日本語とアイヌ語の証拠性

全文

(1)Title. 日本語とアイヌ語の証拠性. Author(s). 井筒, 勝信. Citation. 北海道教育大学紀要. 人文科学・社会科学編, 57(2): 123-136. Issue Date. 2007-02. URL. http://s-ir.sap.hokkyodai.ac.jp/dspace/handle/123456789/844. Rights. Hokkaido University of Education.

(2) 北海道教育大学紀要（人文科学・社会科学編）第57巻第2号 JournalofHokkaidoUniversityofEducation（HumanitiesandSocialSciences）Vol．57，No．2. 平成19年2月 February，2007. EvidentialityinJapaneseandAinu＊ IZUTSU Katsunobu. DepartmentofEnglish，AsahikawaCampus，HokkaidoUniversityofEducation. 日本語とアイヌ語の証拠性井筒勝信. 北海道教育大学旭川枚英語学研究室. ABSTRACT Ainu，anindigenouslanguageinJapan，hasoftenbeencontrastedwithJapaneseinitsrichsystemof. evidentialtyexpressionsalongwithitssystematicuseofpersonalaffixestoverbs／nounsanditsreadiness Ofnounincorporation．ThepresentstudyshowsthatJapaneseandAinumanifestslightlydifferentbut VerySimilarsystemsofevidentiality；eXpreSSionsforevidentialityareconstructedwithformalnounsand a be−typeOrdo−typeformalverbinbothlanguages．Thedifferencemostlyconcernsthesensoryand. inferentialevidentialsinthetwolanguages．Whilethestrengthofjustificationismoreclearlyexpressed thanthekindofjustificationinJapanese，thekindofjustificationismoredominantinAinu．Thestrength. OfjustificationiscodedinsomeinferentialevidentialsinJapanesebutisonlysuggestedsecondarilyby SOmeSenSOryeVidentialsinAinu．. 1．Introduction ThepresentstudyproposesacontrastiveanalysisofevidentialsinJapaneseandAinu，tWOmajor indigenouslanguagesofJapan，fromfunctionalandcognitivepointsofview．Sincethenotion“evidentiality’’. hasmainlybeendevelopedaroundthedescriptivestudiesofindigenouslanguagesinNorthandSouth. ＊ ThisisaslightlyrevisedversionofapaperdeliveredatthelOthInternationalConferenceoftheEuropeanAssociation forJapaneseStudies，heldatWarsawUniversity（Poland）on27through30August2003．IamverygratefultoProfessors AndreWlodarczyk，RomualdHuszcza，andAlfredF．Majewiczfororganizingandmoderatingaworkshopthatprovided mewiththeopportunitytopresentthepaperinEurope．ThanksalsogotoKazuhikoYamaguchi，MitsukoNaritaIzutsu andYoritakaTezuka，andtoThomasJaquesforcommentingonearlierdraftsofthepaper．Itgoeswithoutsayingthat allremainingerrorsaremine．. 123.

(3) 井筒勝信. America，itsintroductionintostudiesofmajorlanguageslikeEnglishandJapanesewasacomparatively laterattempt（ChafeandNichoIs1986）．TheapplicationofthenotiontominorlanguageslikeAinuhas beenslower．. Thelinguisticexpressionsanalyzedinthepresentstudyhavetraditiona11yreceivedprimarilysyntactic， and onlysecondarilysemantic，description and treatment；theyhavebeen exclusivelycategorizedas auxiliariesorparticles（withsomemeaning）．veryfewstudieshavetakentheviewthatsomeofthe auxiliariesand／orparticlesformafunctionalorconceptualcategorywhichcanbetermed“evidentiality’’ （Aoki1986）．. Whileevidentialitycircumscribesexpressionsof“kindsofevidence’’initsnarrowersense（Anderson. 1986），initsbroadersenseitalsoencompassesexpressionsofso−Cal1edmodalitywhichconcern“strength ofevidence．’’InEnglish，differentsyntacticelements（suchasverbsandadverbs，auXiliaryphrases，Or clauses）cancodeevidentiality；thoughsomelinguistsmaydoubtthepresenceofthecategory．In JapaneseandAinu，however，itismoreplausibletoassumethecategorybecauseitiscodedinavery limitednumberofsyntacticelements（auxiliariesorparticles）．. ThepresentstudydemonstratesthatJapaneseandAinumanifestverysimilarbutslightlydiffering SyStemSOfevidentiality．Thedifferenceisbestreflectedonthesensoryandinferentialevidentialsinthe twolanguages．Whilethestrengthofjustificationismoreclearlyexpressedthanthekindofjustification inJapanese，thekindofjustificationismoredominantinAinu．Thestrengthofjustificationiscodedin. SOmeinferentialevidentialsinJapanesebutis，atbest，Onlysuggestedsecondarilybysomesensory evidentialsinAinu．. Afteridentifyingwhatevidentialityandevidentialsarebysurveyingthreemajorearlierstudiesin SeCtion2，IexaminespecificevidentialsinJapaneseandAinuinsections3and4．Finally，insection5，I. makeacomparisonbetweentheJapaneseandAinuevidentialitysystemsbysummarizingthedescriptions Ofevidentialsinthetwosystems．. 2．EvidentialityandEvidentials EvidentialitywasfirstnoticedasagrammaticalcategoryamongthestudentsofNativeAmerican. languages（Boas1938）andlaterbecamerecognizedasanotionalcategorythatdoesnotnecessarilyform aunitarysyntacticcategorybutmaybecodedindifferentsyntacticformsinmanylanguages．Those. specificformsarecalledevidentials．AsAikhenvald（2003：27，n．7）putsit，“Theterm‘evidential’itself appearstohavefirstbeenintroducedbyJakobson（1957／1995）．’’ ChafeandNichoIs（1986）isthefirstattempttoestablishacross−1inguisticortypologicalframework. forevidentiality．Chafe（1986）usesthetermsevidentialityandevidentialintheirbroadestsense．Inhi terminology，“．．．eVerythingdealtwithunderthisbroadinterpretationofevidentialityinvoIvesattitudes. towardknowledge’’（ibid．：262）．Heintroducesfournotionsintohisdescription：（i）reliabilityofknowledge，（ii）modeofknowing（belief，induction，hearsay，anddeduction），（iii）sourceofknowledge（？？？，eVidence， 1anguage，hypothesis），and（it）whatknowledgeismatchedagainst（verbalresourcesandexpectations）．（i）concernsthespeaker’sassessmentoftheknowledgeinquestion，intermsof“itsdegreeofreliability， thelikelihoodofitbeingafact’’（Chafe1986：264），rangingfromhightolow．（ii）and（iii）arebothassociated. 124.

