Title On the frontness opposition in loanwords in Old Uygur

Author(s) Erdal, Marcel

Citation 内陸アジア言語の研究. 17 P.3-P.24

Issue Date 2002-09

Text Version publisher

URL http://hdl.handle.net/11094/18240

DOI

Rights

Osaka University
 Onthe肋ntnessoPPositionin董oanwords

inOldUygur

MarcelErdal

Inthispaperwetrytocomeabitclosertoadetemlinationofhowfbreignwordscopied

intoOldTurkicmayhavebeenpronounced.AfterProf6ssorSh6gaitolaidsolidfbunda-

tionsfb士thisdomainofresearchin1978,itwassubsequentlydealtwithbyanumberof

scholars.11talsoisoneofthefewmattersaboutwhichscholarsworkingonthislan-

guageappeartodisagreesharply,andtranscriptionsoftextsshowthegreatestdiffer-

encesandfluctuationsinthisparticulardomain.Itwouldtherefbreseemusefultoat-

temptaclari且cationofsomeaspectsofthematteratthisstage,especiallyaftermostof

theUygursourcesinBr査hmlhavealreadybeenedited.

Whenlexemesormorphemesarecopiedfromonelanguageintoanother,the

copiesfindthemselvesimmersedinanewphonologicalsystem.Tounderstandwhat

happensduringsuchborrowingswehavetocontrast
、thesystem-inourcasemainlythe

vowelsystem-ofthesource正anguagewiththatofthetargetIanguage,Systematic

frontnessoppositionswerefbreiεntoalllanguageswithwhichりygurwasincontact

(withtheexception-marginalfbrmuchofthecorpus-ofMongolic).TheUygurs

appeartohaveintroducedthemintocopiedlexicalmatehaltovaryingdeg爬es,andin

verycomplicatedconfigurations.

OldUygurborrowedasignificantpartofitslexiconfめmcontactlanguages.

SincemostofourUygurcorpusconsistsoftextsofBuddhistcontent,thegreatm瓠ority

1SeeespeciallySh6gaito1978,M㏄rloose1980,Laut1986:89-148,R6hrbom1988Johanson

1991:83-85,Erdall991:16f,R6hrbom1996.

(3)
ofsuchelementsultimate豆ycameffomSanskrit,theculturelanguageofp叩 一andnon-

IslamicSouthAsia.Wehaveaverygood㎞owledgeofthevowelsystemofSanskrit,

whichwasalsooneofthesourcesfbrdirectlexicalcopiesinOldTurkic.2However,

mostoftheIndicte㎜sgotintoUygurthroughChinese,TokharianorIranianlan-

guages,3inwhichcaseitistheirsoundsystemswhichbecomerelevantfbrロnderstand-

ingtheshapeswhichthesetemlsacquire.Whatweknowofthevowelsystemsof

SogdianandcontemporaryChinese,whichwereamongthemostimportantdirect

sourcelanguagesfbrUygur,isquitesketchy,butthevowelsystemsofthetwovarieties

ofTokharianarereasonablyclear.OurprimaryquestionconcemingTokharianisthe

phoneticstatusofthecharactertranschbedas∂anddescdbedasa5んwαvowel;the

rephonologisationofthisvowelinwordstakenupbyUygurwillhavetobesystemati-

callydealtwith.TheSemiticalphabets,inwhichUygurismostlywritten,areoftenfar

廿omgivingafロllpictureofthevowels.Thecentralconcemofthep爬sentpaper,the

distinctionbetween/i/andIV,isnotatalladdressedbythosew践tingsystems.Norisit

generallyaddressedbytheBr亘hmlsystemasusedfbrUygur(asdetailedbelow),though

Tibetanscript,theotherIndicwritingsysteminwhichUyguriSwritten,doesappearto

makethisparticulardistinctiontoacertaindegree.4

1willhereassumethatphonemicoppositionsdidindeedexistinUygurbetween

2 OrkhonTurkicalreadyhasIndicloanssuchas∫5vαrαor〃taxara6.DirecIborrowingfrom

SanskntcameintofashioninthelateststageofBuddhistUygurscripture.

3 Theactualsourcefbrsuchtemsintheselanguagesiso血ennotSans㎞tbutoneoftheP颪krits

orMiddleIndo-Aryanlanguages;the`directcontactlanguages'ofUygurwere,moreover,also

incontactwitheachother.

4 BuuhentheusewhichUygursmadeofthisscriptisespecia11yinconsistent,besidethefactthat

itsuseforthislanguageissolimited.TheBr五hmims,editedasTI"VIIIIdiffersfromtheothers

inwridng御(aswellas/b/)asEwhereastheIetterEinotherBrdhmimss.representsthevoweI

le/bothinTurkicandbOrrowedwords.InthisitshowsthesamepracticeasBuddhKat,whichis

inTibeセmwriting.Unfb血nately,thisisamedicaltextwhichhasnothingtoofferconceming
ゆ
ourtoplc.

(4)
/Vand〃ontheonehand,andbetween〃andla!ontheother,thoughbothofthese

oppositionsmight,undercertaincircumstances,getneutralised(asphonemicoppQsi-

tionsinalllanguagessometimesdo).5Evenwhenthewritingd㏄snotgivedirectinfbr-

mationonwhetherweshouldread'orちweshouldtherefbrestillbeIbokingfbrclueson

whichleadingtoP爬fbr.

`Signa貰,letters

WhenwritingOttomanTu面sh,certainArabiccharactersfUnctiohedas`signal'letters

showingwhetherthesyllablesinwhichtheyappearedbelongedtotheffontortheback

sehes;Iamhe爬repeatingatermusedbyAndreasTietzeinanunpublishedlectureas

quotedbyR6hrbom1996:176.TheconsonantpairusedmostconsistentlyinOttoman

fbrthispurposewas .gげasasignalfbrbacksyllables,た 可servingasasignalfbrf沁nt

syllables.This`signalling'isfbundalreadyintheUyguralphabet,whichusesthetwo

SemiticcharactersXandKfbrdistinguishingbetweenback-ha㎜onyandffont-har-

monystringsi血odginallyTurkicwords,X6is,intheUyguralphabet,usedalsofbr

representing/glinback-hamlonystrings7butnodistinctionismadebetween/klahdlg!

inTurkic食ont-harmonywords.

DealingwiththeUygurconespondencesofSans㎞tた,8,ぬ,肋and8ぬ,Sh6g樋to

l978:96statedthatXoftheUygur3criptcorrespondsonlytoSanskrithwhiletheother

fburSanskritconsonantstogetherreceiveadiff￠rentrepresentation.R6hrbom1988then

5 IwilldealwiththequestionwhetherOldTurkicdistinguishedbetween'and乙andwhatdle

natureofthisdislinctionmayhavebeeninitsdiffbrentsubcorpuses,inadiff6rentpaperIam

P㈹panngatp爬sent・

6 scholarshavegeneral韮yusedQfbrrefb㎡ngtothischaracterbutlwillrefbrtoitasxbecause

itgoesbacktoSemiticoぬ θ魏(and8伽 θ',alsopronouncedasfhcativeinSogdian,f士omwhere

重hescripIgoεi飢oUygur).

