,todenoteplural),buttheabsenceoftheaffixdoesnot implytheabsenceofthecorrespondinggrammaticalfeature(e.g.,droppingmenleaveswords thatcouldbeeithersingularorplural). 105

Forthemostpart,Chinesewordshavenoinflections."Nouns"generallyhavenonumber,gender,or casefeatures."Verbs"havenoconjugationstoindicatepast,present,orfuturetense."Adjectives" havenocomparativeorsuperlativeforms,andfrequentlycannotbediscriminatedfrom"adverbs" (e.g.,haomeansbothwellandgood,andisoftenconsideredtobeastativeverb"isgood,"Brandt, 1943).However,thisisanextremeview,asallWesternlanguageshave"exception"or"irregular" wordsthatlackthestandardinflectionor,occasionally,anyinflection.Further,assuggestedin subsection(1)above,Chinesehasotherlinguisticdevicestocommunicatethesyntacticinformation carriedbyinflectionsinWesternlanguages.But,asweindicatedearlier,thesedevicesareoptional andappeartobelessreliablecuestothesyntacticstructureofasentence,orofitswords,thanare mostsyntacticcodingdevicesinWesternlanguages. Insum,thepaucityofgrammaticalinflectionsforindividualcontentwordsisyetanotherreason whybilingualsmayperceiveEnglishcontentwordstocontributemorestructuralinformationtothe sentencethanChinesewordsdo.Consequently,thebilinguals'ratingofEnglishcontentwords relativetotheiroveralladaptationlevel,shouldexceedthoseofthemonolingualsubjects. FunctionWords Figure5.3showsthatbilingualsalsorateEnglishfunctionwordsascontributingmoretosentential structurethandomonolinguals.TherearethreeattributesofChinesethatwouldprovide explanationsforthistrend. 1.MultipleSyntacticFunctions.First,aswasthecaseforcontentwords,Chinese"functionwords" canoftenplaymorethanonesyntacticrole,whetherthesefunctorsbeaffixes,freeorbound markers,particles,ortruewords.Forexample,asindicatedearlier,t'ouisamarkertodesignate nouns,aswellasbeinganadverb/prepositionoflocality.Itisalsoa"classifier"(seefootnote1)of mules,oxen,anddonkeys!Anotherexampleofmultiplerolesservedbyfunctorsisprovidedbythe particletioritsvariantte(Brandt,1943).(a)Incombinationwithaverb,andpriortoanoun,ti servesasarelativepronoun(shuohuatijen;thepersonwhospeaks).(b)Itservesasasemantically emptymarkerthataverbisbeingmodifiedbyanadverb(nishuotit'aik'uai;youspeaktoofast).(c) Itindicatesthepossessiveorgenitivecase(wotishu;mybook,thebookofme).(d)Itindicates adjectiveswhentheyarenotfollowedbysubstantives(woyaomaihaoti;Iwanttobuygoodones). (e)Itoftenfollowsaverbtoindicatepasttense(t'atsot'ienhuihaiti;hecamebackyesterday).(f) Tifollowingyuformsakindofadjectiveclause(therearethosewhich/who....).(g)In combinationwithreduplicatedadjectives,tiformsadverbs(haohaoti;diligently,properly).(h) Combinedwithsubstantives,tiindicatesoccupationsorprofessions(beggar,merchant,gatekeeper, etc.).(i)Itisusedwiththenamesofmaterials(cheshihmut'outi;thisismadeofwood).Asmany Chinesefunctors 106

playmultipleroles,theirreliabilityaspsychologicalcuesinsignalinganyonesyntacticattribute islow. 2.Optionality.InadditiontothemultiplerolesplayedbyChinesefunctors,eventheirpresencein sentencesisoptional.Asdiscussedearlier,whenthereislittlechanceofambiguity,characters indicatingsyntacticfunctionsinChinesearegenerallydropped.Fromaninformationstandpoint, thismakesChineseanefficientandlowredundancycode,butfromaninformationprocessing standpoint,itcanyieldunreliablecuestosyntacticstructure.Asanexampleoftheoptionalityor nonreliabilityofsomeChinesefunctorsaslinguisticcues,itisinstructivetoreadthekey sentenceinastandardgrammarlessonontheparticlesma,a,na,ni,andyi,usedtoindicate interrogativeform,sentencetermination,andassertion(i.e.,strongdeclarative):"Norulesare givenfortheirusage,andtheirproperusagecanonlybeacquiredbycloseattentiontothe mannerinwhichtheChineseusethem"(Brandt,1943,p.133). 3.Affixes.Finally,someChinesefunctorsareinrealityaffixesorboundmorphemes,asopposed toindependentwords.Forexample,assuggestedearlier,somewordsthatwouldbefreestanding prepositionsinEnglishoccurasnounsuffixesfortimeandplacewords.Ifthose"affixes"arenot viewedasfullwords,theymaylowerthebilingual'sadaptationlevelforthecontributionof functorstothestructureofasentence. THEROLEOFVERBSINSENTENCESTRUCTURE Figures5.2,5.3,and5.5showthatverbsareanexceptiontothetrendsforstructureratingsof contentwords.Fortheotherthreecategoriesbilingualratingsexceedthoseofmonolinguals.But forverbs,thereverseoccurs(ttest:p<.05).IntermsoftheComparisonHypothesis,ifthe syntacticattributes(andthusthesubjectiveperceptions)ofChineseverbsexceedthoseof Englishverbs,ahigherbilingualadaptationlevelwouldresult.Thiswouldyieldalower structuralratingforEnglishverbsincomparisontothemonolinguals.Thelinguisticliterature supportstheseideas,asindicatedbelow. 1.GrammaticalAffixes.AlthoughChinesecontentwordshavenoinherentgrammatical inflections,"verbals"dohaveanextensivesetofboundparticlestoconveyaspectualandmodal 4(butrecentworkquestionstense).Thus,astandardintroductory information,andpossiblytense textonChinese(Brandt,1943) ____________________ 4Incontrasttoearlierlinguists,LiandThompson(1981)maintainthatChinesehasaspectbut nottensemarkers.Theydefinetense(i.e.,past,present,future)asthetimeofareportedevent relativetothetimeofspeaking,andaspectastimerelativetotheinternalmakeupofthe eventitself(e.g.,progressive:reading),ortootherevents. 107

containsgrammaticallessonsonmarkersandparticlesattheendsof20chapters,and16ofthose aredevotedprimarilytoverbrelatedparticles.Forexample,chiuindicatesimmediateaction(wo chiuch'u,Ishallgonow);chocanindicateanactionorstateinprogress(e.g.,sitting);lai+cho indicatespasttense;taoindicatescompletionofaction;ch'icanindicateeitherthebeginningor completionofaction;chao+pucanindicatetheimpossibilityorfailureofaction(e.g.,can't);pei indicatespassivevoice. 2.MultipleGrammaticalCategories.Theotherlexicalcategorieshavesubstantialoverlapof words,andthatisalsotruefor"verbal"characters.However,linguistshaverecentlyitemizeda largenumberofdistinctsubcategoriesforverbsalone,basedonbothsyntacticandsemantic features.Further,thelinguistsareinmoderateagreementonthedefinitionsofmanyofthese categories.Allofthesecategoriesmeetthegeneralsyntacticdefinitionsof"verbals,"as discussedbelow(e.g.,canserveasasentencenucleus[exceptcoverbs],andcanbenegated). Thiswidevarietyofsubcategoriespermits"verbs"tohandleabroadrangeofsyntactic functions,someofwhichareservedinEnglishbyadjectives,prepositions,andadverbs. Standardtextsdescribetheverbcategoriesasfollows(Henneetal.,1977).Verbsaredivided intotwomaincategories:descriptionverbsandrelationalverbs.Descriptionverbsareoftwo kinds:(a)actionverbs(e.g.,xie=write;yan=perform;dai=wear)and(b)qualityverbs(e.g., hao=isgood;qingchu=isclear;notethatthesewouldberepresentedasacopulaplusa predicateadjectiveinEnglish).Relationalverbsareoffourkinds:(a)existentialverbs(e.g.,you toexist,tobe,tohave);(b)classificatoryverbs(e.g.,shi=tobeequalto;xingtobenamed);(c) coverbs(e.g.,yong=using;xiang=facing;notethatthesewouldsubstitutefortheEnglish prepositions"with"and"toward");(d)modals(e.g.,hui=can).Thus,theChineseverbals,asa largeandsomewhatcohesivegroup,maybeperceivedascomprisingamorepowerfulandwide rangeoflinguisticfeaturesthanother"lexicalcategories." 3.SententialNucleusorCentral.ThemostimportantsyntacticattributeoftheChineseverbis thatitcanserveasthe"central"ofasentence.Everythingelseinthesentenceultimatelydepends ontheverb,andnoother"lexicalcategory"canservethisnuclearsyntacticfunction(Henneet al.,1977).Kratochvil(1968)presentsaschemeforsententialparsingoranalysiswherebyall wordsformakindoftreediagramasafunctionoftheir"linguistic"distancefromthecentral. Thisisakindofdeepstructureanalysis,asreducedsentencescanbeformedbysuccessively droppingwordsthataresyntactically(andsimultaneouslysemantically)mostremotefromthe central.InanyChinesesentenceapredicateisrequired,andtheverbisboththelinguistic(and psychological)centerorfocusofthepredicateandofthesentence.Informingreducedsentences fromtheoriginal,i.e.,allpossiblesentencesderivedfromtheoriginalstring,the 108

centralmustbeapartofeach.Further,theverbitselfistheminimalsentencethatcanbederived fromtheoriginal.Infact,inChinese,singleverbsentencesarequitefrequentandnormal.Some examplesfromHenneetal.(1977)arethefollowing:Jide=(I)remember;Pianyi=(It)ischeap; Shuizhe=(someone)issleeping;Nianwanle=(I)havefinishedreading(it);Chiguo=(He)has (inthepast)tasted(it). Incontrast,Englishrarelyhassingleverbsentences(exceptinunusualsituations,e.g.,the command:Run!)becausesubjectandobjectpronouns,auxiliaryverbs,andsoon,areobligatory ratherthanoptionalinagrammaticallywellformedstring.Thus,"minimal"Englishsentencesare generallythreewords:subject,verb,object.AscompleteChinesesentencesdonotrequirean explicitsubjectand/orobject,theverbalgainsamuchhighersyntacticstatusthananyother lexicalcategory.TheChineseverbisnotonlyaword;itisoftenanentiresentence!Thisis probablythemostimportantreasonwhythetrendsforstructureratingsofverbsdifferfromthose ofothercontentwordsinFigs.5.2,5.3and5.5(S/S+M).Thebilingualadaptationlevelfor verbsishigherthantheperceivedstructuralroleoftheEnglishverb. COMPARISONOFSTRUCTUREANDMEANING Thetypeofrating(structure/meaning)interactswiththelexicalcategoriesandwiththesubject groups.Forbothbilingualandmonolingualsubjectsthemeaningratingsexceedthestructure ratingsforcontentwords,butarelessthanthestructureratingsforfunctionwords(Figs.5.4and 5.5).Thesetrendsareexpectedifthesubjectivemeaningandstructureratingsaredeterminedin largepartbythesemanticandsyntacticattributesofthecontentandfunctionwords.Assuch "languagedetermined"effectsfor18lexicalcategoriesarediscussedbyAaronsonandFerres (1984a),wedonotconsiderthemhere.Instead,wewillfocusontheinteractionswiththe languagebackgroundofthesubjects. Inrelationtothetotal"linguisticvalue"(S+M),theperceivedemphasisofthevariouslexical categoriesonstructureversusmeaningdiffersforthetwosubjectgroups.InFig.5.5bilinguals, incomparisontomonolinguals,indicatefunctionwordsasconveyingrelativelymoremeaning (M/(S+M)),andcontentwordsasconveyingrelativelymorestructure(S/(S+M)).Thetopof 5Theleftgraphshows Fig.5.7showsthesetrendsaveragedforfunctionandcontentcategories. thatChineseperceiveEnglishfunctionwordsascontributingrelativelymoretomeaningthan Americansdo.Butthereverseistrueforcontentwords.TherightgraphshowsthatChinese perceivethecontentwordsascontributingrelativelymoretostructurethanAmericansdo. Again,thereverseistruefor ____________________ 5Verbshavebeenomitted,basedontheirreversalinstructuralattributesdiscussedearlier.The interactionsinthetopandbottomofFig.5.7arestatisticallysignificantbothwithandwithout verbs 109

FIGURE5.7Interactionsbetweenlanguagepopulationsandlexicalcategories.Toppanels:The proportionof"linguisticvalue"devotedtomeaning(left)andstructure(right).Bottompanels: ChineseAmericanratingdifferencesformeaning(left)andstructure(right). 110

functionwords.Thesegraphs,normalizedfortotallinguisticvalue,provideevidencethatthedata cannotbeattributedsimplytoaresponsebiastowardhigheroverallratingsforChinese. AcomparisonofFigs.5.2adwith5.2ehprovidesanotherwayofexaminingtheinteractions betweensubjectgroupandlexicalcategorythatareillustratedinFigs.5.5and5.7(top).Figs.5.2ad showgreaterChinese/Americanmeaningdifferencesforfunctionthancontentwords.Incontrast, Figs.5.2ehshowgreaterChinese/Americanstructuredifferencesforcontentthanfunctionwords. ThisinteractionisshownmoresimplyonthebottomofFig.5.7.Thesegraphsshowtherating differencebetweenChineseandAmericans,averagedforeachofthetwowordgroups(leftpanel). Thedifferenceinmeaningisgreaterforfunctionthancontentwords(rightpanel).Incontrast,the differenceinstructureisgreaterforcontentthanfunctionwords.Theseratingdifferenceswould correspondtodifferencesinadaptationlevelsinthetheoreticalmodelofFig.5.6.Theinteractions arereasonableinlightoftherelativemagnitudesofthevariouslinguisticdifferencesbetween ChineseandEnglish,asdiscussedlaterinthischapter. FunctionWords Thefunctionworddifferencesbetweensubjectgroupsweregreaterforthemeaningthanthe structureratings.Althoughfunctionwordsinbothlanguagesprovidesyntacticcues,Chinese functorshavelesssemanticvalueinpartbecausetheyplaymultiplesemanticroles,andtheiruseis oftenoptionalsemantically.However,formanyfunctionwordsinmanycontextsthereislittlepoint inevendiscussingsuchconsiderations,asthese"empty"wordshavevirtuallynosemanticattributes (Forrest,1973;Kratochvil,1968;personalcommunicationwithNYUstudents). ContentWords Aconsiderationofcontentwordsshowstheoppositetrends.Withtheexceptionofverbs,content worddifferencesbetweensubjectgroupsweregreaterforthestructurethanthemeaningratings. Althoughcontentwordsinbothlanguagesaresemanticallyimportant,syntacticdifferencesbetween ChineseandEnglishcontentwordsarelarge.Asdiscussedearlier,Chinesecontentcharactershave noinherentgrammaticalcategories,andtheyalsohavenoinflectionstoindicatecase,number, gender,orsyntacticagreementwithotherwordsintheirsentences(Forrest,1973;Kratochvil,1968 ;Simon,1937).KratochvilstatesthatChineselinguistshavebeentrying,unsuccessfully,to establishgrammaticalcategoriesforover75years.Hementionsthreecriteriatoclassifywordsinto grammaticalcategoriesinWesternlanguagesthatfailforChinese.First,Chinesewordscannotbe groupedbyorthographicorwordshapefeatures(generallyindicating 111

inflections,e.g.,s,ing).Second,Chinesewordscannotbegroupedreliablyaccordingto syntacticfunctioninrelationtootherwordsinthesentence,asindicatedbythe"fatherwrites" andthe"haircut"examplesgivenearlier.Third,theword'sinherentmeaningdoesnotreliably signalagrammaticalcategory,asobservedinconnectionwiththe"languagegame"examples. Insum,Fig.5.1,Fig.5.2,Fig.5.3,5.5and5.7showthatallEnglishwordscontributetoboth structureandmeaning,butthattheirrelativecontributionoremphasisdiffersamonglexical categories,withfunctionwordsemphasizingstructureandcontentwordsemphasizingmeaning. OfparticularinteresthereisthatbilingualsviewEnglishfunctorsascontributingrelativelymore tomeaning,andcontentwordsascontributingrelativelymoretostructure,incomparisonto monolingualsubjects. DEVELOPMENTALIMPLICATIONSOFCHINESEENGLISHDIFFERENCES OurtheoreticalevaluationoftheliteratureprovidesevidenceforChineseEnglishdifferencesthat haveimplicationsforcognitiveandlinguisticdevelopmentinbilingualChineseEnglishchildren. Inthissectionweprovidesomespeculationsonsuchdevelopment,basedonthepastliterature. Thediscussionprovidesideasaboutthecognitiveandlinguisticbackgroundthatleduptothe performanceofthe18yearoldsinoursentenceperceptionexperiment. 1.LanguageAcquisition.TheChineseEnglishlanguagedifferenceshaveimplicationsforseveral aspectsoflanguageacquisition.Forexample,HuangandHatch(1978)reportdataonthefirst6 monthsoflearningEnglishbya5yearoldTaiwaneseimmigrant,rearedmostlybyhis grandmother,whospokenoEnglish.PaulomittedtheEnglishcopularverbandarticlesin sentenceswheretheyareoptionalandgenerallyomittedinChinese(e.g.,This[isa]kite;That[is a]baby).Althoughtherecanbemanyreasonsforsuchomissions,Paul'spatternsoflanguage acquisitionmaybepartiallyattributedtohislanguagebackground.Analogously,Hakuta(this volume)pointsoutthatchildrenlearningEnglishwithJapaneseorKoreanastheirnative language,neitherofwhichhasanarticlesystem,havedifficultywithEnglisharticles.In contrast,hepointsoutthatEnglishFrenchbilingualchildrenneverconfusedefiniteand indefinitearticles,astheselanguageshavesimilarsystems. Bilingualchildren,suchasPaul,maybedelayedinotheraspectsofEnglishsyntax,relativeto monolingualchildren.Asecondaspectistheacquisitionofcorrectwordorderforsentences,as Chinese(beinga'topicprominent'language)permitsfarwidersyntacticvariationsthandoes English.Inthisrespect,ananalogousexampleisthecomparisonofItalianandEnglish,asItalian alsopermitsmorewordorderflexibilitythandoesEnglish.BatesandMacWhinney(1976, 1981)suggestthatbecause"Italiansapparentlylistentomassivedistor 112

tionsofSVOorderallthetime,"this"actuallydelaystheacquisitionofbasicwordorderinItalian children."Bates,McNew,MacWhinney,Devescovi,&Smith(1982)andMacWhinney,Bates,and Kliegl(1984)showthatthisdifferenceinpermissiblewordorderbetweenthetwolanguagesyields interestingexperimentaldataoncomprehensionofEnglishsentenceshavingambiguouswordorder cues.BilingualItalianAmericansproduceddataonwordorderingchoicesthatwereintermediate mixturesofthedataproducedbymonolingualAmericans(inEnglishsentences)andItalians(in Italiansentences).AsChinesewordorderingismoreflexiblethanItalian,wewouldexpect ChineseAmericanchildrentotakealongtimetoacquirethebasicEnglishwordordering principles,andthen,eveningradeschool,toproducemixtureorderingsthatarenotgrammatically correct.Indeed,thetwostudentswhoprovideddataforourearlierwritinganalysis,JCandJY,both producedafairnumberofsentenceswhosewordorderwasstylisticallyunusual,althoughnotquite ungrammatical. However,bilingualismmightyieldsometradeoffsduringacquisition.Thebilingualchildmight benefitfromthewellorganizedsemanticsysteminChinese,wherelargewordgroupsareformedin relationtocommonrootsorradicals.Hakuta'schapterinthisvolumesuggeststhatsuchbenefits couldoccur.Itwouldbeinterestingtoseedataonvocabularygrowthinbothlanguagesduringthe first15yearsforChineseEnglishbilingualsincomparison(a)tobilingualsintwoIndoEuropean languages(e.g.,FrenchEnglish;GermanEnglish)andalso(b)tomonolingualEnglishspeakers. Indeed,thefactthatChinese"contentwords"representbroadsemanticconceptsthatareusedover andoveragaininthesamelexicalform,regardlessoftheirgrammaticalcategory,mayleadto earlieracquisitionofthoseconcepts. 2.ReadingSkills.Animportantaspectoflinguisticdevelopmentisreading,andagain,abilingual backgroundcouldhavebothpositiveandnegativeimplications.AsChinesecharactersare morphemes,bilingualsmayperformlexicalaccessonmorphemesizedunitsinEnglishsentencesin caseswheremonolingualsmayoperateonmultimorphemicwords.Usingsmallerpsychological unitscouldyieldslowerEnglishreadingratesatlowerlevelsofprocessing.Ontheotherhand, greaterglobalcontextsensitivityandatendencytoskipoverfunctionwordsthatwouldbeoptional inChinesecouldpossiblyspeeduptheEnglishreadingrateforbilingualsathigherlevelsof processing. Ifweconsidertheearlystagesoflearningtoread,thedifferencesbetweenanideographicand phonemiclanguagemayplayastrongroleinacquiringdecodingskills.Analogously,Baronand McKillop(1975)classifytheirmonolingualAmericansubjectsintotwogroupsbasedonindividual differencesindecodingskills.The"Chinese"presumablydowholeworddecodingandhavea "directpath"fromtheorthographytothesemanticsofaword.Incontrastthe"Phonecians" presumablydecodelettersorsyllablesintophonologicalrepresentationsbeforeextractingthe semantics.Pastresearchsuggeststhatthe"Chinese"may 113

progressthroughtheearlystagesofreadingfasterandmayswitchfrom"controlled"to"automatic" decodingsoonerthanthe"Phonecians."But,whenfacedwithnewwords,the"Chinese"willlackor beweakinappropriatephonemicdecodingskills. Ifweconsiderhigherlevellinguisticprocessingduringreading,ourownresearch(Aaronson& Ferres,1984b,1985;Aaronson&Scarborough,1976,1977)suggeststhatthereshouldbe differencesbetween"Chinese"and"Phonecians."Whenreadingforrelativelycompleteandlong termretentionofinformation,asisdonewhenstudyingforexams,the"Phonecians"shouldfare better.Thisisbecauseaphonologicalrepresentationofthesurfacestructure(a)mayyieldamore durablemnemonicrepresentationofindividualwordsand(b)mayaidinparsingthesentenceinto memoryspansizedphraseunitsor"chunks,"assyntacticandphonologicalphrasesgenerally coincideinEnglish.Ontheotherhand,whenreadingforimmediatecomprehensionofgist,asis donewhenskimmingnewspapersornovels,the"Chinese"shouldfarebetter.Aphonological representationhereappearstobeunnecessary,toslowthereadingprocess,andtointerferewith higherlevelintegrationthatcutsacrossphrasesandsentences.Thus,thebilingualsmaydobetter thanthemonolingualsundertheconditionsfavorableto"Chinese"individualdifferencesinreading, atboththedecodingstagesandthehigherlevelstagesoflinguisticprocessing. 3.SocialCommunicationSkills.Ifweconsidermoregeneralaspectsofwrittenandspoken communicationthatoccurasthechildmatures,itappearsthattheChinesedependquiteheavilyon context,bothlinguisticandnonlinguistic,toconveythecompletemeaningunderlyingindividual sentences.IfsuchrelianceoncontextexceedswhatistypicalformonolingualAmericans,itcan havebothpositiveandnegativeresults.Thebilingual'sproduction,particularlyspoken,mayleadto misperceptionbyamonolingualAmericanlistener,assomeofthemessagemaybeexpressed implicitlyviacontext,ratherthanexplicitly.Thiscouldleadtosocialproblemsduringthepreteen andteenageyears.Ontheotherhand,agreatersensitivitytocontextbythebilingualmayleadto bettercomprehensionandintegrationofinformationinEnglishwhenheorshecommunicateswith others. Inamoregeneralsense,L.Bloom(1981)stressestheinteractionbetweencontext,cognition,and languagedevelopment.Shesuggeststhatlanguagedevelopmentisdeterminedby(a)the"target" languageinthechild'shomeandcommunity(inrelationtoasecondlanguage),(b)thechild's environmentalcontext(whichisonesourceofthemeaningsofutterances),(c)thechild'ssocial context(whichisonesourceofcommunicativeintentions),and(d)thechild'scognition(as utterancesarebasedonknowledge).IfweassumethatyoungChineseAmericanbilingualsoften haveChineseastheir"target"language,andarebiculturalaswellasbilingual,wewouldexpect theirsocialcommunicationskillstodifferfromthoseofmonolingualAmericans.Thesedifferences may 114

havebothadvantagesanddisadvantages.InourexperienceteachingatNYU,whichisaboutamile northofNewYorkCity's"Chinatown,"wehavenoticedafewdifferencesinthelinguisticand nonlinguisticaspectsofsocialcommunicationinourstudents.Anecdotally,theChineseAmerican bilingualsareoftenlessverballyfluent,andmoresociallyshy,thantheirmoretypical"NewYorker" stylecohorts.But,theircommunicationappearstobemorecontextsensitive(bothsocialandverbal context)thantheircitywiseclassmates. 4.AcademicPerformance.Animportantpartofchildren'sdevelopmentistheirschoolperformance. Asmuchofthisperformanceiswritten,languageskillscanplayamajorrole.Asindicatedearlier, bothJCandJYhadquitenoticeablewritingproblems.Itappearsthatthismayinpartbe responsibleforlowerSATscoresandcoursegradesthanwouldbeexpectedbasedontheir nonlinguisticabilities.AnespeciallydifficultacademicproblemoccurredforVT,whohadcome frommainlandChinaabout5yearspriortohisNYUgraduation.Allofhisgraduateschool applicationswererejectedbecausehisverbalGREwasabout400(upfrom300theprecedingyear). ThiswasinthecontextofhisstraightA,PhiBetaKapparecord,andreceiptof$500forthebestall universityhonorsthesis.Indeed,onewouldexpectacademicdifficultiestobeevenmoreseverein primaryandsecondaryschool,wherealargeamountoftheacademicperformanceandofthe student'sgradesarebasedonthedevelopmentofwritingskills. However,thefactthatChineseisamoreconcretelanguagethanEnglishmayleadtosomepositive outcomesinacademicwork.Basedonlargenumbersofanecdotes,itappearsthatChinese Americansproduceanunusuallyhighrateofstudentswhoexcelinscienceandengineering, perhapsbecausetheyformconcreteandmeaningfulrealizationsfromabstractideas. Insum,thedifferencesbetweentheChineseandEnglishlanguagescanleadtobothadvantagesand disadvantagesinthelinguisticandcognitivedevelopmentofChineseAmericanbilinguals.These developmentaldifferencesareconsistentwithboththeComparisonHypothesisandtheCognition Hypothesisputforthatthebeginningofthispaper.Cummins'chapterinthisvolumeprovides furtherinformationonfactorsthatcanleadtoadvantagesordisadvantagesforthebilingual's cognitiveandlinguisticdevelopment.Theseincludelinguisticandextralinguisticfeedbackinthe communicationcontext,thenatureofthecognitivetaskdemands,andthechild'sproficiencyin eachlanguage. SUMMARY ThecomparisonofEnglishwithanonIndoEuropeanlanguagerevealedinterestinginformation aboutlanguagespecificdifferencesincognitiveprocessingthathavenotbeenobservedwhenthe bilingual'slanguagesarebothfromthe 115 havebothadvantagesanddisadvantages.InourexperienceteachingatNYU,whichisaboutamile northofNewYorkCity's"Chinatown,"wehavenoticedafewdifferencesinthelinguisticand nonlinguisticaspectsofsocialcommunicationinourstudents.Anecdotally,theChineseAmerican bilingualsareoftenlessverballyfluent,andmoresociallyshy,thantheirmoretypical"NewYorker" stylecohorts.But,theircommunicationappearstobemorecontextsensitive(bothsocialandverbal context)thantheircitywiseclassmates. 4.AcademicPerformance.Animportantpartofchildren'sdevelopmentistheirschoolperformance. Asmuchofthisperformanceiswritten,languageskillscanplayamajorrole.Asindicatedearlier, bothJCandJYhadquitenoticeablewritingproblems.Itappearsthatthismayinpartbe responsibleforlowerSATscoresandcoursegradesthanwouldbeexpectedbasedontheir nonlinguisticabilities.AnespeciallydifficultacademicproblemoccurredforVT,whohadcome frommainlandChinaabout5yearspriortohisNYUgraduation.Allofhisgraduateschool applicationswererejectedbecausehisverbalGREwasabout400(upfrom300theprecedingyear).

ThiswasinthecontextofhisstraightA,PhiBetaKapparecord,andreceiptof$500forthebestall universityhonorsthesis.Indeed,onewouldexpectacademicdifficultiestobeevenmoreseverein primaryandsecondaryschool,wherealargeamountoftheacademicperformanceandofthe student'sgradesarebasedonthedevelopmentofwritingskills. However,thefactthatChineseisamoreconcretelanguagethanEnglishmayleadtosomepositive outcomesinacademicwork.Basedonlargenumbersofanecdotes,itappearsthatChinese Americansproduceanunusuallyhighrateofstudentswhoexcelinscienceandengineering, perhapsbecausetheyformconcreteandmeaningfulrealizationsfromabstractideas. Insum,thedifferencesbetweentheChineseandEnglishlanguagescanleadtobothadvantagesand disadvantagesinthelinguisticandcognitivedevelopmentofChineseAmericanbilinguals.These developmentaldifferencesareconsistentwithboththeComparisonHypothesisandtheCognition Hypothesisputforthatthebeginningofthispaper.Cummins'chapterinthisvolumeprovides furtherinformationonfactorsthatcanleadtoadvantagesordisadvantagesforthebilingual's cognitiveandlinguisticdevelopment.Theseincludelinguisticandextralinguisticfeedbackinthe communicationcontext,thenatureofthecognitivetaskdemands,andthechild'sproficiencyin eachlanguage. SUMMARY ThecomparisonofEnglishwithanonIndoEuropeanlanguagerevealedinterestinginformation aboutlanguagespecificdifferencesincognitiveprocessingthathavenotbeenobservedwhenthe bilingual'slanguagesarebothfromthe 115

samelanguagefamily.EnglishwritingerrorsbyChineseAmericanbilingualscanbeattributed toseverallinguisticattributes,includingthepaucityofinflectionsoncontentwordsandthe optionalityoffunctionwordsinChinese.Further,ChineseEnglishbilingualsshowdifferencesin handlingabstractionthatcanbeattributedtotheabsenceofexplicitlymarkedlinguistic structures,suchascounterfactuals,genericarticles,andentification.Thesedatasupportedthe CognitionHypothesis,thatlanguagespecificdifferencescanleadtoassociatedcognitive performancedifferencesbetweenbilingualsandmonolingualsinprocessingthemeaningand structureofsentences. Inanexperimentontheperceptionoflexicalcategories,wefoundthatChineseAmerican bilinguals,whowerenativespeakersofbothlanguages,ratedthecontributionsofindividual contentandfunctionwordstothemeaningandstructureofEnglishsentencesinwaysthat differedfrommonolingualAmericans,andthatthetrendscouldbeaccountedforintermsof semanticandsyntacticdifferencesbetweenthetwolanguages.Onanabsolutebasis,bilinguals generallyratedEnglishwordsinthevariouslexicalcategoriesascontributingmoretoboth meaning(M)andstructure(S)thandidmonolinguals.Onarelativebasis,bilingualsgenerally ratedEnglishcontentwordsascontributingmoretostructure(i.e.,S/(M+S)scores)thandid monolinguals.Also,bilingualsratedEnglishfunctionwordsascontributingrelativelymoreto meaning(i.e.,M/(M+S)score)thandidmonolinguals.ThesedatasupportedtheComparison Hypothesis:thatbilingualsperceiveEnglishwordsinrelationtotheirpastcombinedlinguistic experienceinbothlanguages,whereasmonolingualsperceiveEnglishwordsinrelationtotheir pastexperiencewiththeEnglishlexicalcategorysystem. Finally,thedevelopmentalimplicationsofChineseEnglishdifferenceswereconsidered. Languagespecificdifferencescouldleadtoqualitativedifferencesbetweenbilingualsand monolingualsinlanguageacquisition,readingskills,socialcommunicationskills,andacademic performance.ThesedifferencesmayarisebecauseChinesediffersfrommostIndoEuropean languagesinwaysthatcastdoubtontheexistenceofcertainlinguistic"universals"suchasthe roleofwordsasbasiclinguisticunits,theexistenceofalexicalcategorysystemthatisinherently markedatthewordlevel,andofreliablewordorderingcuestosentencestructure. ACKNOWLEDGMENTS ThisresearchwassupportedinpartbyPHSGrantMH16,496toNewYorkUniversityandbyan NYUResearchChallengeGrant.WethankMartinBraine,MurrayGlanzer,PeterHomel, MichaelPalij,ChingFanSheu,andBrainWattsforreadingearlierdraftsofthispaperandfor theirhelpfulcomments. 116

REFERENCES Aaronson,D.,&Ferres,S.(1983a).Lexicalcategoriesandreadingtasks.Journalof ExperimentalPsychology:HumanPerception&Performance,9,675699. Aaronson,D.,&Ferres,S.(1983b).Amodelforcodinglexicalcategoriesduringreading. JournalofExperimentalPsychology:HumanPerception&Performance,9,700725. Aaronson,D.,&Ferres,S.(1984a).Astructureandmeaningbasedclassificationoflexical categories.InS.White&V.Teller(Eds.),Discoursesinreadingandlinguistics(pp.2157).New York:NewYorkAcademyofSciences,AnnalsNo.433. Aaronson,D.,&Ferres,S.(1984b).Readingstrategiesforchildrenandadults:Someempirical evidence.JournalofVerbalLearning&VerbalBehavior,23,189220. Aaronson,D.,&Ferres,S.(1985).Readingstrategiesforchildrenandadults:Aquantitative model.PsychologicalReview,93,89112. Aaronson,D.,&Scarborough,H.S.(1976).Performancetheoriesforsentencecoding:Some quantitativeevidence.JournalofExperimentalPsychology:HumanPerceptionand Performance,2,5670. Aaronson,D.,&Scarborough,H.S.(1977).Performancetheoriesforsentencecoding:Some quantitativemodels.JournalofVerbalLearning&VerbalBehavior,16,277303. Au,T.K.(1983).ChineseandEnglishcounterfactuals:TheSapirWhorfhypothesisrevisited. Cognition,15,155187. Au,T.K.(1984).Counterfactuals:InreplytoAlfredBloom.Cognition,17,289302. Au,T.K.(1985).Languageandcognition.InL.L.Loyd&R.L.Schiefelbusch(Eds.), LanguagePerspectivesII.Baltimore,MD:UniversityParkPress. Baron,J.,&McKillop,B.J.(1975).Individualdifferencesinspeedofphonemicanalysis,visual analysisandreading.ActaPsychologica,39,9196. BasicChinesecoursein25lessons(1981).China:EasternNormalUniversityPress. Bates,E.,andMacWhinney,B.(1976).Languageandcontext:Studiesintheacquisitionof pragmatics.NewYork:AcademicPress. Bates,E.,&MacWhinney,B.(1981).Secondlanguageacquisitionfromafunctionalist perspective:Pragmatic,semantic,andperceptualstrategies.InH.Winitz(Ed.),Nativelanguage andforeignlanguageacquisition,AnnalsoftheNewYorkAcademyofSciences(Vol.379,pp. 190214). Bates,E.,McNew,S.,MacWhinney,B.,Devescovi,A.,&Smith,S.(1982).Functional constraintsonsentenceprocessing:Acrosslinguisticstudy.Cognition,11,245299.

Bloom,A.(1981).Thelinguisticshapingofthought:Astudyintheimpactoflanguageon thinkinginChinaandthewest.Hillsdale,NJ:LawrenceErlbaumAssociates. Bloom,A.(1984).CautionThewordsyouusemayaffectwhatyousay:AresponsetoTerry KitfongAu's"ChineseandEnglishcounterfactuals:TheSapirWhorfhypothesisrevisited". Cognition,17,275287. Bloom,L.(1981).Theimportanceoflanguageforlanguagedevelopment:Linguistic determinisminthe1980s.InH.Winitz(Ed.),Nativelanguageandforeignlanguageacquisition, AnnalsoftheNewYorkAcademyofSciences(Vol.379,pp.160171). Bond,Z.S.,&Gray,J.(1973).Subjectivephrasestructure:Anempiricalinvestigation.Journal ofPsycholinguisticResearch,2,259266. Brandt,J.J.(1943).IntroductiontospokenChinese(Americaned.).NorthManchester,IN: HeckmanBindery. Caramazza,A.,Grober,E.,&Garvey,C.(1977).Comprehensionofanaphoricpronouns. JournalofVerbalLearningandVerbalBehavior,16,601609. Chen,Chaoming.(1982).AnalysisofpresentdayMandarin.JournalofChineseLinguistics,10, 282358. 117

Cheng,R.L.(1978).TeachinginChineseoutsideChina.InB.Spolsky&L.Cooper(Eds.), Casestudiesinbilingualeducation.Rowley,MA:NewburyHouse. Clark,H.H.,&Clark,E.V.(1977).Psychologyandlanguage.NewYork:HarcourtBrace Jovanovich. deP.A.Villiers(1974).Imageryandthemeinrecallofconnecteddiscourse.Journalof ExperimentalPsychology,103,263268. Ferres,S.(1981).Awordclassencodingmodelforadultsandchildrenincomprehensionand recalltasks.Doctoraldissertation,NewYorkUniversity. Fincher,B.H.(1978).BilingualismincontemporaryChina:Thecoexistenceoforaldiversity andwrittenuniformity.InB.Spolsky&R.L.Cooper(Eds.),Casestudiesinbilingual education.Rowley,MA:NewburyHouse. Forrest,R.A.D.(1973).TheChineselanguage.London:Faber&Faber. Foss,D.J.,Cirilo,R.K.,&Blank,M.A.(1979).Semanticfacilitationandlexicalaccessduring sentenceprocessing:Aninvestigationofindividualdifferences.MemoryandCognition,5, 346353. Gibson,E.J.,&Levin,H.(1975).Thepsychologyofreading.Cambridge,MA:MITPress. Halliday,M.A.K.,(1956).GrammaticalcategoriesinmodernChinese.Transactionsofthe PhilologicalSociety,London,178224. Hatch,E.M.(1978).Secondlanguageacquisition.Rowley,MA:NewburyHouse. Healy,A.,&Miller,G.A.(1971).Therelativecontributionofnounsandverbstosentence acceptabilityandcomprehensibility.PsychonomicScience,24,9496. Henne,H.,Rongen,O.B.,&Hansen,L.J.(1977).AhandbookonChineselanguagestructure. Oslo,Norway:Universitetsforlaget. Herdan,G.(1964)."ThestructuralisticapproachtoChinesegrammarandvocabulary".The Hague:Mouton&Co. Huang,J.,&Hatch,E.A.(1978).Achinesechild'sacquisitionofEnglish.InE.M.Hatch(Ed.), Secondlanguageacquisition.Rowley,MA:NewburyHouse. Jeremy,R.J.(1978).Useofcoordinatesentenceswiththeconjunctionandfordescribing temporalandlocativerelationsbetweenevents.JournalofPsycholinguisticResearch,7,135150. Karlgren,B.(1962).SoundandsymbolinChinese.HongKong:CathayPress. Kratochvil,P.(1968).TheChineselanguagetoday.London:HutchinsonUniversityPress. Kucera,H.,&Francis,W.(1967).ComputationalanalysisofpresentdayAmericanEnglish.

Providence,RI:BrownUniversityPress. Levelt,W.J.M.(1970).Hierarchicalchunkinginsentenceprocessing.Perceptionand Psychophysics,8,99103. Li,C.N.,&Thompson,S.A.(1981).MandarinChinese:Afunctionalreferencegrammar. Berkeley,CA:UniversityofCaliforniaPress. Li,Y.C.(1971).AninvestigationofcaseinChinesegrammar.SouthOrange,NJ:SetonHall UniversityPress. Liu,I.M.,Chaung,C.J.,Wang,S.C.(1975).Frequencycountof40,000Chinesewords.Taipei, Taiwan:LuckyBooks. Lyons,J.(1979).Introductiontotheoreticallinguistics.London:CambridgeUniversityPress. MacWhinney,B.,Bates,E.,&Kliegl,R.(1984).Cuevalidityandsentenceinterpretationin English,GermanandItalian.JournalofVerbalLearningandVerbalBehavior,23,127150. Mathias,J.,&Kennedy,T.L.(Eds.)(1980).Computers,languagereformandlexicographyin China:AreportbytheCETAdelegation.Pullman,WA:WashingtonStateUniversityPress. Miller,G.A.,&JohnsonLaird,P.N.(1976).Languageandperception.Cambridge,MA: HarvardUniversityPress. Sapir,E.(1921).Language:Anintroductiontothestudyofspeech.NY:Harcourt,Brace& World. Simon,W.(1937).HastheChineselanguagepartsofspeech?TransactionsofthePhilological Society,London,99119. 118

Spolsky,B.,&Cooper,R.L.(1978).Casestudiesinbilingualeducation.Rowley,MA: NewburyHouse. Tewksbury,M.G.(1948).SpeakChinese.NewHaven,CN:FarEasternPublications,Yale University. Venezky,R.(1984).Language,scriptandreadinginChina.InJ.Y.Mei,ReadinginChina.New York:NationalCommitteeonUSChinaRelations. 119

[Thispageintentionallyleftblank.] 120

6 AcquiringandProcessingFirstandSecondLanguages:CommentsonHakuta,Cummins,and AaronsonandFerres MartinD.S.BraineNewYorkUniversity Ifoundallthreechaptersinterestingandoftenprovocative,andalmostallmycommentsare associationstotopicsdiscussed,ratherthandirectcommentsonthechapters. Myfirstpointisaratherobviousonethatisimplicitinmanydiscussions.Itdealswiththequestion ofwhatthelearner'serrorsaredueto,andtherelationofthistointerferencebetweenL1andL2. WhatIsuggestisthatanimportantsourceoferrorsoccurswhenthelearnerhasanideatoexpress butdoesn'thavethemeanstoexpressit;i.e.,thelearneristryingtosaysomethingandlackssome oftherulesofthelanguageneededtomapthedesiredmeaningintospeech.Ofcourse,forthis sourceoferroritdoesn'tmatterwhetheritisfirstorsecondlanguagelearning,buttheavailable strategiestocopewiththesituationarenotquitethesameinthetwocases;inparticular,thesecond languagelearnerhasmoreoptionsthanthefirst. Forexample,itmaybejustsomemorphologicalexceptionsthatthelearnerlacks.Supposethata learnerisatthestageoflearningEnglishwhentheonlywayheorsheknowshowtoformapast tenseistoputa/d/attheendofaverb.Theniftheywanttosaythatsomebodywentsomewhere theyaregoingtohavetogeneratetheformgoed;theyhavenootherwayofsayingit.Thus,theerror isduetothelearner'stakingthelineofleastresistance;heorsheisexpressingtheideaintheonly waytheyknowhow. Nowletmetakeanotherexample,frommyformer24montholdsubjectAndrew.Hisparentswant toputAndrewtobed,buthedoesn'twanttogo,andwantsthemtoknowthat.Hehasonenegation forminhisrepertory,usedfornegativerequests:itconsistsofsayingnofollowedimmediatelybya wordforwhateveritishedoesn'twant,allputunderoneintonationcontour.Hisutterance,Nobed, expresseshisrejectionintheonlywayheknowshow. 121

TheseexamplesoferrorsmaybecomparedwiththosethatHakutacitesfromhissubjectUguisu. ShewantstosaythatshemadeanaccidentalerrorandsheknowstheEnglishwordmistake,but shedoesn'tknowwhetheritisanounoraverb.ItturnsoutthattheJapanesewordfor"mistake" isaverb:youcouldsayImistakeinJapanese,althoughinEnglishyouhavetosayImadea mistake.AndsoUguisucomesoutwiththingslikenotthereImistake;don'tgivememore becauseyou'remistaking;I'vejustmistakeandIjustskipped,inwhichtheuseofmistakeasa verbisclearlyborrowedfromJapanesestructure. NowwhatI'dliketosuggestisthatwhenL1learnersaretryingtoexpressanidea,thereare onlyafewthingstheycando.First,theycanusewhateverrulestheyknowthatareadequate. Theywillmakeerrorswhenevertherearelittlesubrulesorexceptionsthattheydon'tknow,asin thesemorphologicalcaseslikegoedandsingedandthoughtedandthelike. Asecondstrategyistosimplifybyomissions,asinthemany"telegraphic"utterancesofyoung children(Brown&Bellugi,1964;Brown&Fraser,1963).inwhicharticlesandothersmall wordsandmorphemesareomittedinchildrenteamingEnglishasL1.Detailisomittedbecause thelearnerlackscompletecommandofrulesthatwouldmapinthedetail(Braine,1974). Apartfromkeepingquietandnotexpressinganything,theL1isbasicallylimitedtothese choices.However,theL2learnerhasanotherpossibility,tousetherulesofL1forL2.Thatis, ifyoudon'tknowhowtosayyourideainL2,thensayitthewayyouwouldsayitinL1 substitutingL2wordsforthoseofL1.Thatiswherethetransferorinterferenceerrors presumablycomefrom.Giventhemanysimilaritiesofformamonglanguages,thisstrategywill almostalwaysleadtoacomprehensibleutteranceandoftentooneacceptabletonativespeakers. ThepointisthatalthoughL2learnersmakesomeerrorsthatdifferfromthoseofL1learners, theerrorsbasicallyhavethesamesourcehavinganideatoexpresswithincompletecommand ofthemeansforexpressingit.ThedifferencebetweenL1andL2learnerscomesfromthefact thatL2learnershavearesourceforcopingwiththissituationthatisnotavailabletoL1 learnersusingtheirnativelanguageasatemporarytemplate.Bythesametoken,the "interference"errorsarenotevidencefordifferentlearningprocessesinL1andL2learners; rather,theyareevidencethatdifferentsentenceproductionprocessesarepossible. Nevertheless,thereisawayinwhichtheprocessunderlyingthese"templateerrors"couldbring abouttheacquisitionofwrongrules.Ifyouareearnestlytryingtoacquirefluencyinasecond language,oneofthethingsyoucandoisrehearseandpracticethelanguageyourself.Youtryto talktoyourselfsilentlyinthelanguagewhenyouarealoneathome,orwhenyouarewalking alongthestreetorwherever;yougooverconversationsandtrytoworkouthowyoushouldhave saidsomethingyoumadeanobviousblunderabout;insum,youtrytothinkinthelanguageand thatwayyoupracticethelanguage. Thistalkingtooneselfprobablyconstitutesinputtothelanguageacquisition 122

device.Now,thiskindoftalkingtooneselfisakindofrehearsalandweknowtatsofarasordinary learningtasksareconcernedrehearsalisarelativelypowerfulaidtolearning.Ican'tthinkofany theoreticalreasonwhythissilentrehearsalwouldnotbeinputtothelanguageacquisitiondevice. However,afairamountofwhatonesaystooneselfinthesecircumstancesmaybegenerated throughthetemplatestrategyjustdiscussedandinthatcaseitiscertainlygoingtobeavery problematickindofinput.Onecanreallyfeedoneselfsomeunidiomaticandpoorlyformedinput thatway.TheresultwillbetheacquisitionofsomerulesforL2thatarewrongbecausetheyare patternedonrulesofL1.Thentherewillhavebeenaneffectonlearning,andnotjustonsentence productionprocesses. Now,rehearsalasametacognitivestrategyisarelativelylatedevelopmentinchildren(Flavell& Wellman,1977),andthekindoflearningstrategydiscussedisundoubtedlymuchmorelikelyto occurinadolescentsandadultsthaninyoungerchildren.Thestrategycouldthereforeberelevantto agedifferencesinacquisitionmechanismsforsecondlanguages,andtopossibleagedifferencesin pronenesstodifferentkindsoferrors.Note,however,thatthisentirediscussionassumesabasic similarityinL1andL2learning,andinlearningasafunctionofage.DifferencesareduetotheL 2learnerandtheolderlearnerhavingerrorproneresourcesavailablethattheyoungL1learner lacks. Mynextpointalsohastodowithagedifferenceinlanguageacquisition.Iwanttoreviewsome interestingbutlittleknowndatathatcorroborateandfleshoutthedetailsofthecorrelationthat Hakutaarguedforinhispaperthatthereisahighcorrelationbetweenageofsecondlanguage acquisitionandmasteryofthelanguageevenafterpartialingoutlengthofresidenceinthecountry whereL2isspoken. ThedatacomefromIsraelicensuses.ItiswellknownthatHebrewisalanguagethatwasrevived closetotheturnofthecenturybyadultspeakersforwhomitwasnonnative.Byandlarge, immigrantshavebeenhighlymotivatedtoacquireanduseHebrew,and,inanycase,theyhavebeen subjecttoconsiderablesocialpressuredesignedtopersuadethemtoacquireit.Israelicensuseshave containedquestionsrelatingtothelanguagesineverydayuse.Figure6.1isasummaryofsome publishedcensusdata(Bachi,1956)onratingsoftheextentofeverydayuseofHebrewasa 1 functionofageatthetimeofimmigration,andlengthoftimesinceimmigration. Thecurvesarenotreadilyinterpretableaslearningcurves.The"lengthofstay"variableonthe abscissaisconfoundedwithculturaldifferencesbetweenvariousimmigrantwavesreachingIsraelat differenttimespriorto1948.Also, ____________________ 1Figure6.1isaplottingofdatainTable8ofBachi(1956)andisessentiallythesameashisFig.9 withsomeadditionallinesadded.Datafrommorerecentcensusesarealsoavailable(Stateof Israel,1963,1966),andalthoughtheagevariableistabulatedlessconvenientlyforourpurposes, theselaterdataindicatethatlaterimmigrantgroups,predominantlyfromAsianandAfrican countries,followthesamegeneralagepatternastheearlierimmigrantsshowninFig.6.1. 123

FIGURE6.1.ExtentofuseofHebrewasafunctionofyearssinceimmigrationandageat immigration(basedon15,616TelAvivmalesin1948datafromBachi[1956,Table81]).Note Theordinateisbasedonanswerstocensusquestionsaboutthelanguagesineverydayuse:if respondentsaidthatHebrewwashisonlylanguageineverydayuse,theresponsewasweighted100; ifhesaidHebrewwashisprimarylanguageineverydayuse,theresponsewasweighted75;if Hebrewwasineverydayuse,butnotastheprincipallanguage,theresponsewasweighted25; responseswithnomentionofHebrewasalanguageineverydayusewereweightedzero.the ordinate,extentofeverydayuse,isnotameasureofcompetenceinthelanguage,althoughahigh rateofeverydayusemustpresumablyindicateacomfortablecommandofthelanguageasa practicalvehicleofcommunicationwithpeers. Afirstthingtobenotedinthesedataistheabsenceofsupportforacriticalperiodforlanguage acquisition(Lenneberg,1967),atleastfortheusualhypothesisaccordingtowhichthecritical periodendsatthetimeofadolescence.Ifchildrenhavespecialequipmentforlanguagelearning whichtheyloseatadolescence,howmightthechangebemanifestinthesecurves?Oneobvious waywouldbeforthecurvesforthepostadolescentlearnerstobelowerthanthoseforthepre adolescents.However,itisobviousthatthe1519andyoungadultgroupsareverysimilartothe preadolescents;anychangewouldbebetter 124

locatedatpreandpost30thanpreandpost13.Anotherwayachangemightbemanifestisinthe shapeofthecurves:ifthechildrenjustsoakupthelanguagethroughtheporesoftheirskinasthe mythgoeswhereastheadultshavetolearnitinunnaturalandcircuitousways,thenonemight expectadifferenceincurveform.However,itisobviousthatallthecurveshavethesameinitialrise topracticalcommandofthelanguage,accomplishedinthefirstthreeorfouryearsafterarrival, withlittlesubsequentchangeinamountofuse.Inshort,thesepopulationdataprovidenosupport forthehypothesis. However,thereis,obviously,astrongagetrendinthesematerials,andthatstronglyconfirms Hakuta'sargumentforacorrelationwithage.ItisconceivablethatthelowerusageofHebrewinthe 3039andoldergroupsreflectsagraduallossoflearningabilityassociatedwithmiddleage.Other possiblecausesofthedroparefactorsoflesserexposureandreducedmotivationandsocialpressure tolearn.Bachinotesthattheyoungeragegroupspassthroughastandardizingprocessof Hebraizationinkindergartens,elementaryandworkingyouthschools,theArmy,andsoforth;i.e., almosteverybodyyoungenough(males,especially)wentthroughsomethinglikeatotalimmersion procedure.OnethingthatItakeoutofthesedataisthatayoungadultwhogoesthroughtotal immersionwillacquirethelanguage. Totalimmersionisastressfuland,formanypeople,unpleasantprocedure.Childrenwhoareput intothelocalschoolinsomeforeigncountryacquirethelanguage,buttheyareusuallypretty miserableforafewmonths.Theolderyouarethemorepoweryouhavetorefusethiskindof immersion(andifyouareoldenoughsuchimmersionishardtofindevenifyouseekit).Itseems reasonablethatitshouldbethesefactorsratherthandifferentiallearningabilitiesthataccountfor mostofthereportedagedifferencesinadulthood.Butofcoursetherecouldwellbeacombination ofreducedexposurewithreducedlearningability. AfterreadingCummins'schapterIwasveryimpressedwiththedifficultyofknowingwhatmanyof theapparentlymetalinguistictasksthathedescribesarereallymeasuring,andIcanonlyapplaud hisefforttomakesenseofthem.Moreover,Ithinkhisdistinctionbetweencontextembeddedand contextreducedtasksisverymuchalongtherightlines.TheoneideaIwouldaddtothedistinction isthatwhatcontextreducedtasksoftendemandfromthesubjectiscontextflexibility(i.e.,an understandingthatordinaryrealworldconversationisnottheonlypossibleinterpretivecontext) andanabilitytoseetheappropriatecontextinwhichthetaskisembedded.Thusa"context reduced"taskisoftennotwithoutcontext;rather,thecontextinwhichitisembeddedisnotroutine, andpartofthetaskforthesubjectistoconstructtheappropriateembeddingcontext. Forexample,considerorquestions.Yourspouseorafriendisinthekitchenandcallsout"You wantteaorcoffeeorneither?"Itwouldbesociallyobtusetotakethisasalogicaltaskandanswer yes,onthegroundsthat,evidently,you 125

wantoneortheother.Obviously,youarebeingaskedtochooseoneoftheprofferedalternatives. Butnowsupposesomeexperimenterholdsoutafistwithachiphiddeninit(e.g.,Osherson& Markman,1975)andsaystoyou"Thechipinmyhandisblueoritisn'tbluetrue,false,or can'ttell?"Itisnotsociallyastutetointerpretthisliketheteaorcoffeequestionandthus respond"can'ttell"becauseyoucan'tseethecolorofthechip(ortoguessthecolor).Youshould seefromsubtledetailsintheformofthequestion(themostnaturalformforanorquestionin thiscontextwouldpresumablybeIsthechipinmyhandblueornot?)andfromthenatureofthe previousquestionsgiveninthetaskthatthisisnotintendedasarequesttochooseadescriptor foranobject,andthatyoushouldlookforsomeotherinterpretivecontextthatwillmakesenseof thequestion.Thus,theproblemisnotreally"contextreduced";rather,itrequiresyoutouse sensitivitytolinguisticformandimmediatepriorcontexttoconstructanappropriatenonroutine interpretivecontext. ReadingAaronsonandFerres'sprovocativechapterledmetolookmorecarefullythanIhadat theBloomAucontroversy(Au,1983,1984,1985;Bloom,1981,1984);thiscontroversyand AaronsonandFerres'sdiscussionanddatagotmewonderingabouttheSapirWhorfhypothesis tothinking,inparticular,thattheusualformulationofthehypothesisneedstobechanged. Tobeginwith,IammoreinclinedthanareAaronsonandFerrestobeskepticalaboutBloom's claimofaneffectofChinesesyntaxonChinesewaysofthought.Thereareissuesofreplicability ofdatatobeaddressed,asAaronsonandFerresnote.Also,becauseitwouldnodoubtbe possibletodemonstratepoorcomprehensionofideasinthespeakersofanylanguageby renderingtheseideasverballyinunperspicuousand/orunidiomaticways,adjudicationisneeded ofAu'sclaimthatsomeofBloom'sresultsareduetohischoiceofhowtoexpresstheideas testedinChinese. Nevertheless,fromBloom'sdataandtheirown,AaronsonandFerresmakeaninteresting proposalthattherearefarreachinglanguagespecificdifferencesinverbalprocessing strategiesbetweenChineseandEnglishspeakers."Farreaching"ismyword,nottheirs,butI thinkitisappropriatebecausetheyarguethattheprocessingdifferencesaffectperformancesthat wouldordinarilybethoughttoextendbeyondlanguageperse,e.g.,socialcommunicationskills andacademicperformance. Now,theexistenceoflanguagespecificcomprehensionprocessesandstrategieshasbecomewell established,withdifferencesfoundamongseverallanguages,e.g.,English,Italian,andTurkish, andappearingalreadyatearlystagesofdevelopment(e.g.,Batesetal.,1985;Bates,McNew, MacWhinney,Devescovi,&Smith,1982;Slobin,1982).Itfollows,then,thatAaronsonand Ferres'sproposaliswellanchoredinpreviouslyestablishedfact.Theissuethatremainstobe arguedishowfarreachingtheeffectsofthestrategiesare.Ipredict 126

thatitwilltakealongtimetosettlethisissue,andthatintheinterimtheargumentsthatarise willbemany,vigorous,andperhapsbitter,withBloomAuaninitialengagement. AlthoughItakenostandonthesubstanceoftheissue,Ithinkthatthewaythecontroversyis shapingupsuggeststhattheremaybesomethingwronginthewaytheSapirWhorfhypothesis hasbeendiscussedinrecentyears.Asusuallypresented,thehypothesisisgraduatedindegreein onlyoneway:Slobin(1979),forinstance,distinguishesastrongformofthehypothesis,in whichlanguagestructuredeterminesthought,fromaweakform,inwhichlanguagemerely influencesthought.Althoughthatgraduationisappropriate,itleavesthedistinctionbetween languageandthoughtasanuntoucheddichotomy,whereasthatdichotomyespeciallyneedstobe graduated.Indeed,oneoftheusualpointsmadeaboutthehypothesisisthat,inprinciple,it requiresindependentassessmentoflanguageandthoughtinparticular,that"thought"be assessedotherthanthroughlanguage,inordertobesurethatanyeffectfoundisaneffectof languageonthought,notjustaneffectoflanguageonlanguageoracorrelationbetween languageassessments.Ofcourse,thismethodologicalpointstemsfrom,andreinforces,the conceptionoflanguageandthoughtasadichotomy.Itisobvious,however,that"language" coversawiderangeofcognitivephenomena,andthatthedistinctionbetween"language"and" nonlanguage"tasksisfarfromallornone. Wehavelongunderstoodthatspeakersofdifferentlanguageshaveacquireddifferentrule systems.Wenowknowthatspeakersofdifferentlanguageshaveacquireddifferent comprehensionprocessesandstrategies,inadditiontothedifferentrulesystems.Inthecaseof writtenlanguages,itisobviousthattheselanguagespecificprocessesmustincludereadingas wellascomprehensionofspokenlanguage,asAaronsonandFerres'sdataconfirm.Onemay surmisethattheyextendtowritingprocessesandstrategies.Thus,languagespecificcognitive processesextendoverawiderangeoflinguisticperformances,andcouldwellextendquitefar intothatgreyareawherelanguageandthoughtarehardtodistinguish.Itseemstome,therefore, thatratherthanofahypothesisweshouldspeakfirstofallofaSapirWhorfquestion:How deeplyandintohowwidearangeofcognitiveprocessesdolanguagespecificprocessing differencesextend?TheSapirWhorfhypothesisisthattheanswertothatquestionis"Quite deeplyandquitewidely."Ofcourse,theultimateanswertotheSapirWhorfquestionwillbe determinedbyresearch,andwillbemuchmorespecificthanasimpleaffirmationordenialof thehypothesis. REFERENCES Au,T.K.(1983).ChineseandEnglishcounterfactuals:TheSapirWhorfhypothesisrevisited. Cognition,15,155187. 127

Au,T.K.(1984).Counterfactuals:InreplytoAlfredBloom.Cognition,17,289302. Au,T.K.(1985).Languageandcognition.InL.L.Lloyd&R.L.Schiefelbusch(Eds.), LanguageperspectivesII.Baltimore,MD:UniversityParkPress. Bachi,R.(1956).AstatisticalanalysisoftherevivalofHebrewinIsrael.Scripta Hierosolymitana,3,179247. Bates,E.,MacWhinney,B.,Caselli,C.,Devescovi,A.,Natale,F.,&Venza,V.(1985).A crosslinguisticstudyofthedevelopmentofsentenceinterpretationstrategies.Unpublished manuscript. Bates,E.,McNew,S.,MacWhinney,B.,Devescovi,A.,&Smith,S.(1982).Functional constantsonsentenceprocessing:Acrosslinguisticstudy.Cognition,11,245299. Bloom,A.(1981).Thelinguisticshapingofthought:Astudyintheimpactoflanguageon thinkinginChinaandtheWest.Hillsdale,NJ:LawrenceErlbaumAssociates. Bloom,A.(1984).CautionThewordsyouusemayaffectwhatyousay:AresponsetoAu. Cognition,17,275287. Braine,S.M.D.(1974).Lengthconstraints,reductionrules,andholophrasticprocessesin children'swordcombinations.JournalofVerbalLearningandVerbalBehavior,13,448456. Brown,R.,&Bellugi,U.(1964).Threeprocessesinthechild'sacquisitionofsyntax.InE.H. Lenneberg(Ed.),Newdirectionsinthestudyoflanguage.Cambridge,MA:MITPress. Brown,R.,&Fraser,C.(1963).Theacquisitionofsyntax.InC.N.Cofer&B.Musgrave(Eds.), Verbalbehaviorandlearning.NewYork:McGrawHill. Flavell,J.H.,&Wellman,H.M.(1977).Metamemory.InR.V.Kail&J.W.Hagen(Eds.), Perspectivesonthedevelopmentofmemoryandcognition.Hillsdale,NJ:LawrenceErlbaum Associates. Lenneberg,E.H.(1967).Biologicalfoundationsoflanguage.NewYork:Wiley. Osherson,D.N.,&Markman,E.M.(1975).Languageandtheabilitytoevaluatecontradictions andtautologies.Cognition,3,213226. Slobin,D.T.(1979).Psycholinguistics.Glenview,IL:ScottForesman. Slobin,D.T.(1982).Universalandparticularintheacquisitionoflanguage.InE.Wanner&L. R.Gleitman(Eds.),Languageacquisition:Thestateoftheart.Cambridge:Cambridge UniversityPress. StateofIsrael.(1963).Languages,literacy,andeducationalattainment.Part1,PublicationNo. 15,1961PopulationandHousingCensus.Jerusalem:CentralBureauofStatistics. StateofIsrael.(1966).Languages,literacyandeducationalattainment,PartII,PublicationNo.