(4) 日本語とアイヌ語の証拠性. withhowtheknowledgeisacquired．Asfarashisexplicationimplies，（ii）and（iii）1argelycovary；theitems. inonecorrespondstothoseintheother：“belief’’isamodeofknowinginwhichevidenceisunknownor unclear，Or“COnCernforevidenceisdowngraded’’（ibid．266），“hearsay’’isbasedonlanguage（whatthe speakerisinformedof），and“deduction’’isamodebasedon“hypothesisformationandprediction’’（ibid．： 272）．1Tnconnectionwith（ii），henotesthat“modeofknowingimpliessomethingaboutreliability，butnot viceversa’’（ibid．：266）．（i司pertainstohowwelltheexpressionthespeakeruses（verbalresources）or. hisexpectationmatchestheknowledgeheintendstoexpress． Chafe（1986：263−264）makesinitialinstantiationofthefournotionsinreferencetothefollowing examples：. （1）a．Ifeelsomethingcrawlingupmyleg． b．It’sprobablyaspider． C．Itfeelssortofcreepy． d．Oddlyenoughitfeelsgood．. In（1a），theideathat‘somethingiscrawlingupmyleg’iswhathecal1sknowledge，themodeofknowing isinduction，thesourceofknowledgeisabodilyperception，andthereliabilityofknowledgeishigher. thanthedegreein（1b）．Thespeakersignalsthat“creepy’’isanapproximatewayofcharacterizingwhat hewasfeelingin（1c），Whilstheindicatesthatthepieceofknowledgeexpressedby“itfeelsgood’’was contrarytoexpectationin（1d）． Anderson（1986：274）adoptsmorerestrictiveterminology，emphasizingtheimportanceofdistinguishing “trueevidentialcategoriesfromotherformsthatsEEMevidential，butarenot．’’Hedefinesarchetypal. evidentialsas（a）showingthe“justificationforafactualclaimwhichisavailabletothepersonmaking thatclaim，’’（b）being“notthemselvesthemainpredicationoftheclausebutareratheraspecification addedtoafactualclaim，’’（c）having“theindicationofevidence．．．astheirprimarymeaning，nOtOnlyas apragmaticinference，’’and（d）beingcodedin“inflections，Clitics，OrOtherfreesyntacticelements（not compoundsorderivationalforms）．’’. Recognizingthat（a）to（c）arebasicallysound，Aikhenvald（2003）countersthat（d）shouldnotbe. Subsumedunderthedefinitionalpropertiesofevidentialsbecause“thiscriterionwouldnotworkforsystems inwhichthedistinctionbetweeninflectionalandderivationalcategoriesisnotclear−Cut．’’Shepointsout. thatevidentialityisinappropriatelytreatedasasubcategoryoftense−aSpeCtOraSaSubtypeofepistemic. OrSOmeOthermodalityandthattheinappropriatetreatmentsareattributabletothe“facttha evidentialsaretypically（albeitinaccurately）translatedintoEnglishandotherfamiliarEuropean languagesasepistemics‘apparently’or‘obviously”’（ibid．：24）．Sheemphasizesthatevidentialityis“a CrOSS−1inguistica11yvalidgrammaticalcategorywhosemeaningissourceofinformationandnotjust‘1inguistic. codingofepistemology”’（ibid．：25）．ThisphraseinsinglequotesistheverysubtitleofChafeand NichoIs（1986）．. 1 However，hementionsEnglishexamplesinwhichinductiondoesnotindicatewhatthenatureoftheevidence（source ofknowledge）is：1imusthavebeenahid；WdlituJaSjustobviousJcouldn’tuJOrk（1986：266）．. 125.