7 Inback-hamonywordsthedistinctionbetween瓜1andlglisshown-thoughnotquite

consistently-bydouble-dottingtheXwhenit【epresents/k/.Thesoundκoftengetsasingle-

dotdiacritic.

(5)
pointedoutthatthespelliρgsystemappliedtostemsofIndianoriginisdifferent丘om

thatusedfbrTurkicelementsbutidenticaltothesystemappliedtowordsofIndian

odginwhenwhtingSogdian(f士omwhereitwasno-doubtcopied):Whetherthematdx

languagewasSogdianorUygur,wordshavingtheconsonantsた,肋,80r8ぬinSans㎞t

orPrakritwereinthe㈹ceivinglanguagespelledwithK,whethertheadjacent(origi-

nally)Indicvowelswereα,o,配(classifiedasbackinTurkic)or1,6(assignedtothe

Turkicffontseries).81fanelementinawordgoesbacktoIndichandisfbllowedby'or

ε,itisalsospelledasKinUygurwriting.TheletterXisinborrowedIndicelements

usedonlyfbrIndichfbllowedbytheothervowels,α,oor灘.9Stopsandfhcativesare

distinguished,then,butnotvoi㏄dvelars丘omunvoicedones.Allthiswouldmeanthat,

asfaraselementscopiedffomIndicinUygurareconcemed,Xdoeshave`signal'value

(sinceittumsuponlyiffbllowedbyα,oorのbutKdoesnot(asitcanbefbllowedby

anyvowel).且oSans㎞ 腕isnotallthatcommon,ofcourse,sothaトallinall-the`signal'

valueofvelarletterswouldinfactbequitelimitedinsuchborrowi!lgs(wheretheir

8 AsmallselecdonofsuchinstancescanbefbundinR6hrbom1988:236,Counte卜examples,X

usedinstemsofIndicorigintorepresent[k],arequitefew.AllthreeinstanceswithX

mentionedinR6hrbom1996:184arequotedfromtheSt.PetersburgSuvms..In1α κ諏 π`body

mark',oneamongthese,theletterprobablyrepresents[x]andnot[k]:Thereissolidevidence

thaUhestring!k§!waspronouncedas[x蓉]inUygurwordsandthispronunciationappearsto

havegottenextendedIoaboπowinginthiscase.AnotherinstanceappearsinBuyanAvirrntik,
の
ロ
apassageaddedtothetextofthisms.inlhe17thcentuly,whichshouldbynomeansbetakenas

evidencefbrOldUyguralthoughitmanagestoapeIhatlanguagetoalargeextent。TherealSuv

ロ
の
textcanonlybeestablishedbyacriticaleditiontakingallthenumerousoldermss.(mostlyin

Berlin)intoconsiderationandwouldexcludeBuyanAvirmtik.Thephenomenonistherefore

marginalatbest.

9 AfbwexamplesarementionedinR6hrbom1988:240.InSogdiantheletterisalsoexclusively

usedfortherepresentationoffficatives;kethisapharyngeal丘icativewhichinSemitic

languagesattractsvowelstothebackofIheoralcavity,Cf.,ontheotherha玲d,tamηaq<Tib.

8ぬ 〃rs-Oag`instruction'withXattheendofaforeignword。

10 ThepresenceofXas刀 財istdggeredonlyifthevowelfbllowingitislaou1;thetriggering

vowelisthesubsequentone,aswiththeallophonesofModemGreeklx/;nouheonepreceding

th6frica丘ve,aswithGeman[x-91.

(6)
eviden㏄woロldhavebeenmostneededfbrestablishing亡hepronunciation).Thisd㏄s

notapplytosuffixesaddedtotheseIexemes:Therethesignallingsystemshowsthe

no㎜alTuddgvowel+vel訂dichotomy,wi血+gαvs,+妨,+吻1'vs.+磁1,吻or+吻

vs.+倣or+1磁.Dea豆ing.withthespellingofwordsofIndo-AryanorigininUygur,

R6hrboml996:177writes:。WasdieKonsonantenbetrif肚,sokannnurdasVol宝ommen

vonsogenannten,Signalzeichen'alssichererHinweisaufVelaritat(SignalzeichenQll)

bzw.Palatali蹴(SignalzeichenK)einerSilbegelten.``However,contrarytothisstate-

ment,theselettershaveno`signarstatusasfarasstemsofsuchoriginareconcemed:

WhenusingUygurscゆtf6rthesewords,theuseofKisnotautomaticallyconnected

withthehamlonyclassofasyllablewithinwhichitisused。Thereistherefbrenothing

iπegularornon-h㎜onicabouuhespellingSLWKWX(fbunde.g.inHTsVIIIl924)

fbr`stanza(accusative)'withalexemeultimatelycomingf士omSkt.510ん α:TheKonly

showsthatthevelarisastopandnotafdcative.Weknowthat,inTurkish,consonants

suchaslk1,1110reven〃havebackandffontvariantsandthattheirpronunciationalone

can,iftheyappearattheendofastem,dete㎜inethehamlonyclassofsuffixesaddedto

it;thusinTurkish∫ αα∫+eand加'+46,12ReadingOttomanTurkishonecould,ifasyl。

lablecontainedthephonemes/k/andpartlylsland〃,guessbytheuseofArabicletters

whetherthesyllablewouldbepronouncedwithffontorbackvowels.Thisphenomenon

didnotαpr'or'holdinUygur.OurinfbrmationonthefrontnessorbacknessofUygur

syllablestherefbremostlyfbllows丘omthewaythevowelsarespelledinthestemsand

fromtheshapesofthesuffixes.TheSemiticwritingsystemsareratherunhelpfulcon-

cemingvowels:[司100kslike[a]exceptintheabsol磁eonsetand[i]and[i]arespelled

identically(exceptwhen【i]isspelledasA);[6]and[司aregenerallyspelledwiththe

11Whatismeantisthecharac顧 診rIhave㈹ 鉛πedtoasX.

12Weneedn'therediscussthequestionwhether[t'land[1'】,these實ontvariants,shouldbe

consideredtobephonemesalthoughthematterarisesonlyinconnectionwithborrowedwords,
orwhetherasupra-segmentalmodelisbettersuitedfbrthetopic.

(7)
digraphWYinthe且rstsyllablebuttherearenumerousexceptions(suchas∫ δz伽,noble'

beingspelledTWZWN).