29,1961PopulationandHousingCensus.Jerusalem:CentralBureauofStatistics. 128

III BILINGUALISMANDCOGNITIVEDEVELOPMENT 129

[Thispageintentionallyleftblank.] 130

7 TheRelationshipofBilingualismtoCognitiveDevelopment:Historical,Methodologicaland TheoreticalConsiderations 1StateUniversityofNewYorkatStonyBrook MichaelPalij PeterHomelNewYorkUniversity Theabilitytoknowandtousetwoormorelanguagesfluentlyandappropriatelyreflectsthe remarkablefunctioningofacognitivesystemthatmaintainsseparationamongthelanguagesbeing usedwhileatthesametimeallowsfreeandeasyinterchangeamongthem.Theperplexing complexityofbilingualcognitivefunctioningisbeliedbytheapparenteaseofperformancethatis manifestedbytheproficientbilingual.Muchbasicresearchstillremainstobedonebeforewe understandhowbilingualcognitionoperatesbuttherearesignsthatmoreandmoreresearchersare beingattractedtothisarea.Incontrast,withintherelatedareaofhowchildrenbecomebilingualand whetherbilingualismhasanyparticulareffectoncognitivedevelopment,therehasbeenmuch researchanddiscussionthoughnotcompleteagreementonthefindings.Inthischapterweexamine thisresearchfromthreeperspectives:first,ahistoricalexaminationoftheissues,highlightingthe dramaticchangesthathaveoccurredinthepast20years;second,anexaminationofmethodological issuesinvolvedandhowtheyrelatetotheclaimsthathavebeenmade;third,whatroletheoryhas playedinguidingresearchandwhatroleitmayplayinthefuture. AHISTORICALVIEW Theviewthatisusuallypresentedofbilingualismanditseffectoncognitivedevelopmentor intellectualfunctioninginmanytextbooksandreferenceworks ____________________ 1PresentlyaffiliatedwiththeChildPsychopharmacologyUnit,DepartmentofPsychiatry,New YorkUniversityMedicalCenter 131

onchilddevelopmentiscapturedinthefollowingquotationfromAusubel,Sullivan,andIves (1980): Theweightoftheearlierevidenceonbilingualismindicatesthatitisaretarding factorinlanguagedevelopment....Abilingualenvironmentapparentlyhaslittle effectontheacquisitionoflanguagebutdoesleadtolaterconfusioninideaword relationshipsandinlanguagestructureandtolessmatureuseoflanguage....Much ofthislanguageretardationreflectsalossofvocabularyinthefirstlanguagethatis notfullycompensatedforbyacorrespondinggaininthesecondlanguage....Thus, bilingualchildrenpossessbelowaveragevocabulariesinbothlanguagesandeven theircombinedvocabularyisgenerallyinferiortothevocabularyoftheir monolingualcounterparts.... Althoughbilingualismdoesnotinhibitthedevelopmentofnonverbalintelligence... itdoeshaveanadverseeffectonthegrowthoffunctionalintelligenceasmeasuredby verbaltests....Someofthisinfluencecanundoubtedlybeattributedtolanguage handicapandsocioeconomicfactors. Ausubeletal.doqualifythesestatementssomewhat: Inviewofmajorgapsinourknowledge,however,anyconclusionsregardingthe effectsofbilingualismonthedevelopmentoflanguageandintellectualdevelopment canonlybeacceptedonahighlytentativebasis.(pp.370371) Thisbleakviewofchildhoodbilingualismwouldcertainlycauseaparenttohavegreatconcern abouthavingabilingualchild.Whatparentwouldwanttoburdenachildwiththeseavoidable handicaps? However,thepicturewithregardtothecognitiveeffectsofbilingualism,especiallyinchildhood, isnotasdismalastheprecedingaccountwouldleadonetobelieve;indeed,somewouldclaim thatthepictureisquitebrightandoptimistic.Whatdoesemergeisafarmorecomplicated picturethanmostwouldinitiallyassume.Ahistoricalreviewoftheresearchprovidesuswiththe broadoutlineofthepictureandsomeimportantdetails. Theearlieststudiesthatexaminedtherelationshipofcognitiveorintellectualfunctioningto bilingualismwereconductedduringthe1920s(Davies&Hughes,1927;Saer,1923;Smith, 1923).Thesestudiesandothersthatfollowedfoundthatbilingualchildren,withbilingualism definedinvariousways,didmorepoorlyinschoolrelativetotheirmonolingualpeers,aswellas scoringloweronstandardmeasuresofverbalornonverbalintelligence.Thispatternoffindings continuedtobeobtainedforthenext40yearsorso.MajorreviewsoftheresearchbyDarcy (1953),Jensen(1962),andWeinreich(1953)allreportedsomenegativeconsequenceof bilingualism,themainconclusionbeingthatonlyverbalintelligencewasaffectedandthat bilingualchildrenperformedaswellasmonolingualchildrenontasksinvolvingnonverbal intelligence.Ausubeletal. 132

andotherwritersappeartoberelyinguponthesereviewsandsomeoftheearlyresearchwhenthey discussbilingualisminchildhood. Aturningpointwasreached,however,in1962withthepublicationofPealandLambert'sstudyof bilingualchildreninMontrealschools.Theiroriginalintentionforthestudywastodocumenthow bilingualismnegativelyaffectedintellectualperformancesothatappropriateremedialprograms couldbedeveloped.Incontrasttomanypreviousstudies,PealandLambertsoughttocontrolmany factorsthatcouldartificiallyproducedifferencesbetweenbilingualandmonolingualchildren. Subjectsinbothgroupswerematchedonsocioeconomicclass,sex,age,and,wheneverpossible, childrenfromthesameschoolorschoolsystemwereused.Anothernovelfeatureoftheirstudywas thatPealandLambertonlyusedbilingualchildrenwhowerebalancedintheirabilitytouseboth languages.Also,theirmeasuresofintelligenceincludedteststhatwerestandardizedonFrench speakingchildrenfromMontreal. Onthebasisofthepreviousresearchitwasexpectedthatthebilingualchildrenwouldbeinferiorto monolingualchildrenonmostmeasuresofverbalintelligenceand,perhaps,wouldbeequivalenton measuresofnonverbalintelligence.Instead,thebilingualgroupperformedsignificantlybetterthan themonolingualgrouponmostofthemeasures(ontheremaindertherewerenodifferences betweenthegroups).Onthemeasuresofverbalintelligencethebilingualchildrenperformed significantlybetterthanthemonolingualchildren. AmoredetailedanalysisoftheresultsledPealandLamberttoconcludethattheperformance advantagesshownbythebilingualchildrenmightbetheresultofgreatermentalflexibilityanda morediversifiedstructureofintellect(Guilford,1956).Intheirconclusiontheystate: ThepicturethatemergesoftheFrenchEnglishbilingualinMontrealisthatofa youngsterwhosewiderexperiencesintwocultureshavegivenhimadvantagesthata monolingualdoesnotenjoy.Intellectuallyhisexperiencewithtwolanguagesystems seemstohavelefthimwithamentalflexibility,asuperiorityinconceptformation,and amorediversifiedsetofmentalabilities,inthesensethatthepatternsofabilities developedbybilingualsweremoreheterogeneous.Itisnotpossibletostatewhetherthe moreintelligentchildbecamebilingualorwhetherbilingualismaidedhisintellectual development,butthereisnoquestionaboutthefactthatheissuperiorintellectually.In contrast,themonolingualappearstohaveamoreunitarystructureofintelligencewhich hemustuseforalltypesofintellectualtasks.(p.20) Andsoatrendreverses. SubsequentstudieshavetendedtoreplicateandamplifytheresultsofPealandLambert.They roughlyfollowtwolinesofresearch:fieldstudiesinvolvinglargesamplesofchildrenand experimentalstudieswithsmallsamples.Webrieflyreviewexamplesofboth. 133

AmajorfieldstudythatprovidedfurtherconfirmationforPealandLambert'sfindingsandalso establishedthenowwellknownlanguageimmersionprogramofsecondlanguageacquisitionwas theSt.Lambertexperimentalschoolprogram(Lambert&Tucker,1972).InthisprogramEnglish monolingualchildrenwereenrolledintoa5yearprogramofbilingualinstructionandyearly evaluation,startingwithkindergartenandendingwiththefourthgrade.Theywerematchedwith monolingualEnglishspeakingchildreninamonolingualEnglishcurriculumandmonolingual FrenchspeakingchildreninamonolingualFrenchcurriculum.Childreninthethreegroupswere matchedonsocialclassvariablesandnonverbalintelligencewhileinkindergarten.(Note:children wereenteredintotheimmersiongroupbecausetheirparentshadsoughtabilingualeducationfor theirmonolingualchildren;LambertandTuckerhelpedtoestablishsuchaprogramandlocatedthe controlgroups.) Theexperimentalgroup(i.e.,Englishmonolingualchildreninthebilingualinstructioncurriculum) wereinstructedonlyinFrenchinkindergartenandthefirstgrade.Inthesecondgrade,instructionin EnglishLanguageArtswasbegunandservedastheonlyclassroominstructioninEnglish. InstructioninEnglishincreasedfortheexperimentalgroupuntilthefourthgradewherehalfofthe classroomtimewasspentinEnglish.TheEnglishandFrenchmonolingualcontrolsreceivedall instructionintheirnativelanguage.Thecurriculumforthecontrolgroupswasthestandardone usedwithintheschoolsystemwhereastheexperimentalgroupfollowedthestandardFrenchschool curriculumwhenbeinginstructedinFrenchandfollowedthestandardEnglishcurriculumwhen instructedinEnglish. Testingwasconductedyearlytoexaminetheachievementlevelsandattitudesoftheimmersion groupchildrenandtheircounterparts.However,becauseofspacelimitations,weconcentrateonthe mainresultsoftheimmersionprogramattheendofthefifthyear.Testingattheendofthefourth graderevealedthefollowing: 1.Theexperimentalgroup'slevelofEnglishproficiencyinsuchareasaswordknowledge,language usage,andsoforth,wasequivalenttothatoftheEnglishcontrolgroupandbothgroupsscored abovethe80thpercentileonnationalnorms.Theexperimentalgroupalsodidnotdifferfromthe Englishcontrolsinabilitytoformwordassociations,abilitytoretellorinventshortstoriesin English,andcommandofrhythm,intonation,enunciation,andoverallexpression.Thechildrenin theexperimentalgroup,whowereEnglishlanguagenatives,didnotsufferanyimpairmentsintheir Englishlanguageabilityasaresultofthebilingualimmersionprogram.Indeed,theyperformedas wellasmonolingualEnglishchildrenandatarelativelyhighlevelonsometests. 2.WhencomparedtotheFrenchControlGrouptheexperimentalgroupmatchedtheirperformance insuchareasasvocabulary,listeningcomprehension,andknowledgeofFrenchconcepts.The experimentalgroupwassomewhat 134

pooreratoralexpression,inrhythm,intonation,andoverallexpression,whenretellingshortstories inFrench.Thesefactorsareimprovedwhenthechildisaskedtoinventastoryandtotellitin French.Theexperimentalgroup'swordassociationsinFrenchwereasrapid,mature,and appropriateastheFrenchcontrolchildrenandtheyhadacomparabledegreeofcomprehensionof themesandplotsinstories.Althoughtheexperimentalgroupwasmorepronetomakeerrorsof productioninexpressinggenderandcontraction,overalltheirabilitywasquitesimilartotheFrench controls.Inessence,theoriginallyEnglishnativeexperimentalgroupchildrenwerefunctionally bilingualbytheendofthefourthgrade. 3.Performanceinanonlanguagesubjectsuchasmathematicsindicatesthattheexperimentalgroup (whichwastaughtmathinFrench)performedatthesamelevelastheFrenchcontrolgroupwith bothgroupsscoringabovethe80thpercentileonnationalnorms. 4.Onmeasuresofgeneralintelligencetheexperimentalgroupmatchedthemonolingualcontrol groups.Theexperimentalgroupalsoseemedmorecapableofgeneratingmoreimaginativeand unusualusesforeverydayobjects,regardlessofthelanguageofthetest. Itwasalsonotedthatthebilingualchildrencouldreadilytransferaskilllearnedinonelanguageto usageintheotherlanguage.Forexample,thoughchildrenwereinitiallytaughtreadingand arithmeticinFrench,thechildrenwerealmostimmediatelyabletoreadanddoarithmeticinEnglish whenitwasintroduced.Onthispoint,LambertandTuckerspeculatethatahigherlevelofabstract cognitiveprocessingmightexistthatallowsthetransmissionofinformationbetweenlanguages. Thiswassignificantbecauseitshowedthatcognitiveskillsdidnotdependuponthelanguagein whichtheskillwasacquired. TheresultsoftheSt.Lambertprojecthavebeensodramaticandinfluentialthatthistypeof bilingualcurriculum,generallyknownasatotalimmersionprogram,hasbeenimplementedinsome areasoftheUnitedStatesandPuertoRico.Thoughtheresultsfromthesenewprogramsareoften similartothoseoftheSt.Lambertproject,notallofthemhavebeensuccessful.McLaughlin (1978)reviewssomeofthesestudiesandexamineshowtheydifferfromtheSt.LambertProject, bothinmethodandresults. TheoriginalfindingbyLambertandPealofsomeformofcognitiveenhancementorflexibilityon thepartofbilingualchildrenhadspurtedresearchersinCanadaandothercountriestoreplicate thesefindingsundercontrolledexperimentalconditions.Muchofthisresearchhasfocusedon whetherbilingualchildrenhavegreaterabilityinmanipulatinglanguage,ormetalinguistic awareness,relativetomonolingualchildren. IancoWorrall(1972)reportedtwoexperimentsonSouthAfricanmonolingual(eitherEnglishor Afrikaaner)andbilingual(EnglishAfrikaaner)childreninwhichthechildrenweretestedontheir awarenessofthearbitraryassignmentofphonemiclabelstoobjectsandwhethertheyweremore attunedtothe 135

phonemicorsemanticdimensionsofwords.Thechildrenrangedinagefrom4to9yearsandeach bilingualchildwasmatchedwithanEnglishmonolingualandanAfrikaanermonolingualwith respecttointelligence,age,sex,schoolgrade,andsocialclass.Inoneexperimentthechildrenwere givena"standard"wordfollowedbytwo"choice"words:onephoneticallyrelatedtothestandard, theothersemanticallyrelatedtothestandard.Inthesecondexperimentavariationofaword substitutiontaskinitiallydescribedbyVygotskywasusedtodeterminetheextenttowhichthe childrencouldaccepttheinterchangeofnamesforobjects.IancoWorrallfoundthatthebilingual childrenwereawareofthearbitrarynatureofnamesatanearlieragethanthemonolingualcontrol groupsandthebilingualchildrenweremoreattentivetothesemanticrelationshipsbetweenwords thanthemonolingualchildrenwhofocusedmoreonphonemicsimilarities. BenZeev(1977),workingwithchildreninIsraelandtheU.S.whowerebilingualinHebrewand English,againfoundthatbilingualchildrenweresuperiortomonolingualchildrenonname substitutionforobjects(e.g.,usingthelabel"macaroni"torefertotheobject"cow").Further,ona verbaltransformationtaskthatconsistedoftherepeatedpresentationoftwononsensewords, bilingualsubjectsreportedmoreverbaltransformationsandatanearlierpointinthetaskthanthe monolingualchildren(inthiscaseverbaltransformationreferstoperceivedchangesinthepresented nonsensestimulus,whichremainsconstant).Thisfindingisusuallyinterpretedasindicatingability torapidlyreorganizeperceptions.Thebilingualchildren,therefore,werebetterabletoreorganize theirperceptionofthenonsensestimulus,aresponsethatismoresimilartoolderchildrenand adults.Onataskinvolvingthedescriptionofamatrixofformsvaryingontwodimensions(e.g., cylindersvaryinginheightandwidth)andthetranspositionofthematrixtoitsmirrorimage,the bilingualchildrenwerealsobetterabletoidentifytheunderlyingdimensionsofthematrixandwere betterabletodescribetheprocessoftransformingoneformofthematrixtoitstranspose. Cummins,inseveralreports(e.g.,Cummins,1978;Cummins&Mulcahy,1978)andinhischapter inthisvolume,hasconductedmuchimportantresearchonthequestionofwhetherbilingual childrenactuallydohavegreatermetalinguisticawarenessthanmonolingualchildren.One consequenceofthisresearch,however,hasbeentoraisesomeseriousquestionsastowhatone meansbymetalinguisticawarenessandwhetheritcanbeadequatelydefinedandmeasured.The readerisurgedtoreadCummins'chapterforareviewofhisworkandastatementofhispresent theoreticalposition. Themainconclusiondrawnfromthesestudiesisthatthebilingualchildappearstobemoreaware oflanguagefunctionsatanearlierageandhasamoreanalyticalapproachtotheuseand comprehensionoflanguage. Severalotherstudieshaveexaminedthewaysinwhichbilingualchildrenaresuperiorto monolingualchildrenonvariousothermeasuresofcognitiveperformance.Torrance,Wu,Gowan, andAlliotti(1970)wereabletoadminister 136

foreignlanguageversionsoftheTorranceTestofCreativeThinkingtoover1,000Chineseand Malayanmonolingualandbilingualchildreninthethird,fourth,andfifthgradesinSingapore.The secondlanguageofthebilingualswasEnglish.Thechildrenwerescoredforfiguralfluency, flexibility,originality,andelaboration.Themonolingualchildrenweresuperiortothebilingual childrenonfiguralfluencyandflexibilityandthebilingualsweresuperioronfiguraloriginalityand somewhatbetteronfiguralelaboration. Landry(1974)alsousedtheTorranceTestwithaU.S.samplewherethebilingualchildrencame fromaForeignLanguageintheElementarySchool(FLES)program.Childrenweredrawnfromthe first,fourth,andsixthgrades(thelanguagethatthechildrenwerelearningintheclassroomwasnot mentionedinthereport).Thechildrenweregivenbothverbalandfiguralformsofthetasks involvingfluency,flexibility,andoriginality.HefoundthatthesixthgradeFLESchildrenscored significantlyhigheronallofthetasksthanthemonolingualchildrenwhereasthefourthgrade FLESchildrenonlyindicatedanonsignificanttrendinthesamedirection.Nodifferenceswere obtainedbetweentheFLESchildrenandthemonolingualchildreninthefirstgrade.TheFLES experienceappearstoprovidesomeformofcognitiveenrichmentasmeasuredbytheTorranceTest ofCreativeThinking. CumminsandGulutsan(1974)presenttheresultsofanexperimentthattestedfordifferences betweenbilingualsandmonolingualsonaspectsofmemory,reasoning,anddivergentthinking. Theyfoundthatbilingualsdemonstratedgreaterverbalability,performedbetteronmeasuresof conceptformation,andscoredhigheronmeasuresofverbaloriginality.Theonlydifferencefound betweenthegroupsinperformanceonmemorytaskswasthatthemonolingualswerebetterableto recallabstractwordsthanbilinguals.Insum,thesestudiesallmakethepointthatbilingualismhas somesortofenhancingeffectoncognitivedevelopment. Howisitthenthatitisstillcommonlybelievedthatbilingualismhasonlynegativeeffectson cognitivedevelopment?Thereareprobablyseveralreasons.First,thereisalwaysalagbetweenthe publicationofresearchinjournalsanditsappearanceinsecondarysourcessuchastextbooks. AlthoughtheworkofPealandLambertandtheresultsoftheSt.Lambertprojecthavebeen availableforquiteawhile,thisresearchmaybeseenasmorerelevanttoeducationalpsychology, andtheimplicationsfordevelopmentalpsychologymaytakelongertobedrawn.Second,although animportanttopic,researchonchildhoodbilingualismandbiculturalismhasneverbeenamajor componentinmainstreamAmericanpsychology.Ashasbeendiscussedelsewhereinthisvolume (Chapter2)thecultureoftheUnitedStateshashadanambivalentattitudetowardthemaintenance oflanguagesofminoritygroups.Sincethecultureundervaluesbilingualismingeneralitisnot surprisingthatspecificdisciplinesundervaluethestudyofsuchtopics.However,asevidence accumulatesonthespecificeffectsofbilingualismamorebalancedportrayalofitisboundto appear. 137

METHODOLOGICALCONSIDERATIONS Animportantfactoraccountingforthedifferencesintheresultsbetweentheearlierstudiesand theresultsofLambertandPeal(1962)andlaterresearchersisthequalityofthemethodology used.Inmanyoftheearlystudies,variablesotherthanbilingualismcanbeidentifiedthatcan accountfortheinferiorperformanceofbilinguals.Forexample,Saer(1923)hadcollected informationaboutthesocioeconomicstatusofthechildren'sfamily,thelanguageusedathome, andtheageofthechildrenbuttherewasnomatchingofmonolingualandbilingualchildrenon thesevariables.Insomeotherstudiesthedeterminationofwhetherachildwasbilingualornot wasbaseduponeitherthenationalityoftheparent(Pintner&Keller,1922)orthesurnameof thechild(Pintner,1932).ThisisinmarkedcontrasttothemethodsofPealandLambert;they hadattemptedtocontrolforanumberofbackgroundvariablesaswellaslanguageproficiency. Itisaseasytoforestalldiscussionofmethodologyinthisareaasitisinanycontentareain psychology:thoughresearchmethodsformtheskeletonforeverystudy,itistheissuebeing discussedthatisthemeat.Yet,whenonerealizesthatthemajorchangeinourunderstandingof therelationshipofbilingualismtocognitivedevelopmenthingedontheusageofrigorous methods,onemuststopandaskhowmethodology,apartfromtheoreticalinsightintothe problemunderstudy,willbeabletofurtherourunderstanding.Suchanevaluationofourcurrent methodologicaltoolsmightallowustodecidewhetherwecanansweraquestionsuchas"Is thereacasuallinkbetweenbilingualismandcognitivedevelopment?"Ifitisanswerable,what sortofmethodologywillhavetobeused? Firstofall,itmustbenotedthatatrueexperimentcanneverbeconductedtodeterminewhatsort ofrelationshipexistsbetweenbilingualismandcognitivedevelopment.Atrueexperiment,inthe traditionalsense(e.g.,Fisher,1935),isdefinedbythehighdegreeofcontrolaresearcherhasin definingtheexperimentalconditionsandeliminatingunwanted,confoundingfactors.Butthe mostimportantcomponentofthistypeofstudyistherandomassignmentofsubjectsto conditions.Ifweweretoconductthissortofexperiment,themodeoflanguageacquisition wouldbedefinedasanindependentvariable;variousbackgroundvariableswouldbe manipulatedascontrolvariables,andperformanceonmeasuresofcognitiveprocessingwouldbe thedependentvariables.Obviously,childrencouldneverberandomlyassignedtodifferent languagelearningsituations(suchassinglelanguageconditions,separatebilingual,intermixed bilingual,etc.)norcouldanexperimentermanipulatesuchimportantvariablesasthe socioeconomiclevelofthefamily,thedegreeofexposuretothedifferentlanguages,intelligence oftheparents,culturalsupportforlanguageusageinthecommunity,andsoon. Ifwecannotusemodeoflanguageacquisitionasanindependentvariablemightwenotsimply equatesubjectsonmajorbackgroundvariables,attempttoidentifysubjectswithhomogeneous languagelearningbackgrounds,andthen 138

runthemthroughanexperimentaltask?Theproblemhereiswhetherwecanidentifythecrucial backgroundvariablesthatneedtobecontrolled.Evenifwecanidentifythem,whatmethodof controlshouldbeused?Matchingthesubjectsmaynotalwaysbepossible,andstatistical adjustment,whichmaybemadeonalegitimatemathematicalbasis,maynotalwaysmakeeither ecologicalortheoreticalsense.Thistypeofcrosssectionaldesignmayallowustoidentifysources ofdifferencesbetweenmonolingualsandbilingualsonvariouscognitivetasksbutonemaynotbe abletoexplainwhythedifferencesexist. Thetraditionaltypeofexperimentaldesignthatmostpsychologistsuseisthereforeinadequatein assessingtherelationshipofbilingualismtocognitivedevelopment.Thesmallsamplestudiesof researcherslikeIancoWorrall(1972),BenZeev(1977)andothersprovideonlyasnapshotofthe bilingualchild'sdevelopment.Crosssectionalexperimentsofthissortareusefulwhentheyare conductedtotestspecificpredictions,suchasreplicatingthefindingofgreatercognitiveflexibility ormetalinguisticawareness,butinthissituationthevalueofsuchstudiesisbasedontheir relationshiptootherstudiesthathavetakenalongitudinalapproach.Insum,crosssectional experimentscanprovidevaluabledescriptionsofprocessesoperatingatasinglepointintimeas wellasestablishingdifferencesbetweengroups,buttheycannottellushowthoseprocessesor differencesdeveloped. Thoughtrueexperimentsarenotpossibleinthisarea,wecanstilltesthypothesesaboutthe relationshipofbilingualismtocognitivedevelopmentthroughtheuseoflongitudinalquasi experimentssuchastheSt.Lambertproject.Systematicstudyofbilingualandmonolingualchildren overlongperiodsoftimewillprovideimportantdataaboutbothgroups,butdifferencesbetweenthe twogroupscanneverbecompletelyaccountedforbecause,obviously,notallvariablescanbe controlled.Forexample,intheSt.Lambertprojectdifferencesbetweentheexperimentaland controlgroupswereminimizedasmuchaspossible,butthereisnoguaranteethatsomeunknown butinfluentialconfoundingvariabledidnotcausethedifferencestoemergebetweenthegroups. Aresearchermayattempttostatisticallycontrolfortheeffectsofbackgroundvariablesthroughthe analysisofcovarianceormultipleregressionanalysis,butaresearchercanneverbesurethatallof therelevantvariableshavebeenincludedintheanalysis.Noristhereanyguaranteethatresultsfrom suchanalyseshavemuchexternalvalidityorgeneralizabilitysincethestatisticalcontrolof backgroundvariablesmayproduceresultsthataremathematicallycorrectbutecologically irrelevant(e.g.,bilingualsmayoutperformmonolingualsoncertaintasksaftersocioeconomicstatus isfactoredoutbutinreallifewecannotremovetheeffectsofSES). Theuseofmoresophisticatedmathematicalproceduressuchaspathanalysis(Li,1975),socalled causalanalysis(Heise,1975),andstructuralequationmodeling(Bentler,1980;Joreskog& Sorbom,1979)willnotprovideacompletesolution.Inthesemethodsspecificmodelsare postulatedforasetofdata.Withinthesemodelsrelationshipsamongvariablesarespecifiedand thenstatisticaltests 139

areconductedtodeterminewhetherthemodelsagreewithorfitthedata.Whenthemodelindicates alackoffitwiththedata,aresearchercanusethisresulttoeliminatethemodelfromfurther consideration.Becauseamodelcanrepresentasetofhypothesesoratheory,therejectionofthe modelistantamounttotherejectionofaparticulartheory.Butobtainingamodelthatfitsthedata adequatelydoesnotserveasconfirmationoftheunderlyingtheoryforthemodel(remember,you canonlyrejectthenullhypothesis;inabilitytorejectitdoesnotmeanitistrue).Ingeneral, structuralequationmodelingwillnotbeabletotellaresearcherwhatisthetruesetofrelationships amongvariablesbutgivesonetheoptionof"fittingvariousmodelswithdifferentnumbersof parametersandofdecidingwhentostopfitting"(p.48,Joreskog,1979).Enoughdataand theoreticalinsightintorelationshipsamongfactorsmayeventuallygiverisetomodelsthatmirror realitymoreaccurately,butmuchgroundworkneedstobelaidbeforethisisrealized. Nosingletypeofresearchdesign,eithertrueexperimentorquasiexperimentallongitudinalstudy, willprovideuswithinformationaboutthelinkbetweenbilingualismandcognitivedevelopment. However,acombinationofthetwowillhelptoestablishit:longitudinaldesignswillprovide informationaboutthegeneralnatureoftherelationship,suchasthebeneficialeffectsreportedby Lambertandothers,andtrueexperimentswillallowthetestingofspecifichypothesesthatwill providenecessarydetailontherelationship. Onecouldarguethatasmorestudiesofbothtypesareproduced,thegreatertheevidencefora particularmodelortheoryrelatingbilingualismtocognitivedevelopment.Thisargumentwouldbe misleading.JustasstudiespriortoPealandLambertseemedtoleadtoanoverwhelmingconclusion (whichappearstobefalse)itispossiblethatcurrentandfuturestudiesmayseemtoleadtoan equallyoverwhelmingbutfalseconclusion.Theprobleminthiscaseisnotwiththemethodological rigoroftheresearchbutwithhowresearchissuesaredefined,whichinturndirectsattentiontoward onesetofvariablesandawayfromothervariables.Anexampleofsuchaproblemcanbeseenin howsomeresearchershavereactedtothefindingofpositivecognitiveeffectsfrombilingualism. Formethodologicalreasons,MacNab(1979)hastakenissuewiththeresultsofthePealand Lambertstudyandothersthatpurporttoshowacognitiveenhancementeffect.Basically,MacNab hasarguedthatthequasiexperimentalorassociationalmethodologyusedinsomeofthesestudies doesnotallowonetodeterminethedirectionofthecasualarrow(i.e.,doesbilingualismimprove cognitionbyincreasingcertainfunctionsordobrighterpeoplebecomebilingual).Otherfactors, suchaslearningasecondlanguagelaterinlife,whichwouldprevent"balance"inbothlanguages andselfselectionofsubjectsinsomeofthestudies,couldalsocompromisethestudies.MacNab doesnotfeelthatbilingualismisnecessarilyabadthing,ratherhetakesissuewiththecontention thatthereisunambiguousevidenceforpositiveconsequencesfrombilingualism. 140

Fromthediscussionofmethodologyearlierinthissectionthereadershouldbeabletoappreciate twopoints:(a)MacNab'scriticismshavesomevaliditygiventhetypesofdesignsusedand(b) thetypeofevidencethatMacNabseemstobeaskingforbeforeheisconvincedthatbilingualism hassomeeffectoncognitionisalmostimpossibletoobtainatthistime.Onthelatterpoint,atrue experimentmightconvincearesearcherlikeMacNab,butforobviousreasons,suchathing cannotbedone.Ultimately,theconvergenceofresultsfromdifferentrigorousstudieswithsolid theoreticalmodelingwillprovidethebasisforelucidatingtherelationshipbetweencognitive functioningandbilingualism. However,thereadershouldbeawarethatthistypeofargument,ofwhetherbilingualismhas beneficialordetrimentaleffects,causesonetofocusonalogicallysecondaryissue.Theprimary issueishowthecognitivesystemaccommodatestheusageoftwoormorelanguages.Itcouldbe thatabilingualcognitivesystemhastobeinvolvedintypesofprocessingthatamonolingual neverhastoengagein,suchastranslationbetweenlanguages.Howthistranslationprocess operates,alongwithitscostsandbenefits,isofgreaterpracticalandtheoreticalinterestthan whetherthereissomeambiguouscognitivebenefit.Indeed,itisonlybyexaminingthecostsand benefitsofthespecificprocessesneededforbilingualcognitionthatwecanmakeanystatement aboutthe"good"or"bad"effectsofbilingualism.Inalllikelihoodtherewillprobablybeno clearcutevidenceforoverwhelmingpositiveornegativeconsequences;instead,wemayseea subtleinterplaybetweenthecostsandbenefitsfortheprocessesinvolved.Suchaconcernislost ifonegetsmireddowninadebateaboutwhetherbilingualismisgoodorbad. Itmustalsobeacknowledgedthattheargumentaboutthemeritsofbilingualism,itscognitive benefitsandcosts,occurswithinasocialcontextthatisnotneutralinitsviewonthedesirability ofhavingabilingualpopulace(seethechapterbyHomelandPalijinthisvolumeon bilingualismandsocialpolicyforagreaterdiscussionofthisissue).Inthosecommunitieswhere bilingualismisbothdesiredandviewedasbeingnecessary,bilingualismwillbefosteredand maintainedregardlessofthecosts.Similarly,inthosecommunitieswherebilingualismis denigratedandunwanted,bilingualismwillbestifledregardlessofanybenefititmayprovide. Researchersshouldbesensitivetowhetherthequestionstheyareaskingaretrulyfundamentalto thephenomenontheyarestudying,oraremotivatedsolelybythesociopoliticalviewsoftheir community. THEORETICALASPECTS Thetraditionalfocusontherelationshipofbilingualismtocognitivefunctioninghasbeen motivatedbyempiricalfindingsandnotbytheoreticalconsiderations.Severalfactorsseemto havebeenoperatingtocausethis.Onereasonforthishas 141

beenthecontextinwhichtheresearchhasbeenconducted.Manyofthestudiescomefromapplied areassuchaseducationalorclinicalpsychologythathadtoconfronttherealityofhavingtodeal withbilingualchildrenwhowerenotperformingatthesamelevelsasmonolingualchildren. Unfortunately,theforeignlanguagethatthechildspokeortheethnicbackgroundofthechildwas perceivedasbeingthesourceofthedifficultyinsteadofotherfactorssuchassocioeconomicstatus ofthechild'sfamily.Experimentalpsychologywasillpreparedtoaddressthepossibleconnection betweenbilingualismandcognitionbecauseitsimplywasnotpreparedtodosountilrelatively recently.Again,somehistorymayclarifythispoint. Thestudyoflanguagefunctioningandcognitiveprocessingasweknowittodayistheresultof severalfactorsthatappearedinthe1950s.Amongthemostimportantofthesewastheapplication ofinformationtheory,bothitsmathematicsandtheoreticalframework,toproblemsinpsychology, andtheuseofcomputersandprogrammingasameansofstudyinghumancognition.The predominantinfluenceinexperimentalpsychologypriortothistime,aswellasforsometimeto come,hadbeenbehaviorism,invariantforms,withanaccompanyingphilosophicalframework basedonlogicalpositivism.Withthedesiretomodelthepsychologicalsciencesafterthephysical sciences,arigorousimplementationofstrictempiricismforcedmanyresearcherstofocuson behaviorandtoignoreordownplayunobservablecognitiveprocesses.Thereweresomeexceptions, suchasthegestaltpsychologists,whofeltthiswasanunnaturallyrestrictiveviewofthemind,as wellasamistakenunderstandingofhowthephysicalsciencesdeveloped(which,theyargued, allowedformanyunobservables),buttheytendedtobeintheminority.Consequently,insteadof languageacquisitionwehadverballearning,insteadoflanguageprocessingwehadverbalbehavior, andinsteadoftransformationgrammarswehadverbalconditioning. WiththeworkofChomskyinlinguistics,NewellandSimoninhumanproblemsolving,andothers likeGeorgeMiller,UlricNeisser,andsoforth,analternativeframeworkwasslowlyestablishedthat allowedafullerconsiderationofmoretheoreticalissuesinhumancognition.Thisframework providedarigorousmeansforstatingandtestinghypothesesaboutunobservablecognitive processesthatshowedthemtobeneithermetaphysicalnorsupernatural.But,thoughthisframework ofcognitivepsychologyisapparenttoustoday,itsinfluenceinotherareasofpsychologywasslow todevelop.Asshownintheearliersectiononhistoricalconsiderations,itisthebilingualism researchofthe1970sthatbeginstoaskquestionsemanatingfromacontemporaryframeworkof cognitivepsychology.RecallthatLambertandPealwerenotconcernedwithtestinghypotheses aboutthenegativecognitiveconsequencesofbilingualism:rathertheywantedtogetaclear descriptionofwhatthesenegativeconsequencesweresothataprogramofremediationcouldbe developed.AndinthecaseoftheSt.Lambertproject,theoriginalmotivationwastoestablishan educationalcurriculumthatwouldproducebilingualchildren,nottotestanyparticular 142

theoryofsecondlanguageacquisition.However,becauseoftheresultsfromthesestudies,other researcherswentontoreplicatetheresultsandtotesthypothesesbaseduponcognitivetheory. Insum,theabsenceofafulltheoreticaltreatmentoftherelationshipofbilingualismtocognitive processingcanbeattributed,inpart,tothelatedevelopmentofappropriatecognitivetheoriesto researchersinthearea. Asecondreasonisthatthemeasuresoftenemployed,suchasstandardizedintelligencetests,do nottellusaboutthespecificcognitiveprocessesinvolvedinperformanceonthetests.For example,theextractionofaverbalfactorfromaseriesoftests,whichontheirfacerequirethe processingoflinguisticinformation,doesnottelluswhatformsthatprocessingtakes.Allthatit cantellusisthatacommonprocessmightbeunderlyingtheresponsesmadetoquestionsonthe tests.Indeed,theuseofsuchfactorsobscurestheveryrealdifferencesinstrategiesthatpeople mightemploy.Thealgorithmicspecificationofprocesses,theexplicitstatementofcognitive componentsandrelatedprocesses,isarecentinnovationderivedfromtheapplicationof informationprocessingtheorytohumancognition.Althoughsuchanapproachismechanistic andcomputationallyorientedithasthebenefitofidentifyingspecificmechanismsthathopefully canbemanipulatedinexperiments.Insteadofhypothesizingageneralizedverbalfactorwecan nowinquireabouthowapersonprocesseslinguisticinformationandhowtheprocessingvaries asafunctionofthetypeoflinguisticinformation. Athirdreasonforthedelayinthedevelopmentofanadequatetheoreticalframeworkfor bilingualcognitionisculturallybased:ThereisapopularconceptionthattheU.S.isprimarilya culturallyhomogeneous,Englishspeakingnationandthatforeignlanguagesandethnicidentities fallbythewaysideaseachindividualistransformedintoatypicalEnglishspeaking"American." ThisisexemplifiedbytheEnglishfirstgroups,whichwanttopassanamendmenttothe constitutionmandatingthatEnglishberecognizedastheofficiallanguageoftheU.S.(theU.S. doesnothavealegallyrecognizedofficiallanguage,unlikeCanada,whichhasalegally mandatedpolicyofbilingualism).Asnotedelsewhereinthisvolume,theU.S.officiallyviews bilingualismasatransitionalphaseinthelivesofpeoplelivingintheU.S.andthusbilingualism isundervalued.Researchintosuchamarginalareaisnotwellsupportedorencouraged.Because ofsuchacontextbothresearchandtheorybuildingareunlikelytodeveloptoanygreatextent.It isexpectedthatsignificantprogressonissuesrelatingbilingualismtocognitivedevelopmentand processingwillbemadeincountrieslikeCanadawherethereisnotonlyacademicsupportfor researchintosuchissuesbutculturalapprovalofsuchactivities. Withrespecttopresentdaytheory,Segalowitz(1981)outlinessomeoftheframeworksthathave beendevelopedinstudyingbilingualdevelopment:Taylor's(1974)networkmodel,whichis basedontheNormanandRumelhart(1975)modelofsemanticmemory;Slobin's(1973)theory oflanguageacquisi 143

tioninwhichchildrenacquirestrategieswhichorientthemtosalientcuesinlanguageacquisition contexts;andBrown's(1973)workoncognitivefactorsinmonolinguallanguageacquisition. Thedefiningfeatureofatheoryofbilingualdevelopmentisthatittakesasitsstartingpointnot theexperienceofmonolingualdevelopmentbutthespecialcircumstancesassociatedwiththe bilingualexperience.Auniqueaspectofbilingualcognitivedevelopmentthathasnocounterpart inmonolingualexperienceislanguageinteractioninthecognitivesystemandwefocusonthatin thissection. Whatisremarkableaboutbilingualcognitionistheabilitytomaintainlanguageseparationwhile allowingreadyintermixingofthelanguages.Variousschemeshavebeendevelopedtoaccount forthisabilitybutnoneseemtohavebeenparticularlypersuasive.Thebestknownofthese schemesisthecompoundcoordinatedistinction,whichwasoriginallydevelopedbyWeinreich (1953)andelaboratedonbyErvinandOsgood(1954).Acompoundbilingualwasonewho learnedbothlanguagesinthesameenvironmentfrombirth.Theexperiencewassupposedto formasinglesystemforthetwolanguages.Originally,thisformulationwasintermsofverbal mediationtheorybutwaslaterdiscardedinfavorofthenotionofasingleinterdependent memorystoreforbothlanguages.Forthecompoundbilingualthetwolanguagesinthecognitive systemshouldhaverelativelyeasyandfrequentinteractionwitheachother(thoughithasnever beenclearwhethertheinteractiontakesplaceonlyatoneleveloflanguagerepresentation,such asthesemanticlevel,orwhetherittakesplaceatalllevels,e.g.,phonemic,graphemic,syntactic, etc.).Thecoordinatebilingualisonewholearnedonelanguagefrombirthandlearnedthe secondlanguagelaterinlifeincontextsdifferentfromthoseofthefirstlanguage.Thisseparation inlearningcontextswouldhavecausedtheseparationoflanguagesinthecognitivesystem, indicatingthatverylittleinteractionoccursbetweenthetwolanguages;thatis,eachlanguage wouldhaveitsownindependentstorageareainmemory. Supportforthiscompoundcoordinatedistinctionhasbeeninconsistent,andsomeresearchers dismissitasbeingirrelevanttounderstandinghowbilingualismaffectscognition.Recent researchindicatesthatbilingualcognitionmayinfacthavefeaturesofboth,thatis,language intermixtureandseparation,butindifferentareasofthecognitivesystem.Studiesusingthe lexicaldecisionparadigmhaveproducedsomeveryinterestingdatainthisregard.Palij(1980) hasshownthatFrenchEnglishbilingualsproduceasemanticfacilitationeffectonabilingual versionofthelexicaldecisiontaskthatisindependentofthesubject'slearninghistory(i.e., compoundandcoordinatebilingualsbothprovidedsubstantialfacilitationeffects).Thesemantic facilitationeffect,thefasterrecognitionofasecondwordafterfirstseeingasemanticallyrelated word(e.g.,fasterrecognitiontimetodoctorafterjustseeingnurse),inthebilingualsituation (i.e.,firstseeingthetranslationofdoctorandthenseeingtheEnglishworddoctor)indicatesthat thelanguagesshareacommonsemanticstore.Meaningforwordsandotherconceptsseemtobe representedinanabstractformacross 144

languagesandrelatedconceptsas"closertogether"insemanticmemoryregardlessofthelanguage oforigin.ThisfindingdovetailswiththosereportedearlierbyLambertandTuckerinthetransferof skillslearnedinonelanguagetotheotherlanguageofthebilinguals. Otherresearch(Scarborough,Gerard,&Cortese,1984)indicatesthatonelevelatwhichlanguages areseparatedisatthesensoryrepresentationlevel(bysensorylevelrepresentationwefollowthe usagedevelopedbyNelson,1979,wherethevisualandphonemicmentalrepresentationofwords constitutesthesensorylevel).Usingalexicaldecisiontask,butwithoutvaryingsemantic relatedness,Scarboroughhasshownthatrepetitioneffectsoccurmorereadilywithinlanguagesthan theydobetweenlanguages.Repetitioneffectsrefertofasterrecognitiontimesforwordsthatare presentedseveraltimesduringthecourseoftheexperiment.Repetitionsthatareclosetogether(e.g., therepetitionoccurringimmediatelyorafterthepresentationofasingleinterveningdifferentword) producefasterreactiontimesthanrepetitionswithgreaterseparation(morewordsbetween repetitions).Inthebilingualcontextthisfindingsuggeststhatthesensoryfeaturesoflanguagesare organizedonthebasisoffeaturalsimilarity.Sincewordswithinalanguagearemoresimilartoeach other,becauseofcommonrulesofphoneticsandorthography,itisnotsurprisingthatthereis reducedlanguageinteractionatthesensorylevel. Yetanotherleveloflanguageinteractionthatisparticularlyinterestingisthatofsyntactic interaction.Atthislevelthereareseveralquestions:Dotherulesofsyntaxremainseparateforthe twolanguages?Doesthesyntaxofonelanguagehavesomeeffectontheproductionsoftheother language?Forlanguageswithradicallydifferentunderlyingsyntacticsystems(e.g.,Chineseand English)howdoproductionsinonelanguagemapontoproductionsintheotherlanguage(howdoes translationoccur)? Researchontheeffectofonesyntaxuponanotherincognitioncomesfromseveralsources.Bates andMacWhinney(1981)examinedhowthesurfaceroleofthesentencesubjectisdifferentially determinedinEnglish,Italian,andGerman.Forexample,itwasfoundthatItaliansprimarilyused animacyasacueininterpretingsimplesentenceswhereasAmericans(English)primarilyrelied upontheorderingofthewordsinthesentence.Thistypeofstrategyclearlydifferentiatesthetwo groups.Thistypeofinformationcanbeusedtodeterminewhether,say,anItalianAmerican bilingualutilizesanItalianinterpretationstrategywhilefunctioninginEnglish,whichBatesand MacWhinneygoontodemonstrate. Additionalexamplesofsuchinteractionsareprovidedinotherchaptersofthisbook.Hakuta(this volume)dealswithavarietyofissuesrelatedtosyntacticinteraction.AaronsonandFerres(this volume)examinetherelatedissueswithregardtotheinteractionofChineseandEnglish. Theimportanceofthesestudiesisthattheyfocusonthespecificfeaturesoflanguagesandhowthe cognitivesystemadaptstothesefeatures.Inthebilingualchildnotonlydoesthecognitivesystem havetoadapttoeachofthelanguages 145

thatthechildknows,butitmustalsoaccommodatethejointfunctioningofthetwolanguages. Theinteractionofthetwolanguagesmaybespecifictothelanguagesunderconsideration.That is,theinteractionsfoundintheChineseEnglishbilingualmaynotbethesameasthosefoundin theFrenchEnglishbilingual.Thesituationisobviouslymorecomplicatedwhendealingwith trilingualsandmultilinguals.Interactionsamongthelanguagesintheseindividualsmaybe simpleextensionsofthetwowayinteractionsseeninbilingualsormayrepresenthigherorder interactionsamongallofthelanguages. Clearly,theseconcernsgofarbeyondthesimplequestionofwhetherbilingualismhasbeneficial ornegativecognitiveconsequences.Itisperhapsmorereasonabletoassumethatbilingualism willhavebothsomenegativeconsequencesandsomepositiveconsequences;thedetermination ofwhatisnegativeandpositivebeingconditionedbythespecificvariablesbeingstudiedand, perhaps,theageofthechildbeingtested. CONCLUSIONS Whatcanweconcludeabouttherelationshipbetweenbilingualismandcognitivedevelopment? Thismuchseemstobeclear:Bilingualismdoesnotseemtohaveanymajordetrimental cognitiveconsequencesinandofitself.Thereisevidencefortheassertionthatbilingualismmay havesomebeneficialcognitiveconsequencesintheformofenhancedlanguageawarenessand greaterflexibilityinitsusage.Nonetheless,themagnitudeofsuchbenefitsshouldnotbe overemphasizeduntilwehavemoredetailedinformationonitsnature.Itisinarguablethat bilingualismhassomesortofinfluenceontheorganizationofinformationinthecognitive systemandhowthatinformationisprocessed,butmuchmoreresearchisneededtoelaborateon whatthisinfluenceislike.Shouldwethinkoflanguageassomethingthatcategorizescognitive functioning(i.e.,separatesstorageareas,processes)andrequiresuniqueprocessing(e.g., translationprocesses)orhavewemadetoomuchoflanguageasaconceptualentity,onlytofind thatknowledgeofmorethanonelanguagemakesnospecialdemandsonthecognitivesystem? Thatbeingabilingualmakescognitivedemandsthataresimilarto,thoughnotidenticalto, knowingandbeingabletousemathematicsormusic?Andwhenwasthelasttimeyouheard someonediscussingthecognitivecostsandbenefitsofknowingthosesubjects? ACKNOWLEDGMENT WewouldliketothankDorisAaronsonforreadingandcommentingonearlierdraftsofthis chapter. 146

REFERENCES Ausubel,D.P.,Sullivan,E.V.,&IvesS.W.(1980).Theoryandproblemsofchilddevelopment (3rded.).NewYork:Grune&Stratton. Bates,E.,&MacWhinney,B.(1981)."Secondlanguageacquisitionfromafunctionalist perspective:Pragmatic,semantic,andperceptualstrategies".InH.Winitz(Ed.),Nativelanguage andforeignlanguageacquisition,AnnalsoftheNewYorkAcademyofSciences(Vol.379,pp. 190214). BenZeev,S.(1977)."Theinfluenceofbilingualismoncognitivestrategyandcognitive development".ChildDevelopment,48,10091018. Bender,P.M.(1980)."Multivariateanalysiswithlatentvariables:Causalmodelling".Annual ReviewofPsychology,31,419456. Brown,R.(1973).Afirstlanguage:Theearlystages.Cambridge,MA:HarvardUniversity Press. Cummins,J.(1978)."Bilingualismandthedevelopmentofmetalinguisticawareness".Journal ofCrossCulturalPsychology,9,131149. Cummins,J.,&Gulutsan,M.(1974)."Someeffectofbilingualismoncognitivefunctioning".In S.T.Carey(Ed.),Bilingualism,biculturalism,andeducation.Proceedingsoftheconferenceat CollegeUniversitaireSaintJean,TheUniversityofAlberta,Canada. Cummins,J.,&Mulcahy,R.(1978)."OrientationtolanguageinUkrainianEnglishbilingual children".ChildDevelopment,49,12391242. Darcy,N.T.(1953)."Areviewoftheliteratureontheeffectsofbilingualismuponthe measurementofintelligence".JournalofGeneticPsychology,82,2157. Davies,M.,&Hughes,A.G.(1927)."Aninvestigationintothecomparativeintelligenceand attainmentsofJewishandnonJewishschoolchildren".BritishJournalofPsychology,18, 134146. Ervin,S.M.,&Osgood,C.E.(1954)."Secondlanguagelearningandbilingualism".Journalof AbnormalandSocialPsychology,49,139146. Fisher,R.A.(1935).Thedesignofexperiments(3rded.)Edinburgh:Oliver&Boyd. Guilford,J.P.(1956)."Thestructureofintellect".PsychologicalBulletin,53,267293. Heise,D.R.(1975).Causalanalysis.NewYork:JohnWiley. IancoWorrall,A.D.(1972)."Bilingualismandcognitivedevelopment".ChildDevelopment,43, 13901400. Jensen,A.R.(1962)."Theeffectsofchildhoodbilingualism".ElementaryEducation,39,132

143,358366. Joreskog,K.G.(1979)."Analyzingpsychologicaldatabystructuralanalysisofcovariance matrices". InK.G.Joreskog&D.Sorbom(Eds.),Advancesinfiactoranalysisandstructuralequation models.Cambridge,MA:AbtBooks. Joreskog,K.G.,&SorbomD.(1979).Advancesinfactoranalysisandstructuralequation models.Cambridge,MA:AbtBooks. Lambert,W.E.,&Tucker,G.R.(1972).Bilingualeducationofchildren:TheSt.Lambert experiment.Rowley,MA:NewburyHouse. Landry,R.G.(1974)."Acomparisonof2ndlanguagelearnersandmonolingualsondivergent thinkingtasksattheelementaryschoollevel".ModernLanguageJournal,58,1015. Li,C.C.(1975).PathanalysisAprimer.PacificGrove,CA:BoxwoodPress. MacNab,G.L.(1979)."Cognitionandbilingualism:Areanalysisofstudies".Linguistics,17, 213255. McLaughlin,B.(1978).Secondlanguageacquisitioninchildhood.Hillsdale,NJ:Lawrence ErlbaumAssociates. Nelson,D.L.(1979)."Rememberingpicturesandwords:Appearance,significance,andname". InL.S.Cermak&F.I.M.Craik(Eds.)Levelsofprocessinginhumanmemory(pp.4576). Hillsdale,NJ:LawrenceErlbaumAssociates. 147

Norman,D.,&Rumelhart,D.(1975).Explorationsincognition.SanFrancisco:W.H.Freeman. Palij,M.(1980).Semanticfacilitationonabilinguallexicaldecisiontask.StonyBrook:State UniversityofNewYork.(ERICDocumentReproductionServiceNo.ED192611) Peal,E.,&Lambert,W.E.(1962)."Therelationofbilingualismtointelligence".Psychological Monographs,76(27,WholeNo.546). Pinter,R.(1932)."Theinfluenceoflanguagebackgroundinintelligencetests".Journalof SocialPsychology,3,235240. Pinter,R.,&Keller,R.(1922)."Intelligencetestsofforeignchildren".JournalofEducational Psychology,13,214222. Saer,D.J.(1923)."Theeffectsofbilingualismonintelligence".BritishJournalofPsychology, 14,2538. Scarborough,D.L.,Gerard,L.,&Cortese,C.(1984)."Independenceoflexicalaccessin bilingualwordrecognition".JournalofVerbalLearningandVerbalBehavior,23,8499. Segalowitz,N.S.(1981)."Issuesinthecrossculturalstudyofbilingualdevelopment".InH.C. Triandis&A.Heron(Eds.),Handbookofcrossculturalpsychology:Vol.4,Developmental psychology(pp.5592).Boston:Allyn&Bacon. Slobin,D.(1973)."Coginativeprerequisitesforthedevelopmentofgrammar".InC.A. Fergueson&D.Slobin(Eds.),Studiesofchildlanguagedevelopment(pp.175208).NewYork: Holt,Rinehart&Winston. Smith,F.(1923)."Bilingualismandmentaldevelopment".BritishJournalofPsychology,13, 270282. Taylor,M.(1974)."Speculationsonbilingualismandthecognitivenetwork".WorkingPapersin Bilingualism,2,68124. Torrance,E.,Wu,J.M.,Gowan,J.C.,&Alliotti,N.C.(1970)."Creativefunctioningof monolingualandbilingualchildreninSingapore".JournalofEducationalPsychology,61,7275. Weinreich,U.(1953).Languagesincontact.NewYork:LinguisticCircleofNewYork. 148

8 Bilingualism,CognitiveFunction,andLanguageMinorityGroupMembership EdwardDeAvilaLinguametricsGroup,SanRaphael,CA INTRODUCTION Thestudyofchildhoodbilingualismmaybedescribedashavingboththeoreticalandapplied importance.Withrespecttotheoreticalinterests,researchershavestudiedtherelationshipbetween bilingualismandvariouscognitiveprocessessuchasintellectualdevelopmentandcognitivestyle. Ontheappliedside,researchershaveusedresultsfromtheoreticalstudiestodesignandtest hypothesesregardingtheeffectivenessofdifferenttreatmentapproachessuchasthosefoundin bilingualeducation.Thepurposeofthischapteristodiscussanumberofissuesrelatedtoboth theoreticalandappliedinterestsintheintellectualandsocialfunctioningofbilingualchildrenin school.Thechapterispresentedintwomajorsections.Inthefirst,wemakeafewintroductory commentsregardingthestudyofbilingualstudents.Inthesecondsection,wereviewtheresultsofa numberofstudiesconductedattheCenterforEducationalStudies(CERAS)atStanfordUniversity overthepast5yearsthathaveaddressedseveralimportanttheoreticalandappliedquestionswithin thecontextofaneducationalprogramdesignedtoaccommodatethelinguisticandeducational heterogeneityofethnolinguisticminoritystudents. BACKGROUNDOFSTUDIES Asubstantialnumberofsourceshavedocumentedthepooracademicperformanceoflanguage minoritystudentsintheUnitedStates.Similarly,agoodnumberofresearchershaveattemptedto explainthispoorperformance.Of 149

particularinteresttothepresentdiscussionisthewidespreadbeliefthatsomehowlanguage minoritystudentsareatanacademicdisadvantagebyvirtueofintellectual,verbal,motivational, andcognitivestyledifferencesthathavebeenequatedwithbilingualism.Unfortunately,studies offeredinsupportofthiscontentionhavetendedtoconfoundbothpovertyandethnolinguistic groupmembershipwithlinguisticproficiency.DeAvilaandDuncan(1980)reviewedover100 studiesontheeffectsofbilingualismconductedintheUnitedStatesoverthepast50yearsand foundthatinonlyafewcases(four)wastheactualextentof"bilingualism"assessed.Withrare exception,subjectsweregroupedonthebasisofethnicity,assuminglanguageproficiency. AconcreteexampleofthisconfusionisfoundinarecentarticlewhereMestre,Gerace,and Lochhead(1982)foundthateventhoughthebilinguals(i.e.,balanced)were...nearly equivalentinbothSpanishandEnglish,thelevelofproficiencyineitherlanguageforthe bilingualgroupwassubstantiallybelowthelevelofthemonolingualgroupinEnglish.Inother words,totheextentbilingualismmeansproficiencyintwolanguages,thisgroupof"Hispanic bilinguals"wasnotbilingual.Thereviewofproblemsassociatedwiththedefinitionand bilingualismbyDeAvilaandDuncan(1980)citesnumerousexamplesofthistypeinavariety ofresearchcontexts.Inpracticalterms,thisshowsthatwithintheUnitedStatesitisimpossible todiscusstheeffectsofbilingualismwithoutconsiderationoflanguageminoritygroup membership. Asfurtherexampleoftheconfoundingofgroupmembershipwithlinguisticcharacteristics,one findsthefollowinglogic.MexicanAmericanstudentsarefoundtobemore"fielddependent" thanAnglocounterparts(Buriel,1975;Ramirez&PriceWilliams,1974;Sanders,Scholz,& Kagan,1976).TotheextentMexicanAmericanstudentsareassumedtobebilingual,obtained cognitivestyledifferencesareequatedwithlinguisticdifferences.Thus,bilingualscometobe viewedasfielddependent.Inareviewoftheliteratureoncognitivestyle,DeAvilaandDuncan( 1980)concluded"virtuallynostudiesinvolvingSpanishlanguagebackgroundchildrenhave controlledforlanguageproficiencyineitherSpanishorEnglishorforintellectualdevelopment. Giventhefailuretocontrolforthesepotentiallyimportantvariables,anydifferenceswithrespect togroupsmustremainequivocal." UsinganapproachbasedontheearlierworkofDeAvila,HavassyandPascualLeone(1976), Duncan(1979),DuncanandDeAvila(1980),DeAvila,Duncan,andUlibarri(1982)foundthat whendifferencesinlinguisticproficiencywerecontrolledthroughassessmentinbothlanguages, differencesincognitivestylewere,toalargeextent,testspecificwithonlyfullyproficient bilingualstudentsdemonstratingconsistentdifferences(higherlevelsofcognitivedevelopment onbothWitkinandPiagettypetasks).Theseresultsreplicateanumberofsimilarstudies conductedoverthepast15yearswhichhaveshownthatwhenlinguisticandtestdemand characteristicsarecontrolled,manyofthereportedethnolinguisticgroupdifferencesincognitive functioningfailtoemerge(seeDeAvila&Duncan,1980,forareview). 150

Finally,mostoftheethnolinguisticgroupcomparisonstudieshavebeenmotivatedbyanattempt tounderstandtheimpactofgroupvariationonschoolperformance.Findingsofbetweengroup differencesonvariouscognitivetaskssuchastheChildren'sEmbeddedFiguresTask(Witkin& Goodenough,1977)aretakentoexplaindifferencesinacademicachievement(Chan,1983).In anextensive3yearcrossculturalstudyinvolvingover1,200studentsfromsixdifferent ethnolinguisticbackgrounds,DeAvila,Duncan,andUlibarri(1982)foundthatcognitivestyle differencescontributedlessthan10%tothetotalvarianceinpredictingtheschoolperformance oflanguageminoritystudents.Ontheotherhand,linguisticvariationaccountedforasmuchas 70%ofthetotalvariance.Cognitivestyleandintellectualdevelopmentvariablesaccountedfor significantproportionsofthevarianceinpredictingschoolperformanceonlyformainstream Anglostudents. Asaresultofourownstudiesandreviewofvariousapproachestothestudyoftherelationship ofschoolachievement,andsuchfactorsasbilingualismorlinguisticproficiency,intellectual development,andcognitivestyle,wehavecometotheconclusionthatresearcherswoulddowell toconsidertheissueoftreatmentandenvironmentalcircumstancesbeforeattributingdifferences inschoolperformancetoeither"bilingualism"or"cognitivestyle"variation.Inthisconnection, theunderlyingtheorythatdrivesourdesignofprogramscouldbeconsidered.Finally,itcanbe arguedthatunderstandingoftheeffectsofbilingualismoncognitivestyle(orviceversa)cannot beachievedwithoutthestudyofthecontextsandprocessesbywhichitdevelops.Onthislast point,DeAvilaandDuncan(1979)maketheargumentthatcognitivestyleandbilingualism developinafashionthatisindistinguishablefromthatofintelligence. ThemajorpolicyfocusforbilingualeducationintheUnitedStatessincethepassageofthe1968 BilingualEducationAct(TitleVII,ESEA)hasbeenoneofcompensatoryassistance.Theview heldbymosteducatorshasbeenthatthedifficultiesfacedbylanguageminoritystudentsresult fromlowerintellectuallevelsofdevelopmentassumedtobeassociatedwithbilingualism, cognitivestyledifferences,motivationaldeficienciesandahostoffactorsreferredtounderthe rubricofSES(seeRosenthal,Milne,Ginsburg,&Baker,1981).Bymeansofincreased resourcesassociatedwiththeremediationofEnglishlanguagedeficiencies(ESL),itisexpected thatchildrenwillgainEnglishlanguageskillssufficienttofullparticipationinEnglishonly classrooms.Thisperspectivehasbeencoupledwithanincreasedrecognitionofthedemandsfor equityineducationalopportunity(Lauv.Nichols,1974).Similarly,increasingandshifting demographicpatternswithintheUSpopulationhaveforcedaconsiderationoftheincreasing numbersofchildrenwhocomefromhomeswhereEnglishisnottheprimarylanguage.Asa result,Americaneducatorshavebecomeincreasinglyawareofthespecialdemandsplacedonthe schoolsbyethnicallyandlinguisticallyheterogeneousstudents. Unfortunately,however,recentreviewsofthecontentofbilingualprogramssuggestaneven greateremphasisonprogramsthatfosteradependenceon 151

predeterminedapproachestoproblemsolvingthanisfoundinregularclassrooms(Aries,1982; Laosa,1977;NievesSquires,1980).ClarkandPeterson(1976)reportthatteachersuseextremely simplelogicintheirinteractionswithstudents.Thissimplifiedlogicisparticularlyevidentin bilingualclassroomsinthe"modifiedspeech"patternsusedbyEnglishspeakingteacherswhen interactingwithstudentsof"limitedEnglishproficiency"(Takahashi,1982).Notsurprisingly,one findsthatcurrentclassroompracticesfocusontherotelearningoffactstotheexclusionofmore complexformsofinformationprocessing.Taba(1966)arguesthatthemainreasonforthelow statusof"thinking"inAmericanclassroomsingeneralandinbilingualclassroomsinparticular derivesfromthebeliefthatthinkingispredicatedonlearningabodyoffactualknowledgei.e., onemustknowcontentbeforeonecanthink.Withrespecttolanguageminoritystudents,English languageproficiencybecomesanadditionalprerequisitetothinking.Theobservationthatteachers faceadifficulttaskisunderscoredbytheexperienceofanypractitionerwhohasworkedin classrooms,continuallyconfrontedwithcomplexclassroommanagementproblemsthattendto subordinatethinkingtocontent.Thinkingisdifficulttofosterinanyclassroom,letaloneina linguisticallyheterogeneousbilingual(orevenmultilingual)environment. Organizationalsociologists(Cohen,Deal,Meyer,&Scott,1979;Intili,1977;Perrow,1967)point outthatdifficulttasksrequirecomplexsupportsystems.Toorganizationalsociologistsand psychologistsalike,the"teaching"ofthinkingisamuchmorecomplexoperationthanisnormally foundinthetraditionalclassroom.Thisisparticularlythecasewhereteachershavelittle understandingregardingthenatureofthelearningprocessandhavebeenencouragedtotreatall contentasthesame(Berliner&Rosenshine,1976;DeAvila&Cohen,1983),asifitwereacquired bythesameprocess.Moreover,manyapproachestotheteachingofcontentencourageroutineatthe expenseofthemorecomplexprocessesrequiredforconceptformation.ShavelsonandSern(1981) haveanalyzedagooddealofteacherbehaviorasastrongadherencetoroutine.AccordingtoDe AvilaandCohen(1983),moreindividuallyrelevantmaterials,morestaff,andmostimportant, moreinteractionamongparticipantsarerequiredtoadequatelyimplementtheprocessesunderlying conceptformation(seeBourne,1966).Thus,fosteringthedevelopmentof"thinkingskills"would require,accordingtothesewriters,morecompleteinterdependentstaffarrangement.Mostteachers (letaloneaides)havenotbeentrainedforthesemorecomplexarrangementsandmost administratorswhosetrainingwasbasedonthetheoryof"culturaldeprivation"(seeRiessman, 1962)arenotaccustomedtothem. Fromapsychologicalpointofviewitappearsthatthecurriculuminmostclassroomsrequiresrecall skillsforfactualinformationandlittleelse.Thedisadvantageofthisapproach,forallchildren,is thatitlimitsthecommitmenttoindividualgrowthandselfdevelopmentespousedinthegoalsof "goodteaching"(Intili&Flood,1976)andactivelydiscriminatesagainstthoselanguageminority studentswhodonotshareinthecontentreferentembodiedinthe 152

curriculum(DeAvila&Havassy,1975).Inthisway,languageminoritystudents,totheextent theyareseenas"bilingual,"aredeniedaccesstoprogramsthatgobeyondpairedassociateor rotelearning(seeDeAvila&Cohen,1983). SAMPLE Theclassroomsselectedforthisresearchwerepartofalargergroupreceivingbilingual instructionunderagrantfromtheU.S.OfficeofEducation(TitleVII).TheTitleVIIprogram providedremedialservicesinbothEnglishandSpanishwithanemphasisonthebasicskills. Childrenparticipatingintheprogram(FO/D)wereapproximately253second,third,and fourthgradestudentsdrawnfromapopulationofstudentslivinginsuburbanandmetropolitan areassurroundingSanJose,California.Atotalofnineclassroomsfromninedifferentschools wereinvolved.Moredetaileddescriptionsofthesamplecanbefoundinthevariouspublications andreportsresultingfromtheprogram.Forthepresent,notethatthestudentsmakingupthe samplewerepredominantlyfromlowermiddleclassMexicanAmericanbackgrounds.Although therewereseveralBlackandAsianAmericanstudents,thegroupasawholewaswidely heterogeneouswithrespecttolinguisticsproficiencyinbothEnglishandSpanish. InadditiontothestudentswhoparticipatedintheFO/Dprogram,300otherstudentswhowere partofthebilingualprogramwereusedtoconstituteacomparisongroup.Notethattheterm comparisongrouphasbeenusedasopposedtocontrolgroupbecausethiswasafieldstudyand notanexperimentinthetraditionalsense.Forexample,studentswerenotrandomlyassignedto treatmentandcontrolgroups.SeeWilson,DeAvila,andIntili(1982)foramoredetailed discussiononthisissueasitrelatestothepresentstudy.Thecomparisongroupwasquitesimilar tothestudyinallotherrespectssaveexposuretothescienceandmathactivitiesdescribedas follows. DESCRIPTIONOFTREATMENTAPPROACH(FINDINGOUT/DESCUBRIMIENTO) Thespecificclassroomortreatmentactivitiesweretakenfromabilingualscienceandmath programentitledFindingOut/Describrumiento(DeAvila&Duncan,19791982).FO/Dis designedforusewithsmallgroupsandismadeupofapproximately150activitiesthatrequire measuring,counting,estimating,grouping,hypothesizing,analyzing,andreportingresults.Basic skillsarepresentedinthecontextofactivities.Forexample,therearenumeroussituationsin whichstudentsareaskedtomakeestimationsandthentochecktheaccuracyoftheirestimations throughsubtractingestimatedvaluesfromobtainedvalues.Theactivitiesrequirecognitive operationsfoundinproblemsolvingtasksusedin 153

cognitivepsychologyresearch(seeSternberg,1981;Valett,1978).Individualactivitiesare organizedandpresentedinmuchthesamewayasinalaboratoryconceptformationorlearning setexperiment.Thus,thereareagreatnumberoftrialsoractivitiesrequiringtheuseofthesame conceptwhiletheirrelevantaspectsordimensionsarevaried.Inthisway,forexample,students areexposedtovariousaspectsoftheconceptofnumberinatleast12differentactivities,each presentedinslightlydifferentconfigurationswhilethe"principleofsolution"isheldconstant (seeHunt,1961). Studentswererequiredtoworkinsmallgroupstoincreaseverbalinteraction.Bypresentingall materialinbothlanguages,anattemptwasmadetomaketheconceptsunderlyingeachtask invariantwithrespecttolanguage.Studentswerefreetointeractinwhateverlanguageor combinationoflanguageswasneededinordertofacilitatecommunication.Linguistic differenceswereviewedasonemorewayinwhichconceptscouldberepeatedwithoutchanging theunderlyingconcept(i.e.,principleofinvariance).Allstudentswereaskedtocompleteeach activityandacorrespondingworksheet(alsoprovidedinbothlanguages).Studentsworkedin smallgroupswheretheywerefreetointeractordiscusseachactivity,aswellastoassistone anotherinfillingouttheworksheets.Studentflowwasmanagedbytheteacherwhoalso facilitatedunderstandingofthetaskinstructions,whichwereprovidedincartoonformatwith textinbothEnglishandSpanish.Thechildcouldtherebyusethemultipleresourcesprovidedby thepictograph,text,teachers,orpeers. OUTCOMEMEASURES Studentprogressandbehaviorwereassessedatdifferentpointsduringthe14weeksrequiredto coveralloftheactivities.Avarietyofcommerciallyandspeciallydevelopedinstrumentswere used.Takenasawhole,thedatacollectedwerebasedonbothdirectandindirectassessment methodsincluding(1)paperandpenciltests,(2)classroomobservations,(3)dailyperformance onworksheetassignment.Resultsarereportedinthreesections. 1.PaperandPencilOutcomeMeasures Threepaperandpencilmeasureswereadministeredonapre/postbasistoassesslearning outcomes.Theircontentandadministrationaredescribedasfollows. IntellectualDevelopment TheCartoonConservationScales(DeAvila&Pulos,1978;Fleming&DeAvila,1980)area collectionofPiagetianinspiredtasksthataregroupadministeredbymeansofacartoonbooklet. Inall,thereare32itemsthatfallintotheeightsubscalesincluding(a)number,(b)length,(c) substance,(d)distance,(e) 154

horizontality,(f)egocentricity,(g)volume,and(h)probability.TheconceptsembodiedintheCCS morecloselyresemblevariousdefinitionsofthinkingskillsthananyotherofthemeasuresinthe testbattery.However,insofarastheconceptsunderlyingtheCCSarenotdirectlycoveredaspartof theactivities,theyshouldbeviewedasmeasuresoftheextenttowhichconceptscoveredinthe activities"generalized"tobroadermentalprocesses.SeeBruner,Oliver,andGreenfield(1966)for amoregeneraldiscussionbetweenconservationtasksand"thinkingskills." AcademicAchievement TheComprehensiveTestofBasicSkills(CTBS)isanationallynormreferencedtestofschool achievement.Itismadeupoftwosections,mathandreading.Forthepresentpurposes,letusbear inmindthatthemathsectionisbrokendownintothreesubscales:computation,concepts,and applications.Theitemsthatmakeupthesesubscalesdifferfromoneanothernotonlyincontentbut inthemethodsusedtoteachthem.SeeDeAvilaandCohen(1983)foramoredetailedanalysisof thesedifferences.Forexample,mathcomputationisroutinelytaughtthroughmemorization,rote,or pairedassociatelearning.Ontheotherhand,mathconceptsitemsmorecloselyresembleconcepts foundintheliteratureonconceptformation(seeBourne,1966).ItwasexpectedthattheFO/D experiencewouldhaveastrongerimpactonMathApplicationsthanonsubscalesrequiringlessin thewayof"thinkingskills,"i.e.,mathcomputation. ScienceMath TheMINItestwasdesignedasacontentreferencedmultiplechoicetestofthevocabularyand conceptscoveredintheactivities(seeHansen,1980,foradescriptionofitsdevelopmentand psychometricproperties).Itwaspresentedinbothpictorialandwrittenformatsinanattemptto reducetheeffectsofreadingskills.ThetestwasadministeredineitherEnglishorSpanishversions dependingontheproficiencyofthestudent.Scoresarereportedasthetotalnumberofitems correctlyanswereddividedbythenumberattempted.Inall,therewere55items. II.Observational/ProcessMeasures Dataonclassroombehaviorswerecollectedthroughtheuseofthreedifferentobservational instrumentsthatfocusedon(a)theteacherandaide,(b)thestudent,and(c)theclassroomasa whole.Eachisbrieflydescribedasfollows. TeacherAideBehaviors Periodicclassroomobservationsweremadethroughtheuseofanobservationsystemdesignedto assessthefrequencyofthefollowingbehaviors: 155