(5) 井筒勝信. Acomparisonoftheabovethreemajorapproachestoevidentialitywi111etusknowthestateoftheart：. Evidentialityisanotional（andgrammatical）categorywhichservesprimarilytospecifythe justification（typica11y，eVidence）forthecognitionthespeakerintendstoexpress． Thetypesofjustificationdistinguishedinal1theapproachesare：（I）sensoryperception（furtherdivided intovisual，auditory，OrnOn−VisualinAnderson（1986）andAikhenvald（2003）），（II）1anguage（also named“hearsay’’inChafe（1986）andAnderson（1986），and“reported’’inAikhenvald（2003）），and（Ⅲ）. inference（differentiatedintoinductionanddeductioninChafe（1986））．2Evidentialsofeachtypewill hereafterbecal1ed“sensory，1inguistic，andinferentialevidentials．’’ ReliabilityinChafe’s（1986）descriptionisanalyzableasthestrengthofjustification，Whichistreated asbeinginvoIvedinbothevidentialityandepistemicmodality．Anderson’svarioustypesof“inferentials’’. StretChoverthetwocategories．Itfollowsthatexpressionsforthestrengthofjustificationmediate evidentialityandmodalityandthatmostoftheexpressionsaresubtypesofinferentialevidentials．Although itis“notasubcategoryofepistemicorsomeothermodality，OrOftense−aSpeCt’’（Aikhenvald2003：1）， evidentialityshouldbeviewedasoverlappingmodality．Thisviewisusefulatleastforthecontrastive analysisofJapaneseandAinuaswi11beshownbelow．. 3．TypeSOfjustincation ThepresentsectionidentifiesspecificevidentialsinJapaneseandAinu．Asconfirmedabove，atleast. threetypesofjustificationaredistinguishable：（I）sensoryperception，（II）1anguage，and（Ⅲ）inference； WeCanaCCOrdinglydistinguishsensory，1inguistic，andinferentialevidentials．Thesensoryandlinguistic. evidentialsindicatethatthecognitionexpressedisinducedbysomethingperceivedandtold（or. written），reSpeCtively．Inferentialevidentialsconveythatthecognitionissomethinginferred．SinceiI血rence Canbemadefromsomethingperceivedortoldaswellassomeothercognition，thedistinctionbetween. thefirsttwoand（Ⅲ）isnotclear−Cut．Theyshouldbeviewedasoverlappingoneanother． 3．l．Japaneseevidentialsofthejustificationtype. Aoki（1986）distinguishesfourkindsofJapaneseevidentials：“descriptionsofsensation’’or“inferences regardingfeelingsofothers’’（卯ru），“markersoffact’’（n巾ノda），“hearsay’’（sooda，tte），“inference’’or. properevidentialforms，，b，00da，7Wii，SOOda）．3Thefirsttwokindsaredismissedfromourdiscussio becausetheydonotserveprimarilytospecifythejustificationforthecognitionthespeakerintendsto. express．4Thethirdkindcorrespondstoourlinguisticevidential．Thelastkindseemstoincludeoursensory. 2 Whatknowledgeismatchedagainst，（iv）inChafe’sdescription，thusfallsoutsidethe（central）categoryofevidentiality． 3 AlthoughAoki（1986：233）describesthedifEerencebetweentwotypesofsoodainpronunciationalone，theirclearer differenceliesintheverbalendingprecedingthem：−iintheinferentialandイりuand−iainthehearsay． 4 Inmyanalysis，nh）marksthe‘‘hearer−01d”status，aCtualorpresupposed，Ofthecognitionexpressedinitspreceding Clause．Iwouldliketoleavethismatterforanotherstudy．. 126.

(6) 日本語とアイヌ語の証拠性 andinferentialevidentials．. SomemoreobservationsandexplanationsmayhavetobeaddedtoAoki’sdescription．First，yOOda，. mitaida，andrasiiareallusabletocodethepresenceoflinguisticaswellassensoryjustificationasin （2a）．wecanutter（2a）eitherwhenwehavesomeperception（1ikeseeingblackasphaltandgreysky throughthewindow，hearingbeatingnoisesontheroof，OrSmellingwetairanddust）orwhenweare. linguisticallyinformed（1ikelisteningtotheradionewsorreadinganewspaper）．Contrastively，SOOda， whenprecededbytheverbalendingsイりuand−taliketeiruin（2a），isonlyusedforlinguisticjustification asin（2b）．whenprecededbytheverbalending−ilikehuriin（2c），itindicatesthepresenceofperceptual justification；itusuallymeansfurthermorethatthespeakermakesinference．Thus，SOOdainthisusemay haveto beclassifiedinto ourinferentialevidential．. ・ご＝．l川い㌔／〟〟／−／■／〃り・…′・ん／川／／J／／・／J＝1雨／． rain−NOM fall−PROG−EVD. ‘Itlookslike［soundslike／seems］itisraining． ‘Ihearthatitisralnlng． l）．l川い㌔／〟〟／−／■／／てトバ＝′′ん／． rain−NOM fall−PROG−EVD ‘Ihearitisralnlng． l−．l川い㌔／〟〟〟−∫川′・ん／． rain−NOM fall−EVD ‘Itlooks／seemslikerain．. り．l川い㌔J山J／−／■／／てり・…一ヾ仙んJん川ニ／・んJんん〃／・んJ山川／揖・ん∴ rain−NOM fall−PROG一口VO. ‘Itseemsitisraining．し、．l川い㌔／〟〟／−／■／〃トり／・やバJJ／て＝ノ／り料バJJ／てJん〟〟料バJJ／・仇 rain−NOM fall−PROG−EVD. ‘Itsoundslike／sme11slikeitisraining．卜l川い㌔／山J／−／■／／てトム料／．ヾ〃／てJん／ノ／ニ料ⅢJ／てり・りん／ノ／㌔／．ヾ〃／てJ． rain−NOM fall−PROG−EVD. ‘Itfeelslikeitisraining／Ifeelitisralnlng．ゞ．l川い㌔／〟〟／−／■／〃ト／り／仇 rain−NOM fall−PROG−EVD. ‘Theysayitisrainlng． 1l．l川い㌔／〟〟／イ■／〃ト／り／∼ト〟〟JJJバ／・ん／〟J川∫J／・ん／〟り／…ん／． rain−NOM fall−PROG−EVD. ‘Theysayitisrainlng．. Icancomeupwithsomemoresensoryandlinguisticevidentials．Expressionslikeyoosuda，kanzida，. 人■〟−／／・ん／．〟J川／揖・／−／．り／り才一／．ヾ〃／てJ．ノ／／り料／．ヾ〃／てJ．・人川㌔／．ヾ〃／てJ．‥′／ニ料／．ヾ〃／てJ．・ん〟′／料／．ヾ〃／てJ．人心／．ヾ〃／てJ．ん／ノ／ニ料／．ヾ〃／てJ、. andyokangasurualsofunctionasevidentialsofsensoryperceptionin（2d−f）．5Inadditiontosooda. 127.