ConcemingtheIastsyIlableofastemwetherefbrehavetorelyonsuffixes:13

Whenwefindtheaccusativeofthewordfbr`stanza'comingfromSkt.∫10た αas

SLWKWKwithKandnotXweareentitledtoassumethatitwas,beside5'o々,also

pronouncedas51δ 々;営10た+1配8and詑oた+'αrl言(bothendinginX)arealsoquitecommon,

butsois51δ た+1躍g(withtwiceK).ThewordwasprobablyboπowedthroughTokhanan,

whe爬oneofthetwovariantsoftheAdialectisspelledas.5'yoた.Whetherthevowelof

thisdialectwasitselfacmallypronouncedas[6](yomay,afterall,havebeenthe

Tokharianspellingfbr[6]),whetheritsoundedclosestto[司toTurkicears,wh弔therit

wasrephonologisedas/6/becauseofthesequence/蓉1/(brokenapartbytheTurks,as

showninspellingssuchasSWLWK)一thereisnoreasontodoubtthe!6/inthisparticu-

larvahantoftheUygurword.SinceSahskritonlyhadonelowroundedvowel(o)

whereasUygurhadtwo(oandδ),thereisalwaysthepossibilityofamappingo→6also

ifSanskritwerethedirectsource.Therearemanycasesotherthanthefluctuating

ぎω10た ∼ 訂 の'δ たwherelexemesoriginallycomingffomSanskritgetffontvowels.

Tbephonicinnuenceo飼bre三gn左

Whenlookingatallthecaseswherewefindthemappingsa→ 益,u→ 茸oro→6,it

tumsoutthatthegreatm勾oritytakeplaccinthevicinityofK(asstatedbyR6hrbom

1988).Inanumberofinstances,someofthemmentionedinR6hrboml988:237,words

inSanskdthavingた μ。,た距。,々o。,80。etc。intheonsetanespelledasKWY。inUygur,no

doubttobepronouncedasた 寵oorた δo;e.g.Gδ 【1rのvαrゴ(BTVII42)<Go滅7vorlor丸 滋ηπ6∫

<Skt.た 雌 μ4α`lotus'.SomesuchlexemesweresubsequentlyborrowedbyMongolian

withaffontvowel,e.9.履 用 翅justmentioned,orGδ 面 襯(inUygurattestede.9.inBT

13Wilhloans,血iss田tementholdsonlyifthesu雌xeshave實ontha㎜ony,accordingtotherule

(restaledbelow)thatbackha㎜onyisthedefaultclassfbrsuffixesaddedtosuchIexemes.

(8)
VIII)
.・ 伽 ∫
・脚 ・Aversifi・d・ ・α4伽 ・text・dit・dbyP・ ・fess・ ・Sh69・it・h・ ・

KWS'L'MWL(ffomSkt.妬 α1α顧1α)alliteratingwith繊 戯8andtherefbretobepro-

nouncedasた 蔚 α'α醒 〃,ifnotwithadditionalf士ontingelsewhereintheword.The

Br盃hmispellingG勿showsthattheUygurspronounced(or∫ α ηθ∫伽85pronounced)the

vowelinthenameoftheriverGangesasla/.R6hrbom1996:182-3triestoexplainthisas

fbllows:"Wenn_einSch頚eibernichtmitderNo㎜desSubsystemsve並rautist,dann

wirdereinK-ZeichendesSubsystemsalspalatalesSignalzeicheninterp爬tieren._

Verl註u丘dieAnpassungandiet雌rkischeNo㎜inRichtungPalatalisierung,dannscheint

mehrderoptischeEindruckdesK-ZeichensaufdenSchreiberverantwortlichzusein.``

R6hfbomappearstohavebeenthinkingsolelyofcopyistsofholytextsintenton(butnot

quitecapableof)retainingtraditionalspelling.Insteadofassumingunprofbssionallty

onem重ghtratherspeakofthegradualandsporadicalinfluenceofspellingpronunciation.

Hadtheinfluencebeenamerelyvisualone,8で の ㍑+燃`at(theascendanceof)the

planet'inZiemeVorr14datedtotheyear1022wouldnothavehadafrontsuffix:The

丘ontinginthiswordmusthavebeencausedbythe8。.

Moreover,wordsnottransmittedoverthecopyingofreligioustextsbutthrough

generaldiffusionshowthesameffonting,e.g。 頬 ρδz`cottQnseed'coming丘omIndo-

AlyanviaTo㎞adan(た αψ δ5αinBbutゆ 郷inA)wi血theohgina1!a/vowelsIeaised

as1田inTurkicor履r履 海`spiceusedassafffonsurrogate',atemlf6undeverywhere

betweenIndiaandtheMediterraneanwithtwo《11s.Tokharianた 痂cガsesame',athird

te㎜ 丘omthesamenon-religioussemanticdomain,isalsoinUygurrephonologisedas

1廿!,asshownbytheBrahmlspellingasた 伽6捕nTrVIIIM28.Ifん ノ θ航rofTokharian

A(疋wrα 短rinB)gives励 θたαr`ce韻emonialtent'(spelledwithKWY。seマeraltimes)in

14.TIIY37br4asread.byA,v.LeCoqitheactualf士agmentisnowlosしSeetheUWfbrth三sand

otherabbreviated .re{brencestotexts.8rのr`δ ♪
κwouldalsobeapossiblereading.TokhABhas

8rα 之 卍8r畝comingf士omSkt.8rα 加butthedef6ctivespellingmakesSogdianalikelydirect

source.

(9)
Uygur,thisisamatterbetweenthesetwolanguageswhichhasnothingtodowiththe

factthattheultimatesourceisSkt.航 醇8αrα`chaper.

Loansstartingwnh/klwhicharrivedfromorbywayofSogdianappearnotto

haveparticipatedinthcfr()ntingprocess:Suchteirnsarekolti`averygreatnumber'<<

Skt.ko∫i(cf.Laut1986:91,105),丸4ρ`anageindivinereckoning',聴 αη`averyshort

moment',karmaput`precepビorkarte<<Skt.8ψ α5'ha`householder'.Ifthisiscorrect,

itwouldlinkfrontingbyonset/k/toTokharianmediation.

〆Thereappearstobeageneralconelationbetween/k/inthecOdaofborrowed

レ  
nounsandhamlonyfluctuationinsuffixesaddedontothem:Thewordspelledas

C'TYKorC'DYKultimatelycomesfromSkt.戸'aたa`storyaboutthelifeofBuddhain

apreviousexistence'.Ithas,inSuv,threeexampleswiththesuffixes+'8and+"8(one

ofthelatterinSuv235,10)asagainsttwoexampleswith+'諺 吾(oneofthelatterinSuv

 
235,2);δ α4襯18appearsalsoe.g.inTTX4.TheTokharianABcognate/a"takisnotvery

likelytobeitsdirectsourcebecauseofthestem'ssecondvowe1.AswithSkt.Srdvaka

appearinginUygurasSR'VYK,Sogdianmediationmayherebelikelier(cf.Laut

l986:117):TheheighteningofvowelsdoesnotappeartobetypicaloftheTokharian

route.μ 猷!館1た`letter,akSara',aSogdianterm(spelledwith
.both'WYand'Winthe

firstsyllable)perhapsultimatelycomingfro!nChinese,showsboth+YKand+YXas

accusativesuffix.15Acommonvariantis,inSemiticscript,spellcdasuiak,whichalso

musthavehad配 猟asunderlyingfb㎜.Thewordspelledasδ 肋'謙inTrVIIIDl7

(fromSkt.α 肋'5θ んαthroughTo㎞ABα わ傭 θた)showsasimilarsuffixfluctuationbe-

tweenthesuffixes+1聖and+'8(accusative).Withotherboπowed。klexernesIhave

comeonlyacrossinstanceswithbacksuffixation:Thefornsaij8ik÷qaisattestedinSA

15{翅 た`dot,diacriほc'comingfromTibetanc海 θ8isattestedwiththesuf∬x+肋;itistherefbre

anotherinstanceofafb鴨ignwordnotshowingbacksuffixationbecauseof∬nalκ.Like屹 薩

(withwhichitconceivablysharesitsetymology)itisnotofIndicoriginandwastherefbrenot.

consideledbyR6hrbom.