1 Asking . and/or StudentBehaviors respondingto Studentbehaviorswereobservedperiodicallythroughoutthe14weeksthestudy student wasinoperation.Becauseofthedifficultyandexpenseofobtainingthesedata, questions observationsweremadeononlyasubsampleofstudents(N=106).The 2 Facilitating subsampleof"target"studentswasdrawnonthebasisof"relativelanguage . performance proficiency"accordingtothepartitionsreportedinDeAvilaandDuncan(1983). onthe Inthisway,thesubsamplewascomposedofstudentswhowereproficient activities monolingualsinEnglishorSpanish,fullyproficientbilinguals,partiallyproficient (showing inoneortheotherlanguage,somewhatlimitedinboth,ortotallylimitedinboth. studentshow SeeDeAvila,Cohen,andIntili(1981,1982)foramoredetaileddiscussionofthe tocomplete procedures.Inallverbalinteractions,theextentofbothEnglishandSpanishusage theactivity wasrecorded.Thespecificbehaviorsobservedarelistedasfollows: 3 Providing 1. Taskrelatedtalk . feedbackas 2. Requestsforassistance tothe 3. Offersofassistance correctness 4. nontaskrelatedtalk ofworksheet 5. Talktoteacherasopposedtoother and/or 6. Cleaningup activities 7. Workingaloneasopposedtoworkingineithersmallorlargegroups 4 .Disciplining 8. Observingothersworking . unacceptable 9. Waitingfordirectionsfromanadult behavioror 10. Taskrelatedtransitionversus"wandering" providing Statedbriefly,thepurposeofthetargetchildobservationwastoprovide management qualitativeinformationastothelanguage,specificsociallydefinedlearning directions behaviors,taskrelatedinteractions,andlevelsofengagement.Thecollectionof 5 Providing thesedataenabledustolinkclassroomsocialbehaviortolearningoutcomeasa . direct functionofstudentlinguisticcharacteristics.Amoredetaileddiscussionofthe instruction relativereliabilityandvalidityoftheobservationalmeasuresmaybefoundin (seeBerliner CohenandIntili(1981,1982). and Rosenshine, 156 1976)

WholeClassBehaviors Todescribetheorganizationalfeaturesoftheclassroom,aprocessmeasureofgroupingpractices wasemployed.Weeklyobservationwasmadeofeachclassastosizeofgroups,natureof activity,andinteraction.Wholeclassobservationwasalsousedasacheckfortheindividual targetstudentandteacherobservations.Foramoredetaileddescriptionoftheprocedureandits agreementwithindividualobservationsseeCohenandIntili(1982). III.PerformanceMeasures Inadditiontothedataprovidedbypaperandpencilandobservationalmeasures,datawerealso collectedonthelinguisticproficienciesofallparticipatingstudentsinbothEnglishandSpanish. Performancedataonworksheetswasalsocollectedfortargetstudents. LanguageProficiency ProficiencyinbothEnglishandSpanishwasassessedinthebeginningofthestudyandits completionbymeansoftheLanguageAssessmentScales(DeAvilaandDuncan,19811982), FormALevelI.SeeDeAvilaandDuncan(1983)foradetaileddiscussionofthepsychometric propertiesofthetest.TheLASismadeupofthefollowingsubscales: 1. Phonemes 2. Minimalpairs 3. Vocabulary 4. Sentencecomprehension 5. Production(storyretelling) Scoresofthecombinedsubtestsaresummedtoprovideacompositeweightedscorewhich representsthestudent'sleveloforallinguisticproficiency:Scorevaluesrangefrom"no proficiency"to"fullyproficient."Thescoresonthestoryretellingsectionofthetestaretheonly subjectiveratings;therefore,thissectionwasblindscoredbynonprojectstaff.Scoringforallof theotherprojectdevelopedmeasureswasconductedbyprojectstaff.ScoringoftheCTBSwas bythepublisher. StudentWorksheets Worksheetswerecollectedforeach.Theywerescoredfor: 1. Totalnumbercompleted 2. Accuracyofcomputationsand/ordescriptions 157

WorksheetswereavailabletostudentsinbothSpanishandEnglishversions.The interraterreliabilitiesforthecodingcategoriesusedtoscoreworksheet performancearereportedinCohenandIntili(1982). RESULTS Theresultsweregeneratedinseveralseparateseriesofanalyses.Inthefirstseries thegeneralquestionofoutcomeswasaddressed.Inthesecondseriesanattempt wasmadetoexaminetheresultsofthefirstseriesofanalyses;outcomes,student characteristics,andclassroombehaviorswereexaminedinconjunctionwith classroomprocessesandinstructionalmethods.Thus,thefirstserieswasdirected atexaminingoutcomes,whereasthesecondwasaimedatexaminingunderlying processes.Theanalysesofoutcomeswerebasedontheentiresample,whereas theanalysesofprocess(observational)datawerebasedonthetargetstudent subgroups. Series1:Outcomes PrePostGains Treatmenteffectswereexaminedbytime,condition,school,sex,andageforeach setofdependentvariables.Consistentstatisticallysignificanttestgainswere foundonprepostcomparisonsforallmajorvariables(i.e.,CTBS,MINI,LAS, etc.).Gainsinteractedwithtreatmentandschoolvariations.Schoolvariation, however,wasassociatedwithlevelofimplementation(seeAnthonyetal.,1981). Thatis,althoughtheFO/Dperiodwasintendedtobe1hourperday,some teachersallocatedlesstime.Treatmentwassignificantlyrelatedtofiveofthe eightoutcomemeasures.Examinationofsexdifferencesrevealedthattherewere nomeaningfuldifferencesbetweenboysandgirlsastoeitherinitial(pretest)base ratesorgainsovertime. Agedifferenceswerealsoexamined.Inthisseriesofanalysessignificantage effectswerefound,suggestingthattheprogramwasleast"effective"fora subgroupofolderstudents.Examinationoflanguagedataindicatedthatthis groupwasmadeupof14monolingualSpanishspeakers.Althoughtherewere otherSpanishmonolingualsinthesample,these14studentstendedtobe substantiallyolderthanotherstudents,insomecasesbyasmuchas2to3years. Insubsequentanalyses,itwasfoundthatthesestudentswererecentimmigrants andtendedtobeof"lowstatus"intheclassroom(seeCohen&Intili,1982).They 158

3 Useand . qualityof written language 4 Useand . completeness ofdrawings 5 Complexity . ofreasoning anduseof inference

alsohadfewerverbalinteractionswitheitherpeersoradults.Thegeneralimportanceof"talking andworkingtogether"isaddressedlater. ComparisonswithNormsandPostHocGroups Inasecondphaseofthisseriesofanalyses,theperformanceofparticipatingstudentswas comparedtopublishernormsandtoposthoccomparisongroups.Duetoanumberofproblems beyondourcontrol,thesecomparisonswereavailableforCTBSscoresonly.Consistent significantgainswerefoundforbothFO/Dandcomparisongroupsthatslightlyfavoredthe FO/Dgroup.Inasecondsetofcomparisons,expectedgainsintheacademiccontentareas(i.e., readingandmath)werecalculatedbyobtainingthedifferencebetweenpublisherrawscorenorms equivalenttothe50thpercentileforfallandspringtestadministration.Thedifferencesbetween thesetworawscoreswerethentreatedas"expectedgain"scoresandcomparedtopre/postgains scoresobtainedbyFO/Dandcomparisonstudents. ResultsforthecomparisongrouprevealedthatactualgainsdidnotalwaysmatchtheTitleVII programexpectations.Inseveralinstances,aspredictedfromourinitialcontentanalysesofthe CTBSsubscales,comparisonstudentsfellfurtherbehindpublishernormsorsubscalesrequiring conceptualproblemsolvingeventhoughtheyimprovedinanabsolutesense.Whenrelativegains ofthecomparisonstudentswerecomparedtogainsexperiencedbytheFO/Dgroup,significant differenceswerefoundinfavoroftheFO/DgrouponthepredictedCTBSsubscales,math application,andconcepts.Thus,althoughthecomparisongroupwasmakingsubstantialgainsin mathcomputation,askillusuallytaughtthroughrotemethods,studentswereactuallyfalling furtherbehindthenormsinthemoreconceptualskillareas. StudentCharacteristics Inanattempttodeterminetheextentofimprovementasafunctionofstudentcharacteristics, priortotheimplementationofFO/D,asubgroupof"problem"students(seeRosenholtz,1981) wereidentified(N=60)fromthe106targetstudentsinthefollowingway.Duringthepretest period,eachteacherwasaskedtoidentify6to8studentswhotheyfeltwouldbelikelyto experiencedifficultyinmasteringthebasicschoolcurriculumand/orthescience/mathconcepts embodiedintheCO/Dactivities.Bythisprocedurewewereabletoidentifyagroupofstudents forwhomtheteacherheldlowerexpectations.Comparisonsbetweenlowexpectationand remainingFO/Dstudentsshowedlowexpectationstudentstoperformconsistentlyloweratboth preandposttestadministrations.Nevertheless,theabsolutegainsforthetwogroupswere virtuallyidentical.Inotherwords,eventhoughthe"problem"group"scoredloweratbothpoints intime,theirrateofimprovementwasindistinguishablefromtherestofthegroupdespitethe lowerexpectations. 159

Inasecondattempttoexaminegainsasafunctionofstudentcharacteristics,asubgroupwas formedonthebasisofintellectualdevelopment.ForthispurposetheCCSprovidedameansfor identifyingpotentially"gifted"students.Anarbitrarycriterionscoreof28pointsoutofa possible32wasusedtogenerateagroupof40students.Pre/postanalysesontheoutcome measureswerethenrun.Resultsrevealedsignificantlygreatergainsacrosstimeforgifted studentsonmostofthemeasureswherecomparisonswerepossible.Theonlyexceptiontothis findingwasontheCTBSmathcomputationsubscalewhereimprovementwassubstantiallythe sameasfortherestofthesample.Ofparticularsignificancewerethegainsinreading (vocabularyandcomprehension)andmath(conceptsandapplications).Inafinalattemptto examineprogrameffectsasafunctionofstudentcharacteristics,datawerealsoexaminedonthe basisoflinguisticconsiderations.Thisquestionwasexaminedintwoways.Inthefirst,thetotal samplewassubdividedintotwogroupsonthebasisoforalEnglishlanguageproficiencyscores (LAS).Theresultanttwogroupsweredefinedas"limitedEnglishproficient"(LEP)and"fluent Englishproficient"(FEP).Acomparisonoftherelativepre/postgainsforthetwogroups revealedsignificantimprovementforbothgroupsonmostoutcomemeasures.Furtheranalyses revealedslightlystrongergainsfortheFEPgrouponthemoretraditionalmeasuresofschool achievement,whichincludedreadingandmathtotalscoresandseveralsubscales.Ontheother hand,LEPstudentsshowedslightlystrongergainsintheCCS,LAS,andMINItest.Thefactthat theCTBSwasadministeredonlyinEnglishwouldaccountfortheslightlystrongergainsonthe partofthefluentEnglishspeakers.Nevertheless,LEPstudentsdidshowsignificant improvementontheCTBS.Thisimprovementwassimplynotquiteasgreatasthatofthefluent speakers.Furthermore,theimportanceofthelinguisticfactorisrevealedbythefactthatonthe teststhatstudentswereabletotakeineitherlanguage,gainswerestrongerfortheLEPgroup. Finally,thefailuretofindgainsontheLASfortheFEPwascertainlyattributabletoaceiling effectandtothearbitrarydefinitionofgroups.TheanalysesofdatabyLEPandFEPsubgroups wereconductedwithoutconsiderationofhomelanguageproficiency.Togaininsightintothe natureoftheinteractionofthetwolanguages,studentswereregroupedaccordingto"relative linguisticproficiency."UsingLASscores,fivelinguisticsubgroupsweregenerated.They included: 1. Englishmonolinguals 2. Spanishmonolinguals 3. Limitedspeakersofboth 4. Minimalspeakersofboth 5. Bilinguals(i.e.,fullyproficientinboth) 160

Analysesofprogrameffectsbylinguisticgroupshowedthattheproficientbilingualspeakers exhibitedthestrongestandmostconsistentgains.Bilingualswerefollowedbylimitedbilinguals (Group3)andSpanishmonolinguals(Group2)whowerefollowedbyminimalspeakers(Group 4)andEnglishmonolinguals(Group1).Notethateventhoughtherewerebetweengroup variations,allgroupsshowedsignificantimprovementonallthemeasuresexceptontheLAS, whichwasduetoanartifactofgroupdefinition(ceilingeffects). SeriesII:ProcessAnalyses Inanattempttoexaminetherelationshipof"process"variablestolearningoutcomes,anumber ofanalyseswereconductedusing"observationaldata"inconjunctionwiththeoutcomemeasures justdescribed.Twosetsofanalysesareofparticularimportance.Inbothinstances,datawere basedonthetargetstudents.Inthefirstset,UyemuraStevenson(1982)examinedthe relationshipofstudentstudentconsultationtoacademicperformance.Inthesecondset,DeAvila andCohen(1983)comparedteacherbehaviorinregularmathclassroomswithbehavior observedduringtheFO/Dclassperiod,andthentied"methodofinstruction"tospecificlearning outcomes. UyemuraStevenson(1982)examinedthreespecificquestionsderivedfromsociologicaland psychologicalprinciples.Inthefirst,theeffectofstudentstudentconsultationwasexamined. Studentstudentconsultationwasdefinedas"taskrelatedtalkbetweenstudentsorothermeansof exchanginginformationaboutthetask(i.e.,modelingordemonstration)"(p.20).Student consultationwasoperationalizedonthebasisoftheobservedbehaviorscodedonthe"target studentobservationform"justdescribed.Specificbehaviorscodedincludedtaskrelatedtalk, requestsforassistancefromanotherstudent,offersofassistance,andworkingtogetherwithout verbalinteraction.Examinationofthecorrelationsbetween"talkingandworkingtogether"and theoutcomemeasuresjustdescribedrevealedsignificantrelationshipsbetweenstudentstudent consultationandmathconceptualizationasmeasuredonthestudentworksheetandtotalmath scoresasmeasuredontheCTBS.Infact,itwasfoundthat"consultingwithfellowstudentswasa morepowerfulpredictoroflearningoutcomesthanconsultingwithteachers"(p.44). Inaddition,UyemuraStevensonfoundthatstudentinteractionfacilitatedperformanceonother academicbehaviorssuchasthestudent'slevelofaccuracyanduseofwrittenlanguage(as opposedtodrawingpictures)todescribetheproceduresandresultsoftheactivities.Thus, UyemuraStevensonfoundthattheextentofstudentinteractionseemedtoberelatedtothe improvementofafairlybroadrangeofbehaviors.Notethatthestudentdataonwhichthese analyseswereconductedwerebasedonalinguisticallyheterogeneoussampleofstudents includingmonolinguals,proficientbilinguals,andchildrenwhohadlittleproficiencyineither standardSpanishorEnglish.Oneinterpretationofthesefindings 161

suggeststhatthelinguisticvariabilityofthesamplemayhaveactuallycontributedtogainsin conceptualareas. InasecondsetofanalysesUyemuraStevensonexaminedtheextenttowhichconsultationwasmore effectiveforstudentswhoexhibited"lowacademicandproblemsolvingskills"andforthosewith higherskillsintheseareas.ThisquestionwassimilartothataddressedbyDeAvila,Cohen,and Intili(1981),andbyRosenholtz(1981).Rosenholtzfoundthatstudentswhowereseenbythe teachersashavingalowprobabilityforsuccessshowedaboutthesameimprovementasother studentsalthoughatlowerabsolutelevels. Totestthenotionofadifferentialeffect,regressionswereconductedforstudentsdividedatthe medianonprioracademicor"cognitiveresources."Withconcretetaskssuchasfillingoutthe worksheets,analysesshowedthatstudentconsultationwasmoreimportantforstudentswithfewer academicresources(i.e.,lowerprioracademicachievement)thanforstudentswhohadexperienced lessdifficulty.Ontheotherhand,resultsshowedthatstudentconsultationwasmoreeffectivefor highachieversonmoreconceptualmaterial. UyemuraStevensonpointsout,however,thattherewasanimportantrelationshipbetween"status" andstudentinteractionthatwouldhaveanimpactontheextentofconsultationbetweenhighand lowstatuschildren.Inarelatedstudyonthesedata,CohenandAnthony(1982)foundthat"high status"studentsweremorelikelytobefound"talkingandworkingtogether"thanchildrenoflower statuscharacteristics.Itappearsthatsocialstatusaffectsstudentinteractionandthatstudent interactioninturnaffectslearningoutcomes.Ifstatuscharacteristicsincludeteachersandpeer perceptionofthechild'sabilityto"speakEnglish"thenitwouldalsoseemclearthattheproblemof languagedifferencefacedbythesestudentsisexacerbatedbya"socialdistance"thatdeniesthem accesstoverbalinteraction. Inathirdsetofanalyses,UyemuraStevensontestedthepropositionthatcombinedteacherand studentconsultationwouldhavean"interactiveeffect"thatwouldleadto"betterqualitylearning outcomesthanwouldbeexpectedfromaddingtheeffectsofeachtypeofconsultation."Whenthis propositionwastested,regressionanalysesfailedtorevealanystatisticallysignificant"interaction effects"betweenteacherandstudentconsultation;studentstudentconsultationremainedthemost significantpredictorofworksheetperformance.AccordingtoUyemuraStevenson,theresultsof thissetofanalysessuggeststhattalkingandworkingtogetherseemtohavetwospecificfunctions: (a)regardlessoflinguistic"deficiencies,"itallowsstudentstouseoneanotherasresources,and(b) itpromotesinteraction,whichinturnfacilitatesconceptuallearning,particularlywhentheteacheris ascarceresourceandunabletoprovidethetypeof"immediatefeedback"widelyadvocatedbysuch writersasBerlinerandRosenshine(1976). DeAvilaandCohen(1983)conductedsecondaryanalysesofthedatajustdescribedinwhichthe relativeeffectivenessof"direct"and"indirect"instruc 162

tiontechniqueswerecompared.Inthisseriesofanalyses,DeAvilaandCohen(1983)observed teacherandstudentbehaviorsintwosettingsandwereabletolinkthesebehaviorstospecific predictedacademicoutcomes. ThedataconsistedofobservationsmadeduringFO/Dandregularmathperiods.Originaldata collectedbyRosenholtz(1981)hadshownthatthemathclassesandFO/Dclassesweretwovery differenttypesofenvironments.Mathclassesweretypifiedby"directinstruction,"withteacher andaideprovidingdirectsupervisionovertwoorthree"abilitygroups."Insharpcontrasttothe FO/Dperiod,therewasverylittleinthewayof"studentstudentconsultation"orverbal interactionbetweenstudents.Rosenholtz'sdataalsoshowedthatengagementwashigherunder complexity. Aspredicted,observedrateof"directinstruction"wasrelatedtogainsonCTBSMath Computation,whereasratesof"talkingandworkingtogether"wereassociatedwithgainsin CTBSMathApplications.TheirfindingsleadDeAvilaandCohen(1983)toconcludethatthere areinterrelationshipsbetweenthenatureoflearning,typeofclassroomorganization,andmethod ofinstructionthatmustbetakenintoaccount.Wholeclassinstructionseemstobeperfectly adequatefortasksrequiringmemorizationoffactsandfigures.Conceptuallearning,ontheother hand,seemstoimplyamorecomplexclassroomorganization.Inotherwords,giventhe complexityofthesituationbroughtonbythediversityofstudentpopulationandbythenatureof thetask,theteacherhaslittlechoicebutto"delegateauthority"(seeCohen&Intili,1982)when educationalobjectivesrequireconceptuallearning. NievesSquires(1980)examinedacademicandlinguisticgainasafunctionofrelativelinguistic proficiencyandinteraction.Resultsofthisseriesofanalysesrevealedthat,withslightvariation, interactioncovariedwithacademicandlinguisticgainasafunctionoflinguisticsubgroup. Bilingualswerefoundtohavehadmoretotalinteractionthananyothergroupwiththeexception ofSpanishmonolinguals.TheverbalinteractionofSpanishmonolinguals,however,was restrictedtoitself.Moreover,gainstendedtobelessconsistentthanforBilinguals.Lowestlevels ofinteractionwerefoundfortheLimitedgroupwiththeothergroupsfailingintermediatetothe extremes. CONCLUDINGREMARKS Anumberofissuesoftheoreticalandappliedinteresthavejustbeendiscussed.Onthe theoreticalsidewehavereviewedresearchontherelationshipbetweenbilingualismand intellectualfunctioning.Ontheappliedsidewehaveconsideredanumberofpsychosocialissues ascontributorstotheschoolsuccessoflanguageminoritystudents.Ostensibly,thefirstissue addressesthequestionofcognitionandbilingualism,whereasthesecondaddressesthequestion ofappropriatetreatmentforbilinguals.DeAvila(1982)hasrecentlysuggestedthatthe 163

schoolbehavioroflanguageminoritystudentscouldbeunderstoodasafunctionofthreeinteracting factors,including(a)intelligence/style,(b)interest/motivationand(c)opportunity/access.Inthe followingconcludingremarks,ratherthansimplyrestatingtheresults,wediscussourresultswithin theframeworkprovidedbythesethreefactors. Considerthequestionoftheinteractionofintelligence,cognitivestyle,andbilingualismas contributorstolanguageminorityperformanceonacademicsubjectmatter.Formany,poor academicperformanceisattributableto,ifnotcausedby,bilingualism.Thisconclusionseems faultyonthreegrounds. First,inthereviewofvariousliteratures,itwasfoundthatconflictingresultsaretheresultofa failuretodistinguishbetweenethnolinguisticgroupmembershipandrelativelinguisticproficiency. Thatis,thefailuretocontrolfortheabsolutelanguageproficiencyofcomparisongroupshas resultedinaconfoundingoflanguagewithintellectualdevelopmentandcognitivestyle. Second,inpastresearchwherelanguageproficiencydifferenceshavebeencontrolled,fewifany meaningfuldifferencesinintellectualdevelopmentorcognitivestylehavebeenfoundthatcouldbe attributedtobilingualism.InthisconnectionSaarni(1973),usingPiagetinspiredtaskssimilarto thoseusedinthepresentresearch,foundthatdifferencesincognitivestylebetweenboysandgirls didnotpredicttodifferencesinintellectualfunctioning.Similartoourownwork,Hyde(1981) reviewedtheliteratureonsexdifferencesinmathematicalreasoning,conductedametaanalysisof theresults,andconcludedthatreportedsexdifferences,althoughstatisticallysignificant,tendednot toaccountformeaningfulproportionsofthetotalvariance.Finally,wehavefoundlinguistic proficiencyamonglanguageminoritystudentstobeafarstrongerpredictorofacademic performancethaneithercognitivestyleorintellectualdevelopment.However,linguisticproficiency inEnglish,althoughnecessary,doesnotseemtobeasufficientcondition. Third,wefindthatcomparisonsbetweenthetotalgroupofFO/DstudentsandnormsontheCCS collectedoverthepast15years(N=6,000)failedtorevealanystatisticallysignificantdifferences. WithrespecttointellectualdevelopmentasmeasuredbyPiagetiantasks,thisgroupofstudentswas nodifferentfromanyothergroupofstudents.Ontheotherhand,however,thegroupwas substantiallybehindtheirmainstreamcounterpartsinlevelofacademicperformance.Althoughthey seemtopossessnecessarylevelsofintellectualdevelopment,intellectualdevelopmentperse,like linguisticproficiency,isnotsufficientforschoolachievement. Ourconclusiononthequestionoftherelationshipbetweenbilingualismandcognitivestyleor intellectualfunctionsisthreefold:(a)Totheextentthatdifferencesinacademicperformanceexist betweenlanguageminoritystudents,theycannotbeattributedtothegroupspresumedbilingual;(b) Students,totheextenttheyareproficientbilinguals,experienceawiderangeofcognitiveandsocial advantagesoverotherstudents;(c)Althoughlinguisticproficiencyand 164

intellectualdevelopmentarenecessaryforthesuccessoflanguageminoritystudents(andother children),theyarenotsufficientgiventraditionalclassroompracticesandorganization. Thesecondmajorfactorinourframeworkdealswiththeissueofmotivationandinterestwhereby theliteratureondifferencesbetweenlanguageminorityandmajoritypopulationssuffersfromthe sameproblemsastheliteratureonintellectualandcognitivestyledifferences.Althoughthischapter didnotcovertheissueofmotivationinanydirectsense,afewcommentsareinorder.Anumberof recentwritershavefoundstudentengagementtobeastrongpredictorofsuccessinacademic subjectareasandhavereportedengagementratesonacademictaskstobeaslowas10%of allocatedtime.Basedonthefindingthatindividualizedseatworktendstoproducelowlevelsof studentengagement,BerlinerandRosenshine(1976)haveconcludedthatthebestwaytoremediate pooracademicperformanceisbymeansofthemethodof"directinstruction."Themajorthrustof thismethodologicalapproachistoimproveacademicperformancebyincreasing"timeontask" throughdirectteachersupervisionandwholeclassinstruction.However,studentengagementcanbe viewedasanindexofinterestandmotivationtocompleteacademictasks.Inthisconnection, Rosenholtz(1981)showedthatdirectinstructiondidnotalwaysmaximizetimeontask,especially whenitoccurredincomplex(multiplegroupsandmaterials)classroomorganizationalstructures suchasrequiredbyFO/D.Rosenholtzfoundthatpeertasktalkwasrelatedtoengagementinboth mathandFO/Dclasses.However,taskengagementwascorrelatedwithdirectinstructiononlyinthe mathclassandwasuncorrelatedintheFO/Dperiodwherethereweremultiplegroupsandactivities insimultaneousoperation.Ingeneral,itwasfoundthatthelevelofengagementduringtheFO/D periodwasinexcessof80%.Whatcanbeconcludedfromthesedataisthatstudentsareinternally motivatedtocompletetasksthattheyfindinteresting.Wefurtherspeculatethatpartoftheinterest isderivedfromthepeerinteractionrequiredtocompletethetasksandnotfromthedirect supervisionoftheteacher.Finally,itappearsthatdifferenteducationalobjectivesrequiredifferent instructionalmethods,which,inturn,implydifferentclassroomorganizations. Thethirdfactorinthefacilitationofacademicsuccessdealswiththeissuesofopportunityand access.Opportunityandaccessmaybediscussedatseverallevelsincludingsocial,curricularand interpersonal.Theconceptofeducationalequitythatmotivatesthesocialjustificationofthe developmentandfundingofcategoricalprogramssuchasTitleVIIBilingualEducation,isbasedon thenotionof"equalopportunityregardlessofsex,religion,raceornationalorigin."Althoughthe avowedpurposeofcompensatoryeducationprogramshasbeentoprovideequalopportunity throughremedialprograms,itisdoubtfulthatmuchofthesocialscienceresearchbaseunderlying thedesignofmanyprogramsactuallyprovidesequalaccesstoeducationandintellectual developmentonabroaderscale.Opportunityimpliesaccessbutitdoesnotguaranteeit.Forexam 165

ple,thebeliefthatEnglishlanguageandbasicskilldeficienciesprecludethinkingandscientific reasoninghasledtoprogramsforthelanguageminoritiesthat,bydefault,emphasizeroteskills attheexpenseofhigherorderintellectualprocesses.Thus,althoughstudentsareprovidedwith theopportunityforsuccessbyplacementinspecialclasses,presumeddeficiencies,ironically, precludeaccess. Thesocialscienceshaveidentifiedawidevarietyofstudentcharacteristicsthoughttocontribute toacademicsuccess.Whatwehavefoundisthatthewholesaleapplicationofmanyofthese findingsisfraughtwithdanger.Ofparticularimportancetothisdiscussionistherelation betweensocialstatusvariablesoperatingbothwithinandwithouttheschoolsettingandthe extenttowhichdifferences(oftenmorepresumedthanreal)onthesevariablesleadtodifferences inthedesignofprogramswhich,inturn,preemptparticularimportantprocessinlearning.The caseofstudentinteractionisaprimeexample.Whatthepreviousfindingshaveillustratedisthat studentinteractionisasignificantcontributortothelearningprocess,particularlyasrelatedto conceptformation.Interactionwithintheclassroom,however,istosomeextentmodifiedby socialstatusfactors.TotheextentthatEnglishlanguageproficiencyoperateslikeotheracademic statusvariables,suchasreading,itcanbeexpectedfromthesedatathatstudentswithlimited proficiencywillexhibitlowerlevelsofinteractionwithEnglishspeakingclassmates.Finally, whatthesedatahaveshownisthatunderclassroomorganizationalconditionswherelanguage minoritystudentsareprovidedwithaccesstomultipleresources,includinghomelanguage,peer consultation,observation,manipulation,andsoon,theywillacquireconceptsasreadilyas mainstreamstudentsand,atthesametime,acquireEnglishlanguageproficiencyandbasicskills. Infact,whatthedatashowisthatthe"bilinguals"areataheadstartinthisregard. REFERENCES Anthony,B.,Cohen,E.G.,Hanson,S.G.,Intili,J.K.,Mata,S.,Parchment,C.,Stevenson,B.,& Stone,N.(1981,April)."Themeasurementofimplementation:Aproblemofconceptualization". PaperpresentedattheAnnualMeetingoftheAmericanEducationalResearchAssociation, Session40.11. Aries,B.(1982,March).Contextualvariationandtheimplementationofthebilingual curriculum.PaperpresentedattheAnnualMeetingoftheAmericanEducationalResearch Association,NewYork. Berliner,D.,&Rosenshine,B.(1976)."Theacquisitionofknowledgeintheclassroom".inR.C. Anderson,R.J.Spiro,&W.E.Montague(Eds.),Schoolingandtheeducationprocess(pp.375 398).Hillsdale,NJ:LawrenceErlbaumAssociates. Bourne,L.E.(1966).Humanconceptuallearning.Boston:Allyn&Bacon. Bruner,J.S.,Olver,R.R.,&Greenfield,P.M.(1966).Studiesincognitivegrowth:A collaborationoftheCenterforCognitiveStudies.NewYork:Wiley. 166

Buriel,J.W.(1975)."CognitivestylesamongthreegenerationsofMexicanAmericanchildren". JournalofCrossCulturalPsychology,6,417439. Chan,K.S.(1983)."LimitedEnglishspeaking,handicapped,andpoor:Triplethreatinchildhood". InM.ChuChang(Ed.),AsianandPacificAmericanperspectivesinbilingualeducation: Comparativeresearch.NewYork:TeachersCollegePress. Clark,G.M.,&Peterson,P.L.(1976,April).Teacherstimulatedrecallofinteractivedecisions. PaperpresentedattheAnnualMeetingoftheAmericanEducationalResearchAssociation,San Francisco,CA. Cohen,E.G.,&Anthony,B.A.(1982,March).Expectationstatestheoryandclassroomlearning. PaperpresentedattheAnnualMeetingoftheAmericanEducationalResearchAssociation,New York. Cohen,E.G.,Deal,T.E.,Meyer,J.W.,&Scott,W.R.(1976).Organisationandinstructionin elementaryschools.(TechnicalReportNo.50).Stanford,Ca:StanfordCenterforResearchand DevelopmentinTeaching. Cohen,E.G.,&Intili,J.K.(1981).Interdependenceandmanagementofbilingualclassrooms (FinalReporttoNIE).Stanford,CA:StanfordUniversity,SchoolofEducation. Cohen,E.G.,&Intili,L.K.(1982).Interdependenceandmanagementofbilingualclassrooms (FinalReporttoNIE).Stanford,CA:StanfordUniversity,SchoolofEducation. DeE.A.Avila(1982)."KeynoteaddressattheCouncilforExceptionalChildren'sNational ConferenceonTrainingWorkshopsontheExceptionalBilingualChild",Oct.31Nov.2,Phoenix, Arizona. DeE.A.Avila,Cohen,E.G.,&Intili,J.K.(1981).Multiculturalimprovementofcognitive abilities(FinalReport).DepartmentofEducation,StateofCalifornia,Sacramento,CA. DeE.A.Avila,Cohen,E.G.,&Intili,J.K.(1982)."Improvingcognition:Amulticultural approach".InS.S.Seidner(Ed.),Issuesoflanguageassessment:foundationsandresearch.Illinois StateBoardofEducation,Springfield,IL. DeE.A.Avila,&Cohen,E.&(1983).Indirectinstructionandconceptuallearning.Inpreparation. DeE.A.Avila,&Duncan,S.E.(1978).LanguageAssessmentScales(LAS).CorteMadera,CA: LinguametricsGroup. DeE.A.Avila,&Duncan,S.E.(1979,winter)."BilingualismandtheMetaset".NABEJournal, 3(2),120. DeE.A.Avila,&Duncan,S.E.(1980)."Definitionandmeasurementofbilingualstudents".In Bilingualprogram,policy,andassessmentissues.CaliforniaStateDepartmentofEducation, Sacramento,CA. DeE.A.Avila,&Duncan,S.E.(19791982).FindingOut/Descubrimiento.SanRafael,CA: LinguametricsGroup. DeE.A.Avila,&Duncan,S.E.(19811982).Aconvergentapproachtoorallanguageassessment: TheoreticalandtechnicalspecificationsontheLanguageAssessmentScales(LAS),FormA.San Rafael,CA:LinguametricsGroup. DeE.A.Avila,&Duncan,S.E.(1983).Thelanguageminoritychild:Apsychological,linguistic, andeducationalanalysis.SanRafael,CA:LinguametricsGroup. DeE.A.Avila,Duncan,S.E.,&Ulibarri,D.M(1982)."Cognitivedevelopment".InE.GarciaE (Ed.),TheMexicanAmericanchild:Language,cognitionandsocialdevelopment.Tempe,AZ:

CenterforBilingual/BiculturalEducation. DeEA.Avila,&Havassy,B.E.(1975)."PiagetianalternativetoIQ:MexicanAmericanstudy".In N.Hobbas(Ed.),Issuesintheclassificationofexceptionalchildren.SanFrancisco,CA:Jossey Bass. DeE.A.Avila,Havassy,B.E.,&PascualLeone,J.(1976).MexicanAmericanschoolchildren:A neoPiagetiananalysis.Washington,DC:GeorgetownUniversityPress. DeAvila,E.A.,&S.M.Pulos(1978).Developmentalassessmentbypictoriallypresented piagetianmaterial:TheCartoonConservationScales(CCS) 167

piagetianmaterial:TheCartoonConservationScales(CCSO).InG.I.Lubin,M.K.Poulsen,J.F. Magary,&M.SotoMcAlister(Eds..),Piagetiantheoryanditsimplicationsforthehelping professions(pp.124139).LosAngeles,CA:UniversityofSouthernCalifornia. Duncan,S.E.(1979).Childbilingualismandcognitivefunctioning:AstudyoffourHispanic groups.Unpublisheddoctoraldissertation,UnionGraduateSchool. Fleming,J.S.,&DeE.A.Avila(1980)."Scalogramandfactoranalysesoftwotestsofcognitive development".MultivariateBehavioralResearch,15,7393. Hansen,S.(1980).PsychometricpropertiesoftheMICAminitest.Unpublishedmanuscript, StanfordUniversity. Hunt,J.M.(1961).Intelligenceandexperience.NewYork:RonaldPress. Hyde,J.(1981)."Howlargearecognitivegenderdifferences"?AmericanPsychologist,36(8), 892901. Intili,J.K.(1977).Structuralconditionsintheschoolthatfacilitatereflectivedecisionmaking. Unpublisheddoctoraldissertation,SchoolofEducation,StanfordUniversity. Intili,J.K.,&Flood,J.E.(1976,April).Theeffectofselectedstudentcharacteristicson differentiationoninstructioninreading.PaperpresentedattheAnnualMeetingoftheAmerican EducationalResearchAssociation,SanFrancisco,CA. Laosa,L.M.(1977)."Cognitivestylesandlearningstrategiesresearch:Someoftheareasinwhich psychologycancontributetopersonalizedinstructioninmulticulturaleducation".Journalof TeachersEducation,28(3),2630. Lauv.Nichols(1974).No.726520(414U.S.at566,1974). Mestre,J.P.,Gerace,W.J.,&Lochhead,J.(1982)."Theinterdependenceoflanguageand translationalmathskillsamongbilingualHispanicengineeringstudents".JournalofResearchin ScienceTeaching,19(5),399410. NievesSquires,S.(1980).Bilingualinstructionalfeaturesplanningstudy.Cambridge,MA:ABT Associates. Perrow,C.(1967)."Aframeworkforthecomparativeanalysisoforganizations".American SociologicalReview,32,194208. RamirezM.III,&PriceWilliams,D.R.(1974)."Cognitivestylesofchildrenofthreeethnicgroups intheUnitedStates".JournalofCrossCulturalPsychology,5,212219. Riessman,F.(1962).Theculturallydeprivedchild.NewYork:Harper&Row. Rosenholtz,S.(1981).Effectsoftaskarrangementsandmanagementsystemsoftaskengagementof lowachievementstudents.Unpublisheddoctoraldissertation,StanfordUniversity. Rosenthal,A.S.,Milne,A.,Ginsburg,A.,&Baker,K.(1981).Acomparisonoftheeffectsof languagebackgroundandsocioeconomicstatusonachievementamongelementaryschoolstudents. Washington,DC:AUIPolicyResearch,OfficeofPlanningandBudget,DepartmentofEducation. Saami,C.I.(1973)."Piagetianoperationsandfieldindependenceasfactorsinchildren's problemsolvingperformance".ChildDevelopment,44,338345. Sanders,M.,Scholz,J.P.,&Kagan,S.(1976)."Threesocialmotivesandfield independencedependenceinAngloAmericanandMexicanAmericanchildren".JournalofCross CulturalPsychology,7(4),451462. Shavelson,R.J.,&Stern,P.(1981)."Researchonteachers'pedagogicalthoughts,judgments,

decisions,andbehavior"ReviewofEducationalResearch,51,455498. Sternberg,R.J.(1981)."Testingandcognitivepsychology".AmericanPsychologist,36(10), 11811189. Taba,H.(1966).Teachingstrategiesandcognitivefunctioninginelementaryschoolchildren. U.S.D.E.CooperativeResearchProjectNo.2404.SanFranciscoStateCollege. Takahashi,Y.(1982).Theeffectsofmodifiedunputandinteractionpatternsonlanguage acquisition.Unpublishedmanuscript,StanfordUniversity. UyemuraStevenson,B.(1982).Ananalysisoftherelationshipofstudentstudentconsultation toacademicperformanceindifferentiatedclassroomsettings 168

academicperformanceindifferentiatedclassroomsettings.Unpublisheddoctoraldissertation, StanfordUniversity. Valett,R.E.(1978).Developingcognitiveabilities:Teachingchildrentothink,St.Louis:C.V. Mosby. Wilson,B.,DeE.A.Avila,&Intili,J.K.(1982,March).Improvingcognitive,linguisticand academicskillsinbilingualclassrooms.PaperpresentedattheAnnualMeetingoftheAmerican EducationalResearchAssociation,NewYork. Witkin,H.A.,&Goodenough,D.R.(1977).Fielddependencerevisited(ETSRB7716). Princeton,NJ:EducationalTestingService. 169

[Thispageintentionallyleftblank.] 170

9 Bilingualism:CognitiveandSocialAspects JosephGlick CityUniversityofNewYork,GraduateCenter Itis,perhaps,everysocialscientist'sdreamtodoresearchofsuchvitalimportancethatitcan showupinanenlightenedsocialpolicy.Indeed,thehopeisthatinformationfoundoutinthe laboratory(eithernature'sortheuniversity's)willprovetobeofdirectrelevancetothevarious socialillsandsocialproblemsthatsurroundus.Bydoingmoreadvancedandsophisticated researchwehopetoreplacebenightedpolicieswithenlightenedones.Wewouldliketobelieve thatweareabletobuildasocietybasedonrefinedknowledge. Perhapsthisdreamisdestinedtoremainadream.Wemayhavetocometounderstandthatitis entirelypossiblethatresearchdesignandexperimentalthinkingarenotwelladaptedtothe realitiesofsocialaction.Normaytheworldofsocialrealitiesbeamenabletobeingresearchedin thewaywemightwishittobe. Thedifficultiesofdoingsociallyrelevantresearchderivedirectlyfromthefactofitssocial relevance.Socialfactsdonotoperatelikevariablesinresearchdesigns.Whiletheconditionfora researchdesignisthatavariableisisolatableforstudy,itismostoftenthecasethatthesocial worldoffersus"bundles"offactorsthatareintertwined. InthefollowingIidentifysomeoftheproblemsinvolvedintheextensionofstandardresearch methods,particularlyastheybearonissuesofbilingualism. SOMECOMMENTSONBILINGUALISM Consideredfromatheoreticalandexperimental(researchdesign)pointofviewitseemsquite reasonablethatthebasicargumentsadducedbyPalijandHomel 171

(thisvolume)andDeAvila(thisvolume)toshowthenonnegativeandpossiblypositiveimpact ofbilingualismonvariouscognitivefunctionsarecorrect,or,areatleast,notnecessarilywrong. Icanthinkofnostrongcognitivedevelopmentaltheorythatwouldseriouslyargueagainst bilingualismontheoreticalgrounds. Piaget(1967),forexample,wouldtreatbilingualisminmuchthesamewaythathewouldtreat therelationshipbetweenonelanguageandcognition.Theimportantvarianceexistsatthelevel ofthecognitivestructuresthatareavailableandnotinthecodethatisusedtorepresentthemor accessthem.ThegeneralPiagetianpositionmightberepresentedasassertingthatlanguage changesfollowonratherthanleadchangesincognitiveorganization.Primaryemphasisisplaced ontheavailabilityoftheconceptintheindividual'srepertoire. Theconceptitselfcomesfromavarietyoftypesofstructuringactivitieswithintheorganismand isnotsimplyanappropriationofanexistinglinguisticallyencodedsocialmeaning.Withinthis theoreticalpositionbilingualismwouldnotbebad;itwouldbemerelyirrelevanttoissuesof cognitivegrowth. InaveinsimilartoPiaget's,someoftheworkonhumaninformationprocessing(laboratory variety)wouldalsosuggesttheprimacyoftheavailabilityofconceptsinconsideringissuesof language,bilanguage,andcognition(Conrad,1964;Siegler,1983).Muchoftheinformation processingworksuggeststhatthesurfaceformofstatementofaproposition(e.g.,theparticular syntaxortheparticularwordsusedtoexpressit)quicklydisappearsasprocessingproceedsto deeperlevelsforlongertermstorage.Longertermencodingisintermsoftheideaexpressed (semanticorconceptuallevel). Becauseitmaybewellarguedthatdifferentlanguagesexpressthesamebasicsemanticnotions butusedifferentlexicalandsyntacticmeanstodoso,wemightnotexpectanylossesoffunction attendantonbilingualism. Wecouldreasonablyexpectlanguageinterferenceifthereareaspectsoflanguagethatdefine semanticdomainsinaradicallydifferentway.Anynegativeeffectswouldexistonlywithin mismatchedsemanticdomainsandnotatthelevelofgloballanguagedifferencesperse. Vygotsky(1962,1978),whoemphasizesculturalmediationalsystemsasbeingofimportanceto cognitivegrowthandperformance,wouldemphasizetheexistenceofanexternalmediatormore thanwouldPiaget.TheVygotskianpositionwouldbeinclinedtostressthepotentiallyfavorable effectsofbilingualism,butonlyunderconditionswheretheavailabilityofmultiplemediatorsis notconfusing(e.g.,wherethemediatorsmapthesamesemanticandconceptualcontrasts). Indeed,someofthecrossculturalworkofLuriaandVygotskyamongGeorgianpeasantsmight suggestthatthepresenceofmorethanonecodingsystemforagivenconceptcouldgreatly advancetheabstractnessofconceptualorganization(Luria,1971). Totheextentthatmediationisimportant,havingmultiplemediatorsmay 172

allowconceptstobedetachedfromsinglemediatorsand,hence,allowforadistinctiontobemade betweenwordandthing(seealsoWerner&Kaplan,1984,forsimilararguments). Thediscussionofissuesofbilingualismintermsoftheavailabilityofmultipleconcepttomediator matchessuggestsanotherangleofapproachtotheissuesinvolvedinbilingualism.Thereisabody ofworkthatviewstheimportanceofmultiplecodingofinformationasisimplicated,forexample, intheabilitytowriteandread.Writingandreadinginvolveabilingualismlikemultiplecoding situation.Linguisticinformationexistsinbothanauralandawrittencode.Argumentshavebeen made,andsomeevidenceadducedforthecognitiveadvancesthatfollowonhavingthismultiple codingavailable(Luria,1982). However,someoftherecentworkofScribnerandCole(1981)onliteracyinAfricasuggeststhat theissuemustbeconsideredintermsofspecificmappingbetweenspeakingwritingandcognitive performance.Thisrecentworkhasarguedforaspecificityofanalysisthattakesthequestionwell beyondissuesofwhetheritisbetter(ornot)tobeabletoreadandwrite.Themorebasicquestionis therelationshipbetweenataskdemandandtheimpliedability,andaswell,onthemannerinwhich literacyhasbeenachieved.Therearesomeformsofliteracywherereadingisnottreatedasa transparentcognitivetoolforreadinganything.Ratheritisabranchofsacredactivityaccessing scripturaltruths(forexample,asmightbecontainedintheKoranoranyothersacredtext). Insum,fromthepointofviewofmanydominantdevelopmentaltheories(whicharenotnecessarily right,onlydominant)thereseemstobenooverwhelmingindictmentofbilingualism.Ifthefield wascalledontogiveitscollectivewisdomonthismatter,itwouldprobablycomeforthwitha resounding"can'thurt,maybeitcouldhelp." However,mostdevelopmentaltheoristswouldask,ataminimum,thattheresearchquestionsposed ininvestigatingbilingualismmustbeposedwithgreatspecificity.Theissuewouldnotbethebig issueofwhetherhavingmorethanonelanguageisgoodorbad.Rather,thequestionwouldcome downtoananalysisofthespecificsoflanguagestructureandlanguageuse. Wewouldstudytheissuesinhighlyspecificdomainswithpreciselydiagnosticteststoexamine specificpredictionsthatmightbederivedfromareasoflanguagematchandmismatch.And,aswith thediscussionofthetheoreticalsimilaritiesofbilingualismtoreadingandwriting(multiple mediation),thedevelopmentalpsychologistwouldlikelywanttoextendthestudyofbilingualism beyonditscurrentboundaries. Howeverhearteningthistourofdevelopmentaltheoriesmightbetoadvocatesofbilingual education,itislargelyirrelevant.IntheremainderofthissectionItrytocarveoutthedimensionsof theirrelevanceofdevelopmentaltheoriestotheissuesinvolvedinbilingualismandshowother directionsforwhereimportantstudiesareneeded. 173

THESOCIALNATUREOFBILINGUALISM Letmestartwithasimplepopularculturalobservation.Noonetalksabout"bilingualism"when talkingabouttheupperclassesofsociety.Bilingualismseemstobeapoorman's(woman's, child's)affliction.Nooneseemstohavebeenworryingaboutconfusingeverythingwhenitwas deemeddesirabletohave"foreignlanguagerequirements"anenforceablepartofthehighschool curriculum.Indeed,ascollegeshavereturnedtoamorebasic"corecurriculum"sotoohave foreignlanguagescometobeseenasanessentialpartofaroundedandbasiceducation.And, indeed,thereisnotarightthinkinggraduatestudentwhohasnotrebelledattheideaofhavingto passalanguageexaminationinordertogetaPh.Dinsomenonlinguisticandnonforeignfield. Fortheadultornearadulthavingproficiencyinmorethanonelanguageisasignofstatus:being "international,"beingnonparochial,beingapartofthewiderworld.Theproblemof bilingualism,whenthoughtaboutinanadultcontext,isaproblemonlywherethereisan unassimilatedorculturallyseparateethnicgrouplivingwithinadominantculture.Theproblem ofadultbilingualismisclearlyapoliticalproblem.Whatisphrasedasalinguisticphenomenon is,infact,apoliticalphenomenon(Homel&Palij,thisvolume). Ihavebeendoingpoliticalresearchforthelastseveralyears.Issuesofbilingual,bicultural policiesinevitablyarisewithinthepoliticalcontextandtheyclearlysignalamassiveresentment ofculturallydifferentgroupsbytheculturalmajority.Wehaveallheardit."Whydotheywantto betaughtintheirownlanguage...whenmyparentscameheretheylearnedtoassimilateand becomeAmericans.""Theywanttostaydifferent...thatreallymeansthattheydon'twanttobe apartofus.""Whyshouldmytaxdollarspayforagroupofpeoplethatdonotwanttobecome American." Inshort,noneofthediscussionsofbilingualisminthepopularcultureorthepoliticalcontext haveanythingtodowithdeepissuesofcognition.Theydealwithissuesofculturalassimilation andmulticulturalacceptance.Thebilingualismdebateisadebateaboutculturalbelongingand thenonunderstandablerejectionofthemainstreamculturebypeoplewhoinsultusbynot becominglikeus.Formany,thinkingaboutbilingualismisreallythinkingaboutwhethera differentpeopleiswillingtopayitsduesbylearningtobecome"likeus." Thissituationmakesthediscussionofbilingualismincognitivetermsextremelydifficult.Weare dealingwithavariablethathasanenormouslypowerfulsocialculturalmeaning.Theloadedness oftheissueoflinguisticassimilationisofclearlydominantpowerwhenbilingualismisthetopic. Itappearstomethattheculturalsignificanceofbilingualismandbiculturalisminpoliticaland socialtermsisofoverwhelmingimportanceinassessingtheresearchliteratureonbilingualism. Theissueisunlikelytobesimplycognitive. 174

OnthispointItakesomesmallexceptionstotheanalysisofthesituationaspresentedbyPalijand Homel(thisvolume).Whiletheyacknowledgetheculturaldimensionandthedifficultiesofdoinga "trueexperiment"theyseemtofocusmoreonsuchissuesaswhetherthelanguagesareofequal strengthwhenchildrenareassessed,andwhetherchildrenwhoarebilingualdobetterorworseon IQandotherachievementtests.Tobesure,thesearecriticallyimportantissueswhenbilingualismis treatedasanexperimentalorquasiexperimentalvariableandthedesireistoadduceevidencein supportoforagainstbilingualeducationprogramsbymeasuringthisvariableanditsimpact correctly. Mypointisthatbilingualismispreciselynotanexperimentalvariable,notbecauseitcannotbe randomlyassignedtopopulationsandthelike.Itcannotbeavariableevenincovariatedesigns becausebilingualismisacodedexpressionofamultiplicityofexperimentalvariablesthatare packagedtogetherinsuchawaythattheisolationandprecisemeasurementofavariableis impossible. Imightaddthatdependentvariables,suchasIQorachievement,arepartofthesamepackage.IQis measuredbymeansofsomegroup'sideasofwhatintelligenceisallabout.Achievementisaneven moreinstitutionalizedversionofwhatprogressmeans.Ofcourse,asubgrouplivingwithinamain culturalgroupwill,inreality,beheldaccountabletotheexpectationsofthemajority.Surely, however,thisisnotacognitivedevelopmentalissue.Itisinfactasocialissuehavingmoretodo withaccountabilitypracticeswithinasocietythanwithdevelopmentorintelligence. ItisquitelikelythatthepositiveresultsofLambertandhiscoworkersinQuebechaveasmuchtodo withtheofficiallybilingualpolicyofthegovernmentoftheProvinceofQuebecastheyhavetodo withtheexperimentalnicetythatthelanguageproficiencyvariablewasequatedpriortotesting. Similarly,itisnotsurprisingthatDeAvila'sprogramrequiredthepresenceofamassivesupport systemforbilingualbiculturalSpanishEnglishbilingualsinCalifornia.IntheUnitedStatesthe continuedmaintenanceofSpanishspeakingcommunitiesisnotculturallyaccepted.Tosupport SpanishEnglishbilingualismprogramsistoswimupstreaminthefaceofformidableobstaclesofa socialculturalvariety.Indeed,peopleneedagreatdealofsupporttoswimupstream. Thereistoogreatatendencyforsocialscientiststosimplyextendandassimilatethevaluesofsocial sciencewhendealingwithsocialproblems.Thelaboratoryizingofsocialproblemsisof questionablevalueevenwhenappliedinoptimumconditions(thelaboratory).Asweextendsocial sciencethinkingtosocialdomains,theextensionoftheisolationofvariablesandassessmentof theirisolatedimpactscomestobeofevenmorequestionablevalue.Tobesure,knowledgeof preciselymeasuredandcontrolledvariablesisimportant.Equally,thiswayofthinkingintroduces highlyartificialframeworksfordiscussion. Letmecyclebackonthispointandtrytodevelopitabitmore.Letussaythatitisthecasethatthe cognitiveperformanceofbilingualsismuchbetterwhenthe 175

languagesareofclosetoequalstrength.Nowletusconsiderhowthismightbeachievedinthereal world.Inordertoachievetheconditionofequalstrengthitseemsreasonabletoassumethatthe languagesmustbegivenequalstatus.Theequalstatusargumentappliesnomatterwhatversionof languageacquisitiontheorywemightbeusing. Foranativistwhoequipsthechildwithavarietyofpretunedlanguageorconceptacquisition devicesthereisstillthenecessityforlanguagespecificinputssothatuniversalrulesmaybe extractedfromeachlanguagethatisbeingacquired.Thisrequiressomeroughequationofamount ofcontactwiththelanguagestobeacquired.Ifaculturegivesroughlyequalstatustotwo languages,itstandstoreasonthatcontactwiththelanguageswouldberoughlyequatedandthat thereforethevariousacquisitiondevicescanbeappliedtoyieldan"equation"onthestrengthof languagevariable. Forthemorelearningtheoryoriented,thesameargumentsapply,exceptnowtheyareappliedtothe equalityofstressonteachingofthevariouslanguages.Thisagainbearsonissuesofthecultural statusaccordedtothevariouslanguagesinvolved.Alowstatuslanguagemightnotreceivethesame teachingstressasthehighstatuslanguage. Thereareevensomevariantsofthelearningtheoryviewthatwillstress,withthenativists,the importanceofmerecontactasacontrollingvariableindeterminingthestrengthanddepthof languageacquisition.Forexample,Reberandhisassociates(see,forexample,Reber&Glick,1979, forasummary)havebeenstudyingpeople'sabilitytoextractdeepstructurerulesfromartificial languages.Theyhavefoundthatpeoplearemorelikelytoextractrulesunder"incidentallearning" conditions(wheretheyaremerelyexposedtoinput)thantheyareunder"intentionalconditions" wherespecificrulesaresought(ortaught). Inrecentyearsbothnativistsandlearningtheoristsarebeginningtoagreethatlanguageacquisition occurswithgreatercompletenessearlierthanwehadbeforeimagined.Aswefindnewwaysof testingchildren'slinguisticabilities,thealmostinevitablefindingisthattheimportantevidencesof linguisticstructurearevisibleatearlierandearlierages.Ifwetakethistrendseriouslythenit becomeslessandlesstenabletoargueformonolinguistictrainingforalongperiodsothata languagecanbeestablishedbeforeanotheroneisadded.Itseemsmoreandmorereasonableto assumethatlanguageacquisitionbythetimethatchildrenhaveachievedschoolageissufficiently advancedsothatwecannotimaginethatthechildisdeprivedoflearninganylanguagewellby beingforcedtolearntoomanylanguages. Theseargumentswouldsuggestthattherealissuesinbilingualismareneitherdevelopmentalnor cognitive.Rather,themainissueisanargumentabouttherespectablestatusofasubcultural language. Ratherthanbeinganeatacademicpoint,thisisapointthathasprofoundimplicationsforsocial policy.Ifwethinkonlyoflanguagestrengthasbeing 176

equalornot,wemaybelievethatallweneedtodoistrainthelanguagesuptopositionsofnear equalityandoureducationalproblemsaresolved. Buttheproblemsaredeeper.Ibelievethatmuchofwhatisunderlyingthemeasuredchangesin "cognition,"whichseemstobethetopicoftheresearchliterature,isreallyanissueof "performance"underfriendlyorunfriendlyconditions. Ithasbecomealmostatruismofdevelopmentalresearchthatitmostoftenturnsouttobethecase thatmeasuredabilitiesarehighlyimpactedonbybothvariationsintaskdemandsandbythelevelof communicationthatisestablishedbetweenchildandexperimenter.Thelevelofcommunicationwill indeedbedeterminedbylinguisticvariables,butitisalsodeterminedbysocialculturalvariables thatservetoenableordisableeffortsatestablishingsharedmeaningsbetweenchildand experimenter. Incaseswheresharedmeaningsareestablished,measuredperformanceisinevitablybetter.We deludeourselvesbythinkingthatthismerelyreflectscognitiveadvancesorcognitivedeficitsonthe partofthechild.Thepsychologicaltestandtheexperimentalsituationisatransactionbetweentwo partieswhoeitherestablishgroundsforcommunicationordonot.Iftheydonot,performance suffers.ThishasbeenclassicallyexemplifiedbytheworkofZiglerandButterfield(1968)who havebeenabletoproduceinstant10pointgainsinmeasuredIQofculturally"disadvantaged" populationssimplybyvaryingthelevelofcommunicationbetweentesterandchild. Iftheserathergeneraland,bynow,banalpointsareaccepted,thereisaclearandnonbanal conclusion.Researchparadigmsthatfocusonthechild'sperformanceorlackofitmisstheboatin importantways.Ifitisacceptedthatsocialconditionsofcommunicationareofgreatimportancein allowingchildrentodisplaytheircompetencies,thenweprobablyshouldnotfocusonthechildren andtheirperformanceinordertoaddresssocialissues.Theirperformanceisvariabledependingon thewaythattheyaretreated.Thewaythechildistreatedintheschoolandintheexperimental situationitselfdependsonmoregeneralculturalattitudesofteachersandexperimenters. Similarly,ifweacceptthepointthattheverynotionofachievementisculturallydefined,wemay wanttorefocusfromissuesofchildrenandtheirachievementtoissuesoftheculturaldefinitionof achievement.Isthereonestandardornot? Somedevelopmentaltheoristswouldargueforasinglestandard,otherswoulddisagree.Mypoint hereisthatthissortofissueshouldbethefocusofourconcern.Itisthesortofissuethatismore aboutproblemswithinintellectualdisciplinesthanaboutchildrenandtheirabilitiesorlackofthem. Thereissomesensetotheargumentthatfocusingonanissuesuchasbilingualism,andmeasuring itaswedointermsoftheperformancesofchildrenonvariouscognitiveandeducationalindicators, shiftsthedebatefromwhereit 177

shouldbelongonthestructureofsocietyanditspoliciestowardamulticulturalpopulation tothevictimsofthesituationtheminoritygroupsandtheirlinguisticorbilinguisticabilities. Inthiswayweareindangerofblindingourselvestorealproblemsinneedofprofoundsolutions. Thereisadangerinfocusingon"abilities"ratherthanpoliticalandcultural"rights." Thedebateoverbilingualismisanimportantsocialdebate.Becauseitissoimportantwemust clearlythinkthroughtheintendedandunintendedconsequencesofjoiningthatdebatewithour focusonthecognitivestatusofthebilingual. Thedebateoversocialpolicywillnotbewonbypointingtowardlinguisticorcognitivedata.The morefundamentalissueshavetodowithsocialattitudesandsocialpoliciesthemselves. TotheextentthatbilingualeducationservestolegitimizeasubculturalgroupthenIamallforit. Sotoo,probably,arecognitivedevelopmentaltheories. However,asIhaveseeninmypoliticalresearch,bilingualandbiculturalprogramsarecreatinga backlash.Dominantculturemembershavebecomeresentful.Theyarenotswayedbyour prettiestandmostpristinedata.Theyare,instead,angryatpeopleswhowishtomaintaintheir identity,andthepubliccostsinvolved. Inthecurrentpoliticalclimatethereisahardeningofclassandculturaldivisions.Thepolitical idealofameltingpotAmericaisdirectlychallengedbyculturalseparateness.Asthenon militarysectorofthefederalbudgetdecreases,wearebecominglessandlesspronetoinvest scarcedollarsinmaintainingsubculturalidentityandsecuringsubculturaladvancementno matterhowcompellingthedatamightbe. Ifearthattheresearchdataanddevelopmentaltheorymayalignonthesideofbilingual biculturaleducationbutthatthesocialstructuralissuesperhapsalignonthesideofthecultural assimilationists. Itmaywellbearguedthateventhedatabroughtforwardtosupportbilingualprogramsuse evidencestructures(e.g.,focusingonkidsandtheirperformanceasthemajordatasource;using culturallydominantnotionsofintelligenceandachievementasthedependentvariables),which ultimatelyimplytheassimilationistposition. Thecauseimplicitintheorganizationofthisvolumemightindeedbebetterservedbya refocusingofattentiononissuesinvolvedinbuildingamulticulturalsociety.Wemightturnour attentiontotheissuesofstandardsthatareusedtoevaluatesubculturalkidsinourschoolsand ourexperiments,andexamineifthesestandardsarecognitivedevelopmentalinsomeuniversal wayorarethemselvesreflectdominantculturevalues"pretendingtobe"universal.Wemight evenshiftfocusfromissuesof"meetingstandards"toissuesofcivilrights. Weinvitedifficultandsoulwearyingquestionsbythisexamination,butitisonethatmust,at somepoint,bedone.Ifearthattheattempttoarguetheissues 178

out,asiftheywereissuesrelatingsolelytowhetherkidsaresmarterordumberwithonelanguage ortwo,leavessomeofthemajorunderlyingissuesuntouched. REFERENCES Conrad,R.(1964).Accousticconfusionsinimmediatememory.BritishJournalofPsychology.55, 7584. Luria,A.R.(1971).Towardtheproblemofthehistoricalnatureofpsychologicalprocesses. InternationalJournalofPsychology,6,259272. Luria,A.R.(1982).Languageandcognition.Ed.,J.V.Wertsch.NewYork:Wiley. Piaget,J.(1967).Sixpsychologicalstudies.NewYork:Vintage. Reber,A.,&Glick,J.(1979,June).Implicitlearningandstagetheory.PaperreadattheISSBD conference,Lund,Sweden. Scribner,S.,&Cole,M.(1981).Theconsequencesofliteracy.Cambridge,MA:HarvardUniversity Press. Siegler,R.S.(1983).Informationprocessingapproachestodevelopment.Handbookof developmentalpsychology.NewYork:Wiley. Vygotsky,L.S.(1962).Thoughtandlanguage.Cambridge,MA:MITPress. Vygotsky,L.S.(1978).Mindinsociety:Thedevelopmentofhigherpsychologicalprocesses.Ed., M.Cole,V.JohnSteiner,S.Scribner,&E.Souberman.Cambridge,MA.:HarvardUniversityPress. Werner,H.,&Kaplan,B.(1984).Symbolformation:Anorganismicdevelopmentalapproachtothe psychologyoflanguage.Hillsdale,NJ:LawrenceErlbaumAssociates. Zigler,E.,&Butterfield,E.C.(1968).MotivationalaspectsofchangesinIQtestperformanceof culturallydeprivednurseryschoolchildren.ChildDevelopment,39,114. 179

[Thispageintentionallyleftblank.] 180

IV BILINGUALISMANDSOCIALDEVELOPMENT 181 [Thispageintentionallyleftblank.] 182

10 SocialPsychologicalBarrierstoEffectiveChildhoodBilingualism DonaldM.TaylorMcGillUniversity Thestudyofbilingualism,ormoregenerallysecondlanguagelearning,hascomealongway.Thata numberofseriouscontroversiesovertheimplicationsofbilingualismdominatetheliteraturedoes notbeliethisassertion,formuchpioneeringgroundworkisneededbeforedifferentpointsofview surfaceandcrystalizeintoimportant,recognizableissues.MyownconcerniswithwhatIbelieveto beanunderstudiedaspectofsecondlanguagelearningandeffectivebilingualism,thatofsocial variablesingeneral,andtheintergroupcontextwithinwhichlanguagelearningtakesplacein particular. Thisfocusofconcernisviewedasanaturaloutgrowthoftheprogressionintheunderstandingof secondlanguageprocesses.Theinitialinterestwasonabilitiesandaptitudes(Carroll,1958,1974; Carroll&Sapon,1959)involvingsuchvariablesasphoneticcoding,grammaticalsensitivity, memory,inductivelearning,and,ofcourse,theroleoffirstlanguageproficiencyinsecond languagelearning. Attentionthenshiftedtoattitudinal/motivationalvariablespromptedbytheinitialandcontinuing workofLambert,Gardner,andtheirassociates(Gardner,1981;Gardner&Lambert,1972). Includedinthisconstellationofvariablesareattitudestowardthegroupspeakingthesecond language,towardthelanguageitself,andtowardtheinstructionalprocess.Themotivational componentincludestheimplicationsofinstrumental,asopposedtointegrative,reasonsforlearning asecondlanguage.Concomitantwiththedevelopmentofinterestinthesevariablescamequestions aboutthecognitiveandpersonalityimplicationsofbilingualism. Itwouldseemanaturalextentiontofocusattentiononthelargersocialcontextofintergroup relationsandappreciatetheroleitmayplayforchildren 183

enrolledin,andcopingwith,variousformsofbilingualprograms.Theattitudesandmotivations ofthesecondlanguagelearnerhave,todate,tendedtobeviewedasindividuallybased orientationswhichaffectthatparticularindividual'ssuccessatlearningthetargetlanguage.The concernhereiswiththesetofsharedperspectivesthatariseinagroupwhenitfindsitselfina particularpowerrelationshipandhistoryofconflictoraccommodationwithothergroups. Specifically,itmustberecognizedthatone'slanguageplaysacentralroleforselfdefinitionin termsofgroupidentity.Assuch,thegrouponeisidentifiedwithandrelationsbetweenone's owngroupandothergroupsinthesocialenvironmentprovidethecrucialcontextinwhich learningalanguageoperates.Andtheintergroupcontextdoeshaveadirectanddramaticimpact ontheclassroomenvironment.Itisnocoincidence,forexample,thatAnglophonechildrenin QuebecareexposedtoFrenchinstructionasearlyaskindergarten,whereasthereisanexplicit regulationagainsttheteachingofEnglishinFrenchschoolsuntilGrade4,andeventhenitisnot required.Suchnonreciprocalpracticesmayseemirrationaluntilaccountistakenofthe intergroupcontextofFrancophonesandAnglophonesinCanada. Thewholequestionofbilingualismraisesfundamentalquestionsaboutitsimplications.Do childrenlearnthesecondlanguagesuccessfully?Isthehomelanguagereinforcedandkept vibrant?Doesusingthemothertongueasthelanguageofinstructionfacilitatethetransitioninto learningthroughthedominantlanguage?Aretherehiddencoststobilingualismintermsof personal,social,andintellectualdevelopment?Itisclearthattheanswertoeachofthese questionsis"Itdepends."Underthescrutinyofobjectiveanalysis,certainbilingualprograms seemhighlysuccessful,othersmuchlessso.Thepresentperspectivearguesthatcertainkey featuresoftheintergroupcontextinacommunitymayserveasbarrierstoeffectivechildhood bilingualismandthatourawarenessofthesekeyfeaturesmayallowforprogramadjustmentsin ordertomaximizetheexperienceforourchildren. MULTILINGUALIDEOLOGIES Childrendonotacquirelanguageinavacuum.Bytheirveryexposuretoaparticularlanguageor languagestheyarereflectingandinsomewayscreatingoralteringasocialreality.Itcomesasno surprisethatinQuebecAnglophonechildrenparticipateinprogramstolearnFrenchandtheir FrancophonecounterpartsfocusonlearningEnglish,orthatinWales,whereEnglishhascome todominate,theconcerniswithmaintainingWelsh,andthatinthesoutheastandsouthwest U.S.A.SpanishandEnglishareoftenthelanguagesatissue.Allcommunitiesthatconcern themselveswithchildhoodbilingualismhaveonethingincommon:linguisticandcultural diversity.Thereareatleasttwo,andoftenmore,languagesubcommunitiesthatmustsharethe socialenvironment. 184

Howthissharingistotakeplacebecomesthecentralissue.Everyonehasatleastavagueidea ofhowdiversegroupsinacommunityshouldrelatetooneanotheringeneral,andhowthey personallyshouldrelatetotheirowngroupandothergroupsinparticular.Notwogroups(or individualsforthatmatter)willhavethesameidea,ideology,oraiminmind.Theorientation toculturaldiversityheldbyeducators,planners,teachers,andparentsofthegroupsdirectly andindirectlyaffectedbyabilingualprogramwillhaveaprofoundimpactonthefinal producttowhichthechildisexposed.Berry(1975,1983),buildingonearliersociological notions,hasidentifiedfourorientationspeoplemayadopttocopewithdiversity:assimilation, integration,separation,anddeculturation. Theassimilationviewisthatdiversityisharmful,creatingandmaintaininggroupbarriers, fragmentingsocietyattheveryleast,ultimatelysettingupconditionsforprejudiceand discrimination.Thesolutionistobelievethatdiversityingeneral,anddifferentlanguagesin particular,mustbeeradicatedandreplacedwithhomogeneousculturalandlinguisticstandardsand use.Ofcourse,theimpliedassumptionisthatlesspowerfulgroupswillaccommodatetothe languagenormsofthedominantgroup.Thetraditional"meltingpot"descriptionoftheUnited Statesashavingasingular"American"cultureandlanguageisconsistentwithanassimilation orientation. Thisassimilationmodelcanbecontrastedwithoneemphasizingintegration.Culturalgroupswould beencouragedtoretaintheirheritagelanguageandtherebytheiridentity.Thisculturalstatusand securitywouldprovidetheenvironmentnecessaryforcooperativeintergrouprelations.Canada's officialmulticulturalpolicyrepresentssuchanorientation. Aseparationideologyalsoemphasizestheneedtoretainlanguagebutbelievesallinteraction betweenone'sowngroupandothergroupsshouldbesuspended.ExtremesoftheBlacknationalist orQuebecseparationmovementareprimeexamplesofsuchanideology. Finally,adeculturationideologymayprevail.Berry(1983)andothersdescribethisasonewhere thepersonfeelsalackofidentitywithanygroup,asmightarisewhenanimmigrantattemptsto shedalltracesofhisorherheritagelanguageandculture.Theresultmaybethelossofthisidentity withnorealfeelingofidentityforthehostculturetoreplaceit;clearlyanundesirablecondition involvingfeelingsofanomieandmarginality.Adifferent,moreoptimisticformofdeculturation mayarisewhenapersonemphasizesindividualcharacteristicsforhumanrelations,feelingthat interactionwithothersonthebasisoftheirculturalidentitystripsthemoftheirindividuality. Thesefourbroadideologies,andthevariousalternativeswithineachcategory,setthestageforthe intergroupcontextwithinwhichlanguageisacquired.Thefourideologiesaboutdiversityare distinguishedonthebasisoftwofundamentalattitudes;first,thedesiretomaintainornotmaintain one'scultureandlanguageand,second,thedesireforpositiveornegativerelationswithother groups.Answerstothesetwofundamentalattitudescombinetoproducethefourideologies. 185

TABLE10.1 ConceptualAnalysisofIdeology Ownlanguageandcultureshouldbemaintained? __________________________________________________________ YES NO __________________________________________________________ YES Integration Assimilation ___________________________________________________________

Positiverelations withothergroups?