(7) 井筒勝信. （precededbytheverbalendingsイりuand−ta）andtte，tOiuandtoiu−hanasida／uwasada／kotodaalsoserve todesignatelinguisticjustificationasin（2g−h）．Thesehaveamoreliterarytone，amOngWhichtoiu，When Standingalone，isexclusivelyawrittenexpression．. Sentence（2e）allows rasiimarginally as wellasyoodaand mitaidatointervene between OtOgaSuru／niolgaSuru／kehaigasuru and their precedingverbphrasesasin Ame−gahut−teiru− yoona／mitaina／？7Ⅶii−OtPuru／nio函uru／kehaigasuru，Whilesentence（2f）disa1lows77Nii’sintervention， thougha1lowingyooddsandmitaida’s，betweenkigasuru／kanz卿uru／ケOkang71Suruandtheverbphrases asinAme−gtihut−teiru−yOOna／mitaina／？？rasii−k函uru／kanzlgtMuruルokangtiSuru．Thiscontraststrongly SuggeStSthatrasiiis，butyoodaandmitaidaarenot，pureeVidentialsofsensoryperception．Bothrasii. andkigtiSuru／kanzlgtiSuruルokangtiSuruarebothpureevidentialsinparadigmaticrelationships；their. COnCurrenCeprOducessomeredundancyandcontradiction．Incontrast，yOOdaandmitaidahaveanothe. functionofindicatingthatthespeakercomparesoranal0gizeshiscognitiontothereferentoftheexpression theyconclude．Notethat．．．noyooddor．．．miiaiddtypica11ytranslatesintoEnglishlike“islike．．．． Japaneseevidentialsofjusti五cationtypeseeminglydonotservetoexpressthestrengthofjusti五cation． Infact，yOOda，mitaida，andrasiidonotmakeabigdifferenceinthestrengthofjustificationimpliedin Ame−gahuru−yOOda／mitaida／7Wii‘Itwillrain／Itlookslikerain：InJapanese，referentialevidentialsfunction asexpressionsforthestrengthofjustificationaswi11bedealtwithinsection4．. 3．2．Ainuevidentialsofthejustificationtype. Tothebestofmyknowledge，Nakagawa（1995）istheonlyexplicitattempttointroducetheconcep OfevidentialityintoAinu．Heidentifiesfourkindsofevidentials：ruWene／an，Sirine／an，hawene／an，and. humine／an／unintheChitosedialect（asouthwesternHokkaidodialect）asin（3a−d）．Hisexplanationfor thedifferencesamongthesentencesin（3）isthatthespeakertreatsthefineweatheras“anindubitable fact’’in（3a），COmeStOreCOgnitionoftheweatherbywayof“visualinformationlikebrightnessoutside thewindow’’in（3b）andbywayof“otherperceptionslikenonlinguisticsounds，nOise，taSte，Smell，tOuCh， feeling，Orinspiration’’in（3d），andis“informed’’oftheweatherin（3c）（ibid．：13）．. （3）a．肋血・壷〆姑”胴Ⅳ那． todayit：is：fineEVD ‘Itisfinetoday．’. 5 Thechoice betweenotogusuruandhoegasuruhingesonwhetherhumansareinvolvedasmajorparticipantsinthe COgnizedeventorstate；？？Ame−guhut−teinL−hoegzISurubut：（i）オブα和ぶO才一おオγ〟一如聯ぶ〟γ〟／．Pof聯ぶ〟γ〟．. quarrel−PROG−EVD ‘Itsoundslikequarrelingvoices／？sounds．’ Furthermore，aZlg提SuruOnlyhasamarginaluseasanevidential．（ii）助成一刀∂ 乃α点β−g視点オ柁−ねぎγ〟βZ如∫〟γ〟． mouth−GENinside−NOM Cut−PROG EVD. ‘Ittasteslikeitisbleedinginmymouth：. 128.

(8) 日本語とアイヌ語の証拠性. l−．／J／JJ／り ∫可・J／●山／∫J／●J仙． ‘Itlookslikeitisfinetoday．’ l−．／J／JJ／り正小／・山／〟J川＝J■． ‘Ihearitisfinetoday．’ り．／′／JJ／りヾJ小JJ・ん／山／川J川． ‘Itseemsitisfinetoday．’. Whilesirineandhumineareclassifiableintothesensoryevidential，hawenefallsunderthelinguistic evidential．Ontheotherhand，ruWenemightbecategorizedintotheinferentialevidential．Thiswillbe. takenupagaininsection4becauseitservesmainlytoconveythestrengthofjustification． IcanfindmanymoreformswithsimilarfunctionsinmyowndatabaseoftheAsahikawadialect（a. northeasternHokkaidodialect），theAsahikawaAinuCo71）uS．Siran，Sirne，Sirki，hawan，hawas，haw 山川JJ川．山川Jノバ．山川川．．∼＝／∫汗J／JJ．〟り／り川∫／〃／JJ．∼＝／∫／／・揖．山／仙∴J／・揖．〟り／り川〟〟J＝／∫．∫．〟り／・〟〟J＝／∫．〟り／り川J／JJ．. hinekotomanasexemplifiedin（4）．. （4）a．‥．g＝戊才. 5才和〃〃α．（0208285）. you＝do．it EVD ‘Itseemsyoudidit．’ b．g〃g g〃＝血0ざ点g紺g〃ざ才γ〃g（1109107）. thus me＝teaSe. EVD. ‘（She）seemedtobeteasingthatway． C．正邪・i＝0γ少α点. ざグ戎グ γ〟紺β〃β（0104701）. rea11y me＝be．reserved．towardlてVO lてVO. ‘（They）lookreservedtowardme． d．〃g カブ〃g αゐ〟〃ゐα紺α〃（1105505）. that place to enter EVD ‘Ihear（thefootsteps）1edintothere． e．点α∽り′〟わγ才ヲg ゐα紺が αγ才 Jり′乃〟＝α乃（0210222）. god pL me＝Say EVD COMP think＝I ‘IthinkIheardgodste11ittome． f．50〃〃0 α〃 ♪g 〃g ゐα紺〃g（0106014）. rea11y exist thing be EVD ‘Iwasinformed（that）istrue． g．〃ゆ〟ゐり′ ゐ〟∽α玖（1102703）. whatburn EVD ‘Itsme11slikesomethingisburning． h．γg和グ砂丘g ゐ〟∽が（0716915）. Wind strong EVD ‘Itsoundslikeitisblowinghard．’. 129.