(10)
 8;itcomesffomSkt.5α 脚8勉 んα,ana(巧ectivederivate.f止om5α ητ8勿,andaccordingto

Lautl986:145hasSogdian5π た'yたasdirectsource:Itssecondsyllablecouldbe/Vas

wellasli1.Anotherrelevantte㎜ofultimateIranianohginisρ5α た`crown',always

spelledwithKandreceivingthesuffix+'18inUII5922;itsdirectsourcemaybe

Tokharianpδ ∬o々,sincetheSogdiancounterparthasaninitia1α'{ザThequestion

whetherthespellingofthevelarhasanyautomaticrelevancefbrthepronunci3tionof

thesufHxis,Ithink,resolvedbyrunifb㎜nlglwslk211r1=π α80ぶ α〃 αr`laybelievers'in

ms.TM332(K6kt廿Turfp.1047):Thewordisspelledwithaffontk2buttheplural

suffixis+'αrandnot+1∂r.16Thisshowsthatthepresenceofaf止ontvelardidnot,by

itself,haveanyinfluenceonthequalityofthesuffixvoweLWehaveasimilarphenom-

enonalsoinManicha}anscript,'ωrαz融`scalesfbrweighing'spelledwithundo㏄ed

α 励inMIII2,8;inUII77,25we且nditspelledwithKinUygurschptasweli.This

wordofIranianodginhasbackvowelsiballIranianandTurkicsources.17

Loanswithfinal!klspelledasKhaveHuctuatingsuffixationwhichisnotin-

dicativeofthevowelclassofthelastsyllableinthestem.Nocertainchoicebetween

readingssuchasαZガ 戯 た,αわ溶 批orα わ諺

ぎ諺ゐcanbeproposedfbranyparticularinstanceif

thereisnoadditionalevidence。WehavealreadyshownthattheplesenceofKisno

reasonfbrdiscardingthelattertworeadingsfbrthisword.TheKappearstohavesaved

.the行ontpronunciationfbrsomespeakers,againstthemoregeneralrulethattheback

hamonyisthedefaultclassfbrsuffixesadded.toIoan$.Thevowelsofthelaststem

16Zieme2001:217tookMz386(TM333)v1-2tocontainanotherrunifb㎜insIanceofthis

te㎜,reading血epassageas蓉1'k212r2:tlw[gl']rl:r2t21.Thisheinterpretedasμ'go15α た1αr

'梶∫81αrδπ ∫andtranslateditas`[audi]torswerebom'.Thiswouldmeanthatη ∫80瀬 κheregets

theffontvahantofthepluralsuffix.However,the丘rstcharacterafterthelacunad㏄snot,I

think,looklikeslbutratherlikek1.Fu曲emo豹e,theverbphrase'μ8αr∂r"wouldimply

durativeaspectoracontinuousoriterativeevent,whichseemsunlikely;besldes,babiesalenot

bomasauditors.Anotherpossiblereadingis1肋 腋 μrりr瞬 諏1rrδ,r∫'`Theywerecomingto

meet(+dadve)'.

17Clausoniswmnginwdtingitwith。8;theα ΨhofManicha}anscriptcanonlybe側.

(ll)
syllablescanhowever,insomecases,t鴎deteminedbyothermeans:V短ationwithα

(asin配 ぎαた)would,e.g.,speak'againstthereading`砿 ゴた'andfbr屹 薦Iwouldalsoplead

fbr6α4孟 ●
たbeside6α4薩(notexcluding猷 ζ
ガた)becausevowelr撮sing(apparentlylinkedto

Sogdianmediation)doesnotnecessarilyimplyffontingandbecauseweknow丘om

caseslikeρ5α たthatKdoesnotf止ontvowelsbefbreit.

Sh6gaitol978hadlimitedhisscopetowordsofIndicoriginbecausehisaim

wastoanswertheqロest量onoftheirrouteintoUygur.Whendealingwiththespellingof

loanwordsofIndicoriginasagainstthespellingoforiginalTurkicstemsandsuffixes,

R6hrbom1988and1996disregardedthespellingofborrowingscomingfromother

languagesintoUygur,astheonesjustquoted.R6hrbom1988:232-233speaksofthe

wholeSanskritlexiconbeingopentotheleamedUygurusage,whichisusedevenwhen

thesourcelanguageofatextbeingtranslatedisChinese.Thefactis,however,thatsuch

Sanskritte㎜sdonotgenerallytumupintheirohginalshapebutintheshape山eyhave

inSogdianor,morecommonly,To㎞arian.WhileleamedtranslatorsIikeSiOkoSali

tutuqandl翫erUygurscholarswouldhaveadirectknowledgeofSanskht,theno㎜al

sourceofte㎜inologyWouldbeindirect,nottospeakofnon-terminologicalloans。We

couldtherefb㈹speakoftwolexicalpermeabilitiesasfara忠Buddhistte㎜inologyofthe

post-Sogdianstageisconcemed:TheIndic-TokharianpemleabilityandtheTokharian

-uygurpermeability.pealingwiththewayuygurrepresentsloanwQrdsinwriting,we

shouldbelookingatthephonicandgraphicmechanismsofborrowingasawhole,asit

isun正ikelythatalanguageshould,atanyparticularstage,havedealtwithaspecific

ultimate(i.e.indirect)sourceinImywaynotrelatedtotherestofthesources.Wehere

seethatManichaeantextsintwowritingsystemsnotdealtwithbythescholarsmen-

tionedshowthesamephenomenonastheonestheydiddealwith.Suffixfrontingby

borrowedKisclearlynotlimitedtoTokharianmediation.

Athiπdpossibilitywithte㎜sborrowedwithKis,ofcourse,thattheKisnotin

thebeginningortheendbutsomewhe祀inthemiddleofthestem.Thebaseofafbrm

(12)
1ike鰯 ゴ+1ま8(Sh6Agon203u.22;<Skt.話 訳:yαbywayofTokhAB動 燃)could田so,of

course,haveendedin/▽,innuencedbyitsfirstvowel;thisinspiteofthespellingwith

丘ontK,as,infact,ぎ 盈'itselfcould(e.g.inTrX66,136,334)havebeenpronouncedas

5α 離inspiteoftheK.'α 肋 π くTokhAB1α 緬 αη3isalsogenerallyspelledwithK.