NO Separation Deculturation Note:AdaptedfromBerry,Kalin,&Taylor,1977. Thepredominantideologyinaparticularethnolinguisticcommunitywillhaveaprofoundeffect onhowbilingualismisviewedbythatcommunity.Adetailedexaminationofthedeculturation, separation,assimilation,andintegrationalternativesdepictedinTable10.1providessome appreciationforthepotentialbarrierstoeffectivebilingualism.Eachofthefourideologies presentsitsownuniqueproblemsforeffectivebilingualism;however,oneofthepredominant themesthatwillemergeistheimportantroleplayedbythreatstoethnicidentity. DECULTURATIONIDEOLOGY Thisorientation,withitsemphasisonthelackofidentitywithanyculturalorlinguistic collectivity,wouldatfirstseemtobeirrelevanttothepresentdiscussion.Uponcloser examination,however,anunderstandingoftheparticulardeculturationideologythatisprevalent canshedlightonthecontextofbilingualprograms.Earlier,twoformsofdeculturationwere noted.Inthefirstcasepeoplefeelalienatedfromallpotentialculturalidentities.Inthis circumstanceabilingualprogramthatoffersformalrecognitionofaperson'sheritagethrough offeringinstructiontochildreninthemothertonguemaybeanimportantpositiveforcein providingthebasisforasecureidentity.Itmayhelpalleviatethefeelingsofmarginalityand alienationassociatedwiththisformofdeculturation. Thesecondformofdeculturationariseswhenpeoplefeelthatitisinappropriatetointeractwith anotheronthebasisoftheirculturalgrouping,preferring,instead,tofocusontheirindividual characteristics.Suchanorientationignoresreality.Ifone'sethnicgroupistobeignored,thenthe nationalculturalnormistakenastheoperatingstandard.Foreveninteractionsonanindividual levelrequireasharedlanguage,setofassumptions,andnormstoallowforcommunicationtobe meaningful.So,bydefault,anindividualisticallybaseddeculturationideologyultimately becomesoneofassimilation,anideologywhoseimplicationsforbilingualismisaddressedlater. 186

SEPARATIONIDEOLOGY Widespreadbilingualismwillberejectedwithaseparationistideology.Anyovertureatlearning thelanguageoftheothergroupwouldindicateanaccommodatingattitudetotallyinconsistent withtheseparationistideology.Itispreciselythisideologythathelpsexplaincurrentbilingual effortsinQuebec.ThemajorityofEnglishspeakingchildrenattendanimmersionformof bilingualprogramandthesuccessofthisprogramisvisibletoeveryone.Andyet Frenchspeakingchildren,whomightcertainlybenefitfromaknowledgeofEnglishintheNorth Americancontext,areexposedtoverylittleopportunitytolearnEnglishinschool.Thisapparent irrationalitybecomescomprehensiblewhenaccountistakenoftheprevalentideologyinthe Frenchspeakingcommunity:onethatisseparationistnotnecessarilyinapoliticalsense,butin thesenseofadisengagementfromdirectcompetitionwiththehistoricallyprivilegedEnglish community. Thespecificbarriertobilingualisminthiscaseisethnicidentity,adeeplyfeltpersonalneedthat ispotentiallythreatenedwhenthereisinteractionwithmembersofanothergroup.Ithasbeen difficulttoappreciatehowthelearningofanotherlanguagemightbethreatening.Theabilityto speakasecondlanguagehasusuallybeenjudgeddesirable.Debatesaboutthementalconfusion associatedwithbilingualismnotwithstanding,theoverwhelmingscientificconclusion(Ben Zeev,1977;Genesee,1981;Peal&Lambert,1962)andthepopularbeliefhasbeenthatknowing anotherlanguageisanachievement.WhetheritislearningLatinatschool,ortheadmirationwe feelforthosewhoarebiormultilingual,theacquisitionofanotherlanguageisseenasadditive asaddingaskilltotheperson'sexistingrepertoire.Untilrecentlythisnotionhasgone unchallengedbecausemostbilingualprogramshavebeeninitiatedbyfamiliesrepresentingthe dominantgroup,WhiteAmericansorAngloCanadians,groupswhoseethnicidentityissecure. Whentheimplicationsforminoritygroupsareconsidered,however,thereariseimportant instancesofwhatLambert(1974)haslabeled"subtractive"bilingualism.Thatis,asituation whereanypositiveconsequencesoflearningasecondlanguagearefaroutweighedbythe negativeconsequencesinother,moreimportantdomainsoftheperson'slife.Theparticularloss inquestionistoone'sethnicor,morebroadly,groupidentity.Iflearninginthesecondlanguage contributestothedemiseinknowledgeanduseoftheheritagelanguage,theresultscanbe devastating. Itisimportanttonotethatthekeyissuethatdistinguishesanadditivefromasubtractive languageexperienceispowerintheintergroupcontext.Membersofapowerfulgroupwhoare secureintheirpositionwillexperiencenonegativeconsequencestolearningthelanguageof anothergroup.Itiswhere,forwhateverreason,agrouphaslesspowerthatethnicidentityand groupsurvivalbecomeanissue. 187

Theformalideathatfearforone'sethnicidentityisperhapsabarriertosecondlanguage learningwasadvancedinitiallybyTaylorandSimard(1975).Asearlyas1963,however( Lambert,Gardner,Barik,&Tunstall,1963)therewasevidencetosuggestthataspeople begintomasterasecondlanguage,theymaydevelopfeelingsofethnicanomie.Ithasalso beennoted,inthecontextofQuebec,thatFrenchCanadianparentswhosechildrenare educatedinEnglishlanguageschoolsshowpatternsofethnicidentitydifferentfromthose parentswhosechildrenattendthenormalFrenchschools(FrasureSmith,Lambert,& Taylor,1975).FrenchCanadianparentswhosechildrenattendaFrenchschooldisplayclose allegiancetoamonolingualFrenchspeakingreferencegroup.FrenchCanadianparentsof childrenintheEnglishschoolappeardividedintheirethnicidentitybetweenmonolingual FrenchCanadiansontheonehand,andbilingualFrenchandEnglishCanadiansontheother. SuchashiftinidentityassociatedwithexposuretoEnglishcouldhaveseriousnegative consequencesforagroupwhichfeelsthatsurvivaloftheFrenchlanguageinNorthAmericais threatened.Specifically,itispossiblethatwithfurtherexposuretotheEnglishlanguagethe Frenchlanguageandculturemighteventuallybeeradicated. DirectevidenceforthemotivationalimportanceofethnicidentitycomesfromastudybyTaylor, Meynard,andRheault(1977).Twoofthemorewellestablished"additive"motivationsassociated withlearningasecondlanguageareinstrumentalandintegrative.Instrumentallymotivatedpersons learnasecondlanguagemainlyforitspracticalvalue,asinqualifyingforabetterjob;anintegrative motivereflectsapersonalinterestinthepeopleandtheircultureandinvolveslearningasecond languageinordertolearnmoreaboutanothergroup,meetmoreandvariedpeoplewithinthatgroup andtobeabletothinkandbehavelikeitsmembers.Althoughbothmotivationalorientationshave provedessentialforsuccess,anintegrativemotivationhasbeenfoundtobetheorientationthatbest sustainsthelongtermdedicationneededforsecondlanguagelearning.ThepurposeoftheTayloret al.(1977)studywastoestablishwhetherthreattoethnicidentityformsaseparatemotivational cluster,distinctfromthealreadyestablishedinstrumentalandintegrativemotives,andwhethersuch amotivationalclustermightposeabarriertosecondlanguagelearning.Tothisendasampleof246 Francophonejuniorcollegestudentswereaskedtomakeratingson11motivationquestions.The ratingswerethenfactoranalyzedinordertodeterminethestructureofmotivationforlearninga secondlanguage.TheresultsofthisanalysisarepresentedinTable10.2,whichdescribesthe particularitemsaswellasthemotivationstructure. FromTable10.2itisclearthatthreattoethnicidentity(FactorIII)isamotivationrelatedto bilingualism,whichisconceptuallydistinctfromthetraditionalinstrumental(FactorI)and integration(FactorII)motivations.Havingestablishedthreattoethnicidentityasadistinct motivationalcluster,thecrucialquestionremains:Howdoesitaffectsecondlanguagelearning?In ordertoaddressthisquestiontheFrancophonestudentsmadeasubjectiveratingoftheir 188

TABLE10.2 PartialResultsofFactorAnalysisofMotivationforSecondLanguageLearning Consequencesof Factors LearningEnglish I II III Earnmoremoney 77 02 04 Getabetterjob 82 04 12 Improvedcontroloneconomy 74 18 06 TobetterfightAmericanization 40 57 03 Opennesstowardothercultures 08 68 46 InteractmorewithEnglishpeople 08 69 41 Lossofculturalidentity 02 04 71 Increasedchanceofassimilation 05 08 63 Increasedanonymity 00 10 79 Greateffortforlittlereturn 07 15 73 Onlymaterialgainsatgreatcost 11 20 76 Note:AdaptedfromTaylor,Meynard,&Rheault,1977. abilitytofunctioninEnglish.Amultipleregressionanalysiswasthenperformedtodetermine whichfactorsbestpredictedabilityinEnglish.Thepredictorvariablesusedfortheanalysis includedthreattoidentity,integrativemotivation,instrumentalmotivation,personalcontactwith Anglophones,andcontactviathemassmedia,tonameafew.Theanalysisyieldedamultiple regressionof.59andtheonlytwovariablestomakeasignificantcontributiontotheregression were(a)personalcontactwithAnglophonesand(b)threattoethnicidentity. TheresultsrevealthatthosewithmorecontactwithEnglishspeakersarethemselvesbetter speakersofEnglish,andthosewhodonotfeeltheirownethnicidentitytobethreatenedare morecompetentspeakersofEnglish.Thatcontactisimportantcomesasnosurprise.Buttherole ofthreattoethnicidentityisstriking.Ifthisvariablehadplayedevenaminorrole,itwouldargue foritbeingconsideredcarefullyinthecontextofbilingualism;forittooverrideotherwell establishedfactorsassociatedwithbilingualismmakesitanissuethatmustnotbeavoided. Whichgroupsarelikelytoencounterthispowerfulbarriertoeffectivebilingualism?First,aswe havenoted,minoritygroupswhofeeltheirownlanguageandwayoflifethreatenedby bilingualismwouldbeexpectedtoavoidorevensabotageallbilingualprogramsinvolvingthe dominantlanguage.Suchgroupsmaynotonlyfeeldirectlythreatened,inwhichcasebilingual programsmayberejectedoutright,butalsoindirectlywherethemotivesofthedominantgroup areambiguous.Takethecaseofabilingualprogramwheremembersofaminoritygroupreceive someinstructionintheirheritagelanguage.Theideologyunderlyingtheprogrammaybe assimilationistthatis,teachingintheheritage 189

languageisviewedasaneffectivevehicleforminoritystudentstoultimatelymakeasuccessful transitionintoEnglishinstruction.Butwhatiftheminoritygroupinterpretstheideologyasone ofseparation.Thatis,minoritygroupsmaybelievethatinstructionintheheritagelanguageis beingusedbythedominantormainstreamgroupnotasawayoffacilitatingtransitionintothe mainstream,buttoostracizetheminoritygroup.Whatbetterwaythantohavethemworkina languagethatclearlymarksthemasdifferent,andatthesametimerobsthemoftheopportunity todevelopnativelikeskillsinthedominantlanguage. Aseparationistideologymayseemsomewhatextreme.However,culturalsecurity,orinthiscase insecurity,canbeaharshreality.Forexample,certaingroupsfacetherealityoflittlepotential forgrowthinnumberseitherbecauseofastringentsetofimmigrationpracticesorbecause politicalrealitiesmakeitimpossible.TheFrancophonepopulationinCanadaseeslittlechance foritselfgrowingsignificantlyinnumbersinthecontextofanEnglishdominatedNorth America.RecentlanguagelegislationwithinQuebechasledAnglophonestobelievethatthe veryexistenceofanEnglishspeakingcommunityisthreatened.Variousnativegroupsinboth CanadaandtheUnitedStatescannotlookforwardtosignificantgrowthintermsofnumbers. Thisisnottosuggestthataseparationistideologyisorwillbeadoptedbyallofthesegroupsbut onlytosuggestthattheirrelativepositionmaketheissueofculturalidentityaverysalientissue. Insummary,threatstoethnicidentitycanleadtoaseparationistideologyandundersuch circumstances,realorimaginedbilingualprogramswillencounterseriousbarriers. ASSIMILATIONIDEOLOGY Anassimilationideologyhastheappealofproducingstandardization,similarity,andconformity essentialingredientsforasharedunderstandingoftheenvironmentandtherebythepotential foreffectivecommunication.Untilrecentlythepredominantsociologicalprospective(seeGlazer &Moynihan,1963;Gordon,1964;Park,1950)wasthatculturalassimilationwouldprevail,and thatultimatelyaculturalconsensuswouldbeattainedthroughtheabsorptionofminoritygroups intothemajorityortheminglingofalltoformanew,emergenthomogeneoussociety.Ofcourse, inahomogeneouslinguisticenvironmenttheconceptofbilingualismisirrelevant.However,the costofassimilationisthefundamentalpsychologicalneedindividualshaveforaprimary, distinctivesocialidentitywithgroupswhoshareaperspectiveontheenvironmentandwho provideasenseofcontinuity,socialreality,andanchorpointsfortheindividual. Despitethisheavycost,assimilation,asablueprintforharmoniousrelations,cannotbe summarilydismissed.Thereis,afterall,awellestablishedbodyofpsychologicalliteratureto supportthenotionthatsimilarityisrelatedtoattraction(Byrne,1969).Highlightingcultural differences,especiallyonadimension 190

assalientaslanguage,wouldthenapparentlyonlyservetolessenthechancesformutual attraction.Directevidencefortheimplicationsofthesimilarity/attractionmodelinthe contextofethnolinguisticgroupsisprovidedbySimard(1981).InherstudyofFrench speakingandEnglishspeakingMontrealersformingfriendships,shefoundthatforaperson oftheothergrouptobeapotentialfriend,thatpersonhadtobemoresimilaronother dimensionsthanamemberofthesamegroup.Inshort,assimilationresultsintheeradication ofcertaindifferences,whichatleastensuresthatentiregroupsarenoteliminatedaspotential targetsfordesirableformsofhumaninteraction. Thesimilarity/attractionmodelnotwithstanding,thereisagrowingrecognitionthatatboththe individualandcollectivelevel,similaritymaynotalwaysbeassociatedwithattraction.Oneofthe fundamentalpropositionsofsocialidentitytheory,arecenttheoryofintergrouprelationsproposed byTajfelandTurner(1979),isthatgroupsstrivefordistinctivenessalongpositivelyvalued dimensions.Theimplicationoftheirpositionisthatindividualsneedaprimarysocialidentitythat isdistinctive.Whenarivalgroupbecomestoosimilar,groupdistinctivenesscanbethreatenedand actuallyleadtoheightenedconflict.Thereisalreadysomeexperimentalandfieldresearchthat supportsthiscontentioninintergroupcontexts(Brown,1978;Turner,1978).Itisalsointerestingto note,attheindividuallevel,thatdeindividuation,theultimateinsimilaritywhereallindividuals becomevirtuallyanonymous,hasbeendescribedasapotentiallynegativepsychologicalstate,one whichrobsthepersonofhisorherindividuality(Dipboye,1977).Insummary,thepsychological literaturedoesprovidesomebasisforsuggestingthatthelackofprimarysocialidentitymaybea significantcostofassimilation. Beyondthisthereisindirectevidencesupportingtheimportanceofprimarysocialidentityfrom recentresearchonmulticulturalism.Forexample,Canada'smulticulturalpolicysupportsthe maintenanceofethnicidentityforallCanadiangroups.Theassumptionunderlyingthepolicyisthat confidenceinone'sownethnicidentitycreatesaclimateoftoleranceandrespectforothers.Alarge scaleCanadawideinvestigationofmulticulturalismwasconductedbyBerry,Kalin,andTaylor (1977).Itwasfoundthatasecureculturalidentitywasassociatedwithpositiveattitudestoward immigrants,towardthepolicyofmulticulturalism,andtowardanumberofspecificCanadianethnic groups.TheseresultswerepartiallycorroboratedinastudywithamorerestrictedsamplebyTaylor, McKirnan,Christian,andLamarche(1979). Insummary,anassimilationistideologyleavesnoroomforbilingualism.Instead,theaimwouldbe toeradicateanyprogramsdesignedtopromotelinguisticdiversity.Theappealofsuchanideology isinthenotionthatcommunicationwillbefacilitatedwhereeveryoneusesthesamelinguistic system.Soopponentsofbilingualprogramsarenotmerelyprejudicedindividualswhowishto subjugateminoritygroupsmanyaresincereindividualswhobelievethatonlyavigorous assimilationpolicycanprovideethnolinguisticminoritieswitha 191

fairchanceatfullparticipationinsociety.Whatbecomeslostintheprocessistheprimaryneed forsocialidentity.Indeed,itisnoaccidentthatthedecadeoftheseventieshasbeen characterizedbyanunparalleledrevivalinethnicidentity.InCanadatheemergenceof multiculturalism,themilitancyofnativepeoples,andtheriseofFrenchCanadiannationalism areprimeexamples,andineachcaselanguagehasbeenafundamentalissue.TheUnitedStates haswitnessedadecadeofaggressivedemandsforcivilrights,theemergenceofacoherent Hispaniccommunityandageneralizedawarenessofethnolinguisticidentity.Itappearsthatthe simplisticlogicofanassimilationistideologycannotmeetmorefundamentalneeds. INTEGRATIONIDEOLOGY Itwasnecessarytodiscusstheessentialcomponentsofanintegrationistideologyinthecontext ofassimilation.Theneedindividualshaveforaprimarysocialidentityisrecognizedinan integrationistideologyandyettheaimisgoodrelationswithothergroups.Suchanideologyis, ofcourse,entirelyconsistentwiththeconceptofbilingualismanditisinpreciselysuchan ideologicalcontextthatwewouldexpectsuchprogramstoflourish. Atthisstage,then,weshiftattentionfromtheideologiesthatserveasbarrierstoeffective bilingualism,tothesocialpsychologicalconditionsthatcanmaximizetheexperienceoflearning anotherlanguage.Althoughitwouldseemthatanintegrationistorientationwouldprovidethe climatemostconductivetobilingualism,achievingthedelicatebalancebetweenthemaintenance ofgoodintergrouprelationswhileretainingprimarysocialidentitythroughlanguageis problematic.Ifeachgroupinthecommunityguardsitsowncultureandlanguage,anysenseofa largercommunityislosttotheproliferationofseparatesubcommunitieseveniftherearegood relationsbetweengroups. Howcanthe"communality"featureofassimilationbeachievedwhileretainingthe"primary socialidentity"componentofintegration?Perhapstheanswerliesinthefactthatwhen ethnolinguisticgroupsoccupythesamesocialenvironment,itisreasonabletoassumethatthey shareoverridingvaluesandgoals(seeLambert&Taylor,1983;Taylor,1981).Bysharingthe sameenvironment,groupscometoshareimportantattributesderivedfromcommonpolitical, social,educational,andreligiousinstitutions.Tociteoneexample,Isajiw(1977)hasdiscussed thenotionof"technologicalculture,"arguingthatsuchacultureinvolvesasetofvaluesthatis sharedbyvirtuallyallethnicgroupscomprisingtheCanadianmosaic.Thevariousethnic identitiesofgroupsmakingupthelargercommunity,then,representlayersthatbuildupona sharedsetofbroadersocialvalues.Intermsoflanguage,theneedwouldbeforsomeagreed uponsetoflinguisticnormsthatpermitinteractionamongallmembersofacommunity.Beyond thiseachgroupmightbeencouragedtomakevibrantuseofitsown 192

languagewhereitdiffersfromthesharedcode.Thosewhosenativelanguagecoincideswith thenormwouldbeencouragedtoaddalanguagetotheirrepertoire.Thiswouldpermit interactionwithakeyminoritygroupinthecommunitythroughalanguageotherthanthat designatedbythesharednorm. Thefundamentalquestiontobeaddressedinthiscontextis,Whatisthebroadlybasedshared linguisticnormandwhoshoulddefineit?Theanswertothequestionisthatusuallytheshared linguisticnormispreciselywhatthemostpowerfulgroupinthecommunitydecidesshouldbethe norm.Theproblemisthatwhatappearsasasharedlinguisticnormisreallyonethathasbeen imposeduponlesspowerfulgroups.Tobeneglectedintheprocessofarticulatingsharednormswill leavelesspowerfulgroupsonlypartiallycommittedtosuchnorms.Inorderforalinguisticnormor standardtobecometrulyshareditrequiresthatmembersofallgroupsparticipateintheprocessof definingthatnorm.Naturally,larger,morepowerfulandhistoricallyprecedentgroupswillhavea greaterimpactonthedefinitionprocess.So,Englishwillnodoubtemergeasthecommunal languageintheUnitedStates,andEnglish,French,orbothwillrepresentthesharednormin differentregionsofCanada.Nevertheless,theparticularformofthelanguageandtherangeofcode variationsacceptedasnormativemaybebroadenedthroughtheparticipationofothergroups. Communitylanguagenormsdefinedinthisfashionshouldleadallgroupstoagreatercommitment totherealizationoftheselanguagenormsandaffectthemotivationofstudentstoachieve proficiencyinthem. Beyondthesharedlanguagenorms,ethnolinguisticgroupscanbeencouragedtodefinetheirown uniquelanguageneeds,beittokeepthemothertonguevibrant,recoveralanguagenearlylost,or broadensocialcontacts.Encouragementcannotbelimitedtolipservicebutmusttaketheformof anequitabledistributionofcommunityresources,financialandotherwise.Viewedfromthis perspective,bilingualisminamultilingualcommunityisanissueforeveryone.Noonegroupis singledoutforspecialtreatmentthatcanmakethemfeelprivilegedontheonehandorstigmatized ontheother.Witheveryoneclearonthesharedlinguisticaims,programscanthenbedesignedto meetthoseaimswhileatthesametimeattemptingtomeettheuniquelinguisticneedsofeach group. Forexample,anethnolinguisticminoritymaywellrequireabilingualprogramwithadualfunction tosupporttheminoritylanguagewhileatthesametimedevelopingcompetenceintheshared languagenorms.Childrenwhosenativelanguageislargelycoincidentwiththenormmayfocuson thelanguageoftheminority,thereby"adding"alinguisticskillwhileatthesametimeprovidingthe minoritylanguagewithincreasedstatus. Thekeytosuccessfulimplementationofbilingualprogramsistheunderlyingmulticulturalideology representedinparentsandeducators,whichwillbecommunicated,howeverindirectly,tothe children.Thefourmajororientationsaretheresultofsimultaneouslyconsideringtwofundamental motivations:thedesire 193

languagewhereitdiffersfromthesharedcode.Thosewhosenativelanguagecoincideswiththe normwouldbeencouragedtoaddalanguagetotheirrepertoire.Thiswouldpermitinteraction withakeyminoritygroupinthecommunitythroughalanguageotherthanthatdesignatedbythe sharednorm. Thefundamentalquestiontobeaddressedinthiscontextis,Whatisthebroadlybasedshared linguisticnormandwhoshoulddefineit?Theanswertothequestionisthatusuallytheshared linguisticnormispreciselywhatthemostpowerfulgroupinthecommunitydecidesshouldbe thenorm.Theproblemisthatwhatappearsasasharedlinguisticnormisreallyonethathas beenimposeduponlesspowerfulgroups.Tobeneglectedintheprocessofarticulatingshared normswillleavelesspowerfulgroupsonlypartiallycommittedtosuchnorms.Inorderfora linguisticnormorstandardtobecometrulyshareditrequiresthatmembersofallgroups participateintheprocessofdefiningthatnorm.Naturally,larger,morepowerfulandhistorically precedentgroupswillhaveagreaterimpactonthedefinitionprocess.So,Englishwillnodoubt emergeasthecommunallanguageintheUnitedStates,andEnglish,French,orbothwill representthesharednormindifferentregionsofCanada.Nevertheless,theparticularformofthe languageandtherangeofcodevariationsacceptedasnormativemaybebroadenedthroughthe participationofothergroups.Communitylanguagenormsdefinedinthisfashionshouldleadall groupstoagreatercommitmenttotherealizationoftheselanguagenormsandaffectthe motivationofstudentstoachieveproficiencyinthem. Beyondthesharedlanguagenorms,ethnolinguisticgroupscanbeencouragedtodefinetheirown uniquelanguageneeds,beittokeepthemothertonguevibrant,recoveralanguagenearlylost,or broadensocialcontacts.Encouragementcannotbelimitedtolipservicebutmusttaketheform ofanequitabledistributionofcommunityresources,financialandotherwise.Viewedfromthis perspective,bilingualisminamultilingualcommunityisanissueforeveryone.Noonegroupis singledoutforspecialtreatmentthatcanmakethemfeelprivilegedontheonehandor stigmatizedontheother.Witheveryoneclearonthesharedlinguisticaims,programscanthenbe designedtomeetthoseaimswhileatthesametimeattemptingtomeettheuniquelinguistic needsofeachgroup. Forexample,anethnolinguisticminoritymaywellrequireabilingualprogramwithadual functiontosupporttheminoritylanguagewhileatthesametimedevelopingcompetenceinthe sharedlanguagenorms.Childrenwhosenativelanguageislargelycoincidentwiththenormmay focusonthelanguageoftheminority,thereby"adding"alinguisticskillwhileatthesametime providingtheminoritylanguagewithincreasedstatus. Thekeytosuccessfulimplementationofbilingualprogramsistheunderlyingmulticultural ideologyrepresentedinparentsandeducators,whichwillbecommunicated,howeverindirectly, tothechildren.Thefourmajororientationsaretheresultofsimultaneouslyconsideringtwo fundamentalmotivations:thedesire 193

tofostergoodintergrouprelationsandthedesiretoretainone'sethnicidentity.Thesignificance ofhavingaclimateofgoodintergrouprelationshaslongbeenrecognizedandneedsonlytobe underscoredinthepresentcontext.Theimportanceofcreatingormaintainingasecure ethnolinguisticidentityemergesasacentralissuethatisonlybeginningtobeappreciated. Threatstoethnicidentityaroseinthecontextofadeculturationideologyonlyindirectlybutwere centraltotheassimilation,integration,andseparatismideologies.Theprocessofidentityisas yetnotwellenoughunderstoodtoallowaconfidentappreciationofitsroleinsuccessful bilingualprograms.Itisinthisuncertaincontextthatatentativesuggestionismadeaboutthe multiculturalideology,whichoffersthebestchanceforensuringthesuccessofbilingualism. Specifically,anintegrationistorientationbuildingonasharedperspectiveinwhichallconcerned havecontributedtothatperspectivemayeliminatemanyoftheintergroupbarrierstoeffective childhoodbilingualism. ACKNOWLEDGMENT IwishtothankM.P.Walkerforcommentsmadeonanearlierdraftofthepaper. REFERENCES BenZeev,S.(1977).Theinfluenceofbilingualismoncognitivedevelopmentandcognitive strategy.ChildDevelopment,100918. Berry,J.W.(1975).Amerindianattitudestowardassimilation:Multiculturalpolicyandrealityin Canada.JournalofInstituteofSocialResearch,1,4758. Berry,J.W.(1983).Acculturation:Acomparativeanalysisofalternativeforms.InR.J.Samuda &S.L.Woods(Eds.),PerspectivesinImmigrantandMinorityEducation.(pp.6578).Lanham, Maryland.UniversityPressofAmerica. Berry,J.W.,Kalin,R.,&Taylor,D.M.(1977).MulticulturalismandethnicattitudesinCanada. Ottawa:MinisterofSupplyandServices. Brown,R.J.(1978).Dividedwefall:Ananalysisofrelationsbetweensectionsofafactory workforce.InH.Tajfel(Ed.),Differentiationbetweensocialgroups:Studiesinthesocial psychologyofintergrouprelations(pp.395429).London:AcademicPress. Byrne,D.(1969).Attitudesandattraction.InL.Berkowitz(Ed.),Advancesinexperimental socialpsychology(Vol.4)(pp.3589).NewYork:AcademicPress. Carroll,J.B.(1958).Afactoranalysisoftwoforeignlanguageaptitudebatteries.Journalof GeneralPsychology,59,319. Carroll,J.B.(1974).Aptitudeinsecondlanguagelearning.ProceedingsoftheFifthSymposium oftheCanadianAssociationofAppliedLinguistics,823. Carroll,J.B.,&Sapon,S.M.(1959).Modernlanguageaptitudetest,FormA.NewYork: PsychologicalCorporation.

Dipboye,R.L.(1977).Alternativeapproachestodeindividuation.PsychologicalBulletin,84, 10571075. FrasureSmith,N.,Lambert,W.E.,&Taylor,D.M.(1975).Choosingthelanguageof instructionforone'schildren:AQuebecstudy.JournalofCrossCulturalPsychology,6,131 155. 194

Gardner,R.C.(1981).Secondlanguagelearning.InR.C.Gardner&R.Kalin(Eds.),A Canadiansocialpsychologyofethnicrelations(pp.92113).Toronto:Methuen. Gardner,R.C.,&Lambert,W.E.(1972).Attitudesandmotivationinsecondlanguagelearning. Rowley,MA.:NewburvHouse. Genesee,F.(1981).CognitiveandSocialConsequencesofBilingualism.InR.C.Gardner&R. Kalin(Eds.),ACanadiansocialpsychologyofethnicrelations.(pp.114131).Toronto:Methuen. Glazer,N.,&Moynihan,D.P.(1963).Beyondthemeltingpot.Cambridge,MA:MITPress. Gordon,M.M.(1964).AssimilationinAmericanlife.NewYork:OxfordUniversityPress. Isajiw,W.W.(1977).Olgainwonderland:Ethnicityintechnologicalsociety.CanadianEthnic Studies,9,7785. Lambert,W.E.(1974).Cultureandlanguageasfactorsinlearningandeducation.InF.E. Aboud&R.D.Meade(Eds.)CulturalfactorsinLearningandEducation(pp.91122). Bellingham:WesternWashingtonStateCollege. Lambert,W.E.,Gardner,R.C.,Barik,H.C.,&Tunstall,K.(1963).Attitudinalandcognitive aspectsofintensivestudyofasecondlanguage.JournalofAbnormalandSocialPsychology,66, 35868. Lambert,W.E.,&Taylor,D.M.(1983).Languageintheeducationofethnicminority immigrants:Issues,problemsandmethods.InR.J.Samuda&S.L.Woods(Eds.),Perspectives inimmigrantandminorityeducation(pp.267280).Lanham,Maryland:UniversityPressof America. Park,R.E.(1950).RaceandCulture.Glencoe:FreePress. Peal,E.,&Lambert,W.E.(1962).Therelationofbilingualismtointelligence.Psychological Monographs,76,123. Simard,L.M.(1981).Crossculturalinteraction:Potentialinvisiblebarriers.JournalofSocial Psychology,113,171192. Tajfel,H.,&Turner,J.C.(1979).Anintegrativetheoryofintergroupconflict.InW.Austin&S. Worchel(Eds.),Thesocialpsychologyofintergrouprelations(pp.3348),Monterey: Brooks/Cole. Taylor,D.M.(1981).StereotypesandIntergroupRelations.InR.C.Gardner&R.Kalin(Eds.), ACandiansocialpsychologyofethnicrelations(pp.151171).Toronto:Methuen. Taylor,D.M.,Meynard,R.,&Rheault,E.(1977).Threattoethnicidentityandsecondlanguage learning.InH.Giles(Ed.),Language,ethnicityandintergrouprelations(pp.99116).London& NewYork:AcademicPress. Taylor,D.M.,McKirnan,D.J.,Christian,J.,&Lamarche,L.(1979).Culturalinsecurityand

attitudestowardmulticulturalismandethnicgroupsinCanada.CanadianEthnicStudies,16, 1930. Taylor,D.M.,&Simard,L.M.(1975).Socialinteractioninabilingualsetting.Canadian PsychologicalReview,16,230254. Turner,J.C.(1978).Socialcomparison,similarityandingroupfavouritism.InH.Tajfel(Ed.), Differentiationbetweensocialgroups:Studiesinthesocialpsychologyofintergrouprelations (pp.235250).London:AcademicPress. 195

[Thispageintentionallyleftblank.] 196

11 TheEffectsofBilingualandBiculturalExperiencesonChildren'sAttitudesandSocial Perspectives W.E.LambertMcGillUniversity Attitudesarerelatedtosecondlanguagelearningandbilingualisminvariousways.Whatlittlewe knowaboutthisrelationshipthusfarcomesfrombothcommonsenseobservationsandbehavioral research,andusuallycommonsenseandresearcharemutuallysupportive.Forinstance,itmakes goodsensetoexpectthatafavorableattitudetowardanotherethnolinguisticgroupshouldaffect positivelyone'sacquisitionofthatgroup'slanguage,andresearchfindingsgenerallysupportthis expectation(seeClment,Gardner,&Smythe,1977;Gardner,1982;Gardner&Lambert,1972; Gardner&Smythe,1975).Whatgoesbeyondcommonsensearethedetailsofhowattitudesaffect motivationtolearnandhowitinturnaffectsthelanguageacquisitionprocess.Tounderstandthese importantcomplexitieswehavetorelyonelaborateresearch,andthusresearch,morethancommon sense,producesinstructive,oftenunexpectedinformation.Ontheotherhand,commonsense,more thanresearch,givesadirectionandpurposetoresearch.Forinstance,itremindsusthatpeople sometimesmasteran"enemy's"language,anexamplewheretheattitudesinvolvedareanythingbut friendlyandopen.Researchhastoexplainthecomplexitiesofbothextremes,thatis,explainhow attitudesandmotivationtolearnsometimesfunctionquiteindependentlyofoneanotherandother timesquiteinterdependently(seeGardner,1981;Gardner,Glicksman,&Smythe,1978).Itisalso truethatresearchcanstimulatecommonsense.Forinstance,inreviewingvariousearlierresearch findings,Genesee,Rogers,andHolobow(1982)wereledtoanewsetofquestionsaboutattitudes andlanguagelearning.Theyrealizedthatalearner'smotivationmightbeonlyonedeterminerof rateandlevelofachievementinlearninganewlanguage;thelearner'sexpectationsofreceivingor notreceivingmotivationalsupportfromtheothergroupcouldbeaseparateand 197

equallyimportantpredictorofsuccessinsecondlanguagelearning.Andthat'swhatthey foundwhentheyputthisnewcompositeideatotest. Thispreoccupationwithattitudesismuchmorethananacademicissue.Therearesocially importantdecisionstobemadethatrequirereliableinformationaboutattitudes,forexample, decisionsaboutbilingualeducationprogramsorprogramstointegrateorassimilateimmigrants.We needtoknowhowattitudesfunctionbothasdeterminersorpredictorsoftherateofsecondor foreignlanguageachievementattitudesas"input"factorsaswellastheeffectsbilingualskill developmentmighthaveonattitudestowardoutgroupsattitudechangeasan"outcome"effect.In otherwords,tobehelpfultodecisionmakers,researchersneedtodetermine(a)if(andhow)hostile, suspiciousattitudestowardanotherethnolinguisticgrouphampertheacquisitionofthatgroup's language,and(b)if(andhow)successfulacquisitionofanothergroup'slanguagepromotesmore friendly,openattitudestowardthatgroupormoregenerallytowardethnicallydifferentpeoples. Becauseresearchersandeducatorshavedrawnattentiontotheissuesunderlyingsuchquestions, importantchangesinvaluesseemtobetakingplace.Forinstance,itisnowrecognizedthat outlawingalanguageaswhenanimmigrantchild'shomelanguage(thelanguageusedforthe basicdevelopmentofconceptualthinkingfrominfancyon)isprohibitedasamodeofexpressionin schoolsettingsisnothinglessthanasubtleformoftyranny.Thepsychologicalconsequencescan beenormousforthelinguisticgroupinvolvedbecauseattitudestowardowngroup,towardself,and towardsocietyareseriouslyaffectedinsuchinstances.Todayoppressivepoliciesofthissortseem tobegivingwaytomoredemocraticoneswhereinpeoplearealoudtospeakandletspeak,and wheremorethanoneformofspeakingbecomessociallyacceptableandcorrect.Theinterethnic groupattitudesinthesecasesareautomaticallymademorefriendly,permissive,andopen,andasa consequencethesocietiesinvolvedaremademuchmoreinteresting. Forthoseinterestedinexploringfurthertheroleofattitudesinlanguagelearning,thereisan establishedresearchbasetheycanbuildon.Theneedforasocialpsychologyoflanguagelearning becameapparenttoseveralofusinthe1950sand1960s.Atthattime,O.H.Mowrer(1950)began tostudytheemotionalattachmentsthatdevelopbetweentalkingbirdsandtheirtrainers,andthe effectssuchattachmentshadonthebirds'skillattalkdevelopment.Ifattitudesaffectbirdtalk,we argued,therewasmuchtoexpectfromattitudesandhumantalk.Inthisatmosphere,SusanErvin (1954)startedherworkontheroleofemotionsandattitudesinchildren'slanguagedevelopment, andRobertGardnerandIbegantolookatbilingualskilldevelopmentfromthesameperspective (seeGardner&Lambert,1959,1972;Lambert,Gardner,Barik,&Tunstall,1963).TheGardner Lambertstudies,exploratoryandfactoranalyticastheywere,convincedusthatprejudicedattitudes andstereotypesabouttheotherethnolinguisticgroupquiteindependentoflanguagelearning abilitiesorverbalintelligencecanupsetanddisturbthemotivationneededtolearntheother 198

group'slanguage,justasopen,inquisitive,andfriendlyattitudescanenhanceandenliventhe languagelearningprocess.Wealsosawthatparentalattitudes,positiveornegative,arepicked upbychildren,sothatpupilsbringafamilycomplexofattitudestothelanguageclasswith them.GardnerandSmythe(1975)havegonefurtherintotheseaspectsofattitudesandhave foundthatpersistenceinlanguagestudyandeagernesstointeractinlanguageclassesalso hingeontheattitudesandmotivationspupilsbringtoschool(seeGardner,1981). Theseresearchstudieswerejustastartandthereismuchmoretobestudiedinthisdomain.For instance,GeneseeandHamayan(1980),workingwithGrade1AngloCanadianpupils,didnotfind anyneat,simplerelationshipbetweenattitudesandsecondlanguageachievement,indicatingthatwe needtoexploreabroaderagerange.Morerecently,Clmentetal.(1977)andTaylor,Meynard,and Rheault(1977)haveextendedtheresearchdomainbyfocusingontheattitudesofmembersofan ethnolinguisticminoritygroupwhenfacedwithlearningadominantgroup'slanguage.Forexample, Tayloretal.(1977)findthatthelearningofEnglishcanposeathreattopersonalandcultural identityforFrenchCanadianuniversitystudentsandthatthisthreatcanhampertheprogressmade inlearningthatlanguage.Thesociallysignificantfactisthatthreatsofthissortcanleadto suspicionanddistrust,Otherstudieshaveshownthatparents'suspicionsandprejudicesabout outgroupsoraboutowngroupcandeterminetheacademicroutetheirchildrenwillfollowandalso thelanguageidentitytheirchildrenwilldevelop.Thus,FrenchCanadianparentswhoseelittlevalue inbeingFrenchintheNorthAmericanscenearepronetoroutetheirchildrenintoanentirelyAnglo educationalsystem,whereasthosewhohavehopeintheFrenchfactinCanadaandprideinbeing FrenchkeeptheirchildreninFrenchacademicsettingswithouthesitationandcreateforthema comfortableFrenchsocialenvironment(FrasureSmith,Lambert,&Taylor,1975). Ontheoutcomeoreffectside,therearealsosomebasicstudiesforresearcherstobuildon. Commonsensewouldsuggestthatasskillwiththeotherlanguageevolves,attitudestowardthe otherethnolinguisticgroupshouldbecomelesssuspiciousandhostilebecausethelearnersare breakingthroughthelanguagebarrierandreducingtheforeignnessoftheotherlanguagegroup withthehelpofateacherfromthatgroup.Asignoranceoftheothergroupdissipates,attitudes shouldbecomelesssuspiciousandpossiblymorecompassionate,butonemustbecautiousin predictingbecausevaluecontrastsandclashescansurfaceasknowledgeabouttheothergroup grows,andthesecontrastscangenerateanewformofsuspicion.RichardTuckerandIstudiedthe attitudesofAnglophoneCanadianyoungstersastheymovedthrough4or5yearsofimmersionin Frenchschooling;thesepupilswerecomparedwith"control"pupilswhohadhadconventional Englishlanguageschoolingonly(Lambert&Tucker,1972).Asofthestartoftheexperience(i.e., atthekindergartenyear)theattitudestowardFrenchCanadianswerebasicallyidenticalforthe parentsofthe 199

immersionandcontrolgroups.ByGrade5,however,theimmersionpupilsrelativetothe controls"liked"Frenchpeoplemore;weremuchmorepronetosaythattheywouldbe"just ashappy"hadtheybeenbornintoaFrenchfamily;andweremuchmorelikelytosee themselvesasbecomingbothEnglishandFrenchCanadianinmakeup.Apparently,muchof theforeignnessoftheothergroupwasdispelledfortheimmersionpupilsandtheybeganto appreciateboththedistinctiveandthesharedcharacteristicsoftheotherethnolinguistic group.Butotherfollowupstudiesofimmersionpupils,usingdifferentprobingtechniques anddifferentagelevels,sometimesreplicatedandsometimesdidnotreplicatethesefavorable attitudeoutcomes,althoughinnocaseweretheimmersionchildrenlessfavorabletotheother groupthanthecontrols.Itis,ofcourse,ourresponsibilityasresearcherstoexplaincasesthat workoutasexpectedaswellasthosethatdon't. Mypurposehereistoreviewandanalyzeasubsetofresearchstudiesthathavedealtwithattitudes ofthetypejustdescribedinordertohighlightboththesocialimportanceofattitudesinlanguage learningandbilingualismaswellasthedifficultyresearchershaveincapturingandexamining them.ThereviewfocusesondatafromEnglishspeakingpupilsandstudentsinimmersion programs,liketheonejustreferredto,becausetheseprogramsprovideuswithaquasiexperimental formatwhereinarelativelyclearpictureoftheattitudesofpupilsandtheirparentsatthe kindergartenlevelisavailable,andwheresomeoftheseyoungsters(thoseinthe"experimental" groups)areschooledmainlythroughaforeignorsecondlanguage(French)whileothers(thosein the"control"groups)followconventionalEnglishlanguageeducationalprograms.Iwillselect examplesofattitudechange(positiveornegative)andofnochange,aswellaschangesinsocial perspectivesatvariouspointsfromtheearlyschoolgradesupthroughhighschool.Thisreviewthen dealsmainlywith"outcomeeffects,"thatis,theeffectsofdevelopinghighlevelskillsinanother group'slanguageonstudents'attitudesandsocialperspectives.Throughout,comparisonswillbe madeofpupilswhohavebecomefunctionallybilingualthroughimmersionprogramsinschoolor equivalentreallifeexperiences(the"experimental"groups)andthosewhohavenothadsuch experiences(the"control"or"comparison"groups).Inmostcases,theattitudesofthepupilsasof GradeIoroftheirparentsweremeasuredandfoundtobeessentiallyalikeonrelevantdimensions. IncaseswherewehavenoGradeIbaselineattitudedata,weassumethatthefamilyattitudeswere essentiallythesame. ATTITUDESANDTHEIRMEASUREMENT Whatdowemeanby"attitudes"andhowdowegoaboutmeasuringthem?Theconcept"attitude" isacreationofphilosophersandpsychologistswhotriedtoexplainanddescribecertain consistenciesintheirownandotherpeople'sthoughts,feelings,andreactions.Attitudes characterizebiologicalsystemsnot 200

physicalsystems,andso,unlikeatomsandphysicalstructures,theyarepresumedtobehighly variablefrompersontopersonandfromonetimetoanotherforanyperson.Thefuncomesin tryingtocaptureandmeasureattitudesreliablyamuchmoredemandingtaskthanmost measurementproblemsinthephysicalorpurelybiologicalsciences.Thefollowingdefinition( Lambert&Lambert,1973)istypicalofwhatpsychologistsmeanbyanattitude: Anattitudeisanorganizedandconsistentmannerofthinking,feelingandreactingto people,groups,socialissuesor,moregenerally,toanyeventintheenvironment.The essentialcomponentsofattitudesarethoughtsandbeliefs,feelingsoremotions,and tendenciestoreact.Wecansaythatanattitudeisformedwhenthesecomponentsare sointerrelatedthatspecificfeelingsandreactiontendenciesbecomeconsistently associatedwiththeattitudeobject.Ourattitudesdevelopinthecourseofcopingwith andadjustingtooursocialenvironments.Onceattitudesaredeveloped,theylend regularitytoourmodesofreactingandfacilitatesocialadjustment.Intheearly stagesofattitudedevelopment,thecomponentscanbemodifiedbynewexperiences. Later,however,theirorganizationmaybecomeinflexibleandstereotyped,usually becausewehavebeenencouragedoverlongperiodsoftimetoreactinstandardways toparticulareventsorgroups.Asanattitudebecomesfirmlyset,webecometoo readytocategorizepeopleoreventsaccordingtoemotionallytonedpatternsof thoughtssothatwefailtorecognizeindividualityoruniqueness.Fixedor stereotypedattitudesreducethepotentialrichnessofourenvironmentsandconstrict ourreactions.(p.42) Wewillrefertothreemajorcomponents:thestereotype,thefeeling,andthereactiontendency componentsofattitudes.Complexityentersbecausethecomponentsappeartodevelopin differentfashionsandatdifferentrates,and,althoughalwaysintegrated,theycometogetherin unexpectedcombinations.Thusone'sthoughtsandstereotypesaboutanothergrouporeventmay bepositiveornegativewhereasthefeelingsandreactiontendenciestowardthatgrouporevent mayormaynotbepredictablyaligned.Accordingly,attitudessometimescoverlovehate relationships,whereone'sstereotypesareunfavorabletowardasocialgroup,butoneis nonethelessattractedtorepresentativesofthatgroup,orviceversa.Itisalsotruethatsocial constraintscanforcepeopletoreacttoagroupinonewaywhiletheirfeelingsandstereotypes workindifferentemotionaldirections. Intheresearchtobedescribed,typicalmeasurementmethodsareusedandthesereflectmainly onecomponentoranother.Forinstance,thestereotypecomponentismeasuredthroughrating scales,e.g.,respondentsareaskedtogivetheirpersonalestimates,byratingsonscales,ofhow friendly...unfriendly,nice...bad,dependable...undependable,andsoforth,membersofa particulargroupseemtothemtobe.Thefeelingandreactioncomponentsaremeasuredthrough directquestionse.g.,howrespondentswouldfeel(orreact)iftheyhadtointeractwith membersofanothergroupinparticularways.Thefeelingandreactioncomponentsarealso measuredthroughvariousscalingpro 201

urese.g.,respondentsareaskedtoestimatehowsimilarordifferentmembersofanother groupseemwhencomparedtoowngroupmembersortoone'sself. TheStereotypeComponentofAttitudes Since1967,wehavecollecteddatafromnumeroussamplesofEnglishspeakingCanadian(EC) childrenwhowereinFrenchimmersionprogramsinMontrealpublicschools,andfromEC "controls"thoseinconventionalEnglishlanguageschoolprogramsthatprovidedonly30to60 minutesperdayofFrenchtaughtasasecondlanguage(FSL).Formostsamples,parents'attitudes towardFrenchspeakingCanadians(FCs)andtheFCculturewereassessedthroughquestionnaires andinterviewsasofthetimetheirchildrenenteredkindergartenorGrade1.Typically,theparentsof theexperimentalandcontrolchildrenhaveverysimilarattitudestowardFCs,attitudesthatusually reflectlittlepersonalexperiencewithorknowledgeabouttheFCpeopleandculture.Ratherthan beinghostile,theattitudereflectsmorefeelingsofstrangertostranger,coloredwithsomesuspicion (seeLambert&Tucker,1972). The"immersion"experienceisbasedontheinteractionof30to32ECpupilswhohavenoprior knowledgeofFrenchandateacherwho,fromthefirstencounter,speaksonlyFrenchtothem.The teacher,usuallyawoman,wouldbeeitherfromEuropeorfromFrenchCanada,andifnotactually monolingualinFrench,wouldplayasthoughsheknewnoEnglishatall.Frenchisthesolemedium ofinstructionfromthefirstdayonthroughthefirst2or3yearsofschool;Englishisprogressively introduced,byseparateteachers,asasecondmediumofinstructionfromGrade3on,reachinga50 50FrenchEnglishcurriculumbyGrade6.Theoutcomesoftheimmersionexperienceoncognitive development,EnglishandFrenchlanguageprogress,andcontentsubjectachievementare remarkable.Theyhavebeendescribedthoroughlyelsewhere(seeGenesee,1978;Lambert& Tucker,1972;Swain,1974). Informationonattitudesandattitudechangesrelatedtotheimmersionexperienceisbasedlargely onparentandpupilratingson7pointscalessuchasintelligent...stupid;strong...weak; friendly...unfriendly;affectionate...unaffectionate;industrious...lazy;kind...mean; happy...sad;humble...proud;selfconfident...not;goodlooking...ugly;pleasant... unpleasant;calm...emotional;talkative...non.Termsvaryaccordingtotherespondent'sage; ratingsaremadewithregardtovariousreferencegroups,e.g.:EnglishCanadians,French Canadians,EuropeanFrenchpeople,aswellasselfratings,e.g.,Me,Myself. PoliticalandAcademicBackgroundEvents Theattitudeswediscusshereweremeasuredinahighlyemotionallychargedtimeperiodfor Quebec,startinginJuneof1968.Inthisperiodmanypoliticalchangestookplace,allfocusedon progressivelystrongerdemandsonthepartof 202

FrancophoneCanadiansinQuebecforlanguagerightsandpoliticalsovereignty.Therootsof Quebecdemandsforpoliticalindependencearedeep(seeGuindon,1971),butaneffective separatistpoliticalmovementwasstartedonlyinthe1960s.Atthesocialpsychologicallevel, researchersinthelate1950sweredocumentingastrikingtendencyamongFCsinQuebecto denigratetheirownethnicity,portrayingFrenchspeakingCanadiansaslesslikelytosucceed, lesssociallyattractive,andgenerallyinferiortoAnglophoneCanadians(seeLambert,1967). Thiscolonialoutlookwithitsstereotypedselfviewssimilarinmanyrespectstothetraditional viewsblackAmericanshavehadofthemselveswasevidentinthethinkingofpreadolescent FCs(seeLambert,Frankel,&Tucker,1966),suggestingthatadefeatistoutlookhada disturbinglyearlybaseamongFCs.Itislikelythatthesesentimentsofinferiority,touchingeven theyouthoftheFCsociety,wasonefactorpromptingseparatiststoforcefullyandopenlyargue foraseparate,FrenchCanadianstate,onecomfortablyindependentoftherestofCanada. CanadianseparatistshadpowerfulAmericanmodelstodrawon,especiallytheCivilRightsand BlackPowerexamplesofthe1960s.InJuneof1968,webeganoursurveysoftheattitudesof AngloCanadianelementaryschoolpupilstowardFCpeopleandFCculture.Bythistime,the politicalmovementforFCcontrolovertheirownaffairs("matrescheznous"),startingwith languagelegislationfavoringtheuseofFrenchoverEnglish,wasclearlyevidentinaseriesof events. 1. 1969:TheSt.LeonardDemonstrations.St.Leonard,amainlyItalianCanadianregionof Montreal,demonstratedagainstnewlegislationthatforcedsomeItalianCanadianchildren intoFrenchlanguageschools.Nearlyallimmigrantgroups,whengivenachoice,had overwhelminglychosenEnglishlanguageschooling. 2. 1970:TheF.L.Q.Crisis.AradicalundergroundgroupknownasLeFrontpourla LiberationduQubecgainedattentionandprovokedaWarMeasuresActreactionfromthe fedralgovernmentinOttawawhentheFLQkidnappedaQuebecpolitician(andlater murderedhim),aswellasaBritishdiplomat,laterfoundandreleasedbyfederalpolice. 3. 1974:EducationBill22.TheLiberalPartyofQuebecputintolawabillthatrestrictedfree choiceofschooling,especiallylimitingFCchildren'saccesstoEnglishlanguageschools; thebillalsostartedaFrenchasthelanguageofworkmovement. 4. 1976:ThePartiQuebecoisVictory.Inasurprisevictory,theLiberalPartyofQuebecwas replacedbyanewpartyopenlyworkingforindependence,souverainetandseparationof the"StateofQuebec"fromCanada. 5. 1977:EducationBill101.AnocompromisebillwaspassedthatmadeFrenchtheonly officiallanguageofQuebec,Frenchtheonlylanguageofwork,andFrenchthelanguageof schoolingforeveryoneexceptthechildrenofthoseECswhothemselveshadattended EnglishlanguageschoolsinQuebec.Thus,allimmigrantsandAnglophonesfromother partsofCanadawere(andare)required 203

