_So_Best_2010_Cross-language Perception of Non-nat

273Cross-languageCross-languageCross-languageCross-languagePerceptionPerceptionPerceptionPerceptionofofofofNon-nativeNon-nativeNon-nativeNon-nativeTonalTonalTonalTonalContrasts:Contrasts:Contrasts:Contrasts:EffectsEffectsEffectsEffectsofofofofNativeNativeNativeNativePhonologicalPhonologicalPhonologicalPhonologicalandandandandPhoneticPhoneticPhoneticPhoneticInfluencesInfluencesInfluencesInfluencesConnieKSo,CatherineTBest.LanguageLanguageLanguageLanguageandandandandSpeechSpeechSpeechSpeech.ThamesDitton:Jun2010.Vol.53535353Part2.pg.273-91———————————————————————————————————————————AbstractAbstractAbstractAbstractThisstudyexaminedtheperceptionofthefourMandarinlexicaltonesbyMandarin-naïveHongKongCantonese,Japanese,andCanadianEnglishlistenergroups.Theirperformanceonanidentificationtask,followingabrieffamiliarizationtask,wasanalyzedintermsoftonalsensitivities(A-primescoresoncorrectidentifications)andtonalerrors(confusions).TheA-primeresultsrevealedthattheEnglishlisteners'sensitivitytoTone4identificationsspecificallywassignificantlylowerthanthatoftheothertwogroups.Theanalysisoftonalerrorsrevealedthatalllistenergroupsshowedperceptualconfusionoftonepairswithsimilarphoneticfeatures(T1-T2,T1-T4andT2-T3pairs),butnotofthosewithcompletelydissimi15larfeatures(T1-T3,T2-T4,andT3-T4).Language-specificerrorswerealsoobservedintheirperformance,whichmaybeexplainedwithintheframeworkofthePerceptualAssimilationModel(PAM:Best,1995;Best&Tyler,2007).Thefindingsimplythatlinguisticexperiencewithnativetonesdoesnotnecessarilyfacilitatenon-nativetoneperception.Rather,thephonemicstatusandthephoneticfeatures(similaritiesordissimilarities)betweenthetonalsystemsofthetargetlanguageandthelisteners'nativelanguagesplaycriticalrolesintheperceptionofnon-nativetones.———————————————————————————————————————————1111....IntroductionIntroductionIntroductionIntroductionThetermlinguisticexperiencehasbeenwidelyusedinaccountsofhowperceptualperformancewithnon-nativespeechcategoriesisinfluencedbythelistener'snativelanguage(L1).WhilemostpreviousstudieshavereportedL1influenceonspeechperceptionatthesegmentallevel,itsimpactonlisteners'speechperceptionatthesuprasegmentallevelisnotwellunderstood.Inparticular,questionsabouthowandtowhatextentlinguisticexperiencewiththefeatures1oftheprosodicsystems(e.g.,lexicaltones,pitchaccents,lexicalstress,andintonation)ofone'snativelanguageaffectsperceptionofnon-nativetonecategorieshavenotyetbeenaddressed.Tobridgethegap,thisstudyexaminedtheroleofdifferinglinguisticexperienceofnativetonalityontheperceptionofnon-nativeMandarintonesbythreegroupsofMandarin-naïvelisteners:nativespeakersofCantonese,Japanese,andEnglish.Studiesoncross-languageperceptionatthesegmentallevelhavegenerallyfoundthatadultsexperiencegreatdifficultiesproducingandperceivingnon-nativeconsonantand/orvowelcontrasts(Lisker&Abramson,1970;Polka,1995;Werker&Tees,1984)althoughsomenon-nativesegmentalcontrastsarediscriminatedquitewell(e.g.,Best,McRoberts,&Goodell,2001).TheL1perceptualeffectsalsoextendtorelateddifficultieswhenadultslearnanewsecondlanguage(L2)(Flege,McCutcheon,&Smith,1987;Guion,Flege,Akahane-Yamada,&Pruitt,2000;Jamieson&274Morosan,1986;Logan,Lively,&Pisoni,1991;Polka,1995).Ithasbeensuggestedthatperceptionofnon-nativespeechcontrastsisconstrainedbyboththephonologicalandthephoneticpropertiesoftheirnativelanguage(L1)(Best,1995;Flege,1995;Polka,1991,1992;Strange,1995).Phonologicalpropertiesarecrucialtothestructuraltransformationofwords.Theyarecontrastiveandcategoricalandinvolvelanguage-specificrulesgoverningthedistribution,patterningandcontext-determinedrealizationofconsonants,vowels,andtones.Forexample,nativespeakersofJapanesehavedifficultiesinperceivingtheEnglish/r/-/l/distinction,becausethesephonemesarenotcontrastiveintheJapanesephonologicalsystem(Brown,2000;Hume&Johnson,2003).Incontrast,weusethetermphoneticpropertiestorefertocharacteristicsofpronunciationthatarenotphonologicallydistinctiveinalanguage(fine-grained,gradient,within-category,non-contrastivedetailsofspeech).Forinstance,whenAmericanEnglishlistenerscategorizedtheZuluaspiratedvoicelessvelarstop[kh]andejective[k'],theyperceivedbothasthevoicelessstop[kh],butthenon-nativesound[k']wasperceivedashavingoddorunusualvoicequalities(non-contrastivegradientdifference),becauseEnglishhasa[kh]butnoejectives(Bestetal.,2001).Effectsoflinguisticexperiencehavealsobeenreportedinsomestudiesonsuprasegmentalfeatures,suchasstresspatterns,quantitycontrasts(e.g.,vowellength,consonantgemination)andsentence-levelprosodicpatterns,withinfluencesfrombothphonologicalandphoneticlevelsoftheL1.Adults'L1prosodicsystemshaveaprofoundeffectontheirperceptionofthesuprasegmentalcontrastsofnon-nativelanguages(Archibald,1992,1993;Guion,Harada,&Clark,2004;Hirata,2004;McAllister,Flege,&Piske,2002).Forexample,bothPolishandSpanishlearnersofEnglishtendtoapplytheirL1stressassignmentruleswhenperformingastressplacementtaskwithspokenEnglishwords(Archibald,1992,1993;Guionetal.,2004).Studiesontheperceptionofnon-nativelexicaltoneshavealsofoundthatlinguisticexp