3勿 漉 η2くSkt。5α 那81痂 彪1π α`monastery'alsohas/alsinSanskritbut實ontvow-

elsinUygur.18R6hrbom1996:183proposestoaccountforthisbyassumingthat

SNKR'M(thespellingwithoutα'4aftertheS,whichwasinuseinSogdianandis

attestedonceinanUygursourceinSogdianscript)wasmisunderstoodasS'KR'M,α1げ

andη 距ηbeingsimilarinUygurwriting;theKisthensupposedtohavebeeninterpreted

asa`signarcharacterfbrffonting,whileretaining回phonically.Allthisdoesnotseem

verylikelyasitpresupposesasimultaneousinterpletationofacharacterstringasboth

NKand'Kandtheabsenceoforaltransmission.ThetemlappearsinTokharianABas

5α 派rδ η20r5α ヵ8罐 鵬withavelarnasalfbllowedbyavelarstop.IfSogdianwasthe

immediatesource,theprocessremainsobscurebutcanbeaninner-Sogdianproblem;if

itwasTokharian,thevelarstopwithwhichthewordappearsintheTokhariandialects

couldhavebeenresponsiblefbrthefronting.

Frontingwithoutvelars

TheonlysolidlyattestedexamplefbrthefrontingofastemwhichhasnoK(andis

therefbrenotmentionedinR6hrbom'spapers)whichI㎜awareofisδr4'η'`jewel',

whoseultimatesourceisSanskritrα ∫
ηα.Itsfirstvowelisshowntobela!byBrahmI

evidence;the丘ontsuffixvahants+1'8,+舷and+8,withwhichitisffequentlyattested,

proveitslastsyllabletohavebeen〃andnot〃.Ihavecomeacrossasingleinstanceof

+X'inBTVIIIA122.TheUWentryfbrthislexemeallowsustotraceitshistorywithin

IJygurlManicha3an・texts(includingtherelativelylatePothiBook)neverhavetheinitial

vowel,whichisclearlyaTurkicaddition;theyshowthespellingsRTNY,RTNYY,

18`∫ αηrα1ガinUW277bisanen'or.

(13)
RDNYandRDDNY,BothSogdianandParthianarerelevantf6rtheshapeofthis

lexeme:SogdianretainedtheSans㎞t'whileP飢hianshows4insteadandmaypossi-

blyhaveinnuencedthefb㎜.RDNYisfU曲erattestedinms.LinTTVI243(asagainst

sixothermss.extantfbrthepassage)asonlynon-Manich紀anevidencef6rsuchspell-

ing.19δr4ηlisthesecondstagethe飴 ㎜wentthrough,thelaterstagesbeing∂r4ゴ η∫,

occasionalδr加'(fbundintheDLTasavariant)and(rarely)∂r4伽 ∫(perhapsborrowed

backffomMongolian).TheManichaeanvariantwaspresumablypronouncedwitha

vowelafterthe〃butwedonotknowwhatthatvowelwas;ifδ ㎡ η∼cameaboutby

metathesis,itwas/田(andnot/a/,asassumedinLaut1986:144).He爬weclearlyhave

丘ontingwithIrapianinte㎜ed鱒.Wedonotknowwithwhatvowelsthisintermed鱒

waspronounced;thevowelsofboπowingswhichreachedUygurbythatwayneednot

concemusasthef士ontingcouldhaveoccurredwithinthoselanguages.Wehavetostate,

atanyrate,thatnotallwordsofultimateIndico丘ginhavebacksuffixesinUygur,even

iftheydonotendin/k1.

↑hemteofSanskrit『 躍inUygur

Sanskrit。 αstemsrefb1TingtoanimatebeingsgenerallyappearinTokharianwith。 θ

whilethoserefbrringtoinapimateentitiesdropthevowel.AsdiscoveredbySh6gaito

1978(andindependentlyfromthisrestatedbyMoenoose1980),thisdist曲uhonisgen-

erallymirroredbyUygur,showingthatUygurborrowedmuchofitsBuddhistlexical

stockffomTokharian.InSemiticwritingsystems,withwhichthevastmajorityof

Uygursourcesarewritten,thecodavowelisibtheseborrowingsspelledasY.Yisin

Uygurusedfbrtranscribingthreesounds,[i],国andalso[e].FromsourcesinIndic

writingweknowthatle!WasinUyguradmissiblealsoinnon-firstsyllables,bothin

19Manichaeani曲enceontheS嵐zY琶 ㎞ 欲Yarokhasbeenvadouslydiscussedinthe㈹cent

years;thisevidenceofonems.againsta110thers〃2'8乃 ∫speakforaspecificManichaean

backgroundofthisparticularscribe.

(14)
borrowingsandinoriginalTurkicwords.Itmightthereforebeassumedthatthesound

appearingasYwas,inthecodaofsuchwords,pronouncedas[e},fbllowingthesource

language.Thisisnodoubtthei℃asonwhytheUWtranscdbessuchwordswith。 θ,both

whentheauthorcanadduceevidencethatsuchaTokharianwordexistsandwhenno

suchevidenceisadduced:Thewordsattestedas"CYTYas"RCWNYor'RCWNY,
ソ ソ ソ
ソ
andas"YWKS'NYarerespectivelytranscribedasα δ'∫8(withTokhariancognate),as

α踏 麗η6andasαyμ 舷 αηθ(noTokhariancognatementioned).Suchetymologicallymoti-

vatedtranscriptionsarenot,asfarasIcoulddetermine,suppOrtedbyBr豆hmTsources.

ThereismuchevidenceagainstreadingUygurYaseeverytimeTokharianhase,apart

ofitmentionedinthenextsection:ThelexemewrittenasasureinTokharianandcom-

ingfromSkt。asuraisinNeujahr73(aBrahmltexteditedinZiemel984)spelledas

asuri.TheUWquotesthisUygurinstanceasasureandwrites:,,TrotzdesBrahml-

Belegstranskribierenwirasure``。Brahmlevidencedeservingtobetakenseriously,

the爬seemstobenoreasonfornotreadingthewordasasuri'.Insimilardisregardfbrthe

data,Lautl986:140writestirte(<<Skt.tirtika`teacheroffalsedoctrine')becauseof

TokhBtirthe,inspiteoftirthiinT"1'VIIIO3;anotherBrahmlinstance,tirtiinMaueKat

Nr.44av8,appearedsubsequently.Otherexampleswithi〃'<<Skt.ainthecodaare

iit7ttalc'aytasiki<Skt.。caitasikainMaueKatNr.44av3,ni8ranti<.Skt.nirgrantha

ibid.v8andfurthertemlsinthatline.InNeujahr75-78we 、findthepropernames

梅 η助 の α,ルtageibhadra,Va〃abhaandVais'ravageaallinUygurversionswithiandnote

intheend.Itistruethatwefindasa
.nke`averylargenumber;averylongpe藍odoftime'

inMaueKat19Nr,11fromTokhAasamkheandsvarmire`novice'inMaueKat3Nrs.78

and84fromTokhB5armゴrθ.Thesetwowordsarenot,however,instancesofSkt.。a>

To㎞.。 θatall,astheycomefromSkt.asatrikhyeya(andnot`asamkha')and5r碗 αμεrα

respectively.