Justasthesepoliticaleventscouldeasilycolortheattitudesofchildren,whetherEC orFC,sotoocouldcertaineventsthattranspiredintheEnglishlanguageschoolsthat introducedimmersionprograms.Fromthestartoftheimmersionexperimentinthe 196566academicyear,thepublicschoolsinvolved(allpartofthecity'sProtestant SchoolSystem)hadtofindFrenchspeakingteacherswhowereProtestants. ConsequentlytheychoserecentimmigrantsfromEurope(France,Belgiumor Switzerland)oractualresidentsofthesecountrieswhowereoftheProtestantfaith.It wasnotuntil196970thattheProtestantrequirementwasrelaxedsothatequal numbersofFrenchCanadianteacherscouldbeintroducedintothegrowingnumberof immersionclassesinMontreal.ThustherewasastrongEuropeanFrenchcharacterto theearlyyearsofimmersion;asidefromtheteachers,theeducationalmodelwasa nononsense,highlystructured,lycetypeprogramofinstruction.Thesefeaturesmay wellhavestrengthenedtheimmersionteaching,butitmayalsohavecreatedinvidious comparisonsbetweenEuropeanFrenchandCanadianFrenchlanguagesandcultures. toattend Furthermore,intheacademicyear196869,itwasdecidedtointroudceFC French youngsters,insmallnumbers,intotheFrenchimmersionclassessoastoapproximate language evencloseratotalFrenchlearningenvironmentfortheAnglophonepupils.The schools. troublewasthattheFCpupilshadtobeProtestant(meaningnonCatholic)becauseof TheParti thereligiousdivisionofMontrealschoolsatthattime.Thetroublewiththatwasthat Qubecois theFCProtestantyoungstershappenedtocomeinlargepartfromlower was socioeconomichomeandcommunitybackgroundsthantheAnglophonesinthose reelected schools.Differencesinsocialclassofthissortmakesubstantialdifferencesnotonly foranother inmeasuredIQandscholasticattainment,butalso,itturnsout,intypeandmagnitude terminthe ofdisciplineproblems.Itbecameapparentimmediatelytoteachers,principals,and early pupilsalikethattheFCProtestantpupils,asasubgroup,wereclearlypoorerstudents 1980s. inallcontentmatters,Frenchincluded,andveryundisciplined.Aftera2yeartrial period,theintroductionofFCProtestantpupilsintoimmersionclassesstopped,butit isverylikelythatwhileintheclasses,theyleftastrongnegativeimpressiononthe ECpupils. StereotypesofAnglophoneImmersionPupils,19661978.ByJuneof1968,thefirst immersionclassintheMontrealprogram(the"Pilot"class,accordingtoLambert& Tucker,1972)hadreachedGrade2,andtheirratingsofFrenchpeoplewere unambiguouslyfavorable,significantlymorefavorablethantheECcontrols,and constitutingwhatwerefertoasaclear"immersioneffect."Theyalso,relativetothe ECcontrols,presentedamorebalancedperspectiveofFCpeople,ratingthemhigher onsuchscalesasgood,smart,friendly,andnice,andloweronscaleslikeshort,small andfat.Wewillrefertothisasa"balanceeffect,"indicatingthattheimmersion childrenratedFCsatorabovetheneutral 204

pointofthescales,therebybringingtheirperspectivesofFCsmoreinbalancewiththeirviews oftheirowngroup,theECs.Inthiscase,theydidnotachievethebalancebydowngrading theirowngroupinanyway.Rather,comparedtotheECcontrols,theyelevatedtheir perceptionsofFCs.Furthermore,theirselfviews(theirratingsof"me")wereineverywayas favorableandwholesomeaswerethoseoftheECcontrols;bothgroupssawthemselvesas friendly,nice,big,tall,smart,andgood.Thisfavorableselfview,incidentally,holdsforall immersiongroupsexaminedtodate,regardlessoftheagelevel(seeLambert&Tucker,1972). BytheendofGrades3and4(June1969and1970),theattitudepatternhadchanged.Atthattime, thegroup"FrenchpeoplefromFrance"wasincludedinthecomparison,andforthe1969and1970 periods,therewasnoimmersioneffect(i.e.,nodifferenceinratingsbetweenimmersionandEC controlpupils),andthusnobalanceeffectfortheirratingsofFCsorofEuropeanFrenchpeople.All ratingsfellaroundandslightlybelowtheneutralpoint(scaleposition4),meaningthatall AnglophonegroupsviewedFCsasslightlydumb,bad,unkindandmean.Thetrendwassomewhat lessnegativetowardEuropeanFrenchpeople,butnotsignificantlyso. Incontrast,forpupilsinthe"Followup"classwhofinishedGrade2inJune1969,therewerestrong immersionandbalanceeffectsintheirratingsofEuropeanFrenchpeople,butnotofFCs,whowere viewednegatively,particularlyonthegoodbad,andintelligentdumbscales.PupilsreachingGrade 2thefollowingyear(June,1970)presentedthesamepictureastrongimmersionandbalance effectforEuropeanFrenchpeople,butnotforFCs.Inbothcases,theselfviewsofimmersion pupilswereasfavorableandwholesomeasweretheselfviewsoftheECcontrols. ASecondAttitudeSurvey:Grades4,5,and6.Toavoidovertestingthesamepupils,weswitchedto newgroupsofimmersionandECcontrolpupilswhohadfinishedGrades4,5,and6inJuneof 1976or1977(seeCziko,Holobow,&Lambert,March,1977;Cziko,Holobow,Lambert&Tucker, December,1977).Theresultsin197677,some5yearslater,wereessentiallythesameasthose foundin197071withyoungerchildren;thatis,forpupilsinGrade4asof197677,therewasno immersioneffectfortheirratingsofFCs(allgroupsratedFCsaroundtheneutralpoint(4)onthe evaluativescales),andnobalanceeffectbecausetheirratingsoftheirowngroupfellbetweenscale points5and6,ontheaverage.However,withregardtoEuropeanFrenchpeople,therewasbothan immersionandabalanceeffect;thatis,theimmersionpupilssawEuropeanFrenchpeopleas significantlymoreaffectionate,kind,andselfconfidentthandidtheECcontrols.Thusthe immersionpupilshadviewsofEuropeanFrenchpeoplethatwererelativelyclosetotheviewsthey hadoftheirowngroup. TheGrades5and6pupilsinthe197677testing,however,showedavery 205

strongimmersionandbalanceeffectintheirratingsofEuropeanFrenchpeopleandofFrench Canadiansaswell.HeretherewereclearsignsofmorefavorableattitudestowardFrench peopleingeneralamongtheimmersionpupils.However,itwasnotthattheimmersionpupils atgrades5and6weresofavorable(theirratingsalsofellatorslightlyabovetheneutral point),butthattheECcontrolpupilshadverynegativestereotypesofFrenchpeople, particularlyofFCs.Thusinthesecasesitappearsthattheimmersionexperiencehad protectedstudentsfromthestrongantiFrenchsentimentsthatcharacterizedtheECcontrols. AThirdAttitudeSurvey:Grade7.InJuneof1976,threenewgroupsofGrade7studentswere surveyedinthestandardmanner,usingthesamescales.Onegroupcomprisedthosewhohadbeen inaFrenchimmersionprogramfromkindergartenthroughGrade6;asecondgrouphadFSL instructionuntilGrade7atwhichtimetheyfolloweda1yearFrenchimmersionprogram;andthe thirdwasanECcontrolgroupwhohadFSLinstructiononlyfromGradeIon(seeCziko,Holobow, &Lambert,April1977).Heretherewasnoevidenceofanimmersionorbalanceeffect,sinceall threeAnglophonegroups,immersionornot,gavegenerallysimilarratingsforbothFCand EuropeanFrenchpeople.Inthiscaseitisnotthattheimmersionpupilsweremorenegativethan usualintheirviews(theirratingsfellatorslightlyabovetheneutralpoint,ontheaverage),butthat theECcontrolsattheGrade7levelhadviewsofFrenchpeoplethatwerelessnegativethanthose seenattheGrade46levelswithotherECcontrols.ItisthuspossiblethattheFSLexperiencehasa similarultimateeffectofamelioratingAnglophonestudents'viewsofFrenchpeople,butthatitjust takesmuchmoretime,finallyshowingitselfatthehighschoollevel.Thisdoesnotmeanthatwhat appearstohavebeenamoreimmediateameliorativeeffectattributabletoimmersionhadpushedthe pupils'ratingstotheupperlimitsofthescales.Instead,theviewsofFrenchpeoplethatAnglophone studentsarriveat,eitherearlyorlate,areneutraloratbestonlyslightlyfavorable.Thisessentially neutralperspectiveofFrenchpeopleamongtheolderchildrenstandsinsharpcontrasttothemuch morefavorableviewtheyhaveoftheirowngroup. AFourthAttitudeSurvey:Grade7and11Students.InJuneof1974asimilarsurveywasconducted inatotallydifferentschoolsystemwithGrades7and11students.Thestudentshadfolloweda differenttypeofimmersionprograma1yearimmersioninFrenchatGrade7after6yearsof conventionalEnglishlanguageprogramwithanFSLcomponentfromGradeIon(seeGenesee, Morin,&Allister,1974a,1974b).Asfortheothersurveys,comparisonsweremadeatGrades7and 11withECcontrolswhohadnothadthe1yeartotalimmersionexperience.Inthiscase,aclear immersioneffectwasfoundattheGrade7level,i.e.,immersionpupilsdisplayedmorefavorable attitudestobothFCsandEuropeanFrenchpeoplethandidtheECcontrols.Again,thedif 206

ferenceswereduemainlytolessfavorableviewsonthepartoftheECcontrols;theimmersion groupsagainrestrictedtheirratingsofFrenchpeopletothe45scalerange.AttheGrade11 level,however,therewerenosignsofanimmersionorbalanceeffectbecausetheGrade11EC controlsdisplayedessentiallythesameviewsofFrenchpeopleasdidtheearlyimmersionand theGrade7,1yearimmersionstudents. CONCLUSIONSABOUTIMMERSIONPROGRAMS'EFFECTSONSTEREOTYPES Thereareseveralgeneralconclusionsthatcanbesafelydrawnfromthisseriesofattitude surveys,andalthoughwearefortunatetobedealingwithlongitudinaldata,wenonethelessview theseconclusionsastentative,requiringfurtherverificationthroughreplicationsinsimilarand differentsettings. 1.First,wearedealingherewithsurveys,i.e.,groupadministered,checklist,essentially superficialmeasuresofpupils'stereotypes,notindepth,personalizedprobes.Theirsaving featureistheirreliabilityandtheirproceduralclarity.Furthermore,theyaresurveysofthe attitudesofECschoolchildrentowardFrenchpeopleconductedintheintenselyemotional politicalclimateofQuebecinthe1960sand1970swhenFrenchEnglishrelationswerecolored bymutualsuspicion,distrust,andfear.Abaptismunderfireofthissortmaywellhavebeenthe bestpossibleperiodfortestingourmainideas.Oneconsistentfindingisthatchildrenwith immersionexperienceattainedandmaintainedanaverageevaluativeratingofFrenchpeoplein the4to5pointrangeon7pointscales,reflectinganeutraltoslightlypositivestereotypeof Frenchpeople.Whydidn'timmersioninFrenchpushattitudestoahigherlevel?Foronething, gettingtoknowanothergroupthroughlearningtheirlanguageandculturemightnotgenerate extremeaffectionifthegroupinquestionslowlyemergesasbeingessentiallyatoddswithone's ownbasicvaluesandpointsofview.Likemarriagesthatleadtodivorceinoneoutoftwoor threecasesintheWesternworld,gettingtoknowotherpeoplewellisoftenaneyeopener! Furthermore,inthisinstance,theothergroupinquestionisonethathasshareddemocraticallya commonlandwithyourowngroupforgenerationsandhassuddenlyexpressedadesireto separatebyforminganautonomousFrenchCanadianstate.Thus,maintainingratingsinthe45 rangemaybealloneshouldexpectfromthosewhoarebeingdistancedbythatsoughtfor separation. 2.Second,theresultsshowthatchildreninimmersionclassesgenerallyhaveeithermore favorableorlessnegativeviewsofFCsandEuropeanFrenchpeoplethandochildrenwhowere notinimmersionprograms.Eveninthecaseswheretheimmersionandcontrolchildren'sviews donotdiffer,therearenoinstanceswheretheimmersionpupilsexpresslessfavorableviewsthan thecontrols.Thus,immersionismorelikelytofostermorefavorableattitudeprofilestowardthe 207

groupwhoselanguageisbeinglearnedthanisthecaseforchildrenwithoutimmersion experience. 3.Innocasedoesimmersionaffectnegativelythechildren'sviewsoftheirownethnolinguistic group(viewsof"EnglishspeakingCanadians")ortheirselfviews(viewsof"me"). 4.Theimmersioneffectismostlikelytoemergeatthestartoftheimmersionexperience,either attheGrade1levelofearlyimmersionprogramsorattheGrade7levelforlate,1year immersion.Thisisunderstandableifonerealizesthattheimmersionpupilshavevirtuallyno socialcontactwithFrenchspeakingpeopleotherthanthatoftheirteacher.Thus,thereisno socialsupporttomaintainpersonalized,favorableattitudestowardsFrenchpeopleexceptforthe contactpupilshavewiththeirFrenchspeakingteacher. 5.Theimmersionexperience,therefore,doesnotamplifyhighlyfavorablestereotypesofthe groupwhoselanguageisbeinglearnedsomuchasitprotectsstudentsfromthehighlycharged, usuallyuninformed,perspectivesoftheothergroupthataremorelikelytorununchecked throughthemindsofchildreninthecontrolgroups.Inseveralinstanceswesawhowthe immersionchildren,eveninthepoliticalpeakperiods,werelesssour,hostileandnegativein theirviewsoftheothergroupthanwerethenonimmersioncontrolchildren. 6.Anotherpossibleexplanationfortheoccasionallackofanimmersioneffectinour comparisons,especiallythoseatthehighschoollevel,isthefactthatalltheECcontrolshada 3040minuteperiodofFSLinstructionthroughouttheirschooling.TheFSLcomponentis expertlyandsympatheticallytaught,usuallybynativespeakersofFrench,andthusFSLmay simplyrequiremoretimetoliftthecontrolchildren'sattitudesuptothelevelsattainedbythe immersionchildrenatearliergradelevels. 7.Moreoftenthannot,thechildren'sattitudeprofilesshowedamorefavorableviewofFrench peoplefromFrancethanofFrenchCanadians.Thismaybedueinparttothepublicimageof Francethathasbeenprojectedhistorically.ItisalsopossiblethatCanadians,EnglishandFrench, haveinheritedacolonialmentalityabouttheoldcountries(EnglandandFrance)asbeing preferred,morecivilizednations.Wealsobelieve,however,thatotherfactorswereverylikely involvedinthemorefavorableratingsgivenEuropeanFrenchpeople:atendencytohiremore EuropeanbornthanCanadianbornFrenchteachers,andanunhappyadministrativeplanto integratesmallnumbersofFrenchspeakingProtestantchildrenintotheimmersionclasses.The troublewasthatthechildrenbroughtinweredisadvantagedintermsofsocioeconomichome backgrounds,makingthemappearcomparativelydumbacademicallyandunmanageablefroma disciplinepointofview.Untiltheyweretakenoutoftheimmersionclasses,theirpresence appearedtoworkagainstthedevelopmentoffavorablestereotypesofFrenchspeakingCanadians becausethisdecidedlyunrepresentativesampleofFCswasclearlyseenas"dumb"and"not nice." 208

Insummary,thestereotypecomponentofattitudesisseeninfulloperationinthesurveys describedhere,andtheupsanddownsofthechildren'sviewsoftheothergroupseemtoreflect sensitivelythepoliticalbackground,theadministrativepoliciesattemptedbyschoolprincipals, and,mostimportant,thegenerallyfavorableeffectsonstereotypeswhichareattributabletothe immersionexperience,butwhicharenotpronouncedinmagnitudenordurablebeyondthefirst2 yearsofthestartofsuchprograms. THEFEELINGSANDREACTIONTENDENCYCOMPONENTSOFATTITUDES Quitedifferentprobingprocedureswereusedtogetatstudents'feelingsfor(oragainst)French speakingpeopleandtheirwillingnessorreadinesstointeractwiththem. TheUseofDirectQuestioning.Aprocedurewefoundveryinformativewastoaskdirect questionsaboutfeelingsandpsychologicalreactions,forexample:HavinglearnedFrenchand aboutFrenchpeopleforsometime,doyoufeelclosertoFrenchpeoplenowthenyouoncedid? Doyoufeelyoulikethemmorenow?Wouldyouliketogettoknowmoreaboutthemandtheir language?Thefirsttwoclassesofimmersionpupilswereaskedsuchquestionsastheyfinished Grade5(thePilotclass)orGrade4(theFollowupclass)inJuneof1971(Lambert&Tucker, 1972).ThesehappentobethesamepupilsjustreferredtowhosestereotypesofFrenchspeaking peoplewerenodifferentfromthecontrolsasofthe1970and1971surveys(seealsoLambert, Tucker,&d'Anglejan,1973).WhatisinterestingisthattheirfeelingsaboutFrenchspeaking peopleareclearlymorefavorablethanthoseofthecontrols,especiallysoforthoseinGrade5. (Seechapter9ofLambert&Tucker,1972.)Thus,tothequestion:"Sinceyouhavestarted learningaboutFrenchspeakingpeopleatschool,doyoulikeFrenchCanadianpeople(also, EuropeanFrenchpeople)morenow?"theGrade5immersionpupilssaidtheylikedFrench Canadiansmorenowthantheydidatthestartofschooling,muchmoresothandidtheGrade5 controlpupils.(Mentionedherearestatisticallysignificantdifferences.)Similarly,whenasked "SupposeyouhappenedtobebornintoaFrenchCanadianfamily,wouldyoubejustashappy?" theGrade5immersionpupilswereagainmuchmorelikelytosaythattheywouldbe"justas happy."BothGrades4and5childrenshowedmorefavorablereactionswhenasked:"Doyou think,inthecourseofyourstudyingFrench,thatyouhavebecomelessEnglishCanadianinyour thoughtsandfeelings,ordoyouseeyourselfnowbeingbothEnglishandFrenchCanadian,oras moreEnglishCanadian?"Inthisinstanceimmersionpupilsatbothgradelevelsweredecidedly morelikelytothinkofthemselvesasbecomingboth 209

EnglishandFrenchCanadianinmakeup.Whenaskediftheyfelttoomuchemphasiswasplaced onFrenchinschoolsothattheymightlikenowtogotoan"allEnglish"school,againthe immersionpupilsatbothgradelevelsdifferedsignificantlyfromthecontrolsinrejectinganall Englishalternative.Then,whenaskediftheywouldliketocontinuelearningFrench,the immersionpupilsweremuchmorepronetocontinuethanarethecontrols,astherecent theorizingofGardner(1981,1982)wouldpredict.(ThedetailsareinLambert&Tucker,1972,p. 192ff.)DirectQuestioningasofHighSchoolLeaving.Thesamegroups(Pilot,Followup,and EnglishControl)weresurveyedagainastheyfinishedhighschool.Theywereaskedtothink backontheirschoolingandthepartlanguagelearninghadplayedinit.Theywerealsoasked attituderelatedquestions,forinstance,questionsabouttheirwillingnesstointegratewith, developfriendshipswith,workinFrenchsettingswith,orcontinuestudiesinFrenchschools withFrenchspeakingpeople.Responsesweretreatedstatisticallyininstanceswheredirectgroup comparisonscouldbemade,andlessformallyininstanceswhereextendedstatementswere made.Parentsofstudentsineachgroupwerealsosurveyed.(ThedetailsareavailableinCziko, Lambert,Sidoti,&Tucker,1980;andinWolfe&Lambert,1979).Thefollowingconclusions weredrawnfromtheseanalyses. 1. Therewasaveryclearexpressionofappreciationfortheimmersionexperienceonthepart oftheimmersionstudentsaswellastheirparents,themajorityofwhomsaidtheywould choosetheimmersionoptionagainiftheyhadtodoitallover.Thegeneralsatisfactionon thepartofimmersionparentsstoodinsharpcontrasttothecontrolparentswhodirected harshcriticismattheschoolsystemforhavingfailedthemandtheirchildrenbynotteaching themFrench.Controlparentswerealsomorelikelytosendtheirchildrentoprivateschools inthehighschoolyears,perhapsasanassurancefortheirchildren'sfuture. 2. Ingeneral,immersionstudentsexpressedafeelingofwellbeing,selfassurance,and satisfactionwiththeirlevelofattainmentinFrench,whichshoweditselfinvariousways:(a) theimmersionstudentshadalreadyhadmoreparttimeandsummerexperiencesworkingin French;(b)theyfeltmuchmorecapableofworkinginFrenchor(c)studyinginFrenchat thecollegeoruniversitylevel;and(d)theyweremoreeagertostudyotherlanguages,as thoughthetasteofsuccesswithlearningFrenchandtherealizationthatonecanmastera foreignlanguagestimulatedamoregeneralinterestinlearningotherlanguages. 3. TheyalsoasagroupshowedamorepositiveattitudetowardFrenchspeakingCanadiansand agreaterwillingnesstomakecontactswithFrenchspeakingpeoplewithwhomtheyused Frenchincommunication. 4. Alargerproportionoftheimmersionthanthecontrolstudentsalsoexpressedadesireto stayinQuebec(eventhoughamajorityofbothgroupsthoughttermsofleavingthe provinceinthefollowingfewyears). 210

5 Amuchlarger . proportionof immersion studentsalso feltthatthey could,given more occasionsto usethe language, becomefully bilingual. Thesefeelings weresharedas wellbythe respective groupsof parents.The contrastis highlightedby thedelightof twostudents whosaidthey hadbeen highly complimented whentakento benative speakersof French.The control students,on theother hand,gavethe impression thattheyhad littlehopeof really mastering Frenchorthat itwastoolate toexpectthat degreeof competence. 6 Boththe . immersion studentsand theirparents

Atthesametime,ouranalysisbroughttolightvarioushurdlesthatthese wellpreparedandmotivatedyoungpeoplefacedwhentheytriedtopenetratethe FrenchworldaroundtheminQuebec.Someofthesehurdleswereverylikely rootedintheEnglishspeakingsocietyitself,whichprovidedfewexamplesof waystomakecontactwithFrenchspeakingpeople;otherhurdleswerejustas likelyrootedintheFrenchspeakingsocietyofQuebec,whichalsoprovidesfew modelsofhowonemightencourageandfollowupongesturesofappreciation andinterest(likelearningtheFrenchlanguagewell)comingfrom EnglishspeakingCanadians.Furthermore,ECsgraduatingfromsecondary schoolinQuebecinthelateseventieswereenteringasocietythatwasdivided andpolarizedalongethnicandlinguisticlinestoadegreetheirparentsor grandparentsneverexperienced.Themainpointisthatitmaywellbebecauseof theirimmersionexperiencesthatthissubgroupofECstudentscanseeintothe societalproblemstheyhaveinherited.Frustratedastheyare,theymaybethe bestagentstomakesubstantivechangesinthesocietyofthefuture.Theyshow signsofgenuinebewildermentastowhysocietymakesitsodifficultforthemto learntheothergroup'slanguageormakesocialcontactswithmembersofthe othergroup. Asaconsequence,perhapsthemostimportantconclusiontobedrawnfromthis investigationofgraduatesofimmersionprogramsisthattheupcoming generationmayhavefreshideasabouttheproblemsofsocialsegregationand cleavagewithinthesociety.Thebeginningsofasolutionareapparentinwhatthe immersionstudentsseemedtoaskofthesocietyanopportunitytoputtheir 211

weremore inclinedthan thecontrolsto seethemerits ofextended formsof immersion, e.g., submerging themselves withFrench studentsin French language schools. Submersionin atotally Frenchschool wasnotseen asaradical steporone thatmight adversely affectone's identityor native language competence. Rather,itwas seenmoreasa naturaland obvious extensionof early immersionby asubstantial numberof immersion studentsand theirparents. 7 Overallthen, . theearly immersion experiencehas astrong impact:it developshigh degreesof skillinthe

French language;it helpsdevelop confidence thatonecan work,study, andliveina French environment aswellasin anEnglish one;it convinces most graduatesthat theycould, givensimple opportunities touseFrench, becomefully bilingual;and itgeneratesa willingness anddesireto meetand integratewith French speaking Canadians.

competenceintheotherlanguagetomeaningfuluse.Attainingfunctionalbilingualskillin Frenchhadapparentlyprovidedthemwithfeelingsofcompetenceinanewlanguageandculture, andawillingnesstobebiculturalcitizens,eventhoughtheparentrunsocietytheylivedin frustratedtheirattemptstoactualizetheirpotential.Incontrast,theECcontrolchildrenfeltmuch lesscompetentandlessreadytobebicultural,andourguessisthattheirlessfavorablefeelings towardtheothergroupderiveatleastinpartfromalackofopportunitytobecomebilingualand bicultural. TheUseofMultidimensionalScaling.Thereisarelativelyrecentrevivalofinterestinpsycho physicalscalingproceduresthattapthefeelingcomponentofattitudesindirectlybyasking respondentstogaugehow"similar"theyviewvariouscomparisongroups.Forinstance,weasked immersionstudentsandcontrolstoconsidervariousreferencepeople(e.g.,Americans,Italian Canadians,EnglishpeoplefromEngland,FrenchpeoplefromFrance,monolingualorbilingual ECsandFCs)andthentorateon9pointsimilarityscales(runningfromverysimilartovery different)eachpossiblepairofreferencepeople,withpairspresentedinarandomorder(e.g., AmericansandFrenchpeoplefromFrance;ItalianCanadiansandmonolingualECs).Wealso includedinthelist"myteacher"and"me." Multidimensionalscaling(MDS)hasprovenvaluableinseveralCanadianstudiesofsocial perception(cf.Christian,1976;FrasureSmithetal.,1975;Taylor,1976).WithMDS,thebases formakingsimilaritydissimilarityjudgementsarelefttotherespondent.Usingrespondents' dissimilarityjudgmentsasdata,MDSproducesaconfigurationofthevariousreferencepeople thatsummarizesandgraphicallydescribestheperceivedclosenessordistancebetweenand amongreferencegroups.Italsouncoverssomeofthedimensionsrespondentsuseinmaking theirjudgments. In1977,weusedtheMDSprocedurewithECGrades5and6studentsintwotypesofFrench immersionprogramsandcomparedthemwithECcontrols.Ouraimwastoobtainaclearer pictureofthesocialattitudesofthesestudents,inparticular,abetterviewoftheaffective consequencesofFrenchimmersionprograms.ThesameAnglophonestudentshadbeentested fortheirstereotypes,asalreadydescribed;asof1977,wefoundastrongimmersioneffect,due mainlytothesourattitudesoftheECcontrols.TheMDSconfigurationsofthesamerespondents indicatedthattheearlyimmersionexperiencetendedtoreducetheperceiveddissimilarityof EnglishCanadiansandFrenchCanadianstoasignificantdegree.Thisrapprochementtookplace mainlybecause"bilingualFCswerebroughtcloserto"bilingualECs"inthethinkingofthe immersionstudents,relativetotheECcontrols,andbecausethe"self"conceptwasplacedcloser toaclusterformedby"bilingualECs"and"bilingualFCs."Thereisalsoatrend(whichisnot statisticallysignificant)fortheimmersionstudentstobring"monolingualFCs"closerto "bilingualFCs"andthusindirectlycloser 212

to"self."Inshort,theearlyimmersionexperienceseemstohavereducedthefeelingsofsocial distanceandseparationbetweenFrenchandEnglishspeakingCanadians.ThevalueoftheMDS methodologyisthatitprovidesameansofinterpretingdatathatonemightmisswitha commonsenseapproach.Thus,itisinterestingtolearnthattherapprochementofFrenchand Englishgroupstookplacemainlythroughthefeelingonthepartofearlyimmersionpupilsthat theyhavebecomesimilarto"bilingualECs"whointurnareseenassimilarto"bilingualFCs" whointurnaresimilarto"monolingualFCs."Instructiveasthemethodis,weneedtotestit furtheronnewgroupsbecausehereweusedsubjectswhointhe1977testingperiodhadshowna clearimmersioneffectforstereotypesofFCs.Thequestionremains:Wouldwegetsimilar resultswithMDSforgroupswhohadnotshownanimmersioneffectatthelevelofstereotypes? MixedProcedures:ScalingandQuestioning.Thefeelingandreactiontendencyfeaturesof attitudescanalsobemeasuredprofitablythroughopenendedquestionsthatproduceresponses thatcanlaterberatedandscaledbyjudges.Thismixedmethodprocedurewasthebasisofan importantrecentstudyconductedinthespringof1979(Blake,Lambert,Sidoti,&Wolfe,1981 ).Inthiscase,AnglophoneimmersionstudentsatbothGrade6(endofelementaryschool)and Grade11(endofhighschool)levelswerecomparedwithECcontrolswhohadFSLexperiences only.Inaddition,FCstudentsattheGrade6and11levelswereintroducedinthecomparison, halfofthembeingbilingualinEnglishandhalfbeingessentiallymonolingualinFrench.Some 360studentswereincludedinthissurvey,whichwasconductedinEnglishandFrench. Herearesomeexamplesofthequestionsaskedandthescoringproceduresemployed.Forthe "friends"question,eachstudentwasasked:DoyouhaveanyclosefriendswhoareFrench Canadian(Canadienanglais)?Ifyes,howmany?Thencamethe"similarities"question:What,in youropinion,arethemainsimilaritiesbetweenEnglishspeakingyoungpeopleandFrench speakingyoungpeople?Pleasetrytogivethreeexamplesofsimilarities.Nextwasthe "differences"question:What,inyouropinion,arethemaindifferencesbetweenEnglish speakingandFrenchspeakingyoungpeople?Pleasetrytogivethreeexamplesofdifferences. Thenthe"statusofEnglishFrenchrelations"question:InCanadatoday,opinionsdiffergreatly abouthowwellEnglishspeakingandFrenchspeakingpeoplearegettingalong.Somepeople feelthatrelationshipsarebadandthereismuchtensionbetweenthetwogroups.Othersfeelthat EnglishFrenchrelationshipsarejustbecomingcomfortablenow,andareevenimproving.What areyourviewsonthisissue?Thencamethe"problems"question:Inyouropinion,whatarethe twomostimportantproblemsseparatingEnglishspeakingandFrenchspeakingCanadians? Finally,theywereasked:IfyoufeelthattherearedifficultiesbetweenEnglishspeakingand FrenchspeakingCanadians,thenwewouldlikeyourideasaboutwhatcouldbedonesothat 213

EnglishandFrenchspeakingCanadianyoungpeoplewouldgetalongbetter.Twoorthree suggestionswouldbeappreciated. FCrespondentsweretestedforskillinEnglishandwerecategorizedasmonolingualorbilingual, providingtwosubgroupscomparabletotheimmersionandECcontrolsfortheECrespondents. Toscoretheopenendedquestions,allresponseswereanalyzedandscaledbytwojudgesworking independently.Forinstance,responsestothesimilaritiesquestionwerecodedandassignedascore from1to6onascalerangingfromsuperficialsimilaritiestoprofound,deepsimilarities." Interjudgereliabilityofcodeassignmentswasfoundtobe.95forthisscale.Asimilarscalewas usedforthedifferencesquestionwhichalsohadahighinterjudgereliability.Forthestatusof EnglishFrenchrelationsquestion,a7pointscalewasdeveloped,withacontinuumfrombadwith muchtensiontogood;atmospherecomfortable.FortheproblemsfacingEnglishandFrench Canadiansandthesuggestionsforimprovementquestions,theessentialideaswerepreservedwhile accomplishingsomesummarization.Some25categoriesweredevelopedforeachquestion. Theresultsofthisstudyarevaluableandinstructivebecausetheyprovideseveralclearcontrastsin attitudesandperspectivesbetweenEnglishandFrenchspeakingCanadiansontheonehand,and betweenthosewhohadbecomebilingual(whetherFCorEC)andthosewhohadnot.Therewere manyexamplesofsuchcontrasts.Forinstancewhenaskedwhattheysawasthemajorproblems separatingECsandFCs,theAnglophonepupilsatGrade6(bothimmersionandcontrols) concentratedon"languagedomination,""groupdomination"and"separation"asmajordivisive problems.Intoday'sQuebec,theseareclearlytheconcernsofAnglophonesratherthan Francophones,sinceFrenchhasbecometheonlyofficiallanguageandthemainlanguageofwork, andFCsholdnearlyallpoliticalpowerpositions.NotsurprisinglytheFCpupilshadmoretrouble agreeingonwhatconstitutesaproblem.Theymentioned"languagedifferences"especiallyand generalfactorslike"stubbornnessandresistancetogettingalong." ThesuggestedsolutionsattheGrade6levelshowedsimilarcontrasts.TheAnglophones(both immersionandcontrols)emphasized"greaterpoliticaltoleranceforAnglophones,"reflectingtheir (andtheirparents')definitionofthecentralproblem.TheAnglophonesalsosuggested"mixing pupilsinbilingualschools"andinstituting"studentexchanges."Incontrast,theFrancophones suggestedsendingECpupilstoFrenchschoolsasapossiblesolution. ClearlyethnicityitselfshapesquitedifferentoutlooksonCanada'scurrentintergrouptensions,and thisprobingprocedureeffectivelybroughtsomeofthesedifferencestolight.Whatisevenmore interestingforusisthat,evenattheGrade6level,beingbilingualornotismoreimportantthan one'sethnicityinshapingcertainperspectives.Thus,beingbilingual(throughimmersionschooling fortheECstudentsorthroughlessformalmeansfortheFCstudents)seemstoprovidetheinsight thatthesolutiontoCanada'stroublescallsforthe"elim 214

inationofprejudiceanddiscrimination."NosuchsolutionisseenintheresponsesoftheEC controlsortheFCmonolinguals! AttheGrade11level,manymoresubgroupcontrastsemerged.Thesubtledifferencesinthe thinkingstylesoftheselateteenagersarefascinating.Forinstancethemostpopularlycitedproblem forallsubgroupswas"languagedifferences,"stressedparticularlybythemonolingual Francophones.Althoughthereisaconsensusthatlanguagedifferencescausetrouble,therearequite differentunderlyingthemesassociatedwithlanguagedifferences.ForbothECandFCbilinguals whohavealreadyovercomethelanguagebarrierthemajorproblemshiftsfromlanguageto "groupdomination,""groupsegregation"and"stubbornness"and"personalresistancetochange." ForECmonolinguals,Canada'smalaiseistraceableto"separatist"tendenciesand"national disunity"coupledwithlanguagedifferences.ForFCmonolinguals,themajorproblemsare "languagedifferences,""dominationbytheEnglish,"and"differencesincustomsandculture." ThesuggestedsolutionsofferedbytheGrade11studentsdisplayawiderangeofsophistication.For instance,thebilinguals(bothECsandFCs)callforgreaterintergroupcontactbetweenEnglishand FrenchspeakingCanadiansmuchmorefrequentlythandothemonolinguals.Bilinguals,inother words,indicatethatfactorsotherthanlanguagedifferencesareinvolvedandarguethatincreased "contact"betweentheethnicgroupsisrequiredinordertoimprovethesituation.Thesamethemeis seenintherelativestressgivento"travelthroughoutthecountry"asasolution.BothECandFC bilingualsseemorevalueintravelthanmonolingualsdo. AsharpcontrastisseenbetweentheECcontrolsandtheimmersionstudentsatGrade11.The controlsweredistinctivewiththesuggestionthatgreaterpoliticaltolerancebegivenAnglophones,a solutionthatlikelyreflectedtheirconcernsaboutseparationanddisunity.Interestingly,theGrade11 immersionstudentsdidnotlistthisalternativeatall.Theirsuggestionsforthereductionoftensions reflectmoresensitivityandbetterunderstandingofthemeaningoftheseparatistmovementin Quebec,e.g.,theunderlyingfeelingsofinferiorityamongFrenchCanadiansandtheunfairnessof treatmenttheybelievetheyhavereceivedinCanada.Becauseoftheirlanguageskills,thebilingual Anglophonesappeartobelessworriedthanmonolingualsaboutthecurrentsignsofpolitical intolerancetowardAnglophonesbecausetheyaremuchbetterabletocopeinmodernQuebec. Whatisdistinctiveinthethinkingoftheimmersionstudentsistheneedtheyseeforpeopletolearn aboutoneanother'sculture.PresumablybecausetheyhavebeentaughtCanadianhistoryfromboth EnglishandFrenchCanadianperspectivesbyteachersfromthetwobackgrounds,theyapparently seethevalueofdualperspectiveteachingandtheneedformoreofit.Theywerealsodistinctivein theemphasisplacedonmixingAnglophonesandFrancophonesinbilingualschoolsystems.This themewasstressedsubstantiallymoreoftenbytheGrade11immersionstudentsthanbyanyother group.Inci 215

dentally,thissamethemewasnotedinthethinkingofotherGrade11immersionstudentswho participatedinanearlierstudy(Cziko,Lambert,&Gutter,1979). ThedistinctivefeatureoftheFCstudents'thinkingatGrade11wasthesuggestionthatstudent exchanges(wherestudentsfromoneethnicgroupcouldvisitintheothergroup'shome)would helpreducetensions.Theyalsoemphasizedtheneedtodevelopbilingualskills,inparticular mutualbilingualismwhereECswouldlearnFrench.However,theydidnotsuggest,asthe Anglophonestudentsdid,thatbilingualschoolsareasolution,nordidtheymentionmixedFC ECsocialandrecreationalactivities.Theysupportedinsteadexchangeshometohomeor schooltoschoolofatemporary,occasionalnature. Themainconclusionswedrawfromthisinvestigationisthatearlybilingualandbicultural experiencesproducenotonlytheobviouspracticalresults(e.g.,aprofoundimpactonlanguage skills),butalsosomeother,andperhapsmoreimportant,changesintherealmofattitudesand ideasaboutintergroupcoexistence.Wefoundthatourbilingualsubjectswerebetterabletomake friendsintheotherethnolinguisticgroupearlierinlife.Furthermore,thosewithearlyimmersion experienceperceiveFrenchspeakingandEnglishspeakingCanadiansasbeingmoresimilarin termsofhumanqualitiesthanmonolingualsdo.Thismorereceptive,lessethnocentricattitude wasevidentatGrade6andonecouldarguethatitcouldbefurtherdevelopedbyeducatorsand parentsbypromotingoccasionsforintergroupsocialcontactsattheelementaryschoollevel.By Grade11,allgroupssharedacommonviewoftheessentialsimilarityofthetwoethnicgroups. Nonetheless,theolderbilingualchildrenstoodoutintheirmodesofassessingtheproblems facingCanadiansandintheinsightstheyhadforreducingintergrouptensions.Weassumethat earlybilingualbiculturalexperiencespromotesuchattitudesandhelptoopenthemindsofthe bilingualyoungsters.Ourargumentisthatbybecomingfunctionallybilingualintheother group'slanguageandlearningabouttheircultureandvalues,thebilingualseesbeyondthe simplesolutionthatlearningtheothergroup'slanguagewillmakeforintergroupharmony.Their perspectiveismoresophisticated,fortheyseeaneedtolearnabout,togotoschoolwith,andto interactsociallywithmembersoftheothergroupasthecrucialfactorsinpeacefulcoexistence. Ourguessisthatthebilingualsmaybemoresophisticatedinthesolutionstheysuggestfor reducingintergrouptensionsbecauseoftheirknowledgeof,andsensitivityto,theothergroup's pointofview. OVERALLCONCLUSIONS Therearetwointerrelatedissuesdealtwithinthispaper,oneconcernedwiththenatureof attitudesandtheirmeasurement,theotherwithaneducationalinterventionor"treatment"that should,theoretically,improveorchangetheattitudesofyoungpeopleinasociallydesirable direction.Thetreatmentparadigminvolved 216

separatecohortsofAnglophonepupilsatvariousgradelevelsinpublicschoolswhowerein French"immersion"programs."Immersion"pupilswerecomparedwithcarefullymatched Anglophone"control"pupilsfollowingaconventionalEnglishlanguageprogram.Atthestartof theprogramtheparentsoftheimmersionandcontrolpupilswereinsomecasestestedandfound tohavebasicallysimilarpatternsofsocialattitudes;inothercases,wehadtopresumethatthe twoparentalgroupshadequivalentattitudes.Theworkinghypothesiswasthatlearning thoroughlyanotherethnicgroup'slanguageandlearningaboutthatgroupthroughateacherwho isanativeinformantwouldhaveafavorableimpactonpupils'attitudestowardthatgroupandthe associatedculture.Thusweexpectedtofindimmersionversuscontroldifferencesinattitudes, whatwecalledan"immersioneffect."Intestingfortheimmersioneffect,attentionwasgivento theoreticallydifferentcomponentsofattitudes,astereotypecomponent,measuredthrough semanticratingscales,versusafeelingorareactiontendencycomponent,measuredthrough directquestioningandmultidimensionalscalingprocedures.Overall,theactualfindings confirmedourexpectations,butmorethanthat,theyprovidedseveralnewinsightsintotheways attitudesfunction,inparticularhowcertainclustersofattitudescandevelopinsuchawaythat one'sperspectiveonsocietywillbecomemorematureandsophisticated.Itappearsthat becomingbilingualandbiculturalisaneffectivewayofenrichingyoungpeople'ssocial perspectives.Withrespecttopupils'stereotypes,thefollowingfindingsarenoteworthy. 1. Althoughnotpronouncednorconsistent,theimmersionexperiencedidhavestatistically significanteffectsonpupils'stereotypesofFCs:thestereotypesofimmersionpupilswere eitherslightlymorefavorableorlessunfavorablethanthoseofthecontrolpupils,ingeneral. 2. TheratingsassignedtoFCs,however,fellwithinthe4to5pointrangeon7pointscales, indicatingthatAnglophonestereotypesofFCswereneutraltoslightlyfavorableatbest. 3. Theimmersioneffectonstereotypeswasnotnecessarilyanameliorationofperceptionsof theotherethnolinguisticgroup,forjustasoftentheeffectwasanessentiallyneutralviewof FCs,whichnonethelesscontrastedwiththeharsh,sometimesaggressiveviewsexpressedby thecontrolpupils. 4. Thestereotypesofimmersionpupilsreflectedlocaladministrativepoliciesand idiosyncrasies,e.g.,biasestowardEuropeanFrenchpeopleinclasseswheretheteachers wereEuropean;biasesagainstFCsinclasseswherenonrepresentativeFCpupilswere integratedintoimmersionprograms.Stereotypesalsoreflectedbroaderpoliticaltensionsin theprovince. 5. Theimmersionexperiencehadnonegativeordampeningeffectsonpupils'selfviewsor viewsoftheirownethnicgroup. 217

6.Theeffects ofimmersion on stereotypes aremore pronounced atthestartof theprograms (atGrades1 Withrespecttothefeelingandreactiontendencycomponents,thefollowing and2for findingsarenoteworthy. programs 1. Theimmersioneffectismoreevidentinthefeelingsandreactiontendenciesof startingat pupilsthanwasthecaseforstereotypes.Thusevenincaseswheretherewere kindergarten, noimmersioncontroldifferencesinstereotypes(e.g.,atGrades4and5inthe andat 197071testing),therewerelarge,statisticallyreliableimmersioneffectswith Grades7and differentquestioningprocedures.Thus,immersionpupilsweremuchmore 8forthose likelythancontrolstosaytheyhadcometoappreciateandlikeFCs,thatthey startingat couldeasilyimaginethemselvesbeingFC,thattheylikedtheintensivestudy Grade7). ofFrenchandwantedtocontinue,andsoforth. Theeffect 2. Alsoattheothergradelevels,wherethestereotypedifferenceswereless appearsto pronounced,therewereimportantimmersioncontroldifferencesinfeelings wearoffwith andreactiontendenciestowardsFCs.Forinstance,atthetimeofleavinghigh time. school,theimmersionstudentsrelativetocontrolsnotonlyshowedhigh 7.Itisalso degreesofskillinFrenchandaconfidencetowork,studyandliveinmainly possiblethat Frenchsettings,butalsoshowedagreaterwillingnessanddesiretomeetand theless interactwithFCs.Theyalsoshowedmorefrustrationbecausesuchinteraction marked isdifficulttoachieveintheCanadiansociety. differences 3. Weconclude,therefore,thattheimmersionexperiencebroadenedpupils' between "socialperspectives"morethanconventionaleducationdidinthesensethat immersion immersionstudentsaskedmoresearchingquestionsofsociety.Forexample: andcontrol Whyitisthatvarioussocialbarriersexist(religious,ethnic,linguistic)that groupsinthe discourageintergroupcontactsinneighborhoods,schools,andinworkand latergrades socialsettings? isduetoa 4. Thenotionof"socialperspectives"impliesthattheremaybesomehigher gradual orderaspectofattitudesthatisoftenoverlookedinthemoreconventional improvement definitionsofattitudes.Thepresentresearchsuggeststousthatattitudescan overtimeof clusterintohigherorder,generalizedorientationsthatencompassnotonlya the toleranceforanothergroup,butalsoaknowledgeof, stereotypes 218 ofthecontrol pupils.Since allcontrol pupilshad FSLtraining fromGrade1 on,that training couldhave beneficial

effectswhich accumulate overtime.

appreciation for,and interestin peoplefrom thatgroup.A person's "social perspective" wouldbe enhancedby learning enoughabout theother groupandits membersto besensitive totheir values, expectations andwishes, andby comingto realize(a) thattheother peopleare muchlike one'sown peopleand oneself,and (b)thatsocial systemsoften workagainst people gettingto knowand appreciate oneanother. 5 Thedata . fromthe scaling procedures supportedthe notionofa maturationof social perspectives: e.g., immersion

REFERENCES Blake,L.,Lambert,W.E.,Sidoti,N.,Wolfe,D.(1981)."Students'viewsof intergrouptensionsinQuebec.Theeffectsoflanguageimmersionexperience". CanadianJournalofBehavioralScience,13,144160. Christian,J.D.(1976).Psychologicaldifferentiationanddefinitionoftheself: Multidimensionalscalingapproach.Unpublisheddoctoraldissertation,McGill University. Clrment,R.,Gardner,R.C.,&Smythe,P.C.(1977)."Motivationalvariablesin secondlanguageacquisition:AstudyofFrancophoneslearningEnglish". CanadianJournalofBehavioralScience,9,123133. Cziko,G.,Holobow,N.,&Lambert,W.E.(1977,March).Acomparisonofthree elementaryschoolalternativesforlearningFrench:Childrenatgradesfourand five.Unpublishedmanuscript,McGillUniversity,DepartmentofPsychology. Cziko,G.,Holobow,N.,&Lambert,W.E.(1977,April).EarlyandlateFrench immersion:Acomparisonofchildrenatgradeseven.Unpublishedmanuscript, PsychologyDepartment,McGillUniversity. 219

students reducedthe perceived distance betweentheir ethnicgroup andFCs moresothan didthe controls. Likewise,at theearly gradelevels, immersion pupilssaw more similarities betweenFCs andECsthan didthe controls. 6 Finally,the . immersion studentsat thehigh schoollevel showed broader social perspectives whenasked todiagnose andsuggest solutionsfor theproblems thatseparate FCsandECs inCanada today.The immersion experience appearsto have provided themwith matureand productive insightsinto society,e.g.,

that simplistic solutions, likegettingto knowone another's language,are notfinal answers. Theyseemed torealizethat gettingto knowthe other languageand theother culturewell isonlya start,that behindthe other languageand cultureare peoplewho apparently arevery muchlike theirown people,and whoare vulnerableto suspicions, threatsand fearsofother groupsunless real opportunities areprovided bythesociety forequitable groupcontact and interaction, startingat earlyschool years.These insightswere notapparent inthe

thinkingof the monolingual control students,and wesuspect thisisso becausethey hadnotbeen giventhe opportunities tolearnthe other language thoroughly norlearn enoughabout theother group.

Cziko,G.,Holobow,N.,&Lambert,W.E.,&Tucker,G.R.(1977,December).Acomparisonof threeelementaryschoolalternativesforlearningFrench:Childrenatgradesfiveandsix. Unpublishedmanuscript,McGillUniversity,PsychologyDepartment. Cziko,G.,Lambert,W.E.,&Gutter,R.(1979,November)."Theimpactofprogramsina foreignlanguageonpupils'socialattitudes".WorkingPapersinBilingualism(pp.1328). O.I.S.E.,Toronto. Cziko,G.,Lambert,W.E.,Sidoti,N.,&Tucker,G.R.(1980)."Graduatesofearlyimmersion: Retrospectiveviewsofgrade11studentsandtheirparents".InSt.R.Clair&H.Giles(Eds),The socialandpsychologicalcontextsoflanguage.Hillsdale,NJ:LawrenceErlbaumAssociates. Ervin,S.(1954).Identificationandbilingualism.Unpublishedmanuscript,HarvardUniversity. FrasureSmith,N.,Lambert,W.E.,&Taylor,D.M.(1975)."Choosingthelanguageof instructionforone'schildren:AQuebecstudy".JournalofCrossCulturalPsychology,6,131 155. Gardner,R.C.(1981)."Secondlanguagelearning".InR.C.Gardner&R.Kalin(Eds.),A Canadiansocialpsychologyofethnicrelations.Toronto:Methuen. Gardner,R.C.(1982)."Languageattitudesandlanguagelearning".InE.B.Ryan&H.Giles (Eds.),Attitudestowardslanguagevariation.London:EdwardArnold. Gardner,R.C.,Glicksman,L.,&Smythe,P.C.(1978)."Attitudesandbehaviorinsecond languageacquisition:Asocialpsychologicalinterpretation".CanadianPsychologicalReview, 19,173186. Gardner,R.C.,&Lambert,W.E.(1972).Attitudesandmotivationinsecondlanguagelearning. Rowley,MA:NewburyHouse. Gardner,R.C.,&Lambert,W.E.(1959)."Motivationalvariablesinsecondlanguage acquisition".CanadianJournalofPsychology,13,266272. Gardner,R.C.,&Smythe,P.C.(1975).Secondlanguageacquisition:Asocialpsychological approach(ResearchBulletinNo.332).London:UniversityofWesternOntario,Departmentof Psychology. Genesee,F.(1978)."ScholasticeffectsofFrenchimmersion:Anoverviewaftertenyears". Interchange,9,2029. Genesee,F.,&Hamayan,E.(1980)."Individualdifferencesinsecondlanguagelearning". AppliedPsycholinguistics,1,95110. Genesee,F.,Morin,S.,&Allister,T.(1974a).Evaluationofthe197374grade7French immersionclass:June1974.InstructionalServices,ProtestantSchoolBoardofGreaterMontreal. Genesee,F.,Morin,S.,&Allister,T.(1974b).Evaluationofthe197374Pilotgrade11French

immersionclass.AreportsubmittedtotheProtestantSchoolBoardofGreaterMontreal. Genesee,F.,Rogers,R.,&Holobow,N.(1982,July).Thesocialpsychologyofsecondlanguage learning:Anotherpointofview.Unpublishedmanuscript,McGillUniversity. Guindon,H.(1971).Socialunrest,socialclass,andQuebec'sbureaucraticrevolution.InB. Blishenetal.(Eds),Canadiansociety:Sociologicalperspectives.Toronto:Macmillan. Lambert,W.E.(1967)."Thesocialpsychologyofbilingualism".JournalofSocialIssues,23, 91109. Lambert,W.E.,Frankel,H.,&Tucker,G.R.(1966)."Judgingpersonalitythroughspeech:A FrenchCanadianexample".JournalofCommunication,16,305321. Lambert,W.E.,Gardner,R.C.,Barik,H.C.,&Tunstall,K.(1963)."Attitudinalandcognitive aspectsofintensivestudyofasecondlanguage".JournalofAbnormalandSocialPsychology, 66,358368. Lambert,W.W.,&Lambert,W.E.(1973).Socialpsychology.EnglewoodCliffs,NJ: PrenticeHall. Lambert,W.E.,&Tucker,G.R.(1972).Bilingualeducationofchildren:TheSt.Lambertstudy. Rowley,MA:NewburyHouse. 220

Lambert,W.E.,Tucker,G.R.,&d'Anglejan,A.(1973)."Cognitiveandattitudinalconsequences ofbilingualschooling:TheSt.Lambertprojectthroughgradefive".JournalofEducational Psychology,65,141159, Mowrer,O.H.(1950).Learningtheoryandpersonalitydynamics.NewYork:RonaldPress. Swain,M.(1974)."FrenchimmersionprogramsacrossCanada".CanadianModernLanguage Review.31,117128. Taylor,D.M.(1976)."Ethnicidentity:Somecrossculturalcomparisons".InJ.W.Berry&W.J. Lonner(Eds.),Appliedcrossculturalpsychology.Amsterdam:Swets&Zeitlinger. Taylor,D.M.,Meynard,R.,&Rheault,E.(1977)."Threattoethnicidentityandsecond languagelearning".InH.Giles(Ed.),Language,ethnicityandintergrouprelations.London: AcademicPress. Wolfe,D.,&Lambert,W.E.(1979).Graduatesofearlyimmersion:Afollowupstudy. Unpublishedmanuscript,McGillUniversity. 221

[Thispageintentionallyleftblank.] 222

1 12 ASocialCognitivePersectiveonBilingualism:CommentsonLambertandTaylor

E.ToryHigginsNewYorkUniversity ItisapleasureformetobediscussantonthesetwochaptersbecauseIwasborninMontreal,was anundergraduateatMcGill,and,mostofall,becausethefirstsocialpsychologycourseIever tookwaswithW.E.Lambert.Thatcoursewasamajorinfluenceonmyinterestinsocial psychologyand,moregenerally,onmyinterestintherelationamonglanguage,thoughtand society.Soitisfunformetobeinvolvedindiscussingthesechapters.Also,itisfunbecauseof theimpactoftheworkofbothLambertandDonTayloronsecondlanguagelearningafterall, howmanypeoplehavedonesuchimportantresearchthattheynamewholeregionsafterthem,as inWally'scasewithSt.Lambert? InmycommentsItalkabouttwoissuesmotivationalfactorsintheacquisitionofasecond languageandtheconsequencesofbilingualismandmakesomegeneralcommentsabouteach ofthem,relatingtowhathasbeensaidsofarandtosomeideasofmyown. First,withrespecttomotivationalfactorsintheacquisitionofasecondlanguage,Lambertand Taylorhavecertainlydonethepioneeringworkinthisarea,alongwiththeircolleagues.What interestsmeistheconsiderationofthisissueintermsofamodelthatIhavebeendevelopingthat derivesfrommanyareasofsocialsciencepragmatics,anthropology,sociolinguistics,speech andcommunication,andsoon.ItisamodelofcommunicationthatIrefertoasthe "communicationgame,"whichissimplyastealfromWittgenstein's"languagegame,"whowas oneofthefirsttotalkintheseterms.Oneimportantaspectof ____________________ 1Thispaperisatranscript(withsomeminormodifications)oftheauthor'sdiscussionduring theconferenceofthepaperspresentedbyLambertandTaylor,withsomeofLambert's remarksontheauthor'scommentsbeingincludedaswell. 223

thismodelisitsemphasisonthefactthatcommunicationhasavarietyofgoals.Thatwould seemobvious.Butinmanyareasofpsychology,andespeciallyinexperimentalsocial psychology,thediscussionofcommunicationhascenteredaroundonlyonegoalinformation transmission.Certainlythatisoneimportantgoalofcommunication,butthereareavariety ofgoalsincommunication. Anotherimportantsetofgoalsincommunicationissocialrelationshipgoals,theideathatpeople communicateinordertoformormaintainarelationshipwithsomeoneorsomegroup.Butin additiontosocialrelationshipgoals,there'swhatGoffmanwouldcall"face"goalsorimpression managementgoals,theideathatwecommunicateinordertomakeourselveslookgoodandtoget otherstomakefavorablejudgmentsofus.Inadditiontofacegoals,thereare"socialreality"goals. Weoftencommunicatejusttoobtainconsensusonwhateveryonethinksaboutsomethingsowecan knowwhat"reality"is.Oneoftheclassicexamples,ofcourse,isSchachter'searlyworkonthe psychologyofaffiliationwherehefoundthatwhenpeopleareanxiousandtheyhaveachoiceof whotowaitwith,theyprefertowaitwithotherpeoplethantowaitalonebutonlyifthoseother peoplehaveexperiencedthesameeventandonlyiftheyareallowedtotalk.Sopeoplehavea motivationtobeinacommunicationsituationwheretheycansharesomeexperiencetogetsome understandingofwhatitmeans. Thereisanothercommunicationgoalthatisn'ttalkedaboutverymuchandhasn'tbeentalkedabout muchinthisconference.Thatgoalisentertainment.Anotherreasonwhypeoplecommunicateis becauseitisfunforthesheerpleasureofit.Partlybecauseitissomethingelsewecando,just likewecandanceorsingorskiporwhateverwecantalk.Infact,itisoneofthethingswemost enjoydoing. Distinguishingamongthesedifferentkindsofgoalshasalonghistory.Andyettheframeworkthat LambertandTaylorhavebeenworkingwith,Ibelieve,enrichesthisnotionofa"communication game."Ialsothinkthisnotionmightenrichtheirframework.Thetaskgoalisverysimilartowhat theycallthe"instrumental"goal.Thesocialrelationshipgoalsaresimilartotheir"integrative" goals.Thesetwocasesarefairlystraightforwardmappingsbetweenthesetwomodels.Whatis interestingishowthethirdvariable,socialorethnicidentity,fitsin.Ithinkithasfeaturesofboth facegoalsandsocialrealitygoals,aswellassocialrelationshipgoals.Perhapsitisbecauseofits multifunctionalnaturethatsocialidentityissoimportant. Nowitmaybethatsocialorethnicidentitymeanssomethinginadditiontothesegoals,butIfeelit wouldbeinterestingtothinkaboutethnicidentityintermsofhowitrelatestoeachofthesegoals, suchas"face"goals.Ithink,infact,thatoneaspectofthenegativemotivationfromethnicidentity worksinthisway.Thatis,somepeoplemaybemotivatednottoacquireasecondlanguagebecause theydon'twanttobeembarrassedbypoorlyspeakingthelanguageofthat"other"group.Social realitygoalsmayalsocontributetonegativemotivation.Considerthecasewhereaminoritygroup istryingtolearnasecond 224

language.Theymightfeelthatthey'regoingtolosetheirsenseofsocialrealitybylearninga secondlanguage.Thethreattoethnicidentityisinpartthethreatoflossofsocialrealitytheir socialreality. Entertainmentgoalshaven'treallybeenlookedatintermsofthismodel.Thereisacomponentof "integrativegoals"thatseemstosuggestentertainmentthatyou'resortofenjoyingit.Butthere's anaspectofentertainmentthatappearstobeleftoutofthemodelatthispointtheideathat justasitisfuntospeakyourlanguageitisfuntospeakanotherlanguage. (WallyLambert:Thatusuallygoesintotheintegrativepart).Yes.Butitmightbeinteresting fromthatpointofviewtosubdivideintegrativeintosocialrelationshipgoalsandentertainment goalstoseewhatkindofimpactthatwouldhave. Letmeswitchgearsnowandtalkaboutthesecondissuethatcameupinbothtalks,aswellasin earliertalkstodaytheconsequencesofbilingualism.IespeciallyenjoyedWally'sdiscussionof socialaffectiveconsequencesofbilingualismbecauseoneofthepointsIwantedtomakewas thattheliteratureonbilingualismseemedtoemphasizecognitiveratherthansocial consequences,includingWally'searlierpioneeringwork.Ithinkthatitwasaproblemwhen peoplesaidthatlearningasecondlanguagehaspositiveconsequencesaswedidnotknowwhat thesocialconsequenceswere.Andjustasitwasmistakeninthefiftiestosaythatlearninga secondlanguagewasabadthing,tosayitisagoodthingintheeightieswithoutknowingwhat thesocialconsequencesareisjumpingthegun.Weneedtoknowmoreaboutthesocial consequences,whichwasWally'spointandIsecondthatstrongly.Ithinkitisimportantto collectdataonsocialconsequencesinthesenseofattitudes,butIamnotsurethatissufficient. Otherkindsofsocialconsequencesaresocialcognitiveconsequencesconsequencesfor makingjudgmentsofotherpeopleandconsequencesforone'smemoryaboutotherpeople. LetmegiveyoualittlebackgroundontheideathatIwishtodiscuss.Whenyouconsidersocial concepts,especiallytraitwordslikefriendly,aggressive,independent,aloof,smart,andsoon, onemustdistinguishbetweenthemeaningofthewordandthereferent.Nowitmayseem ridiculousformetomakethatpointwhenOgdenandRichardsmadeitmanyyearsago.Butitis surprisinghowoftenweforgettheimplicationsofmaintainingthisdistinction.WhatIamgetting atisthatthereisadifferencebetweenthemeaningofawordandtheactualsocialeventthatwill cometomindwhenyouusetheword.Thisdifferenceisdemonstratedinmyrecentresearch wherethemeaningofawordstaysthesame,butbecausethestandardchanges,thereferentis different. LetmegiveanexampleofwhatImeanbythat.Takeforinstancethewordfriendly.Themeaning ofthewordfriendlyisliterally"morefriendlythanthestandard."Whenyousay"Tomis friendly,"thatisacomparativeassertionandthemeaningoftheassertionis"Tomismore friendlythantheaverageperson,"or"Tomismorefriendlythansomestandard."Nowthe referentialmeaningis 225

goingtodependthereforeonthestandardthatyouhaveforfriendliness.Letussay,forinstance, thatyouare2yearsofage.Youhaven'tmuchexposuretopeople.Tenyearsgobyandthe distributionofinstancesoffriendlinesschanges.Nowyouhaveadifferentstandard.Thepointis thatthemeaningofthewordhasn'tchanged.Themeaningisstill"morefriendlythanthe standard."Butwhathaschangedisthestandard. IfyouseeTomperformcertainbehaviors,yousay:"Oh,Tomisfriendlybecauseheismore friendlythanthestandard."Nowyoudon'tseehimfor10years.Yourememberthathe'sfriendly. "Ohyeah,Tomisfriendly,"Thenyousaytoyourself,"Gee,whywashefriendly?"Whatyoudo isreconstructTom'sbehaviortobeconsistentwiththatlabel.ThisisaclassicBartletttype phenomenon,andwhatyougetisdistortionovertimeifthestandardhaschangedbetweenthe timethatthebehaviorwaslabeledandthetimethatthelabelisusedtorecallthebehavior.This isthekindofresearchthatI'vebeendoing,wherethemeaningofthewordstaysthesamebut onemanipulatesthestandard,whichdoescausemajorerrorsinmemory. Theimplicationsofthisprocess,Ibelieve,arequiteinterestingforbilingualismandespecially forthekindofbilingualismthatisinvolvedinbiculturalism.Infact,itmayprovideawayof distinguishingbetweenthebiculturalversusuniculturalbilingual.ThesituationIaminterestedin isexactlythesituationwhereyoumayfindabiculturalbilingual,forexample,anAmerican Spanishbilingualwhohasatraittermlikefriendlywherethemeaningofthewordisthesamein bothEnglishandSpanish,"morefriendlythantheaverageperson,"butthestandardof "friendliness"isdifferent. (Lambert:Ithinkthere'swhereyoucangetsomechallenge.Itisfascinatingbecausefriendship inFrenchtakesonadifferentform,youknow.Weuse"friend"looselyinEnglish,apparently, andtheFrenchrestrict"friend"toaveryselect,smallgroup.Thus,inSpanishyoumayhavethe same.Soyouhavegottobecarefulabouttheidentityofwordsbetweenlanguages.) Yes!ButI'mgoingtorestrictmyselftocomparativetraitassertionsbecausewhenyouaretalking aboutspecificobjectsthemeaningcanbedifferentsincetheoptionalfeaturesofwhatpeople haveinmindmaybedifferent.Thecaseofcomparativeassertionsisbetterbecausethereare manylanguages,especiallyRomancelanguages,wherecomparativetraitassertionsmeanthe samething.Evenifitisrestrictedtothat,itwouldbeimportantbecausealltraittermsare comparativetraitassertionsandthat'swhatweusewhenwetalkaboutpeople.So,evenifitonly hadtodowithpersonalitytraitsitwouldhavesignificantimplicationsforourimpressionsof people,ourattitudes,andoursocialmemory. NowthesituationIamthinkingofisacasewherethetwowordshavethesamelinguistic meaning,"greaterthanthestandard,"butbecauseitisbicultural,thestandardmightbevery differentinthetwocultures.TheaveragefriendlypersoninSpanishculturemaybevery differentfromtheaveragefriendlypersoninEnglishculture.Myquestionisthiswhathappens whenyougeta 226

childwhoisthekindofbiculturalbilingualwehavebeentalkingabout,whospeaksone languageathomeandonelanguageatschool?Whathappenswhenthischildgoestoschool, meetssomebodyforthefirsttimeandseesasetofbehaviors.ThinkoftheSpanishchildwhois goingtoanEnglishschool.ThesearechildrenwhospeakSpanishathomebutspeakonly Englishatschool.Thinkoftheboywhogoestoschoolandbecomesbilingualinthefirsttwoor threeyears,meetssomeone,Tom,hasarelationshipwithhim,anddecidesthatTomisfriendly inEnglish.Thenhegoeshome.NowheistalkingSpanishandhismothersays: "Didyouhaveaniceday?" "Oh,ImetthisboynamedTorn"spokeninSpanish. "Whatishelike?" "Tomisfriendly." WhatIamsuggestingisthatinordertosay"Tomisfriendly"inSpanishheisactuallydistorting hismemoryofthebehaviorthatheobservedwhenhecalledhimfriendlyinEnglish,becausethe wordshavethesamemeaningbutthestandardsaredifferent.Youactuallyhavetodistortyour memoryforittofit.Thenhegoesbacktoschool.NowhehasadistortedmemoryofTom.He meetsTomagain,Tombehavesinthesamefriendlymannerbutappearstobehaveinadifferent mannerbecauseitisdifferentfromthestoredmemory.Sohere'sachildwhoperceivesan inconsistencyinTom'sbehavior. Ihaven'tcollectedanydataonthisyet.That'sarealadvantageatthispointbecauseIcancreate anyscenarioIwant,andIwanttopointupthefactthatwedon'tknow,withoutdataofthis nature,whetherthesocialcognitiveconsequencesofbilingualismarenegativeorpositive.The negativescenarioistheoneIhavejustdescribed,whichisthatthechildperceivesan inconsistencyinTom'sbehavior.Whatdoesinconsistencycause?Well,weknowthat inconsistencycausesnegativeaffect,uncertainty,asenseofunpredictability,asenseoflackof control. Nowletusconsiderthepositivescenario.Thepositivescenarioisthatthechild,bythefactthat hehasthesedifferentstandards,developsanappreciationforvariability.Hestartsnoticingall kindsofvariabilityamongpeople.Well,whatdoesvariabilitycause?Weknowfromtheresearch oningroupoutgroupperceptionbyTajfelandothers,andIhavedonesomeworkonthisaswell, thatifyouperceivevariability,youbecomelessstereotyped.Infact,thatiswhatastereotypeis itisaperceptionoflackofvariability.So,ifyouperceivevariability,youbecomeless stereotyped.AndIknowfrommyownworkthatifyouperceivevariabilityyoudon'tassumethat othersaredoingthingsbecausethey'reblackorbecausetheyareSpanishorbecausetheyare whiteoranythingelse.Sowhatwehaveisareductioninstereotypedjudgmentsofothers. And,evenmoredirectly,ifthischildhasmorethanonestandardthenthereisareductionin ethnocentricismbecauseoneofthedefinitionsofethnocentricismisthatonemakesjudgments intermsofasinglestandard;namely,thestandard 227

ofone'sculture.AsRogerBrownsaidyearsagoin"WordsandThings,"theworstthingabout ethnocentricismisthatyoualwaysmakejudgmentsintermsofonestandard,yourgroup's standard,andassumethatistheonlystandard.Thatisthedanger.Ifbiculturalbilingualsuse morethanonestandard,theymaybelessethnocentricandthatisthepositivescenario.Butthe pointisthatonlyresearchwilltell. 228

V BIDIALECTISM 229 [Thispageintentionallyleftblank.] 230

13 TheLinguisticandSociolinguisticPositionofBlackEnglishandtheIssueofBidialectalismin Education JohnD.RoyBrooklynCollege,CityUniversityofNewYork ThelinguisticandsociolinguisticfactorsthatsurroundedtheimportationofAfricansasslavesare notparalleledintheimmigrationofothergroups.Theimportationofpeopleswithdifferentnative languagesandthegroupingofthemtogetherinasocialstratumstructurallyremovedfromEnglish resultedintheformationofanemergencylanguageaWestAfricanEnglishPidgin. ThisemergencylanguagedevelopedoutoftheearliestcontactofAfricanswithEnglishspeakers. Spencer(1971)citesevidencethatestablishesthatthisemergencyAfroEnglishPidginwasinuse inWestAfricabarely20yearsaftertheestablishmentofthefirstBritishfortthereandwasthe mediumofcommunicationintheEnglishslavetrade. ThegreatnumberofWestAfricanlanguages,andthepracticeofpurposefullyvaryingtheloading pointsinordertomixthelanguagegroupsandthuslessenthechanceofshipboardinsurrection,are twoofthefactsthatfavoredthedevelopmentanduseofaWestAfricanEnglishemergency languagefunctioningasameansofcommunicationbothbetweentheEnglishandtheAfricans,as wellasbetweenAfricansofdifferentlanguagegroups.Cruickshank(1916)citesWilliamSmith, SurveyortotheRoyalAfricanCompanywhomadeavoyagetoGuineain1726.Henotes: ThelanguagesintheGambiaaresomanyandsodifferentthatthenativesononeside [oftheriver]cannotunderstandthoseontheother,whichisagreatadvantagetothe Europeans,whotradethereforSlaves....Thesafestwayistotradewithdifferent nationsoneitherside,andhavingsomeofeverysortonboardtherewillbelessthe dangerofanyplot.(p.1) 231