(9) 井筒勝信 ⊥．d扇お＝α〃ゐ〟∽〃g 紺α…．（0103901）. dream＝I EVD and ‘IfeltthatIhadhada dream and．．．． j．点0お〃〆祓α 紺α 5才和〃〃α（0308017）. Villagegood cp EVD AP ‘Thevillagelookedwonderfu1．’ k．∽α紺fり′ α 点0わ∽ ざ才和〃（1105316）. die. pERF CP. EVD. ‘Itseemsthat（he）wasalreadydead．’ l．〃JhJJJ／JJ山J∫JJ■〃■J／JJ山／J／〟／J／／イりん小・J川 ∫／／・揖 ∫■〟り／・山J∴l叩ソ川・＝け出いリ. how be for so be place at Ainulive cp EVD COMP I＝think ‘Iwonderedwhy，aSIcouldsee，Ainulivedinsuchabadplace． m．〃gα〃〃り′ 〟5. 点α〃g ざ才祓才（0206301）. that fire disappear cp EVD ‘Thatfireseemstohavegoneout．’ n．βγ坤gγgγg，g〟∽α紺〃g 点βわ∽ ゐα紺が（1114613） roar. die. cp. EVD. ‘（Thebear）roared．Itsoundedlikeitdied．’ 0．飯川イ（戎扉α．馴＝f〟和点〟〃才ざg血γ ゐα紺が（1113907）. war to pASS＝take should coMP EVD ‘Iwasinformedthat（myson）wasgoingtowar． p．〃β ゐ〟Cグ. ノわざ少グ，∽0点0γ 点め∽α〃．（1109206）. that old．woman go．back go．to．bedlてVO ‘Thatoldwomanseemedtohavegonehomeandslept．’ q．が砂カブ〃g 加わ∽α〃（0104201）. go．out CP EVD ‘（They）seemtohavegoneout．’. Nakagawa’sexplanationbasicallyappliestomostoftheexamplesin（4）．Si7un，Sirne，andsirkiin（4a−C） and（4j−m）indicatethepresenceofthevisualjustificationforthespeaker’scognitioncodedintheir. precedingclauses．Hawan，hawas，andhawnein（4d−f）and（4n−0）areinterpretableasdesignatingthe Speakerbeinginformedoftheeventorstatecodedintheirprecedingclauses．Human，humas，and humnein（4g−i）referstothesensoryperceptionsotherthanvisualoneswhichareresponsibleforthe Speaker’scognitioncodedintheclausesprecedingtheevidentials．. Atleasttwopointsremaintobeexplained，however．Thefirstisconcernedwiththemorphological1y Separatedevidentialkotoman．Itservesasasensoryevidential，indicatingthatthespeakerarrivesathis expressedcognitionbywayofsomesensoryperceptionasin（4p−q）．ItdifEersfromtheotherevidentials Ofthesametypeinnotspecifyingthekindofthesensoryperception．Thisnatureisattributabletothe. factthatitlackselementslikesir，haw，andhum．Theseelementsrespectivelyimplythevisualperception， 1inguisticinformation，andnon−Visualperception．Thus，kotomanmaybefunctionallyclosetoJapanese. 130.

(10) 日本語とアイヌ語の証拠性. yooda，mitada，andrasii．Infact，ittranslatesintothemeasilyinJapanesetranslationsofAinutexts．The Othersensoryevidentials，bycontrast，donotoftentranslateintoJapanesefelicitously．. Thesecondpointisthefactthatthekindofjustificationindicatedbyhawan，hawas，andhawneis notnecessarilyl．INGUISTIC．Itcanbesomeothervoiceorsoundproducedbycreatures．In（4n），forexample， thebeargavearoarwhenitwasshotandthisroarletthespeakerknowitsdeath．Tnthiscontext，the Speakerisnotlinguisticallyinformedofthebear’sdeath．Hawasisalsousedwhenthejustificationis. birdssinging．Insomeexamples，therefore，hawan，hawas，andhawnemayfa11underthesensoryevidential ThepurelinguisticevidentialmightbeyakLV，eaSin（5a−b），althoughitsstatusisnottota11yclear．It isclearlyanalyzableasyakanヮe（coMP−PASS＝Say）‘besaidthat’andthuscomparabletoJapanesetoiu， analyzableastoiu（coMP−Say）．Butthereisalsoaclearcontrast．Interestinglyenough，3ukLV，eCanfollow anotherevidentialoflinguisticjustification，hawas，aSin（5b）．InJapanese，WeCanhardlyutterasentence like（5c）．. （5）a ‥郎魚卵′ ノ払＝才お点〕叱虎砂g．（1102516）. COMP I＝Say EVD ‘IhearIsaid so：. b…．ねα椚. ∫g点0γ ゐα紺がプ成の′g（1113903）. be．absent coMP EVD EVD ‘Iheartheysaidtherehadneverbeen（suchaperson）．’ C．？？ざ0〃〃α カグわー紺α 才〃α点. α〃α ざ00dα わ才〟（1113903）. SuCh person−TOP be．absent pAST EVD EVD ‘Iheartheysaidtherehadneverbeen（suchaperson）．’. 4．Strengthofjustification Evidentialsforstrengthofjusti五cationoftencoincidewithmarkersofepistemicmodality（seeexamples inChafe（1986）andAnderson（1986））．Thenumberofsuchevidentialsismuchsmallerthanthatof evidentialsofjusti五cationkindsinbothJapaneseandAinu．Asexemplifiedin（6a），Japanesehasnitbinai， nikimatteru（colloquial），hazuda，daroo，andkamosirenaiindescendingorderoftheirimpliedstrengthof justification．Althoughmikomida，mitoosida，andsoodaasin（6b−C）primarilydenotethepresenceof inferentialjustificationforthespeaker’scognition，theyalsosuggestacertainstrengthofjustification，. Whichapproximatestothatassociatedwithddroo．Sooddasinhurisoodaisremarkableinstronglyevoking. theimminenceoftheeventexpressed（Aoki1986：232）．・■；＝．l川い㌔／〟〟／−／■／〃JJJ／／料／／JJJ／／〟〟ニ仙ん／・ん／／・…′山／川り∫／ハ〃J／／． rain−nOmfa11−PROG EVD ‘Itmust／should／will／mLV，be7uimngl’ l）．l川い㌔／〟〟／てJ川砧りJJ／′ん／川／／＝心′ん／．. rain−NOMfa11 EVD ‘Itisexpectedtorain．. 131.