(15)
Back.vocalicsuf6xationasdefaultin且oans

ItwasPeterZiemewhofirst(1969:57-8)pointedoutthatf6reignlexemesappearingin

Uygurgetbackvowelsuffixesevenifthestemshaveffontvowels;Gabain1950:58had

statedthatthereareirregularitiesinthisdomainbuthadnotdealtwiththem,theonlyone

shementionsbeingthatη1'rwα η(asshespellsthete㎜)gets柳whereas .かηめrα η(<<

pαr'η'rり 勿 のisattestedwith+舷asdativesuffix.Ziemediscussedanumberofex-

amplesinManicha3ansources,mentioninginanotethatthephenomenonispresentalso

inBuddhisttexts.ThematteristhenelaborateduponbyR6hrbom玉988andl996,who

limitshimselftoIndictermsin.Buddhistsources.20Thefactisbestobservedintexts

writtenwithBrahmlwritingwhenthestemendsin[e],whichisspelledasθ;e.g,in

α∫αηたθ+1α π α(MaueKatNr.19,ll),ぎ αη 癬 θ+1αrrandま αrη2'rθ+'α7た α(Nrs.3,78and84).

Thestemof∫ 郎 θ+'α 励 吻,ofhispupils'(AtSOlv4,alsoBr置hmi)showsthesamephe-

nomenonbutcomesfromChinese.InBuddhKat23(Tibetanwriting)4ゴ4伽+118`wear-

ingadiadem'seemstobethepref￠rablereadingagainsttheeditors'`4鋤.溺118',but

4'41}η'18cannotbeexcluded.Thefbmlisattestedwith+118alsoinMaitr2v23,BTVII

andelsewhere。Thereisnoknownreasonf6rノ 万 甑`anger,ofIranianohgin,tohavehad

backvowels;eventhislexemewithtwoli/sshowsfluctuationsinsuffixhannony:

叛5∫ ∫+1αr+た α(MII10;4,TrIX94andelsewhere)vs.声 ぎ'∫+'δr+履(MIIINr。11v3and

elsewhere).SinceweknowfromBrahmlevidencethatthelastsyllableofUygurα5α ηたε

20R6hrbom1996:178refbrstoasingleexception,ぬ 磁+η 吻inMaueKatNr.44ar2;whichhe

considerstobeanemr.Anotherexcepdon,inaManich㊧antextandwithanIranianword,is

脚 乞α8+価withKinthesuf∬x,refbπedtobyZiemel969=57.rψ 訂+4δinMaueKatNr.3,86is

consideredanerroraswell,as中eeditortranscribesthebaseasr￠1㎏8か;R6hrbom1996:178

doesnotagree,thinkingthef}ontnessofthesuffixiscausedbythepresepceofg.Sincewemust

assumethepresenceofanauxiliaryvowelafterthe8(whichR6hrbomdoesnotdo),rulesfor

suchvowels(discussedbelow)shouldapply.Weshouldnot,ontheotherhand,Iimitourview

toborrowings:Sincewefindμ8μr+面inNr.3,90aswellas96(inthefbllowingIineinthems.

and41inesfurtheron),thescribeapPearstohavebeengoingthoughasingularstretchof

confUsion.

(16)
ls!￠1,α5α η々θ+'1吾(documentedintheUW;consistentlyspelledwithXinthesuffix)and

α5αη丸θ+4α φ(DKPAMPeters263,403)canalsoserveasexamplesfbrthephenomenon.

Similarlywith4θ η+た α`tothereligion'inTTIIA46,inviewofthefirstsyllableofwhat

shouldbereadas4θ η吻rinTibetanscriptinBuddhKat26,29and30.Withoutthe

infb㎜ationonthebasevowelffomaninstanceinanIndicscript,wecouldalsohave

readitas`4置'η'(asZieme1969:37infactdoes).

Therulethatborrowingsshouldgenerallybefbllowedbysuffixesofbackhar-

monyhasaquitewidediffusion;inQarakhaniditholdsfbrIslamicloans:TheQBand

theAtabatu'1一 耳aq盗'iqspellstemswithconsonantsindicatingffontvowelsbuthave

theirsuffixeswithback-hamlonyindicatingconsonants;cf.Mansuro舘lul959:90.The

rulestillstands-strangelyenough-inmodemTatarwithRussianloans(thoughnot

withearlierloansffomPersianandArabic),asdescribedinOnerl998:22-23:e.g.

4ρPθ ∫
嫉 繊`apPet孟te'40た 研 π8π∫'α
動 アπ8α η`documented',ρro雄 ω18彪 面`toyourpro-

cess',ρr∫ η飽ψ'αr孟'`theirphnciples'etc.;therearenoirregularitiesamongback-vowel

stems.Onerpointsoutthatthemattercannotbemotivatedbyrece口tspellinghabits,asit

isfbundalsoamongtheTatarcommunityinFinland,whichemigratedduringtherevo-

lutionandusesTatarmainlyasaspokenmedium.ThesourcesoftheTatarphenomenon

aretobef6undinClassicalChagatay.

Wetranscribedtheめfthepossessiveandgenitivesuffixesof疏 吻rl切 ヵwithガ

becauseBrahmidoesnotshow中ed榔erencebetweenthesetwosounds.Ifthelastsyl-

1ableofaborrowedsteminaB士ahmitexthas',thatcan題lsorep爬sent[i]aswellas[i]

(butnot[eD:Thewordspelledasρrαvαr'煽 αinZweiFrag5canbereadasρrαvαr量 凌+4α

aswell,sincethesameffagment耳1sowritesIVas'inyα 漉,海oη 量
◎(accusative),α'118.and

manyotherwords.Wemightthuspossiblyreadr罐 ●+'αinsteadofrα ぎ1+∫αinTrVIII

L32,fromSkt.r誘'`zodiacsign';ノ 雁 ∫
ψ μ〃'+11'81αrinTTVIIIG13mightalsohave診'.For

theaccusativefb㎜spelledゆ α∫∫1+51η`itsemergence'inBrahmlTrVIIIH4theread-

ingμ 砂 α'1否肋ispossibleaswell.ThewordintheUWandinLautl986:144spelledas

(17)
αr砦∫,comingffomSkし ハ3'presumablyvia.somePrakritandthenviaSogdianぜy,was

probab互ypronouncedasαr営 孟'inUygur:君 ∫∫+1αr4αinTTVIIID6(withhistoricalspelling

・fth・ ・t・m,if・ead・ ・nectly)㎝d・'・ 囎,+た ・21et・.i・Uygurschpt卑 ・yb・n・ ロ ・・f,b・t

αr砦εinBuddhKat26and29makesthislikely;seeMaue&R6hrbom1984:294and

f60m.42(alsorefbrdngtotheTibetallitinerary)fbrtheuseofθtorep㈹sentIVinthat

source.Thecodavowelofthiswordisunlike正ytohavebeen8asithas'inthesource

languages。 〃'zvα πε+118`characterisedbymoralblemish'inBr巨hml(TTVIIIE47)and

instancesoftheadditionof+1αrま 吾,+1α 矯 α,+1孟●8and+た αtothisbaseinUygurwritingare

alsogivenasproofthatfbreignwordswithffontcodasarefbllowedbyback-classsuf-

fixes.However,thespellingη'zvα η6inBuddhKat9,27,37and43pointsillthedirection

of厩vα 鴫thereadingproposedbyMaue&R6hrboml984isinfactη 孟
乞vαηi=22

ThenwehaveagroupofUygur。'stemsco廿espondingtoTokharial1。 εstems:

ん研 め αη読+1αrinA婁SUlv6,coming丘omSkt.た 躍 η肋 勿4飢hroughTokhBん 研2わ 緬 轟46,

cansaf￠lybespelledwithl'inthelastsyllableofthestem;notethat∫ α ∫ε+'α 吻1'ηappears

inthesamems.,sothatthescribe.didhave。6inhisinventory.8α η吻 ハノ'+1αrffomSkt.

gα η伽 αη αoverTokhB8α η融 αrv8isquiteunlikelytobeanerror,asitappearstwice

(A￠SO3r2andvl).Norisη'8rα η"+'αr(TTVIIIG13)<TokhB〃'8rα η'舵 くSkt.

η'r8rα 励 α`anchorite',supportedbyπ'8rα π∫
ゴinMaueKatNr.44av8:Bothprobably

have肋 ∫5∫
碗+1αrl'(MaueKatNr.44bB2)presumablybelongstothesamegroupifit

ultimatelycomesffomSkt.η'5擁 伽 α`fbod',astheeditorthinks.

αわ'ぶ'〃 αわ揃 霞mentionedaboveisanotherwordofIndic→Tokharianorigin

withheightenedfbreignθ,attestedwiththebacksuf登x+1量80ntheonehand,withthe

accusativesuffix+∫g(BTVIII355and361)on.theother.Whywasthele/inた μ励 αη4e,

21R6hrboml988:240givesαr砦'+た αand當 αゼ+118asproofthatIndicloanwordsgetback-class

sumxesevenwhentheyendinffontvowels;erroneously,tomymind.

22Sogdian,thesourceIanguage,hasπyzβ'η'々/ηyzβ'の ・6り,probablytobetranscribedasnizβ5ne,

(18)
ρ5μrεorα わ∫
ぷ8々notretained23althougho5α πたθ(<Tokh.α5α 那 肋 ε)doeshaveit?Non-

first-syllablelelsmayhavebeengraduallyeliminatedffomOldTurkic:yα ε`7',

yθ8ケ 切 θ`20'and5伽rθ`pointed'mayhaveoriginallyhavehad!6/andlostitinthe

courseoftime.Sothiswasaninner-Uygurprocess.

Iamnαawareofasingleins亡anceofTokharian。6<Skt.。 αappearingas。6in

Uygur.Ifthisis.reallyso,thentranscription6suchasαr働 η6shouldbecorrected.

Apossibleretractinginnuenceoftheaccusativesuf猶x

Onefactorwhichmayhavechangedcoda〃(oreven/b/aswell)into〃couldhavebeen

theaccusativefbrmaddedtoitintheback-harmonyvahant.Takeη1zvα π」'+g(attestedin

SuvandinBTVIIH20)andα6αrl'+8(attestedinSuv),which爬achedUygurf止om

Sogdian(theultimatesourceofthelatterbeingSanskrit),酌 謝'+8(TA4272),肋 ぷα'rl7+g

(Suv),pωr4ρ'1'+8(sixdmesinTA)or4αrπ 診'+8(Suv484,17),allspelledwithXinthe

suffixandhere-tentatively-transcribedwith〃attheendofthestem.Thefactthatthe

codavowelgetsintothe∫ α刑 θ3y〃 αわ'6asaback-consonantishighlyln(elytohave

爬tractedittobackposition.This粥 αythenhavcir血uencedthewholeparadigm.Iam

awareofnogoodargumentagainstreading4αrη ∫asゴ αr励Oandtaking4αrη1'+'1言,

4αrη 量
●+'αr三:getc.tobeinstancesofstandardvowelharmony.

Auxiliaryvowels

Consonantc1ロstersoffk)1℃ignwordsareinUygur,asinotherTurkiclanguages,often

brokenupbyauxiliaryvowels;theyusuallyinvolveanr,whetherasyllabiconebe-

tweenconsonantsoraconsonantalonebesideavoweLThereisnoreasonwhysuch

auxiliaryvowelsshouldbetakentobeli1,aswrittenbysomescholars,ifadjacentvow-

elsbelongtothebackseries.TheUWisinconsistentinthismatter:Theauthorcorrectly

23TheUWgivesα う漉 たaslemma;thisisunacceptableasitisagainstB語hmidocumenIation.

(19)
writes(珈 三
ア∫α<<Skt・ α1π.πα,α ηα∫か αvα4α ∫
μ<<。 δ5rαvα 。andα η￠ραηrα♪5∼}η
αr"1孟吾<<

。3η3!∫'。
,butα わ14αr∫η∼ofUW(<TokhABの ん'励 αrη2)shouldbeα わ'4αr伽,especially

sinceaMaitrvariantisspelledα わゴ
ぬrα 肱 αわ'ρか αy(UW;<TokhABα わ履 ρrδy;Uygur

Brahmiδ 肋 ψ 短y)mightbetterbeα ∂ψ か αy:Theprecedingvowelis'butρrαy>ρ1ア αy

wouldseemtoconcemadevelopmentwithinthedomainoftheoriginalsecondsyllable.

αηで切 η"r'営inLaut1986:124fbotn.4shouldbeα ηα漉 ア毒

警:OtherUygurvariantsofthis

word(<TokhA伽 伽 面r6,Bα 紹 磁r6くSkt.伽 αη∫αrンのmentionedintheUWare

αηαη∫rαぎ,αηα〃∫診}噂
α営,αηαπ∫αrαぎandα παη〃 踏.TheUW'so1ガrα η1αηθ<<Skt.oprα η2ζ7∼2α

`incommensurate'shouldbe￠ μ γα溺
αηガwhileαr4か<<Skt.δ ㎡rδ(nameofamoon

station)shouldbeα 肋 ア.TheUWentryfbrα 磁(thus!),ultimately<Skt.

.掴,quotes

instanceswheredoubleorsingleα1げofthefirstsyllablearereplacedby'YfromSh6,

Sh6KenkyU,BTVIIIandSuv;theauthorthinkstheyshouldbereadas`6彦 ガ(because

twoofthemaltematewithsingleα'⑳or`lr砦 ∫'.ButδisbynomeansthesameasθandY

cannotbeusedtorepresentit,aspellingwithasingleα'4cannothavebeenusedfbr

representing[e],noristhereanyhistoricaljustificationfbr[e】./Vdoes,ontheother

hand,altematewith/a/asauxiliaryvowelinmanyoftheinstancesquotedhere.Laut

1986:97writes{勤 効rfbr{狛 た〃(くTokharianABc盈 熔randSogd.c脈r<Skt.cα 々rα);the

〃issupportedbythevariants面 たrα♪rand〔¥α ♪たαrandbywhatwe五ndordonot且ndin

thesecondsyllableof.thesourcelanguages.24Similarconsiderationsshouldmakeus

writeρ α'かandnotρ α∫'r(<TokhApatr),ぎ α5'かinsteadof5α ∫∫'r(<<㍍5∫rの,vα δγand

notvα δεr(<To㎞ABv切 αり25andsof6rth、Occasionallyauxiliaryvowelsappearbeside

consonantsotherthanlr1;he爬wejustmentiononeoftheexamples,α 舖 ψαδ'く(TokhA

α5vの'),whichisinLautl986:126quotedasα 舘vα δ∫η.

24LautisagainstM㏄rloose's(1980=69)derivationofthewordffomTokharian,presumably

becauseoftheMai面simit'sdefbctivespellingas(菊r.

25TheUygurvariantvα 澁may,ontheotherhand,reflecuheTokhAvariantwα ∫か.

(20)
Thequahtyofraisedvowels

Laut1986:96takesthewordfbr`sandalwood',ohginallySkt.cα η4α πα,tohaveb㏄n

borrowedintoUygurbywayofSogdian(oη'η)andnotTokharian(Acα η(∫αη1,B

cαη面 那1cα η吻 のbecauseofitsraise面rstVoweL26WhatIhereo切ecttoishisspelling

oftheUygurlexemeas`6'π 短 ガandnotδ 弛昭 η:Raisedvowelsshouldnotautomatically

beassumedtobeffonted.Inthiscasewefindthewordspelledwithnlinthe(runifbrm)

IrqBitig,which(sincethissourcedoesnotbelongtoOrkhonTud(icbuttoUygur)

shouldbesufficientevidence,Apothercaseinpointisthespelling`翫zz加'inLaut

1986:138.Italtemateswith∬ αzαηandcomesffomTochAB面 ∫αη1,both爬asonsfbr

readingthisvariantas営 αz量}2.TheeditoroftheTAhascorrectly(asinsomanyother

cases)transcribedthiswordas諏zlカ.

Condusion

Ihave,inthispaper,triedtopointatclues-someofthemobviousbutmanytoooften

disregarded-concemingthephonicshapeofwordswhichUyguradaptedf止omother

Ianguages;Ihopetohavepointedatsomeoftheproblemsinvolved.Atpresentwe

understandonlyapartofthefbrceswhichwβreatplayinthismatter.Whatisneededin

ordertogetabettergraspofwhatwasreally .goingonisasystematicsurveyofallthe

availablematerial.Thefbremostprincipletobeappliedshould,Ithink,bethetrusttobe

placedinexplicitevidence:inthiscasemostlyffomBrahmiandTibetanscriptsources.

ThesecondprincipleIwouldpropoundisthatUygurisa`㈹al'Ianguagewithspeakers,

possessinganintemallogic,andnotmerelyamediumfbrthetransmissionofscriptures

whosech叙acterisdete㎜inedbyschbes.Therewere,ofcourse,1inguislicdi晩rences

betweenthedialectsofvahouscommunities(notallofthemsu倒cientlydete㎝inedas

26Theterm`SproBvokal',Le.auxiliary(`anaptyctic')vowel,usedbythea戚hor五nthisand

similarcasesisinexactasthewordinthesourcelanguagescanbeexpectedtohavehadthe

samenumberofsyllablesasinUygur(i.e.inthiscasetwo).

(21)
yet),butitshouldbepermissibletodrawconclusionsffomonetextgrouptowardsthe

otherunlesstherearegoodreasonstothecontrary。Thefinalpointisthatthereobviously

wasindividualvariationbetweenusersofthelanguage,linked,amongotherthings,to

thedegreeoferudition.Inthatsense-thewritingsystemsnotbeingquiteadequate丘om

thephoneticpointofview-therewillalwaysbeadegreeofindete㎜inacy.Itisourtask

tobringthatdeg爬edowntoaminimum.ThistaskIhave,inthispaperdedicatedtoa

fbremostcontributorinthisverydomain,beenabletoaccomplishonlytoaveryIimited

extent.

Bib蓋iography

Erdal,M.1991101dTurkicVVordFor〃lation.Wiesbaden:Han・assowitz(Turcologica9).2

volS.

Gabain,A.v.1950=〃tttirkischeGram〃uαtik.2.verbesserteAuflage.Leipzig:Harrassowitz.

Johanson,L.1991:`OnsyllabicfrontnessoppositionsinT町kic.'VariaEurasiatica.

Festschrit}ノ 髭rProfessorAndrdsR6π α一Tas.Szeged, .77-94.

Laut,J.P.1986:DerfrtihettirkischeBuddhismusundseineliterarischenDenkmdi'er.

WiesbadenlHarrassowitz(VdSUA21).

Mansuroglu,M.1959:`DasKarakhanidische.'。Phi'ologiaeTurcicaeFundamentaL

Wiesbaden:Steiner,87-112.

Maue,D.1996:肋'鰍'5c舵HandschrPten,Tei'1.DokumenteinBra'hmiundtibet'scher

Schrift,Beschriebenundheraus8egeと ♪envoηD,ルf.Stuttgart:Steiner.(Verzeichnisder

OrientalischenHandschrifteninDeutschlandXIII9)

MaueD.&R6hrbornK。1984-85:`Ein"buddhistischerKaIechismus"inaltttirkischerSprache

undtibetischerSchrift'.ZD〃Gl34=286-313(TeilI)andl35:68-91(TeilII),Texthe爬

referredtoasBuddhKat.

ノ
Moerloose,E1980:`Sans㎞tloanwordsinUighur'.」 ∬4:61-78.

Moriyasu,T.1985;`Etudesuruncal6chismebouddhiqueouigouren6cdturetib6taine(P.し

1292)'.ルfe〃loirsoftheFacu〃yρ プ∠etters,Osakaこ ノniversit),25:1-85,

Oda,J.1983:`RemarksontheIndic``Lehngut"oftheStikizytikmtikyaruqsUtra'.Sprachen

desBuddhismusinZentralasien,ed.K.R6hrbornandW.Veenker,Wiesbaden:

Harrassowitz,65-72.

Oner,M.1998:Bugtinkde'KipgakTdirkgesi.Ankara:T肛rkDilKurumu.

R6hrbom,K.1977f:Ui8urischesWtirterbuch.Sprachmaterialdervorislamischenttirkischen

TexteausZentralasienぐWiesbaden:Steiner).6fascswithcontinuouspagenumbering,

referredtoasUW.

(22)
一1981:`ZumWanderwegdesaldndischenLehngutesimAlttUrkischen' .HansR.Roemer&

AlbrechtNoth,eds.,3∫ κ4'επz配rGθ5c痂cぬ ∫θ 配π4κ κ〃μr4θ5yor4θr8ηOr∫ θη'5.

Fθ