ThisAfroEnglishemergencylanguagewasalsothemediumofcommunicationonthe plantationsofAmerica.NewlyarrivedAfricanslearned,primarilyfromfellowslaves,the mediumofcommunicationthatwasusedinthenewworldbetweenAfricanswhodidn'tsharea commonlanguage,aswellasbetweenmasterandslave(Dillard,1972). ThesocialandculturalcontextofancestralAfricanlanguageswasreplacedbytheenvironment oftheplantationanditssystemofcommunication.Asaresult,thechildrenoftheWestAfricans learnedtheemergencyhybridlanguagesystemthatcombinedEnglishwordswithelementsof WestAfricanlanguagestructureratherthantheancestralAfricanlanguagesoftheirparentsor thesociallydistantEnglish.Theemergencylanguagesystemthenbecamethenativelanguageof thesefirstgenerationAfroAmericans.Anativelanguagewhichdevelopsinthiswayisdefined asacreole(Hall,1966). ThePlantationEnglishCreolethatdevelopedinAmericawasthecarrieroftheideas,actionsand valuesoftheslavesociety.Itwasacoherentnaturallanguagesystemnotverydifferentin structurefromtheEnglishCreolesthatcanbeheardtodayinruralWestIndianvillagesandin partsoftheSeaIslandsofSouthCarolinaandGeorgia. EnglishCreoleexistedasthelowstatuslanguageina'diglossic'(Ferguson,1959)pairwith English.Contactwiththehighstatuslanguagevariety,English,producedaprocessthrough whichstigmatized,lowstatuslinguisticfeaturesofEnglishCreolebegantogivewaytoEnglish formsashypothesizedbyEnglishCreolespeakers.Varioussegmentsofslavesocietyhad differentlanguageneedsanddifferentaccesstoEnglish.Houseservantshadmoreexposureto Englishthandidfieldhands;towndwellersmoreexposurethandidtheirruralcounterparts.Over time,generationshadmoreandmorecontactwithEnglishthroughlocation,employmentand educationandmoreofthestigmatizedfeaturesweredroppedandmoreoftheforms hypothesizedasEnglishwereadded.Successivegenerationstransmittedtotheirchildrenaless marked,creolesystemexceptinthoseareaswheretherewasonlyminimalcontactwithEnglish. Thisprocessoflanguagechange,aprocessofdifferentiallinguisticacculturationtermed decreolization,isresponsibleforthemovementofPlantationEnglishCreoletowardEnglishand fortherangeofsocial,regionalandindividualdialectsthathavebeendescribedasBlackEnglish (BE)andBlackEnglishVernacularthatcanbeheardintheurbanandruralBlackcommunities today(Roy,1984). Thefabricofthesociolinguisticsituationjustdescribedhasnotchangedsignificantlyformuch oftheBlackpopulation.BlackEnglish,likeEnglishCreole,remainsthelowstatusdialectused inBlackcommunitiesthatremainataconsiderablesocialdistancefromEnglish.Processesof linguisticchangesimilartothosethatonceoccurredintransformationofEnglishCreoleinto BlackEnglisharenowaffectingBlackEnglishasitcontinuestoconvergeonEnglish.These processesinvolvetheacquisitionofStandardEnglish(SE)featuresaswell 232

astheavoidanceofcertainstigmatizedBlackEnglishfeaturesandtheattemptbyparentsto transmita"better"Englishtotheirchildren. Tothosefamiliaronlywithclearlybilingualsituations,thechangefromEnglishCreoletoBlack EnglishtoEnglishmayseemaninordinatelylengthyandpainfulprocess.Successivegenerationsof nonEnglishspeakingimmigrantgroupsgaincompetenceinEnglishrelativelyquicklyeither throughthedevelopmentofbilingualismortheadoptionofEnglishastheironlylanguage.The reasonsfortheslowpaceofthelinguisticacculturationofthedescendantsofAfricanimmigrants arebothlinguisticandsociolinguistic. ImmigrantgroupsthatapproachEnglishfromaclearlydifferentfirstlanguagewithrareexceptions havenotpassedthroughthepidginization,creolizationanddecreolizationprocessesthatare responsibleforthewiderangeoflanguageformsthatarepresentincommunitiesusingBlack English.Inthebilingualsituation,therearecleardifferencesbetweenthevocabularyandthe grammarofthenativelanguageandthetargetlanguage,English.Thisisnotthecaseinthe relationshipbetweenEnglishCreoleandEnglish,ontheonehand,orbetweenBlackEnglishand StandardEnglishontheother.EnglishCreoleandBlackEnglisharebothformedoutofthelexical resourcesofEnglish,buttheselexicalelementsareembeddedwithinhistoricallydifferent grammaticalsystems,manyofthelanguagesubsystemsofwhicharestructurallydifferentfrom English.Becauseofthesimilarityofthevocabularyandelementsofthegrammar,thesystematic differencesbetweenthenativelylearnedvarietyofBlackEnglishandtheStandardEnglishtarget areobscured(Roy,1977). ThelinguisticresourcesavailabletothechildwithinaBlackEnglishspeakingcommunityduring theperiodoflanguageacquisitionmaycontainavarietyoflanguageformsthatrangefromdeep BlackEnglishtoStandardEnglish.Becauseofthesurfacesimilarityandtheopacityofthe structuraldifferencesofthedifferentlanguagevarietiesintheirenvironment,childreninaBlack English/StandardEnglishenvironmentcannotclearlyidentifythesystematicdifferencesbetween thelowandthehighstatuslanguagevarietiespresentaroundthem.BrownandBellugi(1964) notedthatadults,inechoingtheirchildren'sutterances,supplygrammaticalmarkers.Thisprocess of"imitationwithexpansion"provideschildrenwithmodelsforthedevelopmentofgrammarfrom abaseformofcontentwordstoincludetherelationalconceptsandcategoriesoftenseaspect, number,possession,deixis,reference,locationandsoon.Ifintheperiodoflanguageacquisition childrenareexposedtononstandardEnglishgrammaticalmodelsthat,atthelevelofsystematic identityareforthechildren,indistinguishablefromtheStandardEnglishforms,theywillacquire themandbuildagrammarthatcontainsthefeaturesofthevarietyofBlackEnglishthatisclosestto them. WhenspeakersofnonstandardBlackEnglishbecomeawareofthelowsocialvalueofBlack Englishformsincertainpublicorformalsituations,theyseekto 233

avoidtheseformsandtoproducethoseofthehighstatuslanguagevariety.Thewidevarietyof expressionthathasemergedwithinBlackEnglishmaybetheresultoftheavoidanceof stigmatizedforms,aswellasthestructuralinterference(Weinreich,1953)ofBlackEnglish ontheproductionoftheformsofStandardEnglish.BlackEnglishformsthatspeakersmay wanttoavoidmaybestructurallyimportanttotheirgrammars,andthetargetedStandard Englishformsthatspeakersmaywanttoproducemayinterferestructurallywiththeir grammaticalsystems,ormaynotbefullywithintheircontrol.Theresultofattemptsby individualstoaltertheirgrammaticalpatterns,eithertemporarily(torespondtocertain situations)orpermanently,hasbeentheproductionofconsiderablelinguistic hypercorrection.andinnovation,someofwhichbecomesaregularpartofthelanguage varietyusedbytheindividualandcommunity.Theseinnovations,neitherStandardEnglish northeBlackEnglishformstheyweredevelopedtoreplace,formyetanotherlayerobscuring thestructuraldifferencesbetweenthegrammaticalsystemsofthenativeandthetarget languagevarieties. ThehighsocialdistancebetweensomeBlackEnglishspeakersandthecommunityofStandard EnglishspeakerscertainlyaffectsboththeavailabilityofSEgrammaticalmodelsandmayaffectthe attitudesoftheseconddialectlearnerinsuchawayastorestricttheprocessofseconddialect acquisition,asSchuman(1973)demonstratesitdoesinadultsecondlanguagelearning.The linguisticandsociolinguisticfactssurroundingtheimmigrationofAfricanstotheNewWorldand thelanguageexperienceoftheirdescendantsillustratethatthereasonsfortheapparentslownessof thechangefromEnglishCreoletoBlackEnglishtoEnglisharebasedinthecomplexitiesofthe continuingprocessesoflinguisticacculturation. ToexaminethesociolinguisticandeducationalparametersthatsurroundtheissueofBlack English/StandardEnglishbidialectalism,somelinguisticpointsmustfirstbeestablished.Thefirst ofthesepointsisthateachhumanlanguagevarietyhasanextremelycomplexrulegoverned, interdependent'systemofsystems'thatenableshumanstoproduceandinterpretthemeaningsand thesignalsoftheirlanguage.Itcannotbesaidthatanyonelanguagevarietyis,intermsofits structure,anybetterthananyother.Thisdoesnotmeanthatalanguagethatarticulatesthecultural experienceofabandofhuntersandgatherersdeepinAmazoniahasthevocabularytodealwithlife inNewYork.Butneitheristhereversetrue.Humanlanguagesmeettheneedsofthecultural experienceofthesocialgroupthatusesthem. Thesecondpointisthatlanguagevarietiesthatcoexistwithinthesameenvironmentmayhave differentsocialvalues,particularlyifonevarietyisusedasamediumofwidercommunication.The languagevarietythathasthehighersocialvalueiscalleda"Language"andthelanguagevariety withthelowersocialvalueisinpopularparlancecalleda"dialect."Ithasbeensaidwithonlyslight flippancythataLanguageisadialectwithanarmy. 234

ThelanguagevarietiesknownasBlackEnglisharenotdialectsintheordinarysense,inthat unlikeregionaldialects,theyhavenotdivergedfromacommonlanguageform,butratherare convergingonEnglishfromthestructurallydifferentlanguagesystemofEnglishCreole.The differencesbetweenEnglishandBlackEnglisharestructurallyimportant,involvingthetense aspectsystemofverbs,thesystemsofpluralizationandpossessionandassociation,thesystems ofpronounsandcomparatives,relativeclauseandquestionformationandnegativization.These areimportantstructuraldifferencesratherthanthesurfacephoneticfeaturesofaccentthatmark thedifferencebetweenoneregionaldialectandanother(Roy,1986,inpress). ThereisawiderangeofdialectaldifferencesbetweenvarietiesofBlackEnglish,someofwhich arefartherfromthelinguisticsystemsofStandardEnglishthanothers.Speakersalsohave differingabilitiesinswitchingfromBlackEnglishformstowardStandardEnglishincontexts judgedappropriate.Thediversitywithinthecommunitiesisanimportantcompoundingfactorin theprocessesofBlackEnglish/StandardEnglishbidialectalism. ItispossibleforchildrentobebornintoacommunitythatincludesBlackEnglishasan importantpartofitslinguisticrepertoireandforthemtohearandperhapsacquireaspartoftheir communicativecompetencesuchavarietyofformsas"Hehere";"Hebehere";"Hebeshere"; "Hedoesbehere";"Hearehere";and"Heishere."Itisimpossibleforchildrentoknowduring acquisitionwhichoftheseformsarepartofthelanguagesystemassignedahighsocialvalueand whicharenot.Itisimpossibleforchildrentorecognizethelimitsoftheconverginggrammatical systemsflowingpastthem.Theboundariesbetweenthesystemsareopaque,asiseventhefact thattherearedifferentlanguagesystemsinvolved. Thisinterandintradialectvariationistheresultoflanguagechangeinprogressanditproduces asituationthatisatoncebothmoredifficultandlessrewardingthanthatwhichfacesthesecond languagelearner.Itislessrewardingforthestudent,asDennisCraig(1968)haspointedout, becausethesharedlexiconandtheopacityofthesystematicdifferencesmeanthatthebidialectal studentdoesnotreceivetherewardsoflearningsomething"new"asthebilingualstudentdoes.It ismoredifficultbecauselearnersandteachersseldomrecognizeorarepreparedtodealwiththe systematicdifferencesbetweenthegrammarofthestudentandthegrammarofStandardEnglish. AlthoughtheremaybeparentalandeducationalpressuretospeakStandardEnglish,thereislittle effectivehelpavailablefromthesesources;alsotheremaybeconsiderablepeerpressurenotto abandonthelanguageofthegroupinfavorofthelanguageofanoutsidegroupseenbysomeas theoppressor(Roy,1981). Itisonreachingschoolagethatchildrenlearnthatsomeofthelanguageformsacquiredfrom parentsandpeers,thelanguageoffirstexperienceanddiscoveryis,insomeunexplainedway, considered"bad"byaneducational 235

systemthat,ingeneral,lackstheabilityand,insomecases,thedesiretoprovidethemwiththe languagevarietythatis"good." ChildrenwhocomefromhomesandcommunitieswherenonstandardBlackEnglishisspoken canspeak,hear,orreadandinterpretthesentence"Theydrivearedcar"tomeanthatonceinthe pasttheydrovearedcar.ThisinterpretationfollowsagrammaticalrulepresentinmostBlack Englishvarietiesthatunmarkedactionverbssignalasinglepastevent.Teachersandbooks, however,interpretandproducesentenceslike"Theydrivearedcar"tomeanahabitualorusual occurrence.Iftheparentorteacherisaskedaboutthetimemeaning,mostmaystatethat"They drivearedcar"indicatesthepresent.Similarly,thesentence"Hedoesdrive"isfortheteacheran emphaticassertionofability,whereasforthestudent"Hedoesdrive"mayexpressmerelya habitualoccurrence. TheseexamplesillustratethatthesystematicdifferencesbetweenthedialectsofBlackEnglish andStandardEnglishcangenerateidenticalsurfaceformsandthatthesesurfaceformshave differentmeaningsdependingonthegrammaticalsystemthatinterpretsthem.WilliamStewart (inunpublishedremarks)referstothisphenomenonas"pseudocomprehension"becausethe speakersmaynotbeawarethattherehasbeenaslipinthecommunicationbetweenthem. Teachersforthemostpartdonotgetaroundtoexplainingwhat"Theydriveacar"meansnordo theygetaroundtoexplainingthedifferencebetween"Hedroveacar"and"Hehasdrivenacar." Theydonotexplainthetenseaspectsystem,orthepassiveortheuseofquantifiersand negativesorrelativeclauseformationbecausetheydonotthemselvesunderstandthesethings despitebeingexpertspeakersandwritersoftheStandardlanguage.Forlikeotherspeakersof StandardEnglish,theycanrecognizethatsomethingiswronginasentencebuthavelittleideaof howwhatisrightgetstobethatway. Limitedbytrainingandknowledgetoteachingthetechniquesofanalysisandelaborationandto thecorrectionoferror,thetechniquesofthetraditionalEnglishclass,teachersencountera studentpopulationwhosedialectisstructurallydifferentfromEnglishandareunabletoteach whatisneeded.Thetechniqueofcorrectionisheavilyreliedon,bysome,inthehopethatthe correctionoferrorswillenablestudentstolearnthesystemofcategoriesandmarkingofthe Englishverbevenifteachersthemselvesdonotconsciouslyunderstandthem.Thereisperhaps thehopeamongsomethatifstudentsmakeenougherrorstheywilllearnwhatisright.However, spotcorrectionsarenotadequatetoilluminateortoteachtargetformsthatarepartofahighly systematiclanguagestructurethatdiffersincomplexwaysfromthehighlysystematiclanguage structurethatisthestudentsnonstandarddialect.Theresultoftheseeffortsatcorrection,on bothsides,isfrustrationandcapitulation. ContrasttheanxiousseriousnessofthelowerclassBlackstudentsinthefirstandsecondgrades whohavecometoschoolinspiredbythehopeforsuccessofparentsandfamilywiththevisible resentfulness,theresultofadjustmenttoeducationalfailure,ofmostoftheirbrothersandsisters injuniorhighandhigh 236

schoolandyouwillseetheextentofthetragedy.Thefailureoftheeducationalestablishmentto dealeffectivelywithlanguagedifferencesresultsintheappallinglevelsofreadingfailureand incompetenceinthewritingofStandardEnglishthatcharacterizeschoolsthatservestudents whoareinfluencedbyBlackEnglish.Atthelevelofnationalpolicy,thefailureoftheEnglish teachingestablishmenttakesamorebizarreform(seeStewart1974b).Thecurrentpolicyofthe collegewritingarmoftheNationalConferenceoftheTeachersofEnglish(NCTE)wassetforth inaresolutioninCollegeCompositionandCommunication(CCCC,1974).Thisresolutionis entitled"Students'RighttoTheirOwnLanguage"(SROL),andstatesthattheteachingof StandardEnglishtospeakersofnonstandarddialectsisnotonlyunnecessarybutisactually "immoral"(CCCC,p.3).Tosupportthisclaim,theresolutionassertsthat"Languagescholars longagodeniedthatthemythofastandardAmericandialecthasanyvalidity"(CCCC,p. 3).Thisassertionignoresthefactthatalladvancedsocieties,atleastfromthetimeoftheGreeks, havehadstandardlanguages.Itrunscounternotonlytocommonsensebutalsototheopinions oflinguistsspecializinginEnglish.Hallidayetal.(1964)statethegenerallyheldposition regardingstandardAmericanEnglish:Hallidayetal.(1964)presentthegenerallyheldposition thatstandardAmericanEnglishisanormthatreferstothoseformsofinflectionalmorphology andsyntaxwhichareusedwidelybyeducatedAmericanspeakersandthatdiffersonlyinminor detailsfromthenormacceptedelsewhereintheEnglishspeakingworld.This"myth"ofStandard English,theSROLresolutionasserts,wascreatedandisperpetuatedin"anattemptofonesocial grouptoexertitsdominanceoveranother"(CCCC,p.3).TheteachingofStandardEnglishis notamatterofculturalimperialism;rather,standardlanguagesareaproductoftheneedsofa complexsocietytocommunicatepreciselyoverwideareas.Itisimportanttorememberthatthere isnothinginherentlybetteraboutthestandarddialectaslanguage:itismerelythestandard system.Toexemplifytheclaim(CCCC,p.6)thatSEisonlytriviallydifferentfromnonstandard dialect,theresolutionpresentsthefollowingsentences: 1. Mary'sdaddyisathome. 2. Mary'sdaddyistohome. 3. Marydaddyhome. Theresolutionincorrectlyassertsthat"Preferenceforoneformoveranother,then,isnotbased onmeaningorevenexactnessofexpression,butdependsonsocialattitudesandculturalnorms." 237

However,fromthepointofviewofboththespeakeraswellasthehearer,thedifferencesarenot trivial.Therangeofpossiblemeaningsofsentence(3)withinBlackEnglishincludes: a. Mary'sdaddyishome. b. Mary'sdaddy'shome. c. Mary,daddyishome. d. Mary,isdaddyhome? Theindividualwhoproducessentence(3)maynotcontroleitherthephonologicalorthe syntacticprocessesinvolvingwordfinal's,whichisusedinSEforthemarkingofthenontrivial systemsofpluralization,tense,possession,andthecontractionsof"is,was,has."TheStandard Englishspeaker,hearingorreadingsentence(3)wouldprobablysaythatthesentenceis ungrammaticalEnglishandthatthemeaningcannotbearrivedatwithcertaintyusingthe grammaticalprocessesknownandacceptedthroughouttheEnglishspeakingworld. Thedifferencebetweensentences(1)and(2)centersonthemeaningdifferencebetween"at"and "to."Arethesetrivialdifferencesofmeaning?AreAmericanswillingtolosethedistinction between"Getthatatthebank."and"Getthattothebank."merelybecausetheSROLresolution assertsthatthedistinctionbetween"at"and"to"isbasedonanattemptbyonesocialgroupto dominateothers?Theclaimthatthedistinctionsbetweenthesegrammaticalmarkersdoesnot effect"meaning,"andexactcommunicationisnotonlyanabsurdattempttodefineawaythe problembutalsoarefusaltoaddresstheneedsofmanyofthestudentsinourschoolsand colleges. Theclaimthatjudgmentsofgrammaticalityaretheproductonlyof"socialattitudesandcultural norms"runscountertotheassertionofChomskyandtheGenerativeTransformationalschoolof linguisticsthatjudgmentsofgrammaticalityareaninherentpartofthebasichumanapparatus foracquiringandproducinglanguage. Althoughtheresolution'sstatement,"Unfortunately,manyemployershavenarrowlyconceived notionsoftherelationshipbetweenlinguisticperformanceandjobcompetence"(p.14) recognizestherealityofthelabormarket,theirargumentgoesontosuggestanimprobable solution:"SinceEnglishteachersareinalargepartresponsibleforthenarrowattitudesoftoday's employers,changingattitudestowarddialectvariationdoesnotseemanunreasonablegoal,for today'sstudentsaretomorrow'semployers"(p.14). TheNCTEthusassignstoEnglishteachersmoreblamethantheydeserveandmorepowerthan theyhaveeverhadorevercouldhavetochangethelinguisticfabricofthesociety.Teachersare requiredtobecrusadersagainstwhatisincorrectlyseenaslinguisticimperialism,ratherthan becomingmorecompetentatteachingthestandardlanguageasaseconddialecttothosewho needtolearn 238

it.ThisapproachisanattempttoescapetheproblemsoftheteachingofEnglishandhas consignedalargegroupofstudentstothirdclasspositionsinsociety. CarolReed(1981)considersthiscalltochangelanguageattitudesfromthepointofviewofa student,whoasks,"howarewetonegotiateourwaythroughthesocietyduringthehundredyearsor soitmaytake(himandhisfriends)toconvincetheAmericanpublicandespeciallytheAmerican employertoappreciatethelegitimacyofourdialect"(p.141). TheSROLdeclarationassertsthestudents'righttotheirownlanguagebyadvocatinganincorrectly motivatedandimprobableprocessofattitudechange.Itassuresthestudentsoftheirrighttoresist whatisclaimedtobethelinguisticandculturaloppressioninvolvedinthelearningofStandard English,thustragicallyensuringtheirlackofsuccess. ByequatingtheteachingofStandardEnglishwithoppression,theSROLhasachievedapolitically convenient,thoughpatentlydishonestandactuallytragicsolutiontotheproblemoftheinabilityto teachStandardEnglish.Unfortunately,theSROLresolutionhasachievedawarmwelcomeamong somesectionsoftheacademiccommunity.DavidShores(1977/1978),inwritingabouttheSROL, citesseveralarticlesrecommendingtheSROLtoteachers.HestatesthattheSROLstatementhas beenwellreceived. Byrestrictingtheteachingandlearningofwhatthesestudentsneedtoknowinorderto communicateeffectivelyinthissociety,theradicalmyopiaoftheSROLresolutioncombineswith thesystemicincompetenceinteachingStandardEnglishtodenytoevenmorestudentsthechance theymighthavehadofsucceeding.Humansnaturallypossesstheabilitytobemultidialectal. Denyingtheimportanceofteachingstandardlanguagetothosewhoneedtolearnitcondemnsthem unnecessarilytothepositionofbeingunabletocommunicatewiththesocietyatlarge. ThecurrentsituationsurroundingBlackEnglish/StandardEnglishbidialectualismisnotthefaultof theisolatedteachersomuchasitisthefaultoftheeducationalestablishmentasexemplifiedbythe NCTEposition. TheNCTEpositioncanbeseenasaresponsetothe'deficit'theoryinAmericaneducation.During thelate1950sandtheearly1960s,theideawasprevalentthattherewasadeficiencyinthelanguage abilityofBlackchildren.TherewaseventheclaimthatghettoBlackchildrenwereessentially withoutlogicalandorderlylanguage.Theworkoflinguists,WilliamLabovamongothers,didmuch toestablishthevalidityofBlackEnglish,tocountertheassertionthatBEisillogical,andtorefute theassertionthatBlackchildrenhaveadeficitoflanguageability.Atthesametime,Labov'swork onvariationtheoryexaminedBEandonincompletedataandexaminingonlythesoundsystemof BEconcludedthatBEwasindeepstructuresimilartoSE.Hehassincerevisedthatpositionand acknowledgestheCreoleoriginpositionpresentedaboveandadvancedinthe1960sandearly1970s byBailey(1965),Stewart(1964,1969a,1969b,1974a)andDillard(1967,1972). 239

However,theeducationalestablishmentneededandwantedtohearthatBEwasonlytrivially differentfromSEbecauseitwasunabletodealwiththeproblemsposedbydialectdifferences. Atalllevelsofeducationalpolicy,thesimilaritiesbetweenthedialectsareemphasized,the differencestrivialized,andtheassertionthatchildrenarecompetentinallthedialectsofbothBE andSEareallpositionsthatareprevalent.Ingeneral,theproblemhasbeendefinedawayandthe implementingofprogramsthatwouldaddresstheneedsofthestudentsinfluencedbyBlack Englishisavoidedbecausethatwouldresultinclassesthat,inmostareas,wouldcontainmostly Blackstudentsleavingadministratorsopentochargesofsegregation.Yettherehasbeenand continuestobeanintuitiverecognitionthatthedialectaldifferencesbetweenBEandSEcause learningproblems.Currently,thedifficultiesthatBlackstudentsencounterinattemptingto bridgetheculturalandlinguisticgapsbetweentheselanguagesystemshavebeenlabeledas pathological.Thus,ghettoschoolsmayclaimthatupwardsof50%ormoreoftheirstudentsare afflictedwithanynumberofasetofpseudoscientific"learningdisabilities."Thus,treating linguisticdifferenceaspathologyhasbecomethemostpopulartacticformanyschoolsystems.It isasituationthatistragicfortheselfimageofthestudentandtragicinthelongrunforthe societybecausetheremedydoesnotaddressthenatureoftheeducationalproblem.Torectifythis situation,thefollowingpointsmustberecognized: StudentsneedtoachieveproductivecompetenceinthegrammarofStandardEnglishin ordertoachieveeducational,socialandeconomicmobilityinthissociety. TheacquisitionofStandardEnglishasaseconddialectcanbeanadditiveprocessand neednotrejectorexcisethedialectformoffamilyandpeers. Itisessentialthatteacherspossesstheknowledgenecessarytodotheirjob.Theteachers ofEnglishwhodealwithspeakersofnonstandardBlackEnglishneedtobetrainedsothat theyknowenoughaboutStandardEnglishandBlackEnglishtoteachsystematicallythose formsofStandardEnglishthatareneededbytheirstudents. Teachingapproaches,methodsandtechniquesneedtobedevelopedandimplementedthat gobeyondthetechniquesemployedtoteachthosewhoalreadyknowthestandarddialect andaddressthestudentswhopossessthestructurallydifferentformsofBlackEnglish. EnglishteachersmustlearntorecognizethesystematicinfluenceofBlackEnglishintheworkof theirstudentsandtodevelopstrategiesformakingtheirstudentsawareofthefactthatthereare differentlanguagesystemsinvolvedandawareofthesystematicdifferencesbetweenthem. Transformationalexercises,contrastiveexercisesofthetypeusedinEnglishasaSecond Languageclasses,roleplaying,rewriting/translatingfromonedialecttotheother,andthe develop 240

mentofeditingskillsaremethodsthatmightbeincluded.Theobjectiveisforstudentsto learntorecognizewhatisnotstandardintheirlanguageandwritingandtobeableto confidentlyproducetheformsofthestandarddialectinappropriatesituations. TeachersshouldbeabletorecognizeerrorsintheuseofStandardEnglishastheproductof historicallydifferentlanguagesystemsthatareincontact,ratherthansignalsofaninabilitytolearn, toreason,ortobeorganized.Theyalsomustbeabletounderstandthatlanguageformsare themselvesneutraltoolsofhumancommunicationthathaveahistoricexistenceandasocialand psychologicalutility.Thereisaneedtorecognizethatthesefactorshavelittletodowiththesocial valueassignedbythesocietytolanguageformsthataredifferentfromthestandardnormofthe educatedelite.Studentsneednotbedeprivedofthedialectoftheirfamilyandfriends;butthey clearlyneedtolearnStandardEnglishforeducational,social,andeconomicmobility.Theyhavenot donesotoalargeextentbecausetheirschoolsystemsandtheirteachershavenotbeenpreparedor abletoteachthem. Thesituationoutlinedhereisnotfacile.Strongandcourageouseffortsmustbemadeinorderto addressthetragicimcompetenceintheteachingofEnglishandintheEnglishteaching establishment.Forthosemanystudentswhoenterschooleagertolearnbutwhodonotsucceed becausetheirdialectformsaredifferentinacomplexwayfromthelanguageoftheschool,thereisa realneedforabidialectalapproachtotheteachingofstudentswhoareinfluencedbyBlackEnglish. REFERENCES Bailey,B.L.(1965)."TowardaNewPerspectiveinNegroEnglishDialectology".American Speech,40,171177. (CCCC)(1974).CollegeCompositionandCommunication,25,(specialissue). Craig,D.R.(1971)."EducationandCreoleEnglishintheWestIndies:Somesociolinguistic factors".inD.Hymes(Ed.),PidginizationandCreolizationofLanguagesLondon:Cambridge UniversityPress. Brown,R.&Bellugi,U.(1964)."ThreeProcessesintheChild'sAcquisitionofSyntax".Harvard EducationalReview,34,133151. Cruickshank,J.G.(1916).BlackTalk:BeingnotesonNegrodialectinBritishGuianawith (inevitably)achapteronthevernacularofBarbados.Demerara:TheArgosy. Dillard,J.L.(1967)."Negrochildren'sdialectintheinnercity".TheFloridaFLReporter,5(3). Dillard,J.L.(1972).BlackEnglish:ItshistoryandusageintheUnitedStates.NewYork:Random House. Ferguson,C.A.(1959)."Diglossia".Word,15,325340. Hall,R.A.(1966).Pidginandcreolelanguages.Ithaca:CornellUniversityPress. Halliday,M.A.K.,Mcintosh,A.,andStrevens,P.(1964).Thelinguisticsciencesandlanguage teaching.London:Longmans. Labov,W.(1966).ThesocialstratificationofEnglishinNewYorkCity.Washington,DC:Center forAppliedLinguistics. 241

Labov,W.(1969)."ThelogicofnonstandardEnglish".TheFloridaFLReporter7(1)6074. Labov,W.(1972).Sociolinguisticpatterns.Philadelphia:U.ofPennsylvaniaPress. Reed,C.E.(1981)."Teachingteachersaboutteachingwritingtostudentsfromvariedlinguistic, socialandculturalgroups".InM.F.Whiteman(Ed.),Writing:thenature,developmentand teachingofwrittencommunication.Volume1:Variationinwriting:Functionaland linguisticculturaldifferences.Hillsdale,NJ:LawrenceErlbaumAssociates. Roy,J.D.(1986inpress)."TheStructureofTenseandAspectinBarbadianEnglishCreole".In M.GorlachandJ.Holm,ed.Focuson:theCaribbean.AmsterdamandPhiladelphia:John Benjamins. Roy,J.D.(1984).AnInvestigationoftheProcessesofLanguageVariationandChangeina SpeechCommunityinBarbados.DoctoralDissertation,ColumbiaUniversity.AnnArbor: UniversityMicrofilms. Roy,J.D.(1981)."RemediationfortheteachingofremedialEnglish",inC.Schoen(Ed.),The CityUniversityofNewYorkWritingSupervisor'sConferenceProceedings1981.NewYork:City UniversityofNewYork. Roy,J.D.(1977).OntheoriginofEnglishCreole:Evidencefromthelexicalstructure.Masters Thesis,ColumbiaUniversity,NewYork. Schuman,J.H.(1973).Secondlanguageacquisition:Thepidginizationhypothesis.Doctoral Dissertation,Harvard.AnnArbor:UniversityMicrofilms. Shores,D.(1977/78)."Anotherlookatstudents'righttotheirownlanguage".IllinoisSchool Journal,57(No.4)Winter. Spencer,J.(Ed.).(1971).TheEnglishlanguageinWestAfricaLondon:Longmans. Stewart,W.A.(1964)."UrbanNegrospeech:SociolinguisticfactorsaffectingEnglishteaching". InR.Shuy,(Ed.),SocialDialectsandLanguageLearning,NCTE,Bloomington. Stewart,W.A.(1969)."SociopoliticalissuesinthelinguistictreatmentofNegrodialect".InJ. E.Alatis,(ed.),MonographSeriesonLanguageandLinguistics,no.22.Washington,DC: GeorgetownUniversity. Stewart,W.A.(1969b).HistoricalandstructuralbasisfortherecognitionofNegrodialect.InJ. E.Alatis(Ed.),MonographSeriesonLanguageandLinguistics,no.22.Washington,DC: GeorgetownUniversity. Stewart,W.A.(1974a)."AcculturativeprocessesandthelanguageoftheAmericanNegro".In W.W.Gage(Ed.),Languageinitssocialsetting.Washington,DC:TheAnthropologicalSociety ofWashington. Stewart,W.A.(1974b)."ThelaissezfairemovementinEnglishteaching:Advancetotherear?" TheFloridaFLReporter12(1&2):8190,9899.ReprintedinM.A.LourieandN.F.Conklin,

(Eds.),(1978).Apluralisticnation:ThelanguageissueintheUnitedStates(pp.333356). Rowley,MA:NewburyHouse. Weinreich,U.(1953).LanguagesinContact.NewYork:LinguisticCircleofNewYork. 242

14 Continuities/DiscontinuitiesintheFunctionandUseofLanguageasRelatedtoSituationand SocialClass WilliamS.HallUniversityofMaryland WilliamE.NagyUniversityofIllinois,ChampaignUrbana Thereisalongtraditionincognitivesocialsciencelinkinglanguageandthought.Amorerecent traditioninsocialscienceresearchlinkslanguagefunctioningandusetoschooling.Differencesin patternsoflanguageusemayreflectorembodydifferencesinhowcommunicativeorcognitive strategiesarebroughttobearonspecifictypesoftasksinspecificsituations.Thepurposeofthe presentchapteristodescribesocioeconomicstatusdifferencesinoneaspectofcommunication:the useofwordsofinternalstate.Thesearewordsthatwhenusedliterallyreferprimarilytointernal states,processes,orexperiences.Thisincludeswordsaboutcognition(think,remember,know), emotions(happy,afraid,love),perception(see,smell,pain),andintentionsanddesires (intend,want,wish)(SeeHall&Nagy,1986).Internalstatewordshavebothcognitiveand educationalsignificance;theyplayanimportantroleincertaintypesofclassroomdiscussion,and theiruseislinkedtocognitivestrategiesandskillsinvolvedinmetacognition(cf.Baker&Brown, 1980;Flavell,1978). SITUATIONALVARIATION Concurrentwithincreaseinginterestinlanguagedifferencesbetweenindividualsofvarioussocial groupshascomeagrowingfocusondifferencesinthespeechofthesameindividualindifferent situations.AnearlystudybyLabov(1964)illustratesverywellsomeimportantaspectsofthe interactionbetweensituationalandsocialvariation.Figure14.1representsdifferencesinthe pronunciationofrbyspeakersfromdifferentsocioeconomiclevels,atdifferentlevelsofformality. 243

FIGURE14.1.UseofRConstrictioninNewYorkCitySESandSpeechStyleThisfigureis originallyfromWilliamLabov,"PhonologicalCorrelatesofSocialStratification"inJohnJ. GumperzandDellHymes(Eds.),Theethnographyofcommunication,specialissueofthe AmericanAnthropologist,Vol.66,No.6,p.171.ItisreproducedherefromRobbinsBurling, Man'smanyvoices,NewYork:Holt,Rinehart&Winston,1970,p.93. Therearefourparticularaspectsofthisfigurethatwewishtodrawattentionto: 1.Boththesituationalandsocialdifferencesarequantitativeratherthanqualitative;they manifestthemselvesintermsofthepercentageofthetimethatacertainpronunciationisused. Thisdoesnotmeanthattherecannotbesomesocialorsituationaldifferencesthatare qualitative;however,muchofsuchvariationwillconsistintherelativefrequencywithwhicha certainformorpatternoccurs. 2.NotethatthemostcarefulspeechofthelowestgroupinLabov'sstudyshowsmoreR constrictionthanthecasualspeechofeventheuppermiddleclassgroup.Thus,thereisinsome senseareal"overlap"betweenthespeechpatternsofeventheextremeendsofthesocialscale. Thisisofcoursenotnecessarilythecasewithallinstancesofsocialandsituationalvairation;but inmanyinstances, 244

suchvariationwillbeoneofdegree,withmuchoverlapbetweenthepatternsofdifferentgroups, ratherthanconsistingofqualitativeandabsolutedifferences. 3.Notethatinthiscasethesituationaldifferencesareaslargeasthesocialdifferences.The magnitudeofdifferencesbetweenthespeechtypesofasinglesocialgroupindifferentsituations isthesameas,orgreaterthan,thatofthedifferencebetweendifferentsocialgroupsatanyone givenlevelofformality.Again,thisisnotnecessarilythecase;theremaybeinstanceswhere socialvariationinlanguagepatternsisfargreaterthananysituationalvariationwithinthesocial group.However,situationalvariationisatleastpotentiallyasgreatassocialvariationapoint thatwewillreturntoshortly. 4.Notethattherearedifferencesinthewaythatdifferentsocialgroupsrespondtoagiven situation.Inthiscase,allgroupsshowasimilarpattern,usingahigherpercentageofR constrictioninmoreformalcircumstancesbutthelowermiddleclassgroupshowsamuch moreextremedifferenceinthisregardthananyoftheothergroups.Ithasbeensuggestedthat thisisduetotheuncertaintyofbeingthesecondhighestgroup;thelowermiddleclassspeakers, wheninaformalsituation,outdotheuppermiddleclassspeakersintryingtosoundupper middleclass(cf.Laboy,1964).Theexamplejustdiscussedinvolvesvariationinphonology.Not surprisingly,earlyworkinlanguagevariation,e.g.,Labov(1966),tendedtofocusonthisarea; phonology,beingoneofthemoreaccessibleaspectsoflanguagestructure,hasalsobeenthefirst areaofmajorfocusinotherfieldssuchashistoricallinguisticsanddialectgeography.Atitle suchasthatofSankoff(1973),"Aboveandbeyondphonologyinvariablerules,"givesan indicationofthedirectionthatprogressinlanguagevariationstudyhastaken,expanding "upwards"intosyntaxandsemantics.However,thesamesocialandsituationalfactorsthat influencepronunciation(e.g.,theage,sex,socialclass,mood,andpersonalityofthespeaker;the roleandstatusoftheaddressee;theformalityandthetopicoftheconversation)playarolein variationinsyntax,vocabulary,andallotherlevelsoflinguisticpatterning. Cole,Dore,Hall,andDowley(1978)documentsomeofthemorespecificeffectsthatsituational variationcanproduce.Likewise,Matthew,Connelly,andMcCleod(1978),investigatingthe speechofafiveyearoldinthreedifferentsituations(alone,withafriend,andwiththemother) foundsubstantialvariation,bothinsyntacticcomplexity(thechildusedfarmoresixwordor longerutteranceswiththemotherthanaloneorplayingwithafriend)andinspeechacttypes (thechildusedmorecommandsalonethanwiththemother;presumablyitusedcommandsmore oftenwhentakingontheadultroleinthediscourse). ThefourpointsmentionedinregardtophonologicalvariationintheexamplefromLabovare alsofoundtoapplytononphonologicalvariationaswell.Variationinotheraspectsoflinguistic patterningmaybequantitativeratherthanqualitative,andtheremaybesubstantialoverlap, betweengroups;situational 245

variationmaybeofthesamemagnitudesas,orgreaterthan,socialclassbasedvariation;and differentsocialgroupsmanifestdifferentpatternsofsituationalvariation. Thisiscrucialtothestudyofsocialvariationinlanguagepatternsinanumberofways.Firstof all,theremaybegenuinesocialclassbaseddifferencesinlanguagethatappearinsome situationsandnotinothers.Forexample,Snowetal.(1976),studyingthespeechofDutch mothersofthreesocioeconomiclevels,foundsignificantclassdifferencesinonesituation(free play)butnotinanother(tellingastorybasedonpicturesinabook). Second,whatappearstobesocialclassbasedvariationinlanguagemayinfactbeanartifactof thedifferentresponsesdifferentsocialgroupsmayhavetothe"same"situation.Cazden(1970), inareviewofresearchonsocialclassbasedlanguagedifferences,concludesthatinmostofthe casesstudied,thedifferencesthatwerefoundwereduetodifferencesinthewaysthatchildren fromdifferentbackgroundsreacttoasituation,ratherthantoanydifferenceinlinguisticability supposedlycorrelatingwithsocialclass.Thiscallsintoquestionanystudyofsocialclassbased languagedifferencesthatdoesnotcarefullycontrolforpotentialdifferencesintheeffectsofthe situationonthesubjectsinvolved.Cooper(1975),forexample,inacritiqueofBernstein'searly work,pointsoutthatoneofthesettingsusedtocollectdataagroupdiscussionofcapital punishmentmaynothavebeenasinteresting,orasfamiliar,orascomfortable,totheworking classsubjectsastomiddleclasssubjects. Then,theresearchpresentedinthischapterfocusesonsocialclassvariationalongwith situationalones.Specifically,thechapterfocusesonthewaythatdifferencesbetweenthehome andschoolsituationsmayinfluencechildren'suseofinternalstatewordsatschool.Theissues herecanbestatedintermsofthe"mismatchhypothesis."Themismatchhypothesispostulates thatlearningdifficultiesexperiencedbychildrenfromminorityor"nonmainstream" backgroundsarecausedbyadiscontinuitybetweenthehomeandschoolenvironmentsthatis, adiscrepancybetweentheexpectations,strategies,andschematathatworkathome,andthe cognitiveandmotivationaldemandsoftheclassroom.Forchildrenofthemainstreamculture,on theotherhand,thereisafairdegreeofcontinuitybetweenhomeandschoolintermsofculture, patternsoflanguageandbehavior,andtypesofstrategiesforinteractingwithadults. Twomorespecific,independenthypothesesaboutinternalstatewordusagecanbeformulated withinthemismatchmodel. Firstofall,itmightbearguedthatschoolsdemandahighdegreeof"metabehavioralawareness," andthatdifferentculturesorsocialgroupsdonotprovidechildrenwiththesameamountof preparationinthisarea.Bymetabehavioralawarenesswemeantheabilitytoanalyze,and verballydescribe,theemotions,thoughts,andintentionsofapersonorfictionalcharacter.Such analysisistypicalofclassroomdiscussion,especiallyrelatingtoreadingstories,eveninearly grades.Theuseofinternalstatewordstotalkaboutfeelings,thoughts, 246

andintentionsisclearlyanimportantaspectofthechild'spreparationforthistypeofschool activity,anditispossiblethatchildrenfromsomesocioeconomiclevelsorethnicgroups receivemoreofsuchpreparationathomethandoothers.Analysisoftheinternalstateword usageoftheadultsinthechildren'senvironmentsathomeandschoolwillindicatewhetheror notthisisthecase. Thefirsthypothesis,then,isthatthereisamismatchordiscontinuitybetweentheinternalstate wordusageofadultsathomeversusatschoolexperiencedbychildrenfromnonmainstream families.Asecondhypothesiscanbeformulatedconcerningtheinternalstatewordsproducedby thechildren.Ifchildrenfromnonmainstreambackgroundsexperienceamismatchofsomesort betweenthehomeandschoolenvironments,thereislikelytobesomeindicationofthisintermsof theirresponsetotheschoolsituation.Theinternalstatewordusageofthechildrenisameasureof oneaspectoftheirresponsethatisofcleareducationalsignificance. Itshouldbenotedthatthesetwohypothesesareindependent:theremaybeamismatchinthe internalstatewordusageofadultsathomeversusatschool,butthismaynotbereflectedinthe internalstatewordusageofthechildren.Conversely,theremaybeamismatchbetweenthehome andschoolenvironmentsthathasnothingtodowithinternalstatewordusageofadults,which howeverinfluencestheinternalstatewordusageofthechildren. Theresearchdiscussedinthischapterteststhesehypothesesinmeasuringdifferencesinchildren's speechbetweenhomeandschoolsituations.Italsocontributestoamoregeneraltheoryofregister orsituationalvariation(cf.Halliday,1978).Inthefollowingpageswedefineinternalstatewords, discusstherationaleforchoosinginternalstatewordsforinvestigation,andpresenttheresultsfrom theanalysesofinternalstatewordusageinnaturalconversationinthehomeandschool environmentsof41/2to5yearoldchildren. COGNITIVEIMPLICATIONSOFINTERNALSTATEWORDUSE Variationoccursinallaspectsoflanguage;inpronunciation,grammar,choiceoflexicalitems, discourselevelphenomena,andsoon.Weareinterestedinlanguageasatransmitterandreflector ofcultureandcognitivestyles,andare,therefore,interestedinlanguagevariationalongdimensions thatwillbeofdirectcognitiveandeducationalsignificance.Thisexcludesfromourconsideration differencesinpronunciation(e.g.,fas'vsfast)orgrammar(Idon'thaveanyvs.Idon'thavenone). Suchdifferencesmaycorrelatewithimportantsocialdistinctions,andrelateinimportantwaysto attitudesonthepartofbothspeakersandhearers;butthereisnodistinctioninconceptualcontent associatedwithdifferencesinpronunciationorgrammaralone.Wethereforefocusedour investigationonaspectsoflanguage,ordifferencesinpatternsoflanguageusethatwould 247

belikelytoinfluencethesocializationofcognitivemodesinchildren;thesehavetodowiththe contentandfunctionsoflanguage,ratherthanwithformalpropertiesofphonologyorsyntax. Theuseofinternalstatewordsiscruciallylinkedtocognitivestrategiesinthreerelatedareas:the acquisitionandorganizationofinternalstateconcepts;theunderstandingofstoriesand discussionofstories;andmetacognitiveskillsingeneral. Thescopeofthepresentresearchisprimarilytodocumentwhatdifferencesthereareintheuse ofinternalstatewordsbetweendifferentsocialgroups,andthentointerpretthesedifferences,as farasthisispossible,intermsofreasonablehypothesesabouttheeffectsthesedifferences shouldhaveonthechild'scognitivedevelopment. PREVIOUSRESEARCHONSOCIALCLASSBASEDDIFFERENCESININTERNAL STATEWORDUSE Thereisasubstantialbodyofresearchonsocialclassbasedlanguagedifferences(cf.Bernstein, 1971,1973)thathassignificantimplicationsfor,andmakessomespecificpredictionsabout, classdifferencesinpatternsofinternalstateworduse. Inanearlystudyonclassdifferencesinlanguageuse,SchatzmanandStrauss(1955,usingas datatranscriptionsofinterviewswithpeoplewhohadexperiencedatornado,andtakingthe extremecasesoneducationalandincomedistribution)foundthatlowerclasssubjectswereless likely,whendescribinghumanbehavior,totalkaboutmotivation.Theyalsofoundaclass differenceinthefrequencyof"youknow"thisconversationaldevicewasmorecommonalong lowerclassspeakersthanmiddleclassspeakers. Inthe"positional"modeofcontrol(cf.Bernstein,1971,1973)whichisassociatedwithrestricted code,andassertedtobemoretypicaloftheworkingclass,regulationofachild'sbehaviorisin termsofexternalbehaviorandthepositionalstatusoftheparticipants.Inthe"personal"modeof control,associatedwithelaboratedcodeandsupposedlymoretypicalofmiddleclassfamilies, moreemphasisisgiventomotivationandintentionincontrollingthechild'sbehavior.Also,in thepositionalmodeofcontrol,theyaresimplyenforcedratherthandiscussed.Thiswould suggestthatinternalstatewordswouldbeusedwithgreaterfrequencybymembersofthemiddle class(althoughthedifferencemightbemanifestedindifferenttypesofusageratherthanin overallfrequency). InBernstein(1973)arearticlesbyseveraldifferentauthorsexploringtheimplicationsofhis theoriesinmoredetail.Someofthearticlesmakemoreexplicitclaimsaboutinternalstateword usage. Indatabasedonquestionnairesinwhichmothersestimatedtheirownusagepatternsinresponse toaseriesofspecificquestions,Henderson(1973)foundthatmiddleclassmotherstalkedmore aboutcognitivetopicsthanaboutaffec 248

tive/interpersonaltopics,andmoreaboutcognitivetopicsthandidworkingclassmothers.On theotherhand,workingclassmotherstalkedmoreaboutaffective/interpersonaltopicsthan aboutcognitivetopics,andmoreaboutaffective/interpersonaltopicsthandidmiddleclass mothers. IfonecouldassumethatHenderson'smethodsaccuratelyreflectthemothers'actualusage(this issomewhatquestionable),thatEnglishsocialclassdifferencesaresimilartothoseinAmerica, andthatmoretalkaboutcognitiveoraffective/interpersonaltopicswillincreasethefrequencyof thecorrespondinginternalstatewordcategories,thenthisresearchmakesspecificpredictions aboutclassdifferenceswemightexpecttofindintheinternalstatewordusageinourdata. AslightlyearlierstudydoneinAmericahadsomewhatsimilarresults.Inastudyofspontaneous storytellingoffourthgraders,vonRaffler,Engel,andSigelmen(1971)comparethespeechof middle/upperclasswhitesandlower/middleclassblacks(thusconfoundingraceand socioeconomicstatus).Theyfoundthatahigherpercentageofblackchildrenreferredtointernal statesthandidwhitechildren(71%vs.46%),butthatthereweredifferencesinthetypeof internalstatesreferredto.Internalstatereferencesbythewhitechildrenweremostlythinkand know;referencestointernalstatesbyblackchildrenrelatedtoemotionsandambitions. TheseresultsseemtoconfirmtheclassdifferencesinchoiceoftopicnotedinHenderson(1973) butdonotconfirmtheoverallimpressiononegetsfromthedescriptionsofrestrictedand elaboratecodes,thatlowerworkingclasspersonsdonotuselanguagetoexploreintentionsand feelings. Thus,internalstatewordsdofigureinclaimsthathavebeenmadeaboutsocialclassbased languagedifferencesthatmightbeofeducationalandcognitivesignificance.However,the researchuptonowhassufferedfromlackofabroaddatabasethattakessituationalvariation intoaccountandincludesnaturalisticconversationbychildrenandtheircaregiversinthehome. INTERNALSTATEWORDSWHATTHEYARE ThecurrentresearchcentersontheanalysisofdatafromtheprojectoutlinedinHallandNagy (1979).Proceduresweredevelopedforcodinginternalstatewords,thatis,foridentifying instancesofinternalstatewordsinnaturalisticdataandcategorizingcertainaspectsoftheiruse andfunctioninthecontextofdiscourse;theseprocedureswerethenappliedtothelargecorpus ofconversationdescribedearlier,andtheresultingdatasubjectedtoanalysesofvariouskinds. Internalstatewords,orwordsofinternalreport,havemeaningsprimarilyconcernedwith internalprocesses,states,andexperiences.Thisincludeswordsaboutcognition(e.g., think,know,believe,remember,figureout),aboutemotions(e.g.,fear,angry,sad,happy),about perceptionsboththefivesenses 249

(see,hear,etc.)andthemore"internal"perceptions(e.g.,dizzy,thirsty,ache),aboutdesires (want,wish,desire),andaboutintentions,choices,anddecisions. Internalstatewordsarewordsthatbyvirtueoftheirlexicalmeaningrefer,whenusedliterally,to internalstatesandprocesses.Thewordjerk(asin,Icouldkillthatjerk)expressesthespeaker's internalstateorattitude,butdoesnotrefertoit,aswouldawordlikeangryorupset.Totakea differentkindofexample,thewordsdidsomethinginasentencelikeIdon'tknowwhatshesaid, butitsuredidsomethingtohim,maywellrefertoaninternalstateorexperience,butnotby virtueofthelexicalmeaningsofthesewords. Manywordsimplyorpresupposeinformationaboutinternalstates,butarenotprimarilyabout internalstatesthemselves.Forexample,complainpresupposesacertaintypeofattitudeonthe partofthespeaker,butisprimarilyaverbofspeaking.Therearealsowordsaboutcapacities, suchasblindorintelligent,whichrelatetointernalstatesandprocesses,butwhichdonotrefer directlytointernalstatesorprocessesassuch. Lexicalambiguitycomplicatestheprocessofdeterminingwhatshouldbeconsideredaninstance ofaninternalstatewordandwhatshouldnot.See,forexample,isaperceptualinternalstate wordinDidyouseethefiretrucks?andacognitiveinternalstatewordinIdon'tseehowyoucan dothat.However,itispresumablynotaninternalstatewordatallinasentencelikeHewentto seehisgrandmother,sinceitseemstobemoreorlesssynonymouswithvisitinthiscontext. Innaturalisticdatatherewillofcoursebeseveraltypesofborderlinecaseswhereitisnotclear whethersomestate,processes,orexperiencecanbeconsidered"internal"(thatis,"mental"or "psychological").(Thetheoreticalissuesinvolvedinsuchcaseshavebeendiscussedindetailin HallandNagy,1986).However,themajorityofwordsoccurringineverydayconversationare ratherprototypicalexamplesofourbasicinternalstatecategories. Wehavedividedinternalstatewordsintofourmajorcategories: 1.Cognitive.Wordsinthiscategoryareaboutcognition,awareness,consciousness,knowledge, understanding,attention,thinking,belief,orcertainty.Someofthecommonlyoccurring cognitivewordsare: think know remember forget understand figureout belief believe guess assume wonder pretend 2.Affective.Wordsinthiscategoryrelatetoemotions.Someofthecommonlyoccurring affectivewordsare: like sorry love angry hate annoy 250 afraid glad

happy mad mood regret 3.Perceptual.Thesewordsrelateeithertothefivesensesortoa person'sawarenessofhis/herownbody,e.g.,painorhunger.Some prefer sad scared upset commonperceptualwordsare: see look hear watch listen taste hurt pain hungry ache tired thirsty 4.IntentionsandDesires.Thiscategoryincludeswordsforinternal statesthatrelatetogoalswhatapersonintends,chooses,orwants. Somecommonwordsinthiscategoryare: want wish intend wouldlike plan mean decide changeone'smind choose hope Incodingforusage,themostimportantdistinctioninourworkis whatwehavelabeledthe"semantic/pragmatic"distinction,whichcan beexpressedinthefollowingquestion:Isagiveninstanceofan internalstatewordbeingusedtorefertoandcommunicateaboutan internalstate?Morebriefly,isitbeingusedliterally? Semantic,orliteral,usages,arethoseinstanceswheretheinternal statewordisusedtorefertoaninternalstate,asforexamplethe knowinMaybeyouknowtheanswer.Pragmatic,ornonliteralusages, arethoseinstanceswherethelexicalmeaningoftheinternalstate wordcontributesindirectly,ifatall,tothepropositionalcontentof thesentence.AtypicalexampleofthiswouldbetheknowinYa know,thereoughttobealaw.(SeeHall&Nagy,inpress). Semanticusesofinternalstatewordscanbefurthersubdividedinto reflectionsandnonreflections.Reflectionsareassertionsbythe speakerabouthis/herowncurrentinternalstate,orquestionsabout theaddressee'scurrentinternalstate. Onemotivationforthisdistinctionistheconceptofmetacognitive experience.Forourpurposes,wecanadoptthefollowing,somewhat modifieddefinitionofmetacognitiveexperience:ametacognitive experienceisawarenessofone'sowncurrentinternalstate.(This overlapssubstantially,butnotperfectly,withthefollowingdefinition byFlavell,1978:"Metacognitiveexperiencesareconsciouscognitive oraffectiveexperienceswhichoccurduringtheenterprise(thatis, somecognitiveenterprise)andconcernanyaspectofit"(p.233). Reflectionsasdefinedabovehavethefollowingrelationshipto metacognitiveexperiences:Whenaspeakermakesanassertionabout his/herowncurrent

251

internalstate,heorshemustnecessarilybeawareofthatstate;theassertionisanexpressionof thatawareness.Whenaspeakerasksaquestionabouttheaddressee'scurrentinternalstate,this presumablyelicitsawarenessonthepartoftheaddresseeonhis/herowncurrentinternal state.Wehaveexcludedfromourdefinitionofreflectionsassertionsbythespeakeraboutthe addressee'sinternalstate(Youknowwhatshoestheyare)aswellasimperativesrelatingtothe addressee'sinternalstate(GuesswhereI'mhiding).Thesemayofcoursealsoelicitawarenesson thepartoftheaddresseeofhis/herowncurrentinternalstate,buttheconnectionisnotasdirect asinthecaseofquestions,wherethespeakerisexplicitlytryingtoelicitsuchawareness.The definitionofreflectionsgivenabovecanbebrokendownintotwomaincriteria:First,fora semanticusagetoqualifyasareflection,itmustbeanassertionaboutthespeaker'sinternalstate oraquestionabouttheaddressee'sinternalstate.Thus,theinternalstatewordmustbeinthepart ofthesentenceassertedorquestioned.Thislargelyrestrictsreflectionstointernalstatewordsin themainclauseofthesentence;relativeclausesandmanysubordinateclausesarepresupposed ratherthanassertedorquestioned.Thus,example(1)and(2)belowconstitutereflections, whereas(3)and(4)donot: 1. I'mthinkingaboutit. 2. Doyouknowwhattheansweris? 3. SomebodyIknowtoldmeaboutit. 4. TheysaythatIknowtheanswer. Tosummarize,themostgeneralcodingcategoryisInternalStateWords.Therearetwobasic typesofclassificationaccordingtowhichthisgeneralcategoryisfurthersubdivided.Onetype ofdivisionisaccordingtolexicalclasses;internalstatewordscanbedividedintofourgroups Cognitive,Affective,Perceptual,andIntentionsandDesiresonthebasisoftheirlexical meaning.Theseconddivisionofinternalstatewords,intersectingwiththefirst,isaccordingto function.InternalstatewordsarecategorizedasPragmaticorSemanticdependingontheirusein context.Thereisafurtherclassificationbyuse:SemanticInternalStateWordscanbefurther dividedintoReflectionsandNonReflections. Thesecodingcategories,discussedinmoredetailinHallandNagy(1986),arearefinementof thecategoriesdesignedfortheinvestigationofinternalstateworduseinnaturalisticdatagiven inGearhartandHall(1982). ThecategoriesCognitive,Affective,andIntentionsandDesiresareverysimilartothethree subcategoriesoftheInternalresponsecategoryoftheSteinandGlenn(1979)storygrammar: Thoughtsorcognitions(e.g.,"MarythoughtJohnwasobnoxious"),feelingsoraffective responses(e.g.,"Marywasveryangry"),andgoalsordesires(e.g.,"MarywantedtohitJohn"). Forthe 252

purposesofstorygrammar,itdoesseembestnottoincludePerceptualwordsinthecategoryof internalresponses.Butinthisresearch,internalstatewordsareofinterestbecauseoftheir implicationsformetabehavioralawareness.Wewouldconsiderperceptualawareness(e.g.,the abilitytoanalyzeaperceptualarrayintoasetofgeometricalormathematicalrelationships)as beingrelatedtometabehavioralawareness(e.g.,theabilitytoanalyzetheemotionsofapersonor thoseofafictionalcharacter).Mostwordshaveanumberofmeanings,sothefactthataword occursinagivencategorymeansonlythatithasatleastonemeaningthatbelongsinthatgroup. Whetherornotaspecificinstanceofthatwordintheconversationbelongsinthatcategorymust bedeterminedonthebasisofthecontext.Theoverallgoalinthisresearchwastoidentifythose factorsthatareassociatedwithdifferencesininternalstatewordusebythetargetchildren.We alsowanttoanswerquestionssuchasthefollowing,concerningpredictionsmadebyspecific versionsofthemismatchhypothesisthatcanbeformulatedregardingourdata:Dochildrenfrom nonmainstreambackgroundsexperienceadiscontinuityormismatchbetweentheinternalstate worduseofadultsathomeandthatwhichtheyencounterinthespeechofadultsatschool?Does thespeechofchildrenfromnonmainstreambackgroundsreflectadiscontinuityintheuseof internalstatewordsthatmightreflectamismatchofculturebetweenhomeandschool experiencedbythesechildren? SUBJECTS Toanswerquestionssuchasthese,wesearchedacorpusgeneratedby39targetchildrenand thosewithwhomtheyspokeaswellasthosewhospokeinthesituationswheretheywere present.Conversationsoccurringintwosituationsarediscussedhere.Theunitfromwhichthe internalstatewordswerecodedwasaturntospeak.Categorizationoftheinternalstatewordwas carriedoutbytwoindependentjudgeswhoworkedfromtranscriptsoftheoriginaltapes.The transcriptsweredistributedacrossraceandSEStominimizeconfoundingduetopractice,and eachjudgecodedapproximatelyhalfthetranscriptsineachrace/SESgroup.Thetranscripts incorporatedthefollowinginformation: 1. anumberidentifyingthetargetchildandhis/herfamily; 2. acoderepresentingtherace/SESgroupofthatfamily; 3. acodeidentifyingthespeakerofeachparticularlineofthetranscript; 4. acodeidentifyingthesituation; 5. asequentialnumberidentifyingeachlineinthetranscript; 6. anindicationofincidencesofsimultaneoustalk; 253

Inaddition,therewasalso,interspersedthroughthetranscript,thecontextual informationprovidedbytheexperimenterduringthetaping.Thisinformationwas usefulininterpretingthediscourse;forexample,towhomagivenremarkwas addressed,orthetypeofactivitiesinwhichtheparticipantswereinvolved. Toassessagreementbetweenthetwojudges,thetranscriptofarandomlyselected subjectwascodedbybothjudges.Theagreementamongthejudgesastowhat constitutedaninternalstatewordwas94%.Theaddresseeofaparticularinternal statewordwascommonlyidentified77.5%.Thedisagreementshereusuallyinvolved uncertaintyabouttheexactidentitydespiteagreementabouttheageandsexofthe addressee.Only5%ofthetotaljudgmentsinvolveddisagreementaboutwhetherthe targetchildwastheaddressee.Theassignmentofeachinternalstatewordtoa particularlexicalcategoryshowed92.0%agreement.Theidentificationoftheobject oftheinternalstateword(thatis,towhoseinternalstatethewordreferred)resulted incommonjudgmentsin84.4%ofthecases.Finally,thejudgmentastowhetherthe internalstatewordrepresentedaReflectionornotwasmadeincommonfor84.7% oftheinternalstatewords.Ingeneral,theinterrateragreementacrossalljudgments isremarkablyhighfordataofthissort. RESULTS FormationofDependentVariables Sincenoattemptwasmadeduringthetapingphaseofthisstudytocontrolthetarget child'sinteractions,theamountofspeechavailableforeachtargetchildandhis/her principalinteractantsvariesquitewidely.Table14.1showsthemeans,standard deviations,andrangesofboththenumberofwordsandthenumberofturnsforthe targetchildrenandtheirprincipalcaretakers.Itisapparentfromthistablethatthe absolutefrequencyofoccurrenceofeachofthecodingcategorieswillbe,inpart, dependentontheamountofspeechsampled.Infact,frequenciesforeachspeaker haveanaveragecorrelationof..73withthenumberofwordsspokenand.71withthe numberofturnstaken. Therearetwotypicalwaystodealwithmethodologicalproblemsofthissort.One methodistotreatthewordsandturnsascovariates;inessence,topartialouttheir influencefromtheinternalstatewordcategories.Theothermethodistodividethe variablesofinterestasalessdirect(andfrequentlylessprecise)wayofpartialingout theirinfluence.Whenthecorrelationsamongtheinternalstatewordcategorieswere computedaccordingtobothmethodsandcompared,the 254

7 theactual . textofthe utterance; 8 acode . indicating whether ornotthe utterance contained a question; and 9 the . numberof wordsand turns spokenby each speakerin the situation.