(11) 井筒勝信. l−．l川い㌔／〟〟／●J∫…′・／J／．. rain−NOMfa11 EVD ‘Itislikelytorain．. Ontheotherhand，Ainuhas俺wevidentialsforstrengthofjustification．Ttisdifficulttoidentifymarkers. ofepistemicmodalityinAinu．Ndnkorasin（7a−C）couldbethebestcandidate，butsuchauseisfarfrom beingtypical．Thiswordismostfrequentlyusedtomarkacertaintypeofimperativelike（7d）．Though veryrarelyusedforstrengthofjustificationasin（7e），kusneisdominantlyusedasamarkerofdeontic modalityasin（7f）．. （7）a．〃才物′ α〃〃0〃血γ 点〟ざ〟…．（0103712）. woods be lてVO because ‘Becausetherewillbewoods（there）．．．． b．お砂g ♪αfg点ざ0∽0 〃g 〃α〃点0γ（0206177）. this only not be EVD ‘（Whatyouhavetodo）shouldnotbeonlythis： C．点〟＝ノわ∫砂g γ〟∫り′ ゐα紺乃g 乃α乃点0γ∫g点0γgぬ点．（1107213）. I＝gO．back want．to EVD EVD CP Say ‘（She）saidmyvoicewouldmeanthatIwantedtogohome． d．g＝如ざ砂才. 〃α〃血γ．（0105709）. you＝gO．back DEON ‘Go home．. e．5り′ α舶αγg 紺g〃 0舶の′O g＝点0γ 点〟ざ〃g 〃仇（1116210）. again than bad man you＝have EVD AP ‘Youwillgetmarriedtoaworsemanagain． f．〟点0γ＝α〃点∽〃g 〃α…．（0526506）. get．married DEON AP ‘Weshal1getmarried／Let’sgetmarried．’. RuwenemightalsobecountedasanevidentialforstrengthofjustificationasImentionedinthe beginningofsection3．2．Nakagawa（1995：13）describesitasindicatingthatthecognitioncodedinits precedingclauseas“anindubitablefact．’’Asfarasonefollowsthisdescription，ruWeneWillbeanin鎚rential evidentialforthestrongestdegreeofjustification．. 5．AComparison：JapaneseandAinuevidentialitysystems. TheJapaneseandAinuevidentialsobservedintheabovesectionscanbesummarizedasTablel．. Theyarelargelydividedintothreetypes：SenSOry，1inguistic，andinferential．Somesensoryevidential SpeCifythekindofrelevantperception．Theyareboldfacedinthetable．WhileJapanesehasanevidential. Ofolfactoryperception（nioi9asuru）andnotoneofvisualperception，Ainuhasthelatter（sir（i）an，. 132.