TABLE14.1 Means,StandardDeviations,andRangesofNumberofWordsandTurnsbyPrincipalSpeakers Person NumberofWords NumberofTurns ________________________________ ______________________________ _______ __ M Target children Mothers 1809.53 1998.06 Range Dinner 909.38 2303810 1410.42 234955 Directedactivity 235.38 17969 SD M 314.43 272.61 SD 153.30 174.11 Range 54770 5714

Target 505.46 110.81 51.97 8209 children Teachers 1702.81 733.15 4362667 193.65 85.45 60472 meandifferenceswerenegligible:.02forperwordversuswordscontrolledand.09forperturn versusturnscontrolled.Thustheuseofproportionsdoesnotappeartobesubstantiallyless preciseatreducingcontaminationthanpartialing.Basedonthis,aswellasthefactthat proportionsaremoreconceptuallymeaningfulandinterpretable,wedecidedthatdependent variableswouldbeformedasproportionsofthenumberofwords. Proportionvariablestypicallyhavepositivelyskeweddistributions,andourvariablesareno exception.Thusforpurposesofstatisticalanalysis,thevariableswerearcsinetransformedto maketheirdistributionsmorenormal.Allreportedmeansandpatternsofmeans,however,are basedonthemoremeaningfuluntransformedvariables. SelectionofDependentVariables Ourcodingproceduresdescribedindetailearlier,categorizedtheinternalstatewordtokens occurringinthecorpusintermsofthefollowingcategories: SemanticCategories Reflections:

Nonreflections:

CognitiveReflections AffectiveReflections PerceptualReflections IntentionsandDesiresReflections CognitiveNonreflections AffectiveNonreflections PerceptualNonreflections IntentionsandDesiresNonreflections 255

Thefollowingstepsweretakentoreducethissetofcategoriestoa workablenumberofdependentvariables:Firstofall,wechoseto grouptogetherconceptuallysimilarvariablesthatindividuallywere toolowinfrequencytobesubjecttoreliableanalysis.Second,we chosetofocusonfunctionalratherthanlexicalcategories. Quotesand Songs Pragmatic Conversational Categories:Devices Indirect Requests Rhetorical Questions Exam Questions Hedges Opinion Questions Attentional Devices Asaresult,allpragmaticcategoriesexceptattentionaldeviceswere groupedintoasinglecategorylabelednonliteralusages.Attentional deviceswerekeptasaseparatecategory.Nonreflectionswere similarlytreatedasasinglecategory.Onlyinthecaseofreflections didweconstructseparatevariablesbasedonthelexicalclassofthe internalstatewords.Songsandquotesweretooinfrequenttobe analyzed. Wealsoincludedintheanalysissomesupercategories:First, reflections,includingallfourlexicalclassesofreflections;second, semanticusages,includingbothreflectionsandnonreflections;and finally,internalstatewords,whichincludesallkindsofinternalstate words(semantic,pragmatic,orquotesandsongs)occurringinthe transcript.Thevariablesincludedinouranalysesandtheinclusion relationshipsamongthemarerepresentedinFig.14.2. FIGURE14.2.ClassinclusionRelationshipsAmongInternalState WordVariables 256

SELECTIONOFSPEAKERSFORANALYSIS Thevariableswechosecanbedefinedforanyspeakerorgroupofspeakerswithinasituation. Certainspeakersandgroupsofspeakersareofspecialrelevanetoourpurposes. Thetargetchildrenwereofcoursethefocalpointofthedatacollectionandarethefocusofthis analysisaswell.Theyaretheonlyindividualsconsistentlypresentinbothhomeandschool situations,thusallowingforananalysisofsituationalvariationinthespeechofthesameperson. Thespeechoftargetchildrenisalsoofspecialinteresttousbecauseweareinterestedintheprocess ofculturaltransmissionthatis,howachildlearnspatternsoflanguageusagefromthespeechin his/herenvironment.Ourtargetchildren,aged41/2to5years,werestillatanagewhentheir understandingofinternalstateconceptsisdeveloping(cf.Wellman&Johnson,1979). Thespeechofotherparticipantswasofinteresttousprimarilyinsofarasitexplained,orfailedto explain,situationalvariationinthespeechofthetargetchildren. Therearethreesetsofspeakerswhocanbeconsideredtocomprisethelinguisticenvironmentofthe targetchildren.Oneofthesecontainsonlytheprimarycaregiver(presumablythemotherinthe homeenvironmentandtheteacherintheschoolenvironment).Anotherpossiblesetcontainsall adultsinthetargetchildren'senvironment.Athome,thiswouldincludefathers,grandparents,and anyotheradultspresent(eithercoincidentallyoronaregularbasis)duringthedinnersituation.At school,thiscategoryconsistsalmostentirelyofteachers.Theexperimenterispresentinthe classroom,butspeaksrelativelylittle,andthereareveryfewteachers'aides.Athirddefinitionof thetargetchildren'slinguisticenvironmentwouldincludeallspeakersotherthanthetargetchildren themselves. Althoughthethirdofthesedefinitionshasanobviousvalidityforsomepurposes,wehaveexcluded itfromouranalysesforthefollowingreasons:Firstofall,sincethefamiliesinoursampledifferin compositionsomehaveonlyonechild,othershaveseveralcomparisonofinternalstateword usageacrossfamilieswouldbedifficulttointerpret,sincechildrenandadultsmaydiffer substantiallyintheirinternalstatewordusage.Thus,ameasureofthechild'slinguisticenvironment thatincludedotherchildrenmightreflecttheagecompositionofthefamilymorethanconsistent differencesininternalstateworduse.Second,thismeasureoftheenvironmentofthechildwould incorporateadegreeofredundancy.Weareinterestedinmeasuringtheinfluenceofsituational factorsontheinternalstateworduseofthetargetchildren.Thespeechoftheotherchildren,if includedinthemeasureofthetargetchild'slinguisticenvironment,wouldincorporatefactors influencingthetargetchild'sspeech,butalsotheresponseoftheotherchildrentothosesame factors. 257

Becauseofthispotentialproblem,inlookingatthespeechinthetargetchild'senvironment,we concentratedontwomeasures:(a)thespeechoftheprimarycaregivers(theteachersand mothers)and(b)thespeechofalladultsinthetargetchild'senvironment. RESULTSOFDATAANALYSIS WebeginbypresentingourdataintermsoftheresultsofsituationsXraceXSESandraceX SESanalysesofvariance(ANOVAs)ontheinternalstatewordvariablesrepresentingthespeech ofthetargetchildren,thespeechoftheirprimarycaregivers(mothersandteachers),andofall theadultsintheirhomeandschoolenvironments.Thenwetrytoaccountforvariationsinthe targetchildren'sinternalstateworduse. AcrossSituationAnalyses Severaltypesofanalyseswereperformedtoinvestigatetheeffectsofsituationandsocialgroup membershiponthetargetchildren'spatternofinternalstateworduse.Firstofall,situationX raceXSESanalysesofvariancewereperformedforallourdependentvariablesonthefollowing groupsofspeakers: 1.Targetchildrenathomeandatschool.Sincethesettingisobviouslyarepeatedfactorinthis case,thetargetchildrenforwhomdatawerenotavailableatbothschoolanddinnerwere excluded(Nforanalysis=36). 2.Primarycaregivers(teachersatschoolandmothersathome).Fortwotargetchildren,the motherswerenotlivingathomeandhencenotpresentatthedinners.Onemotherspokeonly fiveturns(23words)inthecourseofthetwodinners,usingnointernalstatewordsatall;itwas decidedtoexcludeherfromthisanalysisaswell(N=37teachers,35mothers). Tables14.214.4presentallthesignificanteffectsfoundbytheseanalyses.Becauseitis consistentwiththeorganizationofoursubsequentdiscussion,theresultsaregroupedaccording totheindependentvariablesinvolved,ratherthanbygroupsofspeakers.Table14.2containsall maineffectsofsituation,Table14.3allsignificantinteractionsofsituationwithraceand/orSES, Table14.4alleffectsofraceandSESthatdonotinvolvesituation. SituationMainEffects.Forthetargetchildren,allsignificantmaineffectsofsituationinvolvea differenceinmeansinthesamedirection;thetargetchildrenusemoreAffectiveReflections, PerceptualReflections,Nonreflections,allSemanticUses,andNonliteralUsagesathomethanat school.Itmustbekeptinmindthatthesevariablesrepresentproportionswiththetotalnumber ofwordsspokenasthedenominator.Strictlyspeaking,onemustsaynotthatthechildrenuse moreSemanticInternalStateWordsathomethanatschool,butthatSeman 258

TABLE14.2 MainEffectsofSituationonInternalStateWordVariables Speakers ____________________________________________________ Target Children Variable Cognitivereflections Affectivereflections Perceptualreflections CAPreflections Intentionsanddesires reflections Allreflections Nonreflections Semanticusages Attentional Nonliteralusages Allinternalstatewords (df=1,32) F<1 F=28.96a F=8.29a F=5.16a F=2.92 F=3.47 F=5.53a F=5.04a F<1 F=10.59a F=4.12 Teachersand Mothers (df=1,64) F<1 F=12.91a F<1 F=1 F=16.42a F=13.45a F<1 F=8.11a F=11.07b F<1 F=2.91 AllAdultsin Environment (of=1,67) F=8.25a F=24.93a F<1 F=18.33a F=12.19a F=20.75a F<1 F=9.11a F=6.93b F=12.37a F=6.40a

aHomemeanisgreaterthanschoolmean. bSchoolmeanisgreaterthanhomemean. p<.05 p<.01

ticInternalStateWordsconstitutealargerproportionofthechildren'sspeechathomethanat school. Lookingattheadultspeechthechildrenareexposedtobothintermsofprimarycaregivers (mothersandteachers)andalltheadultsintheenvironmentasimilarpatternisseen.Inthecase ofthespeechofalladults(forwhichmoreofthemaineffectsofsituationaresignificant),there areproportionatelymoreCognitive,Affective,andIntentionsandDesiresReflectionsusedat homethanatschool;thesameholdsforCognitive,Affective,andPerceptualReflectionstaken asagroup(CAPreflections),forallReflections,allSemanticUses,NonliteralUsages,andall InternalStateWords.TheonlyexceptionisforAttentionalDevices,whichareusedmoreby adultsatschoolthanbyadultsathome. Incomparingtherelationshipofthespeechofthetargetchildrenandadults,threedifferent patternsemergefromaninspectionofTable14.2.First,fortwovariablesAffectiveReflections andSemanticUsesthereisasignificanthomeschooldifferencefoundamongadults,anda correspondingsignificantdifferenceamongthetargetchildren.Second,foralmostalltheother variables,

259

TABLE14.3 InteractionsInvolvingSituationonInternalStateWo

Spe TargetChildren(df=1,32) ____________________________ Variable Cognitive reflections Affective reflections Perceptual reflections CAPreflections Intentionsand desiresreflections Allreflections Nonreflections Semanticusages Attentional Nonliteral usages Allinternal statewords
p<.05 p<.01

________________________________________

MothersandTe Race <1 2.34 1.69 1.42 <1 3.64 <1 2.85 <1 4.08 <1

_____________

Race 1.83 2.71 <1 3.63 3.28 5.94 2.73 9.22 1.92 <1 13.11

SES 1.73 <1 2.37 <1 1.34 <1 4.63 <1 1.68 <1 <1

Race&SES 1.03 <1 <1 1.92 <1 3.16 1.71 1.65 3.41 1.51 <1

260

TABLE14.4 RaceandSESMainEffectsandInteractio

Spea TargetChildren(df=1,32) ____________________________ Variable Cognitive reflections Affective reflections Perceptual reflections CAPreflections Intentionsand desiresreflections Allreflections Nonreflections Semanticusages Attentional devices Nonliteral usages Allinternal statewords
a

__________________________________________

TeachersandMoth Race <1 <1 1.08 <1 2.80 <1 <1 <1 4.77a <1 <1

________________

Race <1 <1 2.04 <1 2.34 <1 2.02 1.80 13.94a <1 4.34a Whitemeanisgreaterthanblackmean.

SES <1 <1 1.09 1.32 <1 1.20 <1 1.33 1.29 <1 <1

Race&SES 3.99 <1 <1 3.74 <1 2.45 <1 1.78 4.93 1.69 <1

SE 1.1

<1

2.3

<1

<1

<1 <1 <1 <1

<1

<1

bWorkingclassmeanisgreaterthanmiddleclassmean. p<.05 p<.01

261

thereisasignificanthomeschooldifferenceamongadults,butnosignificantdifferenceamong thechildren.(Amongthevariablesunderdiscussionhere,thepatternofmeansamongtarget children,thoughnonsignificant,wasalwaysinthesamedirectionasthesignificantdifferences amongtheadults.)Thethirdpattern,perhapsthemostinterestingofthethree,involvesthetwo variablesPerceptualReflectionsandNonreflections,forwhichthetargetchildrenshowa significantsituationaldifference,whereasneithertheprimarycaregiversnortheadultsingeneral showsuchadifference.Moreover,inthecaseofNonreflectionsthereisalmostnodifferenceat allinthemeansfortheadultsinthehomeversusschoolenvironments.InthecaseofPerceptual Reflections,wherechildrenhaveasignificantlyhighermeanathomethanatschool,theadults (bothprimarycaregiversandadultsingeneral)showanonsignificantdifferenceintheopposite direction. Thisillustratesapointthatwillbecomeevenclearerinfurtherresults.Specifically,theuseof internalstatewordsbythetargetchildrenreflectstheinternalstateworduseoftheadultsintheir environmenttoacertainextentandincertainrespects,buttherearealsoimportantaspectsofthe targetchildren'sinternalstatewordusethatcannotbeaccountedforintermsoftheinternalstate worduseoftheadultsaroundthem. InteractionsofSituationWithRaceand/orSES.Thesignificantinteractionsinthespeechofthe targetchildrenreportedinTable14.3wereeachfollowedupbyposthoctesting.TheTukeytest wasusedforallpairwisecomparisons,andtheScheffemethodwasusedforthenonpairwise comparisons.Asignificancelevelof.05wasusedforallposthoctesting. ThesignificantsituationXraceinteractionssuggestedthatblackchildrenhadlowermeansat schoolthanathomeforReflections,SemanticUses,andInternalStateWords.Atschool,white childrenhadhighermeansthanblackchildrenforSemanticUsesandInternalStateWords;white childrenathomealsohadahighermeanthanblackchildrenatschoolforSemanticUsesand InternalStateWords.Furthermore,itwasfoundthatthemeanfortheblacktargetchildrenat schoolwassignificantlylowerthanthemeansfortheotherthreecells(blackchildrenathome andwhitechildrenathomeandatschool),takenasagroupforReflectionsandInternalState Words.Also,thecombinedblackhomeandwhiteschoolmeanswerehigherthanthecombined whitehomeandblackschoolmeans;thiswassignificantforSemanticUsesandInternalState Words.ThesignificantsituationXSESinteractionforNonreflectionsindicatesthatthemeanfor middleclasstargetchildrenathomeishigherthanthemeanformiddleclasstargetchildrenat school. ThesituationXraceinteractionthefactthattheblacktargetchildren'smeanatschoolfor Reflections,SemanticUses,andInternalStateWordsislowerthantheirmeanathome,and lowerthanwhitetargetchildren'smeansatbothhomeandschoolbearscruciallyonthe theoreticalissuesandhypothesescentraltothepresentresearch.Thereisapparentlysomething abouttheschool 262

situation,ortheblacktargetchildren'sresponsetoit,thatcausesthemtouseproportionately fewerinternalstatewordsthanthewhitetargetchildrenuseatschool.Oneofthechiefgoalsof furtheranalysesisthereforetoidentifyasfaraspossiblethesituationalfactorsinfluencingthe children'sinternalstateworduse,andtoseetowhatextentthesefactorscontributetothe situationXraceinteraction. Theinteractiontakesonadifferentcharacterfortheanalysesinvolvingteachersandmothers, however.ThesignificantsituationXraceXSESinteractionforReflectionshasthefollowing pattern:athome,thewhitemiddleclassmothershavethehighestmean,andtheblackmiddle classmothershavethelowestmean;atschool,theteachersofblackmiddleclasstargetchildren havethehighestmean,andtheteachersofwhitemiddleclasstargetchildrenhavethelowest mean.ThesamepatternholdsforCAPReflections,andalmostthesamepatternforAffective Reflectionsaswell. ThesituationXraceinteractionforNonliteralUsages,ontheotherhand,indicatesthatthemean forblackmothersishigherthanthemeanforwhitemothers.Thisorderisreversedamong teachers.Thesamepatternobtainsfortheinteractionintheadultenvironmentathomeandat school. Inthecaseofoverallsituationalvariation,therewassimilarity,atleastforsomevariables, betweenthechildrenandtheadultsintheirenvironment.Inbothcases,therewasageneral tendencytousemoreinternalstatewordsinmostcategoriesathomethaninschool,bothonthe partofthetargetchildrenandtheadults.Inthecaseoftheinteractionsofsituationwithraceand SES,however,thereisnoapparentsimilaritybetweenthetargetchildrenandtheadultsintheir environment.Morespecifically,notraceofthesituationXraceinteractioninthetargetchildren's useofReflections,SemanticUses,andInternalStateWordscanbefoundinthespeechoftheir primarycaregiversoroftheadultsintheirenvironments.Therefore,itappearsthatan explanationforthisinteractioninthespeechofthetargetchildrenwillhavetobesought elsewherethanintheinternalstateworduseoftheadultsintheirenvironment. MainEffectsandInteractionofRaceandSES.Finally,wewanttotakenoteofanyoverallrace orSESdifferencesindependentoftheeffectofsituation(seeTable14.4).Amongthetarget children,racemaineffectsindicatethatwhitetargetchildrenusemoreAttentionalDevices,and moreInternalStateWordsingeneral,thandoblacktargetchildren.Thispatternholdsfortheuse ofAttentionalDevicesamongprimarycaregivers,andfortheuseofIntentionsandDesires Reflectionsintheadultenvironments.ThesignificantmaineffectofSESontheuseof PerceptualReflectionsbyadultsintheenvironmentindicatesagreateruseofthiscategoryof internalstatewordsintheenvironmentsofworkingclasstargetchildren. ItshouldbenotedthatthemaineffectsofraceandSESinourdataarebothratherfewandalso ratherspecific.Whitetargetchildrendousemoreinternalstatewordsingeneralthandothe blacktargetchildren,but,asitisseeninthe 263

Followinganalyses,thisdifferenceholdsonlyatschool,notathome.Thewhitetargetchildren alsousemoreAttentionalDevices(thatis,theysay"Look!"moreoften),buttherearenoother maineffectsofracefoundamongthetargetchildren. Amongprimarycaregivers,theonlysignificanteffectofraceorSESisagainintheuseof AttentionalDevices.Andamongadultsintheenvironment,thereisonlyonesignificantmain effecteachforraceandSES;inbothcases,aspecificsubcategoryofreflectionsisinvolved. WithinSituationAnalyses RaceXSESANOVAswerealsoperformedonalldependentvariableswithineachsituationfor eachgroupofspeakersanalyzedinthepreviousANOVAs(thatis,targetchildren,mothersand teachers,andalladults).TheresultsoftheseanalysesaresummarizedinTables14.5and14.6. Therearesignificanteffectsofraceinthespeechofthetargetchildrenatschool.Aswas indicatedbythesituationXraceinteractionintheacrosssituationanalysesabove,theblacktarget childrenshowfewerSemanticUsesandInternalStateWordsthanthewhitetargetchildrenat school,eventhoughthereisnosuchracialdifferenceinthehomesituation. TheInfluenceofAdultInternalStateWordUseonTargetChildren Havingnowpresenteddataonvariationinthetargetchildren'suseofinternalstatewords,we wanttoseetowhatextentthisvariationcanbeaccountedforintermsofothermeasurable aspectsofthetargetchildren'scommunicativeenvironment.Anobviousplacetostartiswiththe internalstateworduseoftheadultswhosespeechthetargetchildrenhearathomeandinschool. Therearetwobasicreasonswhywemightexpecttheinternalstateworduseofthetarget childrentoresemblethatoftheadultsintheirenvironment.First,sincethechildrenarelearning thelanguagelargelyfromtheadultsintheirenvironmentespeciallythehomeenvironment onewouldexpectsimilaritiesbetweenadultsandchildreninatleastsomeaspectsoftheir speech.Second,boththeadultsandchildreninagivensituationmaybesimilarlyaffectedby someaspectofthesituationthatinfluencesinternalstateworduse(forexample,thegeneral topicofconversation). Onemeasureofsimilaritybetweentheinternalstateworduseofchildrenandadultsisto comparethepatternsofmeans.InTable14.4,forexample,itcanbeseenthattheracial differencesinthetargetchildren'suseofattentionaldevices(whitechildrenusemoreofthem thandoblackchildren)isfoundinthespeechoftheirprimarycaregiversaswell.Ontheother hand,theSituationXRaceinteractiononsemanticusagesandinternalstatewordsfoundinthe speechof 264

TABLE14.5 WithinSituationEffectsofRaceandSESatH

Spea TargetChildren(df=1,34) ____________________________ Variable Cognitive reflections Affective reflections Perceptual reflections CAPreflections Intentionsand desiresreflections Allreflections Nonreflections Semanticusages Attentional devices Nonliteralusages Allinternal statewords Race 1.97 2.41 2.56 6.14a <1 2.62 <1 2.59 9.10b <1 1.94 SES 1.13 <1 3.82 <1 1.93 2.27 2.74 <1 <1 <1 <1 Race&SES 1.66 <1 <1 <1 <1 <1 1.50 <1 <1 <1 <1 Race <1 2.48 3.34 <1 3.99 3.87 <1 2.89 1.30 <1 1.01

_________________________________________

Mothers(df

_________________

SES <1 <1 <1 <1 <1

<1 3.69

1.11 <1

<1 2.08

aBlackmeanisgreaterthanwhitemean. bWhitemeanisgreaterthanblackmean. cMiddleclassmeanisgreaterthanworkingclassmean.

p.05 265

p.01

TABLE14.6 WithinSituationEffectsofRaceandSESatS

Spea TargetChildren(df=1,33) ____________________________ Variable Cognitive reflections Affective reflections Perceptual reflections CAPreflections Intentionsand desiresreflections Allreflections Nonreflections Semanticusages Attentional devices Nonliteral usages Allinternal statewords Race <1 <1 <1 <1 3.63 3.68 3.02 5.88a 5.56a <1 9.03a SES <1 1.43 <1 <1 <1 <1 2.19 <1 <1 <1 <1 Race&SES 2.81 1.23 1.31 4.20 <1 3.96 <1 2.45 3.22 2.03 1.04 Race <1 <1 <1 1.01 <1 <1 <1 <1 3.74 5.25a <1

_______________________________________________

Teachers(df=

_________________

SES <1

<1

4.24

<1

<1

<1 <1 <1

<1

<1

<1

aWhitemeanisgreaterthanblackmean. bWorkingclassmeanisgreaterthanmiddleclassmean. p.05

p.01 266

targetchildren(seeTable14.3)isnotpresentatallinthespeechoftheirprimarycaregivers oroftheadultsintheirenvironmentingeneral.Asmightbeexpected,certainaspectsofthe children'sinternalstatewordusemirrorthatoftheadultsintheirenvironmentmoreclosely thanothers. CorrelationalAnalyses Therelationshipbetweenthespeechoftargetchildrenandthespeechintheirenvironmentswas investigatedinmoredetailintermsofcorrelationalanalyses.Foreachvariableindicatinganaspect ofinternalstateworduse,correlationswereperformedtodeterminetowhatextentthespeechofthe targetchildrenresembledthespeechintheirenvironments,asrepresentedby(a)theprimary caregivers(teacherandmother),(b)alladultsintheenvironment,and(c)allspeakersinthe environment.ThesecorrelationsareshowninTable14.7. Itisinterestingtonotethatallthesignificantcorrelationsarepositive,and,withtheexceptionofthe significantpositivecorrelationbetweenthetargetchildandtotalenvironmentathomeforsemantic usages,allthesignificantrelationshipsinvolvereflections,orsomesubcategoryofreflections. Specifically,atschoolthereisasignificantcorrelationbetweenthetargetchildandhisorheradult environmentForaffectivereflections,perceptualreflections,CAPreflections,intentionsanddesires reflections,andallreflections.Athome,thereisasignificantcorrelationforeverycategoryof reflections;thatis,forallthosesignificantfortheschoolenvironmentaswellascognitive reflections.Becausereflections,bydefinition,areintimatelyrelatedtothespeakerand/orhearerin thecurrentsituation,itappearsreasonabletohypothesizethatthe"immediacy"ofreflectionsthat is,thefactthattheyrefertothespeaker'soraddressee'sowncurrentinternalstateleadsthemto showahighdegreeofcorrelationbetweenthetargetchildanditsenvironment.Becausetheyare especiallyrelatedtothe"hereandnow,"theymaybemorestronglyinfluencedbysituationalfactors suchasthegeneralsubjectmatterthataffectsallparticipantsintheconversationinasimilar fashion. AscanbeseeninTable14.7,thespeechofthetargetchildrengenerallycorrelatesmorestrongly withthespeechofalladultsthanwiththespeechoftheprimarycaregiveralone,andmorestrongly stillwiththespeechofallspeakersintheenvironment. Table14.8presentsthesesamecorrelations,computedacrosssituation.Thesecorrelationsreflect thedegreetowhichthespeechofthetargetchildatschoolisinfluencedbythespeechinthehome environment,andthedegreetowhichthespeechofthetargetchildathomeisinfluencedbythe speechintheschoolenvironment. AcomparisonofTables14.7and14.8revealsthat,ingeneral,anysimilaritybetweenadults'and childrens'internalstateworduseiswithinsituations,ratherthanbetweensituations.Theinternal stateworduseathomedoesnothavea 267

TABLE14.7 WithinSituationCorrelationsBetweentheTargetChildanDifferentMeasuresoftheLinguisticE Situation School _________________________________ Teacher InternalStateWord UseVariable Cognitive reflections Affective reflections Perceptual reflections CAPreflections Intentionsand desiresreflections Allreflections Nonreflections Semanticusages Attentional devices Nonliteralusages Allinternal statewords

_______________________________________________________________

Hom All Mother Only (N=35) .217 .311 .035 .177 .033 .065 .193 .100 .131 .331 .196

__________________

Only (N=37) .280 .491 .162 .398 .348 .343 .244 .004 .037 .184 .048

Adults (N=37) .271 .496 .170 .424 .304 .329 .243 .002 .052 .176 .049

Speakers (N=37) .300 .501 .361 .433 .493 .414 .245 .000 .183 .182 .048

Adult

(N=3

.605

.519

.239

.441 .211

.177

.108 .219

.056

.250 .078

p<.05,twotailed

p<.01,twotailed 268

TABLE14.8 AcrossSituationCorrelationsBetweentheTargetChildandDifferentMeasuresoftheLinguistic Situation TargetChildatHome __________________________________ All Teacher InternalStateWord UseVariable Cognitive reflections Affective reflections Perceptual reflections CAPreflections Intentionsand desiresreflections Allreflections Nonreflections Semanticusages Attentional devices Nonliteralusages Allinternal statewords Only (N=33) .366 .296 .078 .099 .073 .046 .154 .181 .262 .157 .326
p<.05,twotailed p<.01,twotailed

_______________________________________________________________

TargetChild

__________________ Mother Only (N=36) .031 .174 .000 .055 .074 .040 .232 .002 .024 .183 .067

Adults atSchool (N=36) .378 .292 .095 .114 .063 .026 .104 .177 .223 .123 .330

Speakers atSchool (N=36) .349 .297 .010 .209 .059 .099 .177 .247 .114 .000 .300

Adult

atHom

(N=3 .212

.006

.057

.214 .092

.116 .287 .037 .115

.303

.076

269

measurableinfluenceonthechild'sspeechatschool;similarly,theinternalstateworduseinthe schoolenvironmentdoesnothavemuchinfluenceonthetargetchild'sspeechathome.One exceptiontothisisthatthetargetchild'suseofcognitivereflectionsathomecorrelatespositively withtheuseofcognitivereflectionsintheschoolenvironment,regardlessoftheparticular definitionoftheenvironment(teachers,adults,orallspeakers).Thetargetchildren'soveralluse ofinternalstatewordsathomealsocorrelateswiththeuseofinternalstatewordsbyadultsin theschoolenvironment.Theonlysignificantinfluenceofthehomeenvironmentuponthetarget children'sspeechatschoolisrepresentedbyapositivecorrelationbetweenthetargetchildren's nonliteralusagesatschoolandthenonliteralusagesofthetotalenvironmentathome. ThesharpcontrastinmagnitudeofcorrelationsbetweenTables14.7and14.8suggeststhattarget children'suseofinternalstatewords,totheextentthatitisinfluencedbythespeechofadultsin thetargetchildren'senvironment,islargelyaccountedforbytheinternalstateworduseofadults withinthesituation.Verylittleofthetargetchildren'suseofinternalstatewordsinschoolcanbe predictedfromthespeechthatthechildrenareexposedtointhehome,norcantheiruseof internalstatewordsathomebepredictedfromtheinternalstatewordusetheyencounterat school. Onespecifichypothesisweareinterestedinevaluatingintermsofourdataisthatchildrenfrom nonmainstreambackgroundsexperiencediscontinuityormismatchbetweentheinternalstate worduseofadultsathomeandthatwhichtheyencounterinthespeechofadultsatschool.This mightbethecase,forexample,ifthereweregreatdifferencesbyraceorSESintheuseof internalstatewordusageencounteredbychildrenfromthedifferentgroupsatschool.Inthis case,somechildren(presumablythe"mainstream"ormiddleclasschildren)wouldexperiencea fairdegreeofcontinuitybetweentheinternalstatewordusageencounteredathomeandat school.Otherchildren,nonmainstreamorminoritypoorchildren,wouldfindthattheinternal stateworduseofadultsatschoolwasquitedifferentfromwhattheyhadexperiencedathome. Ourdatagivenosupportforthishypothesis.AswasseeninTable14.4,thereareonlyafew,very specificeffectsofraceorSESintheinternalstateworduseofadultseitherathomeoratschool. AnyhomeschooldiscontinuitywouldshowupasaSituationXRace,SituationXSES,or SituationXRaceXSESinteractioninthespeechofadults.Thereareafewsuchsignificant interactions(seeTable14.3),butexceptforthoseinvolvingnonliteralusages,thepatternof meansinvolvedisnotconsistentwiththemismatchhypothesis(e.g.,thehomeschooldifference isgreaterforthewhitemiddleclassmothersandteachersthanfortheotherrace/SESgroup). Therefore,exceptforthecaseofnonliteralusages,thereisnoevidenceofahomeschool mismatchintheinternalstateworduseofadultsforthemainstreamchildreninoursample. Indirectevidenceforamismatchcouldbefoundbycomparingthespeechofthetargetchildren athomeandatschooltoassessthedegreeofcontinuity 270

betweenhomeandschoolintermsoftheirinternalstateworduse.Table14.9presentsthe correlationsbetweenthetargetchildren'shomeandschoolinternalstateworduse,calculated separatelyforthetworacialgroups. Table14.9doesinfactshowapronounceddifferencebetweenthetworacialgroups.Forfour variablesintentionsanddesiresreflections,reflections,semanticusages,andinternalstatewords thewhitetargetchildrenshowasignificantsimilarityintheirspeechbetweenhomeand school.Itcanbenotedthatthelastthreeofthesefouraresuperordinatecategoriesrepresenting anincreasinglygeneralpictureofinternalstateworduse.However,therearenosignificant correlationsbetweenhomeandschoolfortheblacktargetchildren,andthedifferencesinthe magnitudeofcorrelationsbetweenhomeandschoolforblacksandwhitesaresignificantfor thesesamefourvariables.Forthesevariables,then,thewhitetargetchildrenshowahighdegree ofsimilaritybetweentheirinternalstateworduseathomeandatschool.Thissuggeststhatfor thosefactorscontrollinginternalstatewordusetheyarealsoexperiencingadefinitecontinuity betweenthehomeandschoolenvironments.Fortheblacktargetchildren,ontheotherhand, thereisnomeasurablesimilaritybetweentheirinternalstateworduseathomeandatschool. Thisseemstosuggestthat,withrespecttothoseaspectsoftheenvironmentthatinfluence internalstateworduse,theyareexperiencingsomediscontinuityormismatchbetweenthehome andschoolsituation. TABLE14.9 CorrelationsbyRacialGroupBetweenTargetChildren'sInternalStateWordUseatHomeand atSchool TABLE14.9 CorrelationsbyRacialGroupBetweenTargetChildren'sInternalStateWordUseatHomeand atSchool TABLE14.9 CorrelationsbyRacialGroupBetweenTargetChildren'sInternalStateWordUseatHomeand atSchool Variable Cognitive reflections Affective reflections Perceptual reflections Intentionsand desiresreflections Allreflections Nonreflections Semanticusages Attentional devices Nonliteralusages Internalstate Whites(N=17) Blacks(N=19) .005 .108 .021 .045 .705 .273 .699 .338 .342 .714

Z .28 1.21 85 2.96

.434 .261 .202 .422 .228 .081 .145 .284 .013

.613

3.18 1.40 2.59 .56 .18 2.44

Variable words
p<.05,twotailed p<.01,twotailed

Whites(N=17) Blacks(N=19)

271

SESDIFFERENCESINCORRELATIONS Inlightoftheaboveresults(Table14.9),parallelanalyseswereperformedcomparing correlationscalculatedseparatelyforthetwoSESgroups.Tables14.10and14.11showthe correlationsbetweeninternalstatewordusepatternsoftargetchildrenandtheadultsintheir environment,athomeandatschool,respectively.Ascanbeseeninthesetables,noneofthe differencesincorrelationsbetweenthemiddleandworkingclassesaresignificanteitherathome oratschool.Thus,SESdoesnotseemtoplayaroleindeterminingthewayinwhichthetarget childrenrelatetotheinternalstateworduseoftheadultsintheirenvironment,eitherathomeor atschool.Itmustbenoted,ofcourse,thatintwocasestherearecorrelationsthatreach significanceforthemiddleclassbutnottheworkingclass.Athome,thereisasignificant positiverelationshipbetweentheuseofreflectionsbythemiddleclasstargetchildrenandthe adultsintheirclassrooms.However,inneithercaseisthedifferencebetweenthecorrelationsfor themiddleclassandtheworkingclasssignificant. Alsoofinterestisthedegreeofsimilaritybetweenthetargetchildren'sinternalstateworduseat homeandatschool.Table14.12presentsthecorrelationsforthedifferentinternalstateword variablesbetweenthechildren'shomeandschoolinternalstatewordusepatterns.Inthiscase, noneofthedifferences TABLE14.10 CorrelationsbySESGroupBetweenInternalStateWordUseofTargetChildrenatHomeand AdultsintheHome MiddleClass Variable Cognitivereflections Affectivereflections Perceptualreflections Intentionsanddesiresreflections Allreflections Nonreflections Semanticusages Attentionaldevices Nonliteralusages Internalstatewords
p<.05,twotailed p<.01,twotailed

WorkingClass (N=19) .647 .590 .145 .234 .162 .013 .165 .075 .326 .072 Z .25 .17 .36 .19 .39 .53 1.33 1.22 .50 .86

(N=17) .590 .629 .272 .298 .296 .203 .575 .356 .154 .368

272

TABLE14.11 CorrelationsbySESGroupBetweenInternalStateWordUseofTargetChildrenatSchooland AdultsinSchoolEnvironment MiddleClass Variable Cognitivereflections Affectivereflections Perceptualreflections Intentionsanddesiresreflections Allreflections Nonreflections Semanticusages Attentionaldevices Nonliteralusages Internalstatewords p<.05,twotailed p<.01,twotailed (N=17) .257 .624 .398 .392 .521 .220 .095 .058 .193 .041 WorkingClass (N=19) .306 .434 .285 .262 .240 .352 .084 .154 .146 .101 Z .15 .27 1.95 .40 .91 .39 .49 .27 .13 .39

TABLE14.12 CorrelationsbySESGroupBetweentheHomeandSchoolInternalStateWordUseoftheTarget Children MiddleClass Variable Cognitivereflections Affectivereflections Perceptualreflections Intentionsanddesiresreflections Allreflections Nonreflections Semanticusages Attentionaldevices Nonliteralusages Internalstatewords p<.05,twotailed p<.01,twotailed (N=17) .301 .503 .075 .030 .073 .242 .227 546 .112 .299 WorkingClass (N=19) .097 .059 .206 .575 .051 .086 .007 .101 .565 .002 Z 1.11 1.35 .78 1.87 .34 .44 .64 1.95 1.44 .85

273

betweencorrelationsforthetwoSESgroupsaresignificant.However,thepatternsof correlationsdifferinaninterestingway:themiddleclasschildrenshowasimilaritybetween homeandschoolintheiruseofaffectivereflectionsandattentionaldevices,andtheworking classchildrenhaveasimilaritybetweenhomeandschoolintheiruseofintentionsanddesires reflectionsandnonliteralusages. Ingeneral,theseresults,takentogetherwiththoseinTables14.914.11,suggestthatraceisamore importantfactorthanSESindeterminingthetargetchildren'sadjustmenttotheclassroomsituation. EffectsofHomeandSchoolEnvironmentsonChildren'sSpeechinSchool Thefindingsjustdiscusseddescribetheeffectsofcertainaspectsofcommunicationpatternsina situationonthespeechofthetargetchildreninthatsituation. TABLE14.13 ComparingCorrelationsBetweenTargetChildren'sInternalStateWordUseatSchoolwithIndices ofConversationalStructureatHomeandatSchool IndicesofConversationalStructure ____________________________________________________ PercentofInternal PercentofPrimaryCaregiver'sInternal StateWordsAddressedto ______________________________________ TargetChild InternalState WordUse Cognitivereflections Affectivereflections Perceptualreflections Intentionsanddesires reflections Allreflections Nonreflections Semanticusages Attentionaldevices Nonliteralusages Allinternalstatewords p<.05,twotailed p<.01,twotailed 274 TargetChild Children StateWordsbyall Speakers Addressedto TargetChild

___________________ ___________________ ___________________ _ _ Home .181 .173 .073 .053 .014 .531 .185 .244 .218 .291 School .198 .040 .078 .009 .092 .258 .185 .396 .396 .350 Home .221 .179 .040 .042 .091 .400 .217 .433 .165 .375 School .003 .141 .154 .191 .122 .023 .122 .064 .228 .079 Home .127 .287 .020 .135 .107 .294 .008 .326 .174 .124 School .289 .103 .147 .125 .036 .161 .098 .324 .378 .240

TABLE14.14 ComparingCorrelationsBetweenTargetChildren'sInternalStateWordUseatSchoolwith IndicesofConversationalStructureatHomeandatSchool IndicesofConversationalStructure ___________________________________________ PercentofTurns Numberof TargetChild InternalState WordUse Cognitivereflections Affectivereflections Perceptualreflections Intentionsanddesires reflections Allreflections Nonreflections Semanticusages Attentionaldevices Nonliteralusages Allinternalstatewords
p<.05,twotailed p<.01,twotailed

Percentof TurnsbyChild

byPrimary Caregiver

TurnsbyChild

___________________ ___________________ ___________________ ____ ___ __ Home .031 .149 .176 .126 .147 .063 .162 .042 .179 .172 School .376 .147 .105 .106 .192 .473 .332 .104 .295 .351 Home .231 .016 .113 .014 .114 .413 .243 .290 .222 .356 School .296 .181 .171 .012 .154 .465 .293 .378 .364 .432 Home .072 .233 .086 .035 .060 .213 .011 .047 .201 .062 School .118 .365 .192 .470 .237 .380 .340 .253 .052 .357

However,themismatchmodelpresumesthattherearealsocrosssituationaleffects.Specifically,we wanttolookforwaysinwhichpatternsofcommunicationathomemayinfluencethechild's responsetotheschoolsituation.Tables14.13,14.14,and14.15presentcorrelationsbetweenindices ofthepatternsofcommunicationathomeandschoolandthetargetchildren'sinternalstateword useatschool. Forsomevariables,ofcourse,thereareonlywithinsituationeffects.Forexample,thetargetchild's internalstateworduseatschoolcorrelateswiththenumberofturnsspokenbythetargetchildat school,butnotwiththenumberofturnsthetargetchildspeaksathome.Similarly,thetargetchild's schoolinternalstatewordusecorrelateswiththepercentageofturnstakenbytheteacher,butnot withthepercentageofturnsatdinnerspokenbythemother.Nordothenumberofadultspresentat dinner,thepercentageofturnsatdinnerspokenbyadults,orthepercentageofturnsatdinner spokenbytheexperimenterhaveanyinfluenceonthetargetchildren'sinternalstateworduseat school. Theremainingfourvariables,however,doevidencecrosssituationalcorrelations.Thehigherthe percentageofthemother'sinternalstatewordsaddressedto 275

TABLE14.15 CorrelationsBetweenTargetChildren'sInternalStateWordUseatSchoolwithIndicesof ConversationalStructureatHome TABLE14.15 CorrelationsBetweenTargetChildren'sInternalStateWordUseatSchoolwithIndicesof ConversationalStructureatHome IndicesofConversationalStructure ___________________________________________________ ______________ PercentofTurnsat TargetChild InternalState WordUse Cognitive reflections Affective reflections Perceptual reflections Intentionsand desiresreflections Allreflections Nonreflections Semanticusages Attentionaldevices Nonliteralusages Allinternal statewords

Numberof AdultsPresentat Dinner .298 .072 .108 .107 .182 .037 .164 .014 .007 .149

DinnerSpokenby ____________________________________ _____ Experimenter .059 .068 .085 .184 .160 .253 .236 .124 .091 .286 Adults .007 .152 .105 .155 .105 .214 .032 .106 .170 .029

p<.05,twotailed

p<.01,twotailed

thetargetchildathome,themorenonreflectionsthetargetchildusesinschool.Thehigherthe percentageofmothers'internalstatewordsaddressedtochildreningeneral,themore nonreflections,attentionaldevices,andinternalstatewordsthetargetchildusesatschool.And thegreaterproportionoftheturnsathomethatthetargetchildhas,themorenonreflectionsand internalstatewordsthechildusesatschool. Theinfluenceofthehomeenvironmentonthechild'suseofnonreflectionsinschoolis noteworthy.Themoreattentionthetargetchildreceives(intermsofinternalstatewords),and thelargerthetargetchild'sshareoftheconversationathome,themorenonreflectionsheorshe usesatschool.Thisisinteresting,firstofallbecausethesesamefactorsdonotcorrelate significantlywiththetargetchild'suseofnonreflectionsathome,andsecondbecauseofthe possibledevelopmentalimplicationsofuseofnonreflections.Nonreflectionsinvolvemore

displacementorabstractionthanreflections,andhencemayindicateamoreadvancedtypeof usage. Animportantimplicationofthepatternofcorrelationsinthesetablesisthespecificnatureofthe influencesofhomeandschoolenvironmentsonchildren'sspeech.Firstofall,mostofthe influencesaresituationspecific;forexample,the 276

percentageofthemother'sinternalstatewordsaddressedtothetargetchildinfluencesthe targetchild'suseofattentionaldevicesathome,butnonreflectionsatschool.Second,asthis sameexamplealsoillustrates,specificaspectsofthecommunicationpatternsinasituation canaffectveryspecificaspectsofthechild'sspeech.Anymodeloftheinfluenceofhomeand schoolenvironmentsonchildren'sspeechmustthereforeallowthisdegreeofspecificityand detail.OverallmeasurementslikeMeanLengthofUtterance(MLU)willbefartoocrudeto assesstheeffectsofsituationalfactorsonpatternsoflanguageuse(cf.Coleetal.,1978). IMPLICATIONSFORTHEMISMATCHHYPOTHESIS Noevidencewasfoundforanymismatchbetweenhomeandschoolenvironmentsforminorityor poorchildren,asfarastheinternalstateworduseofadultswasconcerned.However,thespeechof whitetargetchildrenatschoolisrelatedstronglybothtotheirspeechathomeandtothespeechof theadultsintheschoolenvironment.Thespeechoftheblacktargetchildrenatschool,however, showsnorelationshipeithertotheirownspeechathomeortothespeechoftheadultsinthe classroom.Thissuggestsverystronglythat(a)theblacktargetchildren'srelationshiptotheschool situationissignificantlydifferentfromthatofthewhitetargetchildren;and(b)therelationship betweenthehomeandschoolenvironmentsisdifferentintheexperienceoftheblackandwhite targetchildren.Thiscouldbetheresulteitherofahomeschoolmismatchexperiencedbytheblack targetchildrenthatisnotreflectedinthevariableswehaveinvestigated,orofdifferentialtreatment ofthetworacialgroupsinschoolintermsofteachingstylesorpatternsofinteraction.Ourfindings suggeststronglythatsuchamismatchordifferencedoesexist,butnotintheinternalstateworduse oftheadultsassuch.Rather,itappearsthatpatternsofinteractionbothathomeandinthe classroomhavesignificanteffectsontheuseofinternalstatewordsbythechildrenatschool. Likewise,ourdatafailedtosupportoneparticularversionofthemismatchhypothesis,namely,the hypothesisthatthehomeenvironmentsofsomenonmainstreamchildrenwouldprovidethemwith lessexperienceinmetabehavioralanalysisthandidthehomeenvironmentsofwhitemiddleclass children.However,theresultsofouranalysesdosuggestthattheblackchildreninoursample experiencedsomesortofmismatchbetweenhomeandschool,althoughthismismatchdidnot directlyinvolvetheinternalstateworduseofadults.Threetypesofevidencepointtotheexistence ofsomesortofmismatch: First,thereisthereduceduseofinternalstatewordsbytheblacktargetchildrenintheschool situation.Althoughthisfindingpermitsmorethanoneinterpretation,itiscertainlyconsistentwith thehypothesisthattheblackchildrenfoundtheschoolsituationdifferentfromthehomesituationin importantways. Secondaretheracialdifferencesintheadultchildcorrelationsathomeandat 277

school.Athome,therearesomeracialdifferences,anditisintheblackfamiliesthatthereare thehighestcorrelationsbetweenchildren'sandadults'internalstateworduse.Atschool,on theotherhand,racialdifferencesarefoundforalargernumberofvariablesanditisthe whitechildrenwhoshowthehighestcorrelationwiththeirteachers.Lackofcorrelationin internalstatewordusepatternsdoesnotmeanlackofcommunication.However,differences inthedegreeofcorrelationbetweenadultandchildinternalstatewordusedoessuggest differencesinthenatureofthecommunication.Ahighcorrelationbetweenadultandchild internalstatewordusesuggeststhatbothadultsandchildrenarebeingaffectedbysituational factorse.g.,thetopicofconversationandtheactivitiesthespeakersareengagedinin parallelways.Alackofcorrelationindicatesthatthesituationalfactorsaffectadultsand childrendifferently.Thusahighercorrelationcouldbeinterpretedasindicatingagreater similarityorreciprocitybetweentherolesofadultsandchildreninthesituation.Alackof correlationcouldindicatesomebreakdownincommunication,orsimplyagreaterdegreeof distinctionbetweentherolesofadultsandchildrenintheconversation. Athirdtypeofevidenceforsomesortofhomeschoolmismatchisfoundinthecorrelations betweenchildren'sspeechathomeandchildren'sspeechatschool.Forfourinternalstatevariables, thewhitechildrenshowasignificanthomeschoolcorrelationwhiletheblackchildrenshownone. Thewhitechildren'sspeechatschool,then,islargelypredictablefromtheirspeechathome;asfar asinternalstatewordsareconcerned,theyseemtobeusingsimilarspeechpatternsinthetwo situations.Thelackofcorrelationonthepartoftheblackchildrenindicatesthatnotonlydothese childrentendtousefewerinternalstatewordsinschool,butalsothattheirschoolinternalstate worduseisnotatallpredictablefromtheirspeechathome.Thisindicatesthatfortheblack children,theschoolsituationdemandsasubstantiallydifferentsetofverbalstrategiesandskills. Analysesofthoseindicesofthepatternsofcommunicationforwhichwehadinformationindicated noclearexplanationforthespecificcauseoftheblackchildren'sreduceduseofinternalstatewords atschool.However,ourdatashowthatthedifferentfunctionalcategoriesofinternalstatewordsare sensitivetoveryspecifickindsofsituationalfactors.Theracialdifferenceinthechildren'sspeechat schoolshouldultimatelybetraceabletoparticularteachingstrategiesorstyles,ortosome differencesbetweenhomeandschoolcommunicationpatterns.Toidentifythese,though,would requiredetailedcategoriesforcodingmothers'andteachers'speechandbehaviorbeyondthescore ofthepresentwork. ACKNOWLEDGMENT TheresearchonwhichthischapterisbasedwassupportedbyagrantfromTheCarnegie CorporationofNewYorktoWilliamS.Hall.Thepreparationofthis 278

manuscriptwassupportedbytheNationalInstituteofEducationunderContractNo.USNIEC 400760116,TheDepartmentofPsychology,andtheComputerScienceCenterofThe UniversityOfMaryland,CollegePark. REFERENCES Baker,L.,&Brown,A.L.(1980).Metacognitiveskillsandreading(Tech,Rep.No.188) Urbana:UniversityofIllinois,CenterfortheStudyofReading.(ERICDocumentReproduction ServiceNo.ED195932). Bernstein,B.(1971).Class,codesandcontrol:Vol.1.Theoreticalstudiestowardsasociology oflanguage.London:Routledge&KeganPaul. Bernstein,B.(Ed.)(1973).Class,codes,andcontrol.:Vol.2:Appliedstudiestowarda sociologyoflanguage.London:Routledge&KeganPaul. Cazden,C.(1970).Thesituation:Aneglectedsourceofsocialclassdifferencesinlanguageuse. JournalofSocialIssues,26(2),3560. Cole,M.,Dore,J.,Hall,W.,&Dowley,G.(1978).Situationalvariabilityinthespeechof preschoolchildren.InM.Ebihar&R.Gianutsos(Eds.),Papersinanthropologyandlinguistics (pp.65105).NewYork:NewYorkAcademyofSciences. Cooper,B.(1975).Languagedifferencesandeducationalfailure(OccasionalPaper3). UniversityofSussex,EducationArea.(ERICDocumentReproductionServiceNo.ED121067). Flavell,J.H.(1978).Metacognitivedevelopment.InJ.M.Scandura&C.J.Brainerd(Eds.), Structural/processtheoriesofcomplexhumanbehavior(pp.213245).Alphena.d.Rijn,The Netherlands:Sijthoff&Noordhoff. Gearhart,M.,&Hall,W.S.(1982).Internalstatewords:Culturalandsituationalvariationin vocabularyusage.InK.Borman(Ed.),Childreninachangingsociety.Hillsdale,NJ:Lawrence ErlbaumAssociates. Halliday,M.A.K.(1978).Languageassocialsemiotics.Baltimore:UniversityParkPress. Hall,W.S.,&Nagy,W.E.(1979).Theoreticalissuesintheinvestigationofwordsofinternal report(Tech.Rep.No.146).Urbana:UniversityofIllinoisCenterforTheStudyofReading. (ERICDocumentReproductionServiceNo.ED177526). Hall,W.S.,&Nagy,W.E.(1986,pp.2665).Theoreticalissuesintheinvestigationofinternal statewords.InIrwin&MyrnaGopnik(Eds.),Frommodelstomodules:Studiesincognitive sciencesfromtheMcGillworkshops.Norwood,NJ:Ablex. Hall,W.S.,&Nagy,W.E.(inpress,1987).Thesemanticpragmaticdistinctionininternalstate words.DiscourseProcesses. Henderson,D.(1973).Contextualspecificity,discretionandcognitivesocialization:With

specialreferencetolanguage.InB.Bernstein(Ed.),Class,codes,andcontrol(Vol.2,pp.48 80).London:Routledge&KeganPaul. Labov,W.(1964).Phonologicalcorrelatesofsocialstratification.InJ.Gumperz&D.Hymes (Eds.),Theethnographyofcommunication[Specialissue].AmericanAnthropologist,66(6),164 176. Labov,W.(1966).ThesocialstratificationofenglishinNewYorkCity.Washington,DC:Center forAppliedLinguistics. Matthew,M.,Connolly,K.,&McCleod,C.(1978).Languageuse,role,andcontextina fiveyearold.JournalofChildLanguage,5,81,99. Sankoff,G.(1973).pp.4459.Aboveandbeyondphonologyinvariablerules.InC.Bailey&R. Shuy(Eds.),Newwaysofanalyzingvariationinenglish.Washington,DC:Georgetown UniversityPress. Schatzman,L.,&Strauss,A.(1955).Socialclassandmodesofcommunication.American JournalofSociology,60,329338. 279

Snow,C.E.,ArlmanA.Rupp,Hassing,Y.,Jobse,J.,Joosten,J.,&Vorster,J.(1976). Mother'sspeechinthreesocialclasses.JournalofPsycholinguisticResearch,5,120. Stein,N.L.,&Glenn,C.G.(1979).Ananalysisofstorycomprehensioninelementaryschool children.InR.Freedle(Ed.),Newdirectionsindiscourseprocessing(Vol.2).Hillsdale,NJ: LawrenceErlbaumAssociates. vonRafflerW.Engel,&Sigelman,C.(1971).Rhythm,narration,descriptioninthespeechofblack andwhiteschoolchildren,LanguageSciences,18,914. Wellman,J.,&Johnson,C.(1979).Understandingofmentalprocesses:Adevelopmentalstudyof "remember"and"forget."ChildDevelopment,50,7988. 280

15 CopingorGroping?PsycholinguisticProblemsintheAcquisitionofReceptiveandProductive CompetenceAcrossDialects WilliamA.StewartTheGraduateSchoolandUniversityCenteroftheCityUniversityofNewYork DespiteacertainamountofgoodnaturedcomplainingaboutthedifferencesbetweenBritishand AmericanEnglish,perhapsbestillustratedbythevariouslyphrasedandattributedwitticismthat EnglandandAmericaaredividedbyacommonlanguage,itisneverthelesspossibleforAmericans andBritonstofollowevenfairlyrapiddiscourseineachother'sdialectswithrelativeeaseand 1 apparentaccuracy. ThisisundoubtedlymosttruejustsolongasstandardvarietiesofBritishor AmericanEnglishareinvolved,butwiththatqualificationthesameinternationaleaseof comprehensionwouldextendtostandardIrish,Canadian,Australian,NewZealand,andSouth AfricanEnglishaswell,andtoonlyaslightlylesserextenttothestandardEnglishofIndia,Hong Kong,Singapore,andthoseAfricanandCaribbeanstateshavingEnglishastheirofficiallanguage. 2 Anobviousreasonforthishighdegreeofcomprehensibilitybetweenwidelydispersedvarietiesof standardEnglishisthatstandardizationandliteracyhaveminimizedstructuraldifferencesbetween them.Oncenonstandarddialectsofvariouskindsareincludedinthepicture,however,structural uniformitybeginstobreakdownovergeographicalandsocialdistance,andthe intercomprehensibilityofdialectsbecomesincreasinglylimitedasthestructuraldifferencesbetween themgrowgreater.Insomecases,twodialectsmaybestructurally ____________________ 1Thetermdialectisusedhereinthelinguist'ssenseofastructurallyidentifiablesubvarietyofa language.Inthisuse,thereisnoconnotationoflowprestigeorstructuraldeviationfroma literaryorofficialnorm. 2Astandarddialect(seefootnote1)orlanguagevarietyisoneinconformitywithcodifiednorms defining"correct"usage,andistheformmostoftenspokenbyeducatedpeopleandusedin officialdocuments,literaryworks,andformalsituations.Populardialectsthatdeviatefromthe standarddialect(s)arecallednonstandarddialects. 281

differentenoughsothateitherisvirtuallyimpossibleforspeakersoftheothertounderstand. Wherethisisso,therelationshipbetweenthetwodialectsofthesamelanguageapproaches thatbetweentwodifferentbutrelatedlanguages.Atthispoint,twomorecomplicationsmust beaddedtothenotionofstructuraldifferencesbetweendialectsandbetweenlanguages,for thatmatter.Thefirstisthat,insteadofbeingdistributedmoreorlessevenlythroughoutthe variousstructuraldomainsofdialecfsorlanguages,thedifferencesbetweenanytwomaybe concentratedinonestructuraldomainorlevel.Thustwodialects(orlanguages)maydiffer primarilyinaspectsofpronunciation,orofvocabulary,orofsentencestructureorother featuresoftheirrespectivegrammars.Moreover,thesedifferentkindsofstructural differences(phonological,lexical,syntactic,morphological,semantictousethemoreprecise linguistictermsforthem)caneachhaveadifferentialeffectoncrossdialectalcomprehension. Indeed,asisdemonstratedlater,theyseemalmosttoscaleinthisrespect,withphonological differencesmosteasilyresolved,semanticdifferencesposinggreatdifficulties,andtheothers inbetween.Thesecondcomplicationinthenotionofstructuraldifferencesbetweendialectsor relatedlanguagesisanoccasional(perhapsnotsorareafterall)comprehensionasymmetry. Thatis,givendialectsorrelatedlanguagesAandB,speakersofAfinditintrinsicallyeasierto understandBthantheotherwayaround.(What"intrinsically'meanshereisthatthe asymmetryisnotduetodifferentdegreesofcrossdialectalexposureforthetwopopulations, ortosuchextralinguisticfactorsasadifferenceinthesocialstatusofAandB.)Awellknown caseofsuchanasymmetryinvolvingrelatedlanguagesisthemarkedlygreatereasewith whichspeakersofPortugueseunderstandSpanishthanviceversa.Thecomprehension asymmetryinthiscaseseemstoarisefromthegreatercomplexityofPortuguesephonology andmorphology.Asimpleillustrationofhowthisworksisfurnishedbyamorphological exampleinvolvingtheSpanish(1)andPortuguese(2)formsof"general"(adjective)andthe derivationallyassociated"generality"(noun): (1)general:generalidade (2)geral:generalidade ItwillbenotedthatthePortuguesederivationalparadigmcontainsallthemorphological informationfortheSpanishforms,butnotviceversa,sothat(ignoringdifferencesinpronunciation) thePortuguesespeakerwillrecognizeSpanishgeneralinhisowngeneralidadeandthusbeableto equateitwithPortuguesegeral(orsoequateitthroughsomeunderlyingshapefromwhichboth geralandgeneralidadearederived),buttheSpanishspeaker'sderivationalparadigmgiveshimno informationthatwillrelatePortuguesegeraleitherdirectlyorindirectlytohisownadjective general.SuchstructuraldifferencesbetweenSpanishandPortuguese,existinginthedomainsof phonologyandsyntaxaswellasmorphology,andcompoundedmanytimesoverinnumberand complexity,could 282

accountinlargepartfortheoverallcomprehensionasymmetrythatexistsacrossthetwo languages. Followingthroughontheimplicationsoftheforegoingexample,itisnotatallevidentthatthe Portuguesespeaker'sgreaterabilitytounderstandSpanish,ifindeeditderivesfromthesekinds ofcrosslanguagestructuralrelationships,willbematchedbyagreaterabilitytoproduce accurateSpanishaswell.For,whilethatfragmentofthePortuguesegrammarspecifying geral:generalidadecontains(forhistoricalreasons)informationallowingthecorrect identificationofSpanishgeneral,noSpanishspecificinformationassuchisincluded.This meansthatthereisnothingtoconstraintheproductionforSpanishofaPortugueselike 3AndsofortherestofSpanish, adjectivalformgeral,whichisinfactnonexistentinSpanish. insofarasthepossibilitygoesofsimulatingitfromthecontentofPortuguesegrammar. Whatoneisseeinghereisanindicationofstillanotherkindofasymmetrythistimenotbetween theabilitiesofdifferentspeakersofdifferentlanguagesordialects,butratherbetweendifferent abilitiesforoneandthesamespeakerwithrespecttoasinglelanguageordialect.Ingeneral,it involvestheabilityofanindividualtocomprehendamuchgreaterrangeofdialectalvariation thanthatindividualuses,orprobablyevencanuse.Oneestablishedwayofreferringtothis phenomenonisastheasymmetrybetweenreceptiveandproductivecompetence,where competenceisone'sabilitytouselanguageanylanguage,oraparticularlanguage. SuchanasymmetrywaspositedinthecaseofthePortuguesespeaker'sreceptiveandproductive competenceinSpanish(assuming,ofcourse,aPortuguesespeakerwhohadnotactuallylearned tospeakSpanish);andwhiletheSpanishspeaker'sreceptivecompetenceinPortuguese(again assumingnospecificlearningofit)wouldbelessthanthePortuguesespeaker'sreceptive competenceinSpanish,itwouldbeconsiderablygreaterthantheSpanishspeaker'sproduction competenceinPortuguese.Norneedseparate(ifrelated)languagesbeinvolved;returningtothe illustrationwhichopenedthispaper,thehighlymutuallyintelligiblestandarddialectsofEnglish usedthroughouttheworld,anAmericannotspecificallyinstructedinBritishEnglishwould certainlynotbeabletospeakitwitheventheremotestapproximationtotheeaseandaccuracy 4 withwhich,asaspeakerofAmericanEnglish,hecouldunderstandit. ____________________ 3 Thewriter,whohasspokenbothlanguagessincechildhood,hasactuallyheardgeral(though withasuccessfullySpanishpronunciationoftheg)inBrazilianattemptstospeakSpanishin UruguayandArgentina. 4Thisdoesnotmeanthatitisimpossibletobeproductivelybidialectal,andcompetentlyso. Individualswhoarefluentintheproductionoftwoorevenmoredialectsofthesame languageabound,justasbilingualsandmultilingualsdo.Thepointisthatapparentreceptive competenceisnotnecessarilyevidenceofproductivecompetenceinthesamedialector language,whichleavesitanopenquestionwhetherapparentcrossdialectalreceptive competenceinitselfamountstobidialectalisminanyrealsenseespeciallyincases,suchas betweenstandardvarietiesofBritishandAmericanEnglish,wherethedialectsare structurallyquitesimilarthroughout. 283

Animportantpointneedstobemadeatthisjunctureconcerningthereceptive/productive competenceasymmetryasitappliesoverdialects(orrelatedlanguages)whichareincreasingly structurallydifferentfromthevarietyinwhichamonodialectal/monolingualspeakerhasboth productiveandreceptivecompetence.Ifamonodialectalspeakerofastandardvarietyof AmericanEnglishweretobetestedforreceptivecompetence(asdeterminedbyeaseand accuracyofcomprehension)forthefollowingdialects,inorder:standardBritishEnglish, standardIrishEnglish,standardIndianEnglish,Yorkshiredialect,SouthernCountiesScottish, OrkneyandShetlandIslandsScottish,onecouldassumewithconfidencethathisreceptive competencescorewouldstartquitehigh,declinewithsomethingofanabruptdropfromstandard IndianEnglishtoYorkshiredialect(correspondingtoanincreaseinstructuraldifferences),and endquitelowforShetlandIslandsScottish,whichcansoundforalltheworldlikeaforeign languagetoanuninitiatedAmerican.Then,ifthesameAmericanweretestedonproductive competenceinthesameseriesofdialects(using,say,anativespeaker'sjudgmentofthe structuralaccuracyoftheattempt),thesubjectcouldbeexpectedtoscoreabysmallyonthemore deviantdialects,butmightreceiveasurprisinglyhighscoreontheonesstructurallyclosetohis own.Sucharesultcouldgivetheimpressionthatsuchanindividual'sproductivecompetence acrossunacquireddialectsisscalar,andinfactfarfromlowfordialectsstructurallysimilartohis own.Andyetthesubjecthasbydefinitionlowproductivecompetenceinallthedialectstested for.Whatishappeninginsuchacaseisthateitherrealorapparentstructuralsimilaritiesbetween thesubjectsowndialectandthatofthetaskcanmasklowproductivecompetencewhenthe subjectcompensatesforitbyproducingtheformsheknows. Althoughtheillustrationjustgivenofthisproblemwashypothetical,beingcontrivedwitha widerstructuralrangeofdialectsthanistobefoundintheexperimentalliterature,asomewhat moresubtleversionofitistobefoundinthepsycholinguisticliteratureonlanguage developmentinlowerclassblackAmericanchildren(e.g.,ErvinTripp,1972),whenthe researcherisattemptingtoassessaBlackEnglishspeaker'sacquisitionoftherulesofStandard Englishgrammar.InadiscussionoftheacquisitionoftheStandardEnglishrulesfortense markingandpersonagreementinthecopula,somethinglikethefollowingdatamightbegiven 5 foratypicalsubject: a. Shemysister. b. Hebemean. c. Youisanicelady. d. Theybeesinmyschool. ____________________ 5 Thedatadisplayissomewhatreconstructedforbrevity,thoughtherangeofcopulavariants remainstypical. 284 (3)

Concerningsuchadisplay,theresearcherpointsoutthat,while(3ad)areclearly nonstandardintheformofthecopulaemployed,(3e,f)areactuallyinconformity withStandardEnglishrulesinthisrespect.Thisisthenofferedasevidencethatthe subjectshaveincorporatedStandardEnglishgrammaticalrulesintotheirlinguistic repertory,i.e.,areacquiringproductivecompetenceinStandardEnglishandare thereforebecoming(orhavealreadybecome)bidialectal.Variationbetween NonstandardandStandardEnglishcopula,asin(3af),isnotnecessarily inconsistentwiththisconclusion,sinceitcouldconceivablyrepresentcode switching. Infact,thedatain(3af)isanalyticallyfarmoreambiguousthanitmightappear atfirstglance,andcouldactuallybegivingafalseimpressionofbidialectal productivecompetencethatthespeakerof(3af)inrealitydoesnothave.An interpretationofthedataasconsistentwithatleastincipientbidialectalproductive competencedependsupon(3ef)trulybeingproducedbyinternalizedStandard Englishrulesforcopulause,whichmeansthattheyshouldnotonlysatisfyStandard Englishgrammarbutbeunaccountableforbythenonstandardgrammarwhich produced(3ad).Yetitisnotatallcertainthateitheroftheseconditionsreally obtains.Foronething,thereisthepossibilityonemightevensaytheprobability that(3e)hasbeengeneratedbythesamerule(s)ofnonstandardBlackEnglish Thisis grammarthatproduced(3c),theapparentStandardEnglishagreementinperson e. mydoll. andformin(3e)beingentirelyfortuitous.Indeed,ifthenonstandardrule(s) involvedherewereofthevariablekindproposedbyLabov(1969;alsoaschapterI I'm f. hungry. inLabov,1972),withvariabilityinthiscasebetweenisand(noovertcopula), thenatleast(3a,c,e)couldhavebeenproducedbythesamenonstandard grammaticalrule(s).AndifI'mcaninsomevarietiesofBlackEnglishbe,nota combinationofpronounpluscontractedam,butratheramorphemicvariantofI borrowedwholefromStandardEnglish(asproposedinStewart,1966),then(3f) 6 couldactuallybeanotherinstanceofthe"zerocopula"constructionof(3a). Theremainingsentences(3b,d)areunlikelytobemistakenforStandardEnglish, sincetheycontainan"invariant"bewhichisusedinmanyvarietiesofBlack Englishtomarkanaspectualdistinctionbetween(byitspresence)[+ITERATIVE], indicatingrepeatedactionorextendedstate,and(byitsabsence) ____________________ 6AIthoughaBlackEnglishstructurallyunitary(monomorphemic)treatmentof I'm(diagnostic:I'mhere:HereIam=polymorphemicorcontractionalI'm;I'm here:HereI'mis/am=monomorphemicI'm)mightappeartobeconcentratedin children'sspeech,thewidespreadoccurrenceofmonomorphemicit'sandthat's (diagnostic:It'sOK:It'scan'thappen;That'sallright:That'sdon'tmatter)inthe speechofevenelderlyadultsshouldcautionagainstassumingapurely developmentalcause. 285