(12) 日本語とアイヌ語の証拠性. sir付ne，Sir付ki）andnottheformer．. Theevidentialsforstrengthofjustificationareassumedtofa11withintheinferentialevidential．They arespecialmembersthatputmoreemphasisonthestrengththanthekindofjustification．Suchmembers areunderscoredinthetable．Itiseasytorecognizetheorderofthosemembersonthescaleofstrength OfjustificationinJapanese：nit如inai，nikimatteiru＞hazuda＞daroo＞kamosirenai．Nosuchorderis. ObservableamongAinuinferentialevidentials，Whosestatusasexpressionsforthestrengthofjustification isitselfunclear．. SENSORY. LINGUISTIC. JAPANESE. AINU. ．l・りり（山．ノ扇／〝吊れγ侶亙. β汗（恥〝，扇イ功闇，誼㈹軋. ．l・りりJJ／（山．ム〝JJこ∫（山．. ゐ〟∽何α〃，ゐ〟∽何が，ゐ〟∽何〃g，. 点gゐαf血，ゐ〟〃戊f血，. 鳥α∽（りα〝，鳥α∽（りαち鳥α∽在ノ〝e，. 0わgαβ〃rlら〝”好αβ〃rlら. 肌Ⅷ軒如叫伽血肝油相即. 鳥oegαβ〃rll，点gゐαな㍑ざ〟γ〟，. 肌用叫御紘鳥肌＝由相場. ム如才／J・軋ん川こ加JJ／J・仇. 点∂わ∽ゐαび匝ノが，ノわわ∽α〃，. ．l・りム〝JJ辟JJJ／川. 〟∫JJ‘一々り／川JJ〝〃. Jりり（山．／／（−．／扉仇／り∫Jト. 〟〝J丹（リ〝仇力〟両日肌力〝J丹（リJJ（−．. 〟〝〃〟品血J川−〟∫肌山ムり／り（山. ムり／川〃力〝J（イり〝J． J（一片りJ・〟〝J亘り〝J．．l・〝叫l・（−. INFERENTIAL nitigamai，nikimatteiru，〟〝こ∼／（山，（山川り，. （竺些些竺竺？），（〃α〃点βγ？）（点〟ぶ侮ノ〃g7）. ム〟川侶江川〝∫．ノ山々りノ扇（山．川∫／りりJ∫（山．Jりり（山. TABLEl：JAPANESEANDAINUEVIDENTIALS. AsfarasTablelisconcerned，thetwolanguagesdonothavegreatlydifferentsystemsofevidentiality． Mostexpressionsforsensoryandlinguisticevidentialityarecomprisedofaformalnoun b，00，mitai，SOO， etc．inJapaneseandsira），huma），hawk）etc．inAinu）andabe−typeVerb（JapanesedaandAinuneand. an）ordo−typeformalverb（JapanesesuruandAinuki）．Apparentcontrastsbetweenthetwosystems COnCernthekindsandstrengthofjustificationcodableinevidentials．Morekindsofsensoryjustification （visual，auditory，andolfactoryperception）andmoredegreesofstrengthofjustification（high，medial， andlow）appeartobedifferentiableinJapanese．6Inthissense，aCCOrdingly，Japanesewouldhavea. moreelaborateevidentialitysystem，WhichseemscontrarytothegeneralassumptionthatAinuhasa richerevidentialitysystem． Thisapparentcontradictionwillbeaccountedforasfollows．Ihavesofarassumedthatthe evidentialsforstrengthofjustificationfallwithintheinferentialevidential．Theyarespecialmembers thatsuggestthestrengthofthespeaker’spresupposedinferentialjustificationmorestronglythanother membersdo．However，aSimilarassumptionmayapplytoAinusensoryevidentials：SOmeOfthemmay. 6 JapaneseappearstolackevidentialsparticulartovisualperceptionlikeAinusirn）an，Sirn）ne，andsirn）ki．Butsucha kindofjustificationisexpressiblewithyoonimieru，Whichmaynotbeapureevidentialbecauseyooniobligatorilycomes beforemieru．Ⅰ句onihikoeruisalsoasimilarpseudo−eVidentialofauditoryperception．. 133.

(13) 井筒勝信. SuggeStthestrengthofthespeaker’spresupposedsensoryjustificationmorestronglythanothermembers，. establishingtheorderofthestrengthlikethatobtainsamongtheJapaneseinferentialevidentialsni物αinai， JJ／山川J／／／．／／てJ．〟〟ニ仙／J／．′ん／／・＝′．川1（川J／川り∫／ハ〃J／／．. Thishypothesiswi11beborneoutpartlybythefollowingexamples．Sincethespeakercanbeinformed aboutthecognitiontheotherpersonarrivesatbywayofvisual（e．g．，Sirne）ornon−Visual（e．g．，hum） perceptionsandbywayofinferencebasedonsomestrongjustification（ruwene），thesequenceofsi OrruWenandhawanearepossibleasin（8a）and（8d）．Bycontrast，thesequenceofhuma）an／as／neand ruweneisnotattested，Whilethatofsira）an／ne／kiandruweneisobservableasin（8b−C）．Moreover，the sequenceofhawk）an／as／ne（asasensoryevidential）andruweneisalsounattested．Itiswellknownthat. expressionsinsomeparadigmaticrelationdonotconcurwitheachother．Thus，WeCanhypothesizetha huma）an／as／ne，hawk）an／as／ne（sensory），andruwene，nOtappearinginsequence，difEerfromoneanother intheirsuggestedstrengthofjustification，eStablishingtheordersimilartotheJapanesecounterpart： JJ／／料／／JJJ／いノ／仙川／／／■／〃J・〟〟ニ仙ん／・・ん／川り・ん／川り∫／ハ〃J／／．. ・バ；l．〟〟J小J／JJJ／〟 ■∫J／川J／JJ／†川JJ／JJ／■ 山J∫JJJ〟血バ ∫／川■ 〟〟J…JJ. dad top otter hunt. to walk EVD EVD. （1102503）. ‘Iheardadwasseentogohuntingotters． b．ダブⅥ∽∽0舶α＝α〃 5才γ〃g γ〟紺g〃g 点〟5払‥．（0526418）. because. play．a．trick＝I EVD EVD. ‘becauseyouknowIseemtohaveplayedatrick．．．． C．正邪・i＝0γ少α点. ざグ戎グ γ〟紺β〃β（0104701）（4c）. rea11y me＝be．reserved．towardlてVOlてVO ‘（They）lookreservedtowardme． d．別働川南＝α〃 α γ〟紺g〃g ゐα紺α〃．（0206130）. feel．pain＝I pERF EVD. EVD. ‘IwasinformedthatIhadsufEeredanacutepain．. （9）＊5才γ付α〃／〃g／鬼才ゐ〟∽似α〃／が／〃ピー山川J√ル／JJノバ仙 ∫／／・√／加JJ仙山. 一／・〟〃・．〃．山川J√ハJ〃JJバ仙 −／・〟〃・．〃．扉／・√ル／JJ仙山. 一山J川両J〃JJバ仙JⅥJ（・．〃．一山／∼抽〟／〃Jバ仙／て川・■仙. Thehypothesizedorderwillappriximateto：ruWene（＞kusne）＞haw存）an／as／ne持enso7つり（＞ nankor）＞huma）an／as／ne．AinusensoryevidentialsaremorefrequentlyusedthanJapanesesensory evidentialsare．ThisaccountsforthegeneralassumptionofthericherevidentialitysysteminAinu．The. frequentusesofAinusensoryevidentialswillbeanaturalconsequence，ifsomeoftheevidentials. （haw存）an／as／ne（sensory）andhum（i）an／as／ne）servetocodethestrengthalongsidethekindof. 134.