[ITERATIVE],indicatingaunitaryactionorrestrictedstate(Fasold,1969,1972;Labov,1969, 7Ifthisappliestosentence(3f),itwouldbeunambiguously[ 1972;Stewart,1966,1967). ITERATIVE],ratherthanambiguously[ITERATIVE]asitsStandardEnglishequivalentwould be. Finally,theformwrittenbeesin(3d)representsinvariantbewithanaddedswhichcertainly doesnotfollowStandardEnglishrulesofgrammaticalaccord,sinceitoccurswithaplural subject.ButaverbalsuffixsdoesexistinsomevarietiesofBlackEnglish,whereitcanfunction asamarkerof[+HABITUATIVE](Stewart,1969a)orasaphoneticallysalientmarkerof[ 8 PAST]. Fromtheforegoinganalysisitshouldbeclearthatdataofthetype(3af)offersvirtuallyno clearevidenceoftheacquisitionofStandardEnglishgrammaticalrulesbyaBlackEnglish speakerwhomightproducethem.EventhemostapparentcasesofStandardEnglishgrammar beingused,sentences(3e)and(3f),turnouttobequestionableinthisregard.IfBlackEnglishis adecreolizingcreolelanguage,ashasbeenargued(Dillard,1972;Stewart,1967,1968,1969b, 1974a),andcopularstructuresofthekindevidentin(3af)werecomparedwithearliercreole equivalents,evidenceofashifttowardStandardEnglishgrammaticalformsandruleswouldbe readilyapparent.Butthischangehasbeenlonginprocess,andsuchStandardEnglishlike featuresasI'm(especiallymonomorphemicI'm)andfullformcopulais(especiallywithoutthird personsingularagreement)couldwellhavebeenincorporatedintoBlackEnglishusage generationsaheadofpresentdayspeakersofthetyperepresentedbythehypotheticalproducerof sentences(3af). SomepresentdayspeakersofBlackEnglishdodisplayaknowledgeofsomeoralloftheforms andrulesassociatedwiththeStandardEnglishcopula.Othersdonotseemtohavedoneso,but maywellhaveacquiredStandardEnglishformsandrulespertainingtootherareasofgrammar; tothese,theymayperhapslateraddStandardEnglishcopulaformsandrules.Noclaimisbeing madethatspeakersofBlackEnglishcannotacquireproductivecompetenceinStandardEnglish; butitisbeingsuggestedthatthestructuralrelationshipbetweenBlackEnglishandStandard Englishissuchthat,atanygivenpoint,theapparentproductivecompetenceinStandardEnglish displayedbyBlackEnglishspeakersmaynotalwaysbereal.Whatismore,thesesamecross dialectalstructural ____________________ 7 ThisBlackEnglish[+ITERATIVE]markerisoftenreferredtointhelinguisticliteratureas "invariantbe"because,unlikeStandardEnglishbe,itisneverinternallyinflectedtois,was, etc.AlthoughisandwascertainlyexistascopularformsinmostvarietiesofBlackEnglish, theydonotbehavesyntacticallylikeinflectedformsofbe.Atthesametime,socalled invariantbecantakeexternalinflections,especiallys. 8The[+HABITUATIVE]senseofsismorelikelywithactiveverbs(e.g.,Iwalkstoschool), whileits[PAST]functionisparticularlyusefulwithstativeverbs(IknowswhatIknows).In somecases,the[+HABITUATIVE]useofscanbeclearlydistinguishedfromits[PAST] usebyoccurrencein[+PAST]grammaticalcontexts,aswhenitcombineswiththeedending (IstaysedupinFloridaforoneyear"IremaineddowninFloridaforayear"). 286

relationshipsmustmaketheacquisitionofproductivecompetenceinStandardEnglishbya BlackEnglishspeaker(orviceversa)anextremelydifficultenterprise.For,giventhelinguistic ambiguityofgrammaticalfunctionwordsandinflectionswhoseformsaresharedacrossthe dialectsbutwhosemeaningsdiffer,howisthespeakeroftheonedialecttomakecorrect inferencesabouttheratherdifferentgrammarbehindthesometimesquitefamiliarverbaloutput intheotherdialect?Afterall,ifaStandardEnglishspeakingresearchpsychologistor psycholinguistcanmistakesomeBlackEnglishusesofisasevidenceofStandardEnglish grammar,thenwhymightnotaBlackEnglishspeakertakeaStandardEnglishuseofisasin conformitywithhisowngrammarhenceevidencethatthetwodialectsarealikeinthisregard sothathe,theBlackEnglishspeaker,alreadyknowsStandardEnglishtothatextent?Insucha casetheseconddialectlearningprocess,slowanddifficultatbest,canbeshortcircuited entirely. Inkind,thestructuralrelationshipsjustconsideredbetweenAmericanBlackEnglishand StandardEnglish,posingsuchdifficultiesforcrossdialectallearning,donotappeartobetypical ofstructuralrelationshipsbetweendialectsaroundtheworldorevenintheUnitedStates,for thatmatter.InthecaseoftherelationshipbetweenlocaldialectsandStandardEnglishin England,whichmaybetakenasanexampleofthemoretraditionalone,crossdialectal differencesareforthemostpartphonologicalandlexical.Thatis,thedialectsaredifferentiated fromeachotherbytheuseofdifferentwordsforsomethings,butprimarilybythewayinwhich sharedwordsandthisincludesthevastmajorityofwordsarepronounced.Apartfromthe effectofdifferencesinpronunciation,grammaticaldifferencesbetweentraditionallyrelated 9Withrespecttothegrammaticalfunctionwordsinparticular,thesame dialectsarefewindeed. formdifferentmeaningrelationshipwhichheldintheAmericanBlackEnglishcaseisalmost nonexistent.Inshort,onecouldsaythattraditionaldialectrelationshipsinvolvesuperficial differencesontopofanunderlyingunity,whereastheBlackEnglishcaseshowsevidenceof superficialsimilaritiesmaskingunderlyingdifferences. ThereareindeedinstancesoftheAmericanBlackEnglishkindofdialectrelationshipinother partsoftheworld,orcasesthatshowtracesofsucharelationship,butthesealmostalways involveahistoryofthelanguagehavingbeenlearnedasaforeignlanguagebyapartofthe speakerpopulation.IrishdialectsofEnglishareamildinstanceofthismildinpartbecauseof thedynamicsofthespreadofEnglishinIrelandoverthelastthreeandahalfcenturies,andmild inpartbecauseIrishandEnglishwerehistoricallyrelatedlanguagestobeginwith.Morewell definedcasesexistwheretheoriginallan ____________________ 9Wright(1905)devotedfivesixthsofadiscussionofthestructuraldifferencesbetweenBritish dialectsofEnglishtophonologyandtheremainingsixthtowhathecalledaccidence,mostly morphologywithonlyabitofsyntax.Whilesomeofthisdifferenceisundoubtedlyduetothe underdevelopedstateofsyntactictheoryandknowledgeinhisday,italsomustinpartreflect theactualnatureofthestructuralrelationshipsbetweenBritishdialects. 287

guageorlanguagesoftheforeignpopulationborenorelationshiptothelanguagetheysetout toacquire.Insomeofthesecases,particularlywherethesecondlanguagelearningwas entirelyinformal,enoughinadvertentrestructuringofthesecondlanguagecouldtakeplaceto createnewvarietiesofitwhichhadsomethingofthecharacterofrelatedforeignlanguages ratherthandialects.ThesearethesocalledpidginandcreolevarietiesofEnglishandother

languagesofmercantile,military,andcolonialexpansion(Hall,1966;Todd,1974,1984).10 Whereapidginorcreolehascontinuedtocoexistwiththerelatedstandardlanguage,asinthe caseofcreoleandStandardEnglishinJamaica,thecreole(becauseofitslowsocialstatus) mayundergoacertainamountofstructuralchange,calleddecreolization,inthedirectionof thestandardlanguage,whilethelatter(becauseitisusuallylearnedasasecondlanguageby creolespeakers)mayundergocreoleinfluence.Intermediatevarietiesarecreatedthroughthis reciprocalinfluence,andtheseinturnblendwitheachothertogiverisetoacreoleto standardcontinuum(Bickerton,1975). Movementacrosssuchacontinuumcharacteristicallyconsistsofwordformchanges,whichmay proceedfairlyrapidlyandintheprocessmerelymasksomefairlysubstantivegrammatical differencesbetweenthetwoends.Becauseofthismaskingeffect,changesintheunderlying grammaticalstructurebecomeincreasinglydifficultforthenaivespeakerandeducatoraliketo monitor.Inotherwords,exceptforthedegreeofstructuraldifferentiationcoveredbythespanfrom thevarietymostunlikethestandarddialecttothestandarddialectitself,structuralrelationships withinthecreolecontinuumareverymuchlikethosethatholdbetweenAmericanBlackEnglish andStandardEnglish.Norshouldthisbeparticularlysurprisingif,asalreadysuggested,Black Englishisalatedecreolizingcreole.Assuch,itwouldrepresenttheremnantofaNorthAmerican colonialorplantationcreolecontinuumfromwhichthemorecreoleendhadeitherdisappeared completelyorbecomerestrictedtotheSouthCarolinaandGeorgiacoastasanotherwisemysterious dialectofundocumentedoriginknownlocallyasGullahorGeechee. Bytheearly1950sithadbecomeamatterofpublicknowledgeandprofessionalconcernthatthe academicperformanceofblackchildrenasagroupwassignificantlybelowthatofwhitechildrenin mostAmericanpublicschools.Psychologistsandlinguistssoonbegantofocusononeparticular areaofacademicachievementinwhichblackchildren(again,asagroup)seemedtoperform especiallypoorly:languageuse.Forsomeofthelinguistswhohadalreadyhadexperiencewith educationalproblemsintheCaribbean,thedetailsofAmerican ____________________ 10Therehasbeenmuchdiscussionconcerningtheprocessesthatmighthavegivenrisetopidgin andcreolelanguages,aswellastypologicalissuesrelatingtotheirclassification.Althoughthese questionsarenotpertinenttothepresentpaper,itisworthmentioningthataninterestingdebate hasdevelopedconcerningtheroleofanallegedinnateuniversalgrammarintheformationof creolelanguages(Bickerton,1981). 288

blackchildren'sproblemswithschoollanguagelookedstrikinglyfamiliar.Ethnicallycorrelated dialectdifferenceswerethereforesuspected,lookedfor,andsomefoundratherquickly.Sincethe notionofacreolepredecessortoAmericanBlackEnglishhadnotpreviouslybeenapartofthe historyofblackspeechasassumedbyAmericandialectgeographersandhistoriansofthe EnglishlanguageinNorthAmerica,adebatearoseastowhetherthenonstandardspeechof blackAmericanshadvestigialcreolefeatures,orwasmerelyavariantofthenonstandardspeech ofsouthernwhites.Mostlinguistsdidagree,however,thattherewassomedegreeoflanguage conflictintheclassroom.Thosewhodidnotacceptacreoleoriginforthenonstandardspeechof blackAmericans,andwhothereforedidnotexpectstructuraldifferencesbetweenitandStandard Englishtobecognitivelyproblematic,simplyurgedteacherstobemoretolerantand understanding.ThoselinguistswhowereconvincedthatBlackEnglishwasapostcreole,andas suchcouldgiverisetocreolecontinuumlikecopingproblemsintheclassroom,proposedthe developmentofspecialpedagogicalmaterialstohelpBlackEnglishspeakerslearnStandard Englishwithaminimumoflinguisticuncertainty. AlthoughsomeeducatorsexpressedtentativeinterestintheproposalforspecialStandardEnglish curriculummaterialsforBlackEnglishspeakers,mostblackeducatorsandcommunityleaders stronglyopposedtheidea.Inlargepart,theiroppositionarosefromafearthatspecialcurriculum materialsforblackdhileren,nomatterhowlinguisticallyjustifiable,wouldbelikelytoresultin theresegregationofthosechildren.Bythelate1960s,blackoppositiontocrossdialectal curriculummaterialshadbeenfannedintoconsiderableheatbyamisunderstanding(perhaps createdorreinforcedbyearlydiscussionsonbilingualeducationforHispanics)thatthepurpose ofsuchmaterialswasactuallytoteachblackchildrentospeak,read,andwriteBlackEnglish. Atthispoint,theeducationalestablishmentwasreadytobackoffcompletelyfromtheissue.But backofftowhere?Itwas,afterall,thelate1960s,bywhichtimeithadbecomeevenclearerthat blackchildrentendedtohavespecialdifficultieswiththeStandardEnglishcomponentofthe curriculum.Moreover,linguistshadbythendemonstratedthattherereallywasalinguisticentity thattheycalledBlackEnglish(sometimestheBlackEnglishVernacular)andthatitsstructural featuresmightmakestandardEnglishdifficultforitsspeakerstolearn.Clearly,withregardto theschoollanguageachievementproblemsofblackchildren,theeducationalestablishmentwas caughtbetweentheScyllaofanineffectualtraditionalcurriculumandtheCharybdisof controversialcurriculuminnovation.Butifonelinguisticnotionhadgottentheeducatorsinto potentialtrouble,anotherwastorescuethemfromit.Fortheysoonfoundrecourseinthe receptive/productivecompetenceasymmetry. Ironyintheturnofeventsisoneofthelittlethingsthatmakelifeinteresting,soitisinteresting that,iftheperceptionofaneedforspecialcurriculummaterialsforAmericanBlackEnglish speakershaditsinspirationinthestudyof 289

educationacrossacreolecontinuumintheCaribbean,sotoodidthecounteringnotionof receptivecompetencebetweenBlackEnglishandStandardEnglishhaveaprecursorinaclaim madebyastudentofCaribbeancreoles.Touchingonthequestionofintelligibilityacrossthe Jamaicancreolecontinuum,DavidDeCamp(1971)notes:"ThevarietiesofJamaican[creoleto standard]Englishthemselvesdiffertothepointofunintelligibility;butsomeJamaicanEnglishis mutuallyintelligiblewithstandardEnglish,"(p.350). AndofcoursehenotestheevenmoreobviousfactthatnoJamaicansseemabletospeakallthe varietiesacrossthecontinuum,andsomeappearquitelimitedintherangetheycanproduce.But, operatingwithinagenerativetheoreticalnotionofahomogeneousspeechcommunityandan "idealspeakerlistener"whoseproductiveandreceptivecompetenceareidentical,DeCamp (1971)attemptstoarguethatparticipantsintheJamaicancreolecontinuumhavenotonly receptivebutalsoproductivecompetenceacrosstheentirecontinuumrange.Theearlierreported comprehensionfailurebetweenvarietiesisleftunexplained,butproductionfailureisascribedto sociolinguisticfactors: Onecouldwellarguethattheindividualspeaker'slimitationtoaspanofthe continuum,thefactthatnoonespeakercancommandtheentirerangeofvarieties,is amatterofperformanceratherthanofcompetence,analogoustohisinabilityto speaksentencesbeyandacertainlevelofcomplexity.Thetheoreticalcompetenceof the'idealspeakerlistener'couldbedefinedasspanningtheentirecontinuum, includingfullcommandofalltheswitchingrulesbetweenanyonepointandanother. Wecouldthenlookforsocioeconomicexplanationsofthemannerinwhichthe actualperformanceofrealspeakersfallsshortofthisidealcompetence.(p.368) DescribingtheGuyanesecreolecontinuumahalfdecadelater,DerekBickerton(1975)differs withDeCampconcerningintelligibilitybetweenthevarieties:"Indeed,onemayoftenobserve speakersinthesameconversationproducingatwidelydifferentlevelsandyetcontinuingto understandoneanotherperfectly"(p.196). Yettheexamplesofthisthathecitesonthesamepagearefarfromconvincing.Inanyevent, Bickertonnotesindividuallimitedproductiverange,asDeCampdidinJamaica.And,like DeCamp,hewantstoclaimthatthelimitationisamatterofperformance,ratherthan competence.Buthewantstodosoforadifferentreason:toexplainclaimedcrosscontinuum intelligibility. Itwouldseemmoreplausibletoarguethataspeakerinacreolesystemboth producesandunderstandsbyvirtueofthesamesetofinternalisedrules,andthatthe strikingasymmetrywhichoftenexistsbetweenreceptiveandproductivecapacity arisesthroughperformancefactors:inparticular,limitationsonopportunitiesfor actualuseof(andhence,fullfamiliarisationwith)particularvarieties,imposedby particularlifestyles,andovertsocialrestrictionsonwhatoccupantsofgivenroles mayormaynotsay.(pp.197198) 290

Inotherwords,thecontinuumparticipantcanunderstandallvarietiesforthesimplereasonthat hecanspeakallvarieties.Ifheisneverobservedtospeakanythingremotelyapproachingall varieties,well,itisjustthathehasnowishtospeaksome,andnoonewilllethimspeakothers. Animplicitclaiminallthisisthatonecanknowalanguagevarietywithouteverusingit.Yet, rightinthemiddleofthelastpassagequotedfromBickerton,thereisthatseeminglycontrary statementaboutlimitationsonuseofalanguagevarietygivingrisetoalackof"full familiarisation"withit. TheparticulartheoryoflinguisticcompetencetowhichbothDeCampandBickertonare beholdenisindeedunfriendlytothepossibilityofaproductive/receptivecompetenceasymmetry, becausethetheoryholdstheseabilitiesasequivalent.Itisnotthecasethat,givenapparent evidenceofgreaterindividualreceptivecompetencethanproductivecompetenceacrossthe creolecontinuum,DeCampandBickertonneededtoresolvetheasymmetrybyexpandingtheir claimsforindividualproductivecompetencetomatchtheobservedreceptiverangemuchless byexpandingbothtocovertheentirecontinuum.Theycouldhaveremainedjustasfaithfultoa theoryofcompetencethatisneutralonproductiveandreceptiveabilitiesbyreactingskeptically toapparentevidenceofreceptivecompetencewherethisseemedtorangefarbeyondthe individual'sapparentproductivecompetence.Why,then,didDeCampandBickertonoptfor expandingobservedproductivecompetencetoatheoreticallymuchwiderrangingability?The answerwouldseemtobethat,forthem,productivecompetencewasequatedwiththe performanceaspectofthetheoreticalcompetenceperformancedistinction(seeValian,1979,for ahistoricaloverview)intermsofwhichtheyareattemptingtodescribeindividualcompetence withinacreolecontinuum.Receptivecompetence,manifestedintermsofcomprehensionability, wouldthenmostaccuratelyrepresenttruecompetenceinsuchacompetenceperformance distinction,andassuchitwouldbeamorereliablemeasureofreal(evenifunattested) productiveabilityaswell.DeCampandBickertonthusallowedanassumedrangeofindividual receptivecompetenceforwhichtherewasonlyequivocalevidencetodefineanindividual productiverangeforwhichtherewasnodirectevidenceatall. MovingdirectlytothequestionofcomprehensibilitybetweenAmericanBlackEnglishand StandardEnglish,WilliamLabov,inhisstudyofthelanguageofHarlemteenagers,cametoa similarconclusion.Inanowfamousexperiment,heaskedhisBlackEnglishspeakingsubjectsto repeatStandardEnglishsentencescontainingembeddedyesnoquestions,suchas (4)IaskedAlvinifheknowshowtoplaybasketball. inwhichyesnoquestionembeddingiscarriedoutwithacomplementizer,if,andrereversalof questionwordorder.Approximatelyhalfofthesubjectsgaveresponsesoftheequivalentform 291

(5)IaskAlvindoheknowhowtoplaybasketball. whichhasthemoredirectBlackEnglishembeddedconstruction.Describingtheexperimentin somedetail,Labov(1972)commentsontheseresults: Inthemostobviousview,wecanobservethatthesubjectfailedtoperformthetask required.Butwecannotoverlookthefactthat[5]isthecorrectequivalentof[4];ithas thesamemeaningandisproducedbythenonstandardrulewhichisthenearest equivalenttothestandardrule.(p.62) Accordingly,heconcludes: Sincethe[subject]doesperformthetranslation,itisclearthathedoesunderstandthe standardsentence.Hethenrapidlyproducesthecorrectnonstandardequivalent[5]. Understandingheremustmeanperception,analysis,andstorageofthesentenceinsome relativelyabstractform... Fromtheseconsiderations,itisclearthatthe[subject]isperfectlycompetentin(atleast this)aspectofthestandardgrammar.(p.63) Finally,hegoesontogeneralizefromthisconclusiontothetheme,nowalreadyfamiliarfrom DeCampandBickertoninothercontexts,that"thegearsandaxlesofEnglishgrammatical machineryareavailabletospeakersofalldialects,whetherornottheyuseallofthemineveryday speech"(p.64). Butif"Englishgrammaticalmachinery"isthereforalltouse,Labov'sexperimentcanhardlybe saidtohavedemonstratedthefact.Forthereremainsthecrucialquestionofwhethersentenceslike (4)mightnotbequitetransparentastomeaningbecauseofhighstructuralredundancyandlow semanticambiguity.Indeed,itisdifficulttoseewhyalistenerwouldneedaknowledgeofthe StandardEnglishrulesforembeddingquestionsinordertodeterminethatthereisaquestioninside (4),sinceitstartsoutwithask.Oneprobablywouldneedminimalgrammaticalcues,suchas SubjectVerbObjectwordorder,butlittleelse.Itwouldbeinterestingtoseeacomprehensiontest ofsentenceslike(4)strippedofalltheirdialectspecificmorphologyandsyntax,e.g.,to (6)IASKALVINHEKNOWHOWPLAYBASKETBALL. Onceaprobablemeaningisinferredforsuchastring,theappropriategrammaticalstructurefrom anydialectcanbeusedinthe"repetition."Inasense,thisistranslation;butitisakindof 11 translationthatdoesnotdependonaknowledgeofmorethanonegrammar. ____________________ 11 Thosewhohavelearnedortaughtaforeignlanguageinclassroomsettings,especiallywhere writtentranslationexercieswereinvolved,willrecognizethisstrategyasresponsibleforsome fairlyaccuratetranslationsbyrankbeginners. 292

Itwasundoubtedlythiscurrentofexpressedscientificopiniononcrossdialectalcompetence, emanatingfromrespectedlinguistswhohaddoneresearchonblacksubjects,thateventually reachedAmericaneducatorsandbecametheauthorityforadvocacyofalaissezfairepolicyon theteachingofStandardEnglishtoBlackEnglishspeakers.Thispolicywasarticulatedinthe formofapositionpaper,Students'RighttoTheirOwnLanguage,whichwasdraftedinthe courseof19711974byaspecialcommitteeoftheConferenceonCollegeCompositionand Communication,finallypublishedbythemin1974(CommitteeonCCCCLanguageStatement, 12 1974)andcosupportedbytheNationalCouncilofTeachersofEnglish. ThepositiontakenoncrossdialectalcompetencebytheCommitteeonCCCCLanguage StatementisessentiallythatofDeCamp(1971),Bickerton(1975),andLabov(1972),except thatitislesscompetentlyarticulated,dogmaticintone,andseekstotrivializetheobservable grammaticaldifferencesbeteenBlackEnglish(neveractuallymentioneddirectly)andStandard English: Althoughtheyvaryinphonology,invocabulary,andinsurfacegrammaticalpatterns, thedifferencesbetweenneighboringdialectsarenotsufficientlywidetopreventfull mutualcomprehensionamongspeakersofthosedialects.(CommitteeonCCCC LanguageStatement,1974,p.4) If"neighboringdialects"istakenasmeaningcloselyrelateddialectsofthetraditionalkind(i.e., thathaveevolvedoutofacommonancestorandexistinadjacentterritories)thenthestatementis essentiallycorrect.Butitispreciselyforthisreasonthattheissueofthehistoricaloriginof BlackEnglishisimportant;adecreolizedcreoleanditsrelatedstandardarecertainlynot "neighboringdialects"inthishistoricalsenseandthereforeshouldnotbeexpectedtohavethe characteristicsofsucharelationship. Followinguponthisallegedcloserelationshipbetweenthevariousdialectsthatcometogetherin Americanpublicschools,butforsomereasontreatingarestatementoftheclaimjustmadeasif itwerenewinformation,thepositionpapergoesontosay: Anotherinsightfromlinguisticstudyisthatdifferencesamongdialectsinagiven languagearealwaysconfinedtoalimitedrangeofsurfacefeatureswhichhaveno effectonwhatlinguistscalldeepstructure,atermthatmightroughlybetranslatedas "meaning".(CommitteeonCCCCLanguageStatement,1974,p.6) CertainlyBlackEnglishandStandardEnglishhavesurfacestructuraldifferencesbetweenthem phonologicaldifferencesandmorphologicaldifferences.Butthatiscertainlynotall;theBlack Englishinvariantbehasbeen ____________________ 12ThehistoryofthepositionpaperisdiscussedinanindepthreviewofitbyStewart(1974b). 293

showntomarkanaspectualdifferenceforwhichStandardEnglishhasnoequivalent.Nordoesit appeartobereadilycomprehendedaccuratelybypersonswhodonotactuallyknowBlack Englishgrammar.Thepublicschoolsarefullofmiddleclass,monodialectal,StandardEnglish speakingteacherswhodailyhearblackchildrenmakingthebe/nonbe([+ITERATIVE]/[ ITERATIVE])distinctionconsistentlyandrepeatedly,andyethavenoideaofitsmeaning. WouldtheCCCCLanguageStatementCommitteeconsiderBlackEnglishinvariantbeamere surfacestructuralfeature?Thenwhyisitsgrammaticalmeaningnotunderstoodbyspeakersof otherdialects?WastheCommitteeperhapsunawareofitsexistence?Longdiscussionsofithad alreadybeeninthelinguisticandEnglishteachingliteratureforhalfadecade.So,ifthe CommitteewasawareoftheexistenceofBlackEnglishinvariantbeanditsgrammatical function,whywasitnotmentioned?Or,atleast,whydidtheCommitteecontinuetoclaimthat differencesbetweendialectswerealwaysconfinedtosurfacestructure?Whatevertheanswerto thesequestionsmaybe,thegeneraleffectofthepositionpaperwastoputtheteachingof StandardEnglishonthedefensive,andtoconsiderablydiscouragefurtherresearchonthe pedagogicaleffectofdialectdifferences.Totalreceptivecompetenceacrossdialectsbecame educationaldogma.WhentheCCCC/NCTElaissezfairepolicyonEnglishteachingwasfinally abandonedbytheprofessionatlarge(and,itwouldseem,somewhatbytheorganizationsthat originallysponsoredit),itwasnotbecausethenotionoftotalcrossdialectalreceptive competencebegantobequestioned;itwassimplybecauseashiftinthevaluesofthelarger societyhadbroughtaboutthereimpositionintheEnglishclassofwhattheCCCCLanguage StatementCommitteewouldhavereferredtoas"surfacestructure"norms. Earlierinthispaper,itwassuggestedindividualproductivecompetenceacrossdialectsmay sometimesappeartobemoreso(morecompetent,thatistosay)thanitprovesactuallytobe. Nowthatcertainclaimsforcrossdialectalreceptivecompetenceparticularlyoverthevariants inacreolecontinuumhavebeenexaminedandfoundnottotallyconvincing,itisperhapsworth wonderingwhetherreceptivecompetenceacrossdialectsmightnotsometimesalsoappearmore competentthanitreallyis.Onemustrememberthatinmostrealworld,crossdialectal interactions,impressionisticevidenceofcomprehensionconsistsforthemostpartofanabsence ofanyindicationparticularlybythelistenerofitsfailure.Andthismeansthatthelistener mustbeawarethathefailedtocomprehendatsomepoint.Onelikelycauseofthisawareness wouldbetheusebythespeakerofsomeunfamiliarword(orunfamiliarpronunciationofa familiarword,therebycausingitnottoberecognized)orexpression.Theuseofanunfamiliar grammaticalconstructionmighthavethesameeffect,thoughinsuchacasethelistenerwould haverecoursetothekindofcomprehensionbylexicalscanningmentionedinthediscussionof Labov'sexperiment.Thereis,however,akindofcrossdialectalunfamiliaritythatwouldbequite likelytocausecomprehensionfailure,yethighlyunlikelytobenoticedoratleast 294

recognizedasalinguisticproblembythelistener:interdialectalambiguity.Awordor grammaticalinflectionorsyntacticconstructioncanbesaidtobeinterdialectallyambiguous ifitssurfaceformissharedbythetwodialectsbutitsmeaningisnot,or,moreprecisely,ifit hasatleastonemeaninginonedialectthatitdoesnothaveintheother.Alexicalexampleof interdialectalambiguityinvolvingStandardEnglishandGullah(orGeechee),adecreolizing formofcreoleEnglishspokenincoastalSouthCarolinaandGeorgia.Sentence(7a)is StandardEnglish;sentence(7b)isGullahinStandardEnglishorthography: a. Thegirlchasedtheboyandkissedhimback. b. Thegirlchasetheboy,kisshimback. Thelinguisticallysophisticatedreaderwillperhapsguessthattenseoftheverbsin(7b)isprobably ambiguous,i.e.,[PAST],butsincethisistruewithinGullahitselftheambiguityisnot interdialectal.Theinterdialectallyambiguousitemisthewordback,interpretedineitherdirection. InStandardEnglish,itssemanticspecificationisroughly[+REPEATED,+RECIPROCAL], whereasinGullahitis[+REPEATED,RECIPROCAL.Thismeansthat,ignoringtheverbtense problem,(7b)hasbothaStandardEnglishinterpretation(7a)andanother,(8): (8)Thegirlchasedtheboyandkissedhimagain. Now,theinterestingthingabout(8)isthatitisnotaninterpretationof(7b)thataStandardEnglish speaker(thatis,onewhodoesnotalsoknowGullahorsomeotherrelatedvarietyofcreoleEnglish) wouldbelikelytocomeupwith,evenifthecontextsuggestedthattherewassomethingwrongwith interpretation(7a).Afterall,theStandardEnglishspeakerknowswhatadverbialbackmeans,ifhe 13 knowsnothingelse! Becauseofdecreolization,syntacticinterdialectalambiguitybetweenadecreolizingcreoleandits relatedstandardlanguageismorelikelythanbetweenanyotherkindofdialectrelationship.And syntacticinterdialectalambiguitymust,ifanything,beevenmoreproblematicforcrossdialectal receptivecompetencethanlexicalinterdialectalambiguity,becausesomuchmorecangoonatthe syntacticlevel.Thefollowingexampleactuallyoccurred,andinvolveda ____________________ 13Evenifthelinguisticcontextmakestheuseofaparticularwordseembizarretoonewhodoes notspeakthatdialect,thereisnoguaranteethateitherthelinguisticornonlinguisticcontextwill suggestthecorrectinterpretation.TheIrishEnglishconstructionbeafter+Verb+ingis,bythe presentwriter'sexperience,consistentlymisinterpretedas"beinpursuitof+Noun,"hence"be aboutthebusinessof+Verb+ing"byAmericanEnglishspeakers(provided,ofcourse,that theyhavenothadagreatdealofexposuretoIrishEnglish),nomatterhowclearthecontext makesitthatsuchaninterpretationisproblematic.Theconstructionusuallymeans"have(just)+ Verb+en"(Bliss,1979;pp.299300). 295 (7)

question(9a)posedinStandardEnglishtoaGullahspeaker,whoansweredwith(9b).Emphatic stressisindicatedbyitalics: a. HadJoeleftwhenyouleft? b. No,JoeleavebeforeIleave. Since(9b)translatesas(10a),theGullahspeakinglistenerhadobviouslymiscomprehended(9a) asmeaning(10b): a. No,JoeleftbeforeIdid. b. DidJoeleavewhenyouleft? Thismiscomprehension(whichcouldalsohaveoccurredwithspeakersofothervarietiesof BlackEnglishoutsideGullahterritory)isaninevitableconsequenceofthefactthat,atthe particularstageofdecreolizationofthelistener,had+Verbisnotunambiguouslyequivalentto StandardEnglishhad+Verb+en,butrathercanalsobeequivalenttoStandardEnglishVerb+ ed. InterdialectalambiguitybetweenBlackEnglishandstandardEnglishwasnotedalmosthalfa centuryagobytheanthropologistMelvilleHerskovits,who,seeingitinthecontextofageneral theoryofNewWorldNegroacculturation,termeditmasking(Herskovits,1941;seealsoDillard, 1972,foracontinuationoftheterm).Subsequently,thetermcamouflaginghasbeenusedforthe samephenomenon(Spears,1982;Stewart,1968).Inanycase,whetherinreferencetocultureor language,thephenomenonprobablyoriginatedasinadvertentreinterpretationinthe acculturativeorsecondlanguagelearningprocess,andneverwastheresultofanintentional efforttoconcealordeceive,despitetheromanticappealofsuchanotion. Becausethiskindofinterdialectalambiguityinvolvessharedsurfacepatternswhosemeanings arefalselyequated,theresultantcommunicationbreakdownisnotlikelytobeperceivedas linguisticinnature.Rather,itmaybeattributedtofaultyinformationoranerroneousconclusion, aswasevidentlyassumedfor(9a)bythespeakerof(9b).Evenwhenthereisanawarenessthat theproblemislinguistic,itisnotlikelytobeclearwhichsharedstructuresareinterdialectally ambiguousandwhicharenot.Ineithercase,theresultantcomprehensionproblemswillnotbe easilyresolvable,andmaycontinuetomakethemselvesfelteitherassemanticbizarrenessor 14 structuralindeterminacygeneralizedthroughoutthediscourse. Insummary,thereseemtobegoodreasonstoquestiontheuncriticalassumptionofgeneral receptivecompetenceacrossdialects,particularlyiftheinter ____________________ 14 EIsewhere(Stewart,1984)Ihavereferredtothiskindofunperceivedinterdialectical miscomprehensionaspseudocomprehension. 296 (10) (9)

dialectalrelationshipexistsacrossacreolecontinuum,orintheaftermathofone.Since AmericanBlackEnglishisinalllikelihoodadecreolizedcreole,andsinceatleastone controlledstudyhasturnedupstatisticallydemonstratedevidenceofcomprehension problemswithStandardEnglishonthepartofBlackEnglishspeakingchildren(Berdan, 1977)and,mostimportantofall,sincethesuccessofpubliceducationwiththosewhoneed itmosthangsinthebalance,theempiricalstudyofthelinguisticandpsycholinguistic mechanisminvolvedinproductiveandreceptivecompetenceacrossdialectsoughttobean importantresearchagenda. ACKNOWLEDGMENT Thematerialinthispaperisinpartderivedfromresearchsupportedbyagrant(GS39814X)from theNationalScienceFoundation. REFERENCES Berdan,R.(1977).Polylectalcomprehensionandthepolylectalgrammar.InR.Fasold&R.W. Shuy(Eds.),Studiesinlanguagevariation.Washington,DC:GeorgetownUniversityPress. Bickerton,D.(1975).Dynamicsofacreolesystem.Cambridge:UniversityPress. Bickerton,D.(1981).Rootsoflanguage.AnnArbor:KaromaPress. Bliss,A.(1979).SpokenEnglishinIreland:16001740,Dublin:DolmenPress. CommitteeonCCCCLanguageStatement.(1974).Students'righttotheirownlanguage.Urbana: NationalCouncilofTeachersofEnglish.[CollegeCompositionandCommunication:25(No.3) Specialissue] DeCamp,D.(1971).Towardagenerativeanalysisofapostcreolespeechcontinuum.InD.Hymes (Ed.),Pidginizationandcreolizationoflanguages.Cambridge:UniversityPress. Dillard,JL.(1972).BlackEnglishintheUnitedStates.NewYork:RandomHouse. ErvinTripp,S.(1972).Children'ssociolinguisticcompetenceanddialectdiversity.Earlychildhood education:SeventyfirstYearbookoftheNationalSocietyfortheStudyofEducation,Part2. Chicago:NationalSocietyfortheStudyofEducation. Fasold,R.(1969).TenseandtheformbeinBlackEnglish.Langauge45, Fasold,R.(1972).TensemarkinginBlackEnglish.Washington,DC:CenterforApplied Linguistics. Hall,R.A.,Jr.(1966).Pidginandcreolelanguages.Ithaca:CornellUniversityPress. Herskovits,M.J.(1941).Themythofthenegropast.NewYork:HarperBrothers. Labov,W.(1969).Contraction,deletion,andinherentvariabilityintheEnglishcopula.Language, 45. Labov,W.(1972).Languageintheinnercity:StudiesintheBlackEnglishvernacular. Philadelphia:UniversityofPennsylvaniaPress. Spears,A.K.(1982).TheBlackEnglishsemiauxiliarycome.Language58:850872. Stewart,W.A.(1966).Socialdialect.InResearchandplanningconferenceonlanguage developmentindisadvantagedchildren.NewYork:YeshivaUniversity. Stewart,W.A.(1967).SociolinguisticfactorsinthehistoryofAmericanNegrodialect.FloridaFL Reporter,5.[ReprintedinH.B.Allen&G.N.Underwood(Eds.),ReadingsinAmerican dialectology.NewYork:AppletonCenturyCrofts,1971]

297

Stewart,W.A.(1968).ContinuityandchangeinAmericanNegrodialects.FloridaFLReporter, 6.[ReprintedinAllen&Underwood1971(seeatStewart1967)] Stewart,W.A.(1969a).OntheuseofNegrodialectintheteachingofreading.InJ.C.Baratz& R.W.Shuy(Eds.),Teachingblackchildrentoread(pp.156219).Washington,DC:Centerfor AppliedLinguistics. Stewart,W.A.(1969b).HistoricalandstructuralbasesfortherecognitionofNegrodialect.In MonographSeriesforLanguagesandLinguistics,22.Washington,DC:GeorgetownUniversity. Stewart,W.A.(1974a).AcculturativeprocessesandthelanguageoftheAmericanNegro.InW. W.Gage(Ed.),Languageinitssocialsetting,(pp.146).Washington,DC:Anthropological SocietyofWashington. Stewart,W.A.(1974b).ThelaissezfairemovementinEnglishteaching:Advancetotherear?In A.C.Aarons(Ed.),IssuesintheteachingofstandardEnglish.[FloridaFLReporter:12. ReprintedinM.A.Lourie,&N.F.Conklin(Eds.),Apluralisticnation:Thelanguageissuein theUnitedStates.Rowley,MA:NewburyHouse.] Stewart,W.A.(1984,January).FromxenlecttomimolecttoPseudocomprehension:Structural mimicryanditsfunctionalconsequencesindecreolization.Paperpresentedatthemeetingofthe SectionofLinguistics,NewYorkAcademyofSciences.(ToappearintheAnnalsofthe Academy.) Todd,L.(1974).Pidginsandcreoles.London:Rutledge&KeganPaul. Todd,L.(1984).ModernEnglishes:Pidginsandcreoles.Oxford:BasilBlackwell. Valian,V.(1979).Thewhereforesandthereforesofthecompetenceperformancedistinction.In W.E.Cooper&E.C.T.Walker,(Eds.),Sentenceprocessing:Psycholinguisticstudiespresented toMerrillGarrett.Hillsdale,NJ:LawrenceErlbaumAssociates. Wright,J.(1905).Englishdialectgrammar.Oxford:HenryFrowde. 298 AuthorIndex A Aaronson,D., 86,98,109, 114,117 Adams,M., 40,52 Aliotti,N.C., 67,72,136, 148

Allister,T., 206,220 Anthony,B., 158,166 Anthony,B. A.,162,167 Ard,J.,32, 38,53 Aries,B.,152, 166 Arlmann Rupp,A.,246, 280 Au,T.K.,84, 117,126,127 Ausubel,D. P.,132,147 B Bachi,R., 123,124,128 Bactens Beardsmore, H.,58,72 Bailey,B.L., 239,241 Bailey,N.,39, 52 Baker,K., 151,168 Baker,K.A., 61,71 Baker,L., 243,279 Balkan,L., 69,71 Barik,H.C., 188,195,198, 220 Baron,J.,113, 117 Bates,E.,112, 113,117,118, 126,128,145, 147 Beaujot,R., 11,12,13,26 Beilin,H.,34, 52 Bellugi,U.,

122,128,233, 241 Bentler,P.M., 139,147 BenZeev,S., 67,68,71, 136,139,147, 187,194 Berdan,R., 296,297 Bereiter,C., 63,71 Berliner,D., 152,156,162, 165,166 Bernstein,B., 248,279 Berry,J.W., 185,186,191, 194 Bialystok,E., 71,71 Bickerton,D., 288,290,291, 292,293,297 Bilinsky,Y., 17,26 Blake,L.,213, 219 Blank,M.A., 87,118 Bliss,A.,295, 297 Bloom,A., 79,83,84, 117,126,128 Bloom,L., 114,117 Bond,Z.S., 86,117 Bourne,L.E., 152,155,166 Braine,M.D. S.,122,128 Brandt,J.J., 96,97,99, 102,106,117 Brown,A,L., 243,279

Brown,R.,33, 34,39,52, 122,128,144, 147,233,241 Brown,R.J., 191,194 299

C Cancino,H., 32,33,38, 39,40,42, 52,53 Caramazza, A.,87,117 Carroll,J.B., 183,194 Caselli,C., 126,128 Cattell,R., 45,52 Cazden,C., 40,52,246, Bruck,M., 279 69,71 Cazden,C. Bruner,J.S., B.,37,40,52 155,166 Chalkley,M. Brunner,E., A.,42,54 18,19,26 Champagne, Buriel,J.W., M.,68,72 150,167 Chan,K.S., Burt,M.,38, 150,167 39,52 Chaung,C. Butterfield, J.,77,96,97, E.C.,177, 118 179 Chen,Chao Butterworth, ming,77,96, G.,40,52 97,117 Byrne,D., Cheng,R.L., 190,194 75,118 Chomsky,C., 38,52 Christian,J., 191,195 Christian,J. D.,212,219 Cirillo,R. K.,87,118 Clark,E.V., 76,86,118 Clark,G.M., 152,167 Clark,H.H., 76,86,118 Clark,R.,35, 52

Clement,R., 197,199,219 Cohen,A. D.,38,52 Cohen,E.G., 152,153, 155,156, 157,158, 161,162, 163,166,167 Cohen,M., 38,52 Cohen,S.P., 67,71 Cole,M., 173,179, 245,277, 279 Comrie,B., 15,16,18,26 Connolly,K., 245,279 Conrad,R., 172,179 Cooper,B., 246,279 Cooper,R. L.,75,119 Corder,S.P., 37,52 Cordes,C., 24,26 Cortese,C., 145,148 Craig,D.R., 235,241 Cruikshank, J.G.,231, 241 Cummins,J., 58,59,61, 67,71,71, 136,147 Cziko,G., 205,206, 210,216,219, 220 D d'Anglejan,

A.,38,52, 209,220 Darcy,N.T., 132,147 Davies,M., 132,147 DeAvila,E. A.,150,151, 152,153, 154,155, 156,157, 161,162, 163,167,169 Deal,T.E., 152,167 DeCamp,D., 290,291, 292,293, 297 deKanter,A. A.,61,71 Devescovi, A.,113,117, 126,128 deVilliers,P. A.,87,118 deVilliers, J.,34,38,52, 53,54 Diaz,R.M., 69,70,72 Dillard,J.L., 232,239, 241,286, 296,297 Donaldson, M.,63,72 Dipboye,R. L.,191,194 Dore,J.,245, 277,279 Dowley,G., 245,277, 279 Downing,B., 45,52 Doyle,A.B., 68,72 Dreyer,J.T.,

23,26 Dulay,H., 38,39,52 Dumas,G., 38,54 Duncan,S. E.,150,151, 153,156, 157,167,168 Duskova,L., 38,52 E Ervin,S., 198,220 Ervin,S.M., 144,147 ErvinTripp, S.,38,52, 284,297 F Fasold,R., 286,297 Fathman,A., 40,52 Feldman,C., 64,65,72 Ferguson,C. A.,232,241 Ferres,S., 86,98,109, 114,117,118 300

G Garcia,R.,21, 26 Gardner,R. C.,183,188, 195,197,198, 199,210,219, Fincher,B. 220 H.,77,118 Garvey,C., Fisher,R.A., 87,117 138,147 Gass,S.,32, Fishman,J. 38,53 A.,20,26 Gearhart,M., Fiske,E.B., 252,279 21,26 Genesee,F., Flavell,J.H., 69,72,187, 123,128, 195,197,199, 243,251,279 202,206,220 Fleming,J. Gerace,W.J., S.,154,168 150,168 Flood,J.E., Gerard,L., 152,168 145,148 Forrest,R.A. Gibson,E.J., D.,97,98, 86,118 99,100,101, Gillis,M.,37, 111,118 39,53 Foss,D.J., Ginsburg,A., 87,118 151,168 Francis,W., Givon,T.,45, 97,118 53 Frankel,H., Glazer,N.,19, 203,220 26,190,195 Fraser,C., Gleitman,L. 122,128 R.,31,40,53 Frasure Glenn,C.G., Smith,N., 252,280 188,194, Glick,J.,176, 199,212,220 179 Frauenfelder, Glicksman, U.,40,53 L.,197,220 Goodenough, D.R.,150, 169 Gordon,M. M.,190,195 Gould,S.J., 48,53 Gowan,J.,67,

72,136,148 Gray,J.,86, 117 Green,D.,64, 72 Greenberg,J. H.,45,48,53 Greenfield,P. M.,155,166 Greer,C.,19, 26 Grober,E.,87, 117 Grosjean,F., 13,26 Guilford,J.P., 133,147 Guindon,H., 203,220 Gulutsan,M., 137,147 Gutter,R., 216,220 H Hakuta,K., 32,33,34,35, 36,37,38,39, 40,42,43,44, 52,53,54,69, 70,72 Hall,R.A., 232,241 Hall,R.A., Jr.,288,297 Hall,W,S., 243,245,249, 250,251,252, 277,279 Halliday,M. A.K.,77,78, 96,103,104, 118,237,241, 247,279 Hamayan,E., 199,220 Handlin,O., 19,27 Handscombe, J.,64,72

Hansen,L.J., 77,78,96,98, 99,100,101, 102,108,109 118 Hansen,S., 155,168 Hanson,S.G., 158,166 Hassing,Y., 246,280 Hatch,E.M., 33,53,75, 112,118 Havassy,B. E.,150,153, 167 Healy,A.,87, 118 Heise,D.R., 139,147 Henderson, D.,248,279 Henne,H.,77, 78,96,98,99, 100,101,102, 108,109,118 Herdan,G., 98,104,118 Herskovits, M.J.,296, 297 Hoefnagel Hohle,M.,37, 44,54 Holobow,N., 197,205,206, 219,220 Howe,H.,21, 26 Huang,J.,35, 53,112,118 Hughes,A. G.,132,147 Hunt,J.M., 154,168 Hyde,J.,164, 168 I

IancoWorrall, A.D.,65,72, 135,139,147 Imedadze,N. V.,57,72 Intili,J.K., 152,153,156, 157,158,162, 163,166,167, 168,169 301

K Kaalt,J.,12, 26 Kagan,S., 150,168 Kalin,R.,186, 191,194 Kang,H.,40, Isajiw,W. 53 W.,192, Kaplan,B., 194 173,179 Isayev, Karlgren,B., M.I.,15, 96,97,98,118 18,26 Keefe,J.,21, Ives,S. 26 W.,132, Keller,R.,138, 147 148 Jensen, KellerCohen, A.R., D.,35,53 132,147 Kennedy,T. Jeremy, L.,23,26,77, R.J.,86, 104,118 118 Kessel,F.,66, Jobse,J., 72 246,280 Kliegl,R.,113, Johnson, 118 C.,257, Krashen,S., 280 39,52 Johnson Kratochvil,P., Laird,P. 96,99,101, N.,76, 103,104,111, 118 118 Joosten, Kreindler,I., J.,246, 16,18,19,24, 280 26 Joreskog, Kucera,H., K.G., 97,118 139,140, Kuno,S.,45, 147 53 L Labov,W., 239,241,242, 243,244,245, 279,285,286, 291,292,293, 297 Lamarche,L., 191,195

Lambert,W. E.,3,9,25, 26,57,58,67, 69,71,72, 133,134,138, 147,183,187, 188,192,194, 195,197,198, 199,201,202, 203,204,205, 206,209,210, 213,216,219, 220,221 Lambert,W. W.,201,220 Landry,R.G., 137,147 Laosa,L.M., 152,168 Larsen Freeman,D., 39,53 Ledger,S.,63, 72 Lehman,W.P., 22,23,26 Lehmann,W., 43,48,53 Lenneberg,E. H.,32,43,54, 124,128 Leopold,W. F.,57,72 Levelt,W.J. M.,86,118 Levin,H.,86, 118 Lewis,E.G., 18,26 Li,C.,48,54 Li,C.C.,139, 147 Li,C.N.,77, 78,105,118 Li,Y.C.,101, 118 Lightbown,P., 33,54 Liu,I.M.,77,

96,97,118 Lockhead,J., 150,168 Luckyj,G.S. N.,17,26 Luria,A.R., 172,173,179 Lyons,J.,76, 77,118 M MacNab,G. L.,140,147 MacNamara, J.,51,54 MacWhinney, B.,112,113, 117,118,126, 128,145,147 Madden,C., 39,52 Maratsos,M. P.,42,54 Markman,E. M.,126,128 Mata,S.,158, 166 Mathias,J., 23,26,77, 104,118 Matthew,M., 245,279 McCleod,C., 245,279 Mcintosh,A., 237,241 McKillop,B. J.,113,117 McKirnan,D. J.,191,195 McLaughlin, B.,24,25,26, 33,54,135, 147 McNew,S., 113,117,126, 128 McQuillan,K., 11,12,13,26 Mei,J.Y.,23,

26 Mestre,J.P., 150,168 Meyer,J.W., 152,167 Meynard,R., 188,189,195, 199,221 Miller,G.A., 76,87,118 Milne,A.,151, 168 Mohanty,A. K.,70,72 302

Morin,S., 206,220 Moser,L. J.,22,27 Mowrer, O.H.,198, 220 Moynihan, D.P.,19, 26,190, 195 Mulcahy, R.,65,72, 136,147

N Nagy,W.E., 243,249, 250,251, 252,279 Nakajima,K., 64,72 Natale,F., 126,128 Nelson,D.L., 145,147 Nemser,W., 37,54 Nieves Squires,S., 152,163,168 Norman,D., 143,148 Novak,M., 19,27 O Oroller,J., 33,37,54 Oiler,J.W. Jr.,63,72 Olson,D.R., 63,72 Olver,R.R., 155,166 Orlov,A.,19, 27 Osgood,C. E.,144,147 Osherson,D. N.,126,128 Oyama,S., 43,44,54 P Palij,M.,144, 148 Parchment, C.,158,166 Park,R.E., 190,195 Pascual Leone,J., 150,167 Patkowski,

M.,44,54 Peal,E.,3,9, 133,138,148, 187,195 Penfield,W., 32,54 Perrow,C., 152,168 Peterson,P. L.,152,167 Piaget,J., 172,179 Pinter,R., 138,148 Politzer,R., 38,54 Price Williams,D. R.,150,168 Pulos,S.M., 154,167 Purbhoo,M., 66,72 R RamirezIII, M.,150,168 Ramirez,A., 38,54 Reber,A., 176,179 Reed,C.E., 239,242 Rheault,E., 188,189,195, 199,221 Richards,J., 33,37,54 Riessman,F., 152,168 Roberts,L., 32,54 Rogers,R., 197,220 Rongen,O. B.,77,78,96, 98,99,100, 101,102,108, 109,118 Rosansky,E.,

39,40,52,54 Rosenholtz, S.,159,163, 165,168 Rosenshine, B.,152,155, 162,165,166 Rosenthal,A. S.,151,168 Roy,J.D., 232,233, 235,242 Rumelhart, D.,143,148 Ryan,E.B., 63,71,71,72 S Saarni,C.I., 164,168 Saer,D.J., 132,138,148 Sanders,M., 150,168 Sankoff,G., 245,279 Sapir,E.,77, 118 Sapon,S.M., 183,194 Savchenko, F.,16,27 Scarborough, D.L.,145, 148 Scarborough, H.S.,114, 117 Scardamelia, M.,63,71 Schacter,J., 42,54 Schatzman, L.,248,279 Schlesinger, I.M.,39,54 Scholz,J.P., 150,168 Schuman,J. H.,234,242

Schumann,J., 33,37,40,52 Scott,W.R., 152,167 Scribner,S., 173,179 Segalowitz, N.,68,72 Segalowitz, N.S.,143, 148 Seitz,M.,69, 72 Selinker,L., 37,38,54 Shapson,S., 66,72 Shavelson,R. J.,152,168 303

Shen,M., 64,65,72 Shores,D., 239,242 Sidoti,N., 210,213, 219,220 Siegler,R. S.,172, 179 Sigelman, C.,249, 280 Simirenko, A.,17,27 Simon,W., 77,78, 104,111, 118 Sinclair deZwart, H.,38,54 Skutnabb Kangas,T., 59,72 Slobin,D. I.,39,54, 126,127, 128,143, 148 Smith,F., 132,148 Smith,S., 113,117, 126,128 Smythe,P. C.,197, 199,219, 220 Snow,C., 37,44,54 Snow,C. E.,246, 280 Sorbom, D.,139, 147 Spears,A., T Taba,H., 152,168 Tager Flusberg,H., 34,38,52, 53,54 Takahaski, Y.,152,168 Taylor,D. M.,199,212, 220,221 Taylor,M., 143,148 Tewksbury, M.G.,102, 119 Thernstorm, S.,19,27 Thompson, S.,48,54 Thompson, S.A.,77,78, 105,118 Tickle,C., 50,55 Todd,L., 288,298 Torrance,E. P.,67,72, 136,148 Toukomaa, P.,59,72 Tran,C.,64, 72 Tucker,G. R.,25,26, 38,52,57, 67,69,71, 72,134,147, 198,199, 202,203, 204,205, 209,210, 220 U Ulibarri,D. M.,150,151,

167 Uyemura 296,297 Spencer,J., Stevenson, B.,161,168 231,242 Spolsky, V B.,75,119 Valett,R.E., Starck,R., 154,169 69,72 Valian,V., Steele,S., 291,298 45,54 Venezky,R., Stein,N., 77,98,119 252,280 Venza,V., Stenson, 126,128 N.,33,37, vonRaffler 54 Engel,E., Stern,P., 249,280 152,168 Vorster,J., Stemberg, 246,280 R.J.,154, Vygotsky,L. 168 S.,57,72, Stevenson, 73,172,179 B.,158, W 166 Wang,S.C., Stewart, 77,96,97, W.A.,237, 119 239,242, Weinreich, 285,286, U.,40,55, 293,296, 132,148, 297 144,242 Stone,N., Wellman,H. 152,158, M.,123,128 166 Wellman,J., Strauss, 257,280 A.,248, Werner,H., 279 173,179 Strevens, Whitaker,R. P.,237, A.,12,14,27 241 Williams,L., Sullivan, 43,55 E.V.,132, Wilson,B., 147 153,169 Svartvik, Witkin,H. J.,37,54 A.,151,169 Swain,M., Wolfe,D., 38,54,64, 210,213, 72,202, 219,221 221 Wong Fillmore,L., 35,40,55

Wright,J., 287,298 Wu,J.,67, 72,136,148 Z Zigler,E., 177,179 304

SubjectIndex A Affixes,105108 AfroEnglishpidgin,231232 DevelopmentintoCreole,232 Mediumofcommunication231232 forEnglishslavetrade,231 inPlantationsofAmerica,232 Age,4344,125127 andlanguageacquisition,125127 Criticalperiod,4344 EffectsonL2acquisition,4344 Positiveeffectsofage,44 Ambiguity,seeInterdialecticalambiguity Asymmetry,betweenproductionandreceptivecompetence,282284 inComprehension,282286 Acrossdialects,283286 Acrosslanguages,282284 Attitudes,201216 Componentsof,201 Definitionof,201 Measurementof,200202 byMultidimensionalscaling,212213 byQuestionnaires,209210,213216 byStereotypescales,202 Attitudesandimmersionprograms,202,207209 Attitudesandmotivation,197 Relationtolanguageacquisition,197200 Attitudesasdeterminants,198 Attitudesaseffects,198,199,202 B Bidialectism,234236 BlackEnglish/StandardEnglishbidialectism,235236 Dialectaldifferences,235 Difficultiesforbidialectalstudent,235236 Psuedocomprehension,236 andLinguisticbackground,234235 Languageversusdialect,235 Bilingualcognition,modelsof,143146 BilingualEducationAct,1968(TitleVII,ESEA),20,151 Bilingualism,5771 Additive,5859,187 Childhood,39,5771,131147,171179,183230 Criticismsofcognitiveenhancementresultingfrom,140141 Evaluationofeffectindevelopmentalpsychology,131133,171173 305

Greater metalinguistic ability developing during,135 136 Historical reviewof research involving,131 137 Human information processing analysisof,172 Impactof cognitive psychologyon theoriesof,142 143 Methodological problemsin researchon, 138141 Multivariate analysisinthe studyof,139 140 Negative impacton intelligence from,131133 Piagetian analysisof,172 Positiveimpact onintelligence from,133137 Relationto cognitivestyle, 164165 Researchwith quasi experimental methods involving,139 140 Researchwith true

C Canada,1114 Canadian Constitutionof 1982,effecton languagepolicy, 13 Charterof Rightsand Freedoms,13 Confederation Actof1867,12 Englishand French Colonies,12 SevenYears War,12 Government actionagainst Manitobain 1983,14 PartiQuebecois, 13 Recentdecline of,14 Political movement amongFrench Canadians,12 Populationof, 1112 Language diversityin,11 Ratesof bilingualismin, 1112 QuebecActof 1774,12 Royal Commissionon Bilingualism and Biculturalism, 1213 Fearsexpressed amongFrench Canadians,13 Resultsof,13

experiments involving,138 139 Theoretical frameworks usedinthe studyof,141 146 TorranceTest ofCreative Thinkingand, 136137 Vygotskian analysisof,172 173 Limited abilityin,59 60 Longterm effectsof,60 61,67 Partial,5960 Proficient,59 60 Shortterm effectsof,60 61,67 Socialcontext of,174179 Subtractive, 58,187 BlackEnglish, 232234,284 294 Acquisitionin childhoodof, 233 Innovations in,234 Processof linguistic changeof, 232233 Comparison with bilingual situations, 233

Socialand economic discrimination againstFrench speakers,12 Supremecourt decisionagainst Quebecin1984, 1314 Chinaand languagepolicy, seePeople's Republicof China Chinese,75116 Abstractnessin, 8386,9899, 102 Bilingualism involving,75 116 Comparison hypothesisof, 79,9495,116 Dialectsin,77 Differences fromEnglish,78 Linguistic attributesof,77 78,8385 Writingin,79 83 Cognitive Maturity,3337 EffectsonL1 acquisition,33 EffectsonL2 acquisition,33 37 Lengthof utterance,37 Prefabricated patterns,3537 Development ofabstract structure,35 37 Semantic relations,34,

39 Sentence coordination, 34 Vocabulary development, Slownessof 37 linguistic Communication acculturation, game,223225 234 Goalsof communication, 224 Socialand ethnicidentity, 224225 306

Communications, D 114115 Decreolization, Communicative 288 activities,6164 Deepstructure, incontext 293 embedded Dialects,281 situations,6163 284,287293, incontextreduced 295 situations,6164 Definitionof, Competence,282 281fn1 284,290291 Structural Asymmetry relationship between between,281 productiveand 282,284,287 receptive,282 293,295 284 ofEnglish, Productive,282 comprehension 284,290291 across,281, Competence, 283284 receptive,282284, E 290293,296 Emptyandfull Comprehension, words,78,100, seealso 107 Miscomprehension, Englishteaching, Pseudocomprehensi 236241 on BlackEnglish Acrossdialects, speaking 281282,291292 studentsand, Betweenrelated 236237 languages,281 Limitationsof 282,291292 teaching Consequencesof techniquewith, bilingualism,225 236237 Socialaffective Policyof consequences,225 National Traditionaland Councilofthe currentappraisals Teachersof of,131138 English Contentwords,80 (NCTE),237 82,96100,103 240 106,109111 Criticisms Context madeby,238 dependency,9798, 240 102 Student'sRight Continuum,see toTheirOwn CreoletoEnglish Language continuum (SROL),237 CreoleEnglish,290 Suggestionsfor

Guyanese,290 Jamaican,290 Creolelanguage, 288 CreoletoEnglish continuum,288, 296 Guyanese,290 Jamaican,290 Criticalperiod,126 127,seealsoAge, EffectsonL2 Acquisition Culturaldiversity, 184,185 Culturalidentity, seeEthnicidentity

reformby,240 Environmental effects,274277 Homeand school,274277 Erroranalysis, seealso SimilaritiesinL1 andL2 acquisition,37 40 Errors,sourceof, 123125 Interference, 124 Ethnicidentity, 184194 Threatsto,187 194 F Finding Out/Descubrimie nto,153166 Outcome measuresused inresearchon, 154158 Observational measures,155 156 Paperand pencil measures,154 155 Performance measures,157 158 Resultsand analysisof researchon,158 163 Outcome measuresused, 158161 Process analysesused, 161163 Formulaic utterances,see

Cognitive maturity EffectsonL2 acquisition, Prefabricated patterns Functionwords, 8082,100103, 106107,109 112 G Geechee (Dialect),see Gullah(Dialect) Grammatical morphemes,see alsoSimilarities inL1andL2 Acquisition,37 40andNative Language Transfer,4043 Groupidentity, seeEthnic identity 307

H Hebrew,acquisition of,125127 I Idealspeaker listener,290 Inflections,see Affixes Interdialectal ambiguity,285,294 296 Interdialectal miscomprehension, see Miscomprehension Intergroupcontext, Gullah 183194 (Dialect), Defined,184 288,294 Internalstatewords, 295 247253 Guyanese Cognitive Creole, implicationsof, continuum 247248 ,see Definition,249 Creole 253 toEnglish Socialclassbased continuum differences,248 Guyanese 249 Creole,see J Creole JamaicanCreole, English seeCreoleEnglish JamaicanCreole, continuum,see CreoletoEnglish continuum L Language development,112 115 Language immersionprogram, seeSt.Lambert project Language interactionin cognition,see Bilingualcognition Languagepolicy

andbilingual development,2425 Additiveversus subtractive bilingualism,25 Psychological implications,25 Comparisonof bilingualprograms inCanada,U.S., SovietUnion,and PRC,2425 Language proficiency,57,60, 6164 Confoundedwith minoritygroup membership,150 151,164 Confoundedwith socioeconomic status,149150 Languageand thought,asa dichotomy,129 Lexicalcategories, 8994,100109 Adjectives,8994, 103105 Adverbs,8994, 103105 Articles,84,89 94,100 Conjunctions,89 94,101 Nouns,8994,103 105 Prepositions,89 94,101102 Pronouns,8994, 100 Verbs,8994,103 105,107109 Linguistic competence,see Competence M Meaning,76,89 94,96103,109

112 Metalinguistic development,57, 58,61,63,6471 Metalinguistic tasks,127128 Miscomprehension acrossdialects,294 296,296fn14 Mismatch hypothesis,277278 Racialdifferences, 277 Socialclass differences,277 278 Morphemes,7778 Multilingual ideologies,184194 Assimilation ideology,185,186, 190192 Deculturation ideology,185,186 Integration ideology,185,192 194 Separation ideology,185,187 190 N Nativelanguage transfer,4043 Avoidanceof structures,4243 Formclass,42 Grammatical Morphemes,4041 Articles,40 Negation,4041 inLaterstagesof L2development, 42 P People'sRepublicof China,2223 Autonomous nationaldistricts in,23

Historyof languagepolicyin, 23 Effectofcultural revolution,23 Recentpolicy,23 308

Q Qubec,203207 Effectof,on Canadian stereotypesof Language hilinguals,204 policy,22 207 Languages, Politicaland 2223 demographic Dialectsof characteristics, Han,22 203204 NonHan R languages, Reading,113 2223 114 Putonghua Receptive andPinyin, competence,see 22 Competence S SapirWhorf Pidginlanguage,288 hypothesis,128 PlantationEnglish 129 Creole,232 Semantics,see Decreolization,233 Meaning Development,232 InfluenceofEnglish, Similaritiesin L1andL2 232 Portuguese(language), acquisition,37 40 282283 Comprehensionof,by in Spanishspeakers,282 Comprehension ofparticular 283 Prefabricatedpatterns, structures,38 seeCognitivematurity, 39 Erroranalysis, EffectsonL2 3738 acquisition Grammatical Processingstrategies, morphemes,40 languagespecific differencesin,128129 41 Frequency,39 Productive 40 competence,see Semantics,39 Competence Situational Pseudocomprehension, variation,243 296,fn14 247 UseofR construction, 244245 Socialconcepts, 225228

Implicationsfor bilingualism, 226228 Bicultural versus unicultural bilinguals,226 228 Meaningand referent,225 226 Socialidentity, seealsoEthnic identityTheory of,191 Social perspectives,216 219 Developmentin bilingual children,216 218 Relationto attitudes,218 219 SovietUnion,14 19 Bolshevik policiestoward nonRussian languages,16 17 Demographic projectionsof Soviet population,18 19 Developmentof nonRussian languagesin the1920s,17 Developmentof pre revolutionary Russianempire, 15 Linguistic diversity,15 Language

families,15 Languagesby percentage speakers,15 Population,14 PostStalinist policyinthe 1950s,1960s, 1718 Republics,14 Secondworld war,17 Linguistic reforms,17 Stalin'svictory toast,17 Sovietlanguage policysincethe late1960s,18 Stalin'spolicies inthe1930s,17 Purges,17 Tsarist language policies,16 Estoniaand Finland,16 NonRussian languages,16 Ukraine,16 Spanish (language),282 283 comprehension of,by Portuguese speakers,282 283 St.Lambert project,134 135 Standard English,284 289 Standard languageor dialect, definitionof, 281,fn2

Structure,76, 9094,103 107,109112 Surface structure,293 Syntax,see Structure 309

T Thresholdhypothesis,5960 Transfer,seeNativelanguagetransfer Typology,seeUniversals U UnitedStates,1921 BilingualEducationAct1968,20 BilingualEducationAct1974,21 DemographicprojectionsofnonEnglishspeakers,21 Englishas"officiallanguage,"19 EthnicandlinguisticdiversityintheU.S.,1920 1980census,20 Lauv.Nichols(1974),20 Oppositiontomaintenanceofbilingualism,21 Reappraisalof"meltingpot"imageintheU.S.,19 Universals,4451 Factoranalysis,4648 Goalsoflanguagetypology,4546 Psychologicalcorrespondence,4851 inLanguageacquisition,5051 W Whorfhypothesis,128129 Wordfrequency,9899,103 310

Vous aimerez peut-être aussi