(14) 日本語とアイヌ語の証拠性. justi五cation．Ainumayseemtocompensatethelackorshortageofinferentialevidentialswithsomesensory. evidentialswhichputmoreemphasisonthestrengthofjustification．InJapanese，SOmeinferential evidentialsarechosentoexpressthestrengthofjustification，butinAinu，SOmeSenSOryeVidentialsmay servethatfunction．. Tfwelookatthe Ainuevidentialitysystemthisway，Japanesecan hardlybeseenashavingamore elaboratesystemofevidentiality．AlthoughJapanesemaydistinguishmorekindsofsensoryjustification （visual，auditory，andolfactoryperception），theevidentialsforthekindsofjustificationaremuchless frequentlyusedthanAinucomparableevidentials．Differentdegreesofstrengthofjustification（high， medial，andlow）areexpressedwithsomeinferentialevidentialsinJapanese；inAinu，theymaynotbe. inexpressiblebutbecodedinsomeofsensoryevidentials．. 6．Conclusion Chafe（1986：261）arguesthatthedifferencebetweenCalifornianIndianlanguagesandEnglish“is notamatterofevidentialsvs．noevidentials’’butispartly“aquestionofhowevidentialityisgrammatica11y expressed：isitbystlfEixes，auXiliaries，adverbs，OrWhat？’’Thedifferenceispartly“aquestionofwhatkinds Ofevidentialityareexpressedmostreadily：aredegreesofreliability，Orinference，OrSenSOryeVidence，Or hearsay，forexample，eSpeCia11ycodable？’’HeremarksthatthedifEerencealsoliespartlyin“aquestionof. Whichevidentialsaremostoftenused；differentlanguagesfocusondifferentkindsofevidentialitymore orlessofthetime．. Asfarasthepresentstudyshows，theevidentialitysystemsofJapaneseandAinudonotdiffer greatly．Bothlanguageshavearichrepertoireofevidentialdevices：SenSOry，1inguistic，andinferential evidentials．Theymayonlybeslightlydifferentinhowevidentialityisgrammaticallyexpressed，What kindsofjustificationareexpressedmostreadily，andwhicharemostoftenused．AlthoughJapanesemay distinguishmorekindsofsensoryjustification（visual，auditory，andolfactoryperception），theevidentials. forthekindsofjustificationaremuchlessfrequentlyusedthanAinucomparableevidentials．Different. degreesofstrengthofjustification（high，medial，andlow）areexpressedwithsomeinferential. evidentialsinJapanesebutseemtobesuggestedbysomesensoryevidentialsinAinu．Whilethe StrengthofjustificationiscodedmoreclearlyinJapanese，thekindofjustificationiscodeddominantlyin Ainu．. ThetwosystemscanbediagramedasinFigurel．Thearrowsrepresenteachtypeofjustification，. theirthicknessthedominanceofthejustificationtype，andtheleftandrightballoonslinguisticinformation andinference．Asseeninthediagram，thedifferencebetweenthetwoevidentialitysystemslieslargely inthethicknessratherthantheconfigurationofarrows，i．e．，inthedominanceratherthanthepresence. andabsenceoftherelevanttypesofjustification．. 135.

(15) 井筒勝信. JAPANIiSIi. AINIl. FIGUREl：JAPANESEANDAINUEVIDENTIALS. KEYTO ABBREVIATIONS COMP：COmplementizer；CP：COnneCtiveparticle；DEON：deonticmodality；EVD：eVidentiality；NOM：nOminative；PASS：paSSive； PAST：preterit：PERF：perfect：PROG：prOgreSSive：TOP：tOpic. REFERENCES Aikhenvald，AlexandraY．2003．EvidentialityinTypologicalPerspective．InAlexandraY．AikhenvaldandR．M．W．Dixon （eds．），StudiesinEvidenthlli＆．Amsterdam／Philadelphia：JohnBenjaminsPublishingCompany． Anderson，LloydB．1986．Evidentials，PathsofChange，andMentalMaps：TypologicallyRegularAsymmetries．InChafe andNichoIs1986，pp．273−312． Aoki，Haruo．1986．EvidentialsinJapanese．InChafeandNichoIs1986，pp．223−238．. Boas，F．1938．Language．InF．Boas（ed．），Gene7711Anthr坤O10g）′，pp．124−145．Boston，NewYork：D．C．HeathandC Chafe，Wallace．1986．EvidentialityinEnglishConversationandAcademicWriting．InChafeandNichoIs1986，261−272． Chafe，WallaceandJohannaNichoIs（eds．）．1986．Evidentiality：TheLinguisticCodingqfEpistemology．Norwood，New Jersey：AblexPublishingCompany． Izutsu，Katsunobu．2003．AinugonoShokosei（EvidentialityinAinu）．InIzutsu（ed．），Ai71ugOAsahikau，aHqge71K呼asu71i MbtozukuBu74）OShoH云nsanno7bmenoKisokenkyu（APreliminaryStudyforGrammarBookCompilationBasedonthe AsahikawaAinuCorpus）．HokkaidoUniversityofEducation，Asahikawa． Jakobson，RomanO．1957／1995．ShiftersandVerbalCategories．OnLanguage，T（pp．386−392．Massachusetts：Harvard. UniversityPress． Nakagawa，Hiroshi．1995．TheAinu−JapaneseDictionary：ChitoseDialect．Tokyo：Sofukan． Okuda，Osami．1995．ConjunctionalizersintheShizunaiDialectofAinu．TheBulletinjbrHokkaidoAinuCultureResearch Cg〃おγ1：139−159．. Okuda，Osami．1997．AdverbialparticlesandSentenceFinalParticlesintheShizunaiDialectofAinu．77uzBulletinjbrHokkaido 』滋∽C〟Jf〟7でj？ゼぶゼαγCゐCg〃おγ3：195−214．. Katsunobu Izutsu idutsu＠asa．hokkyodai．ac．jp （旭川校助教授）. 136.

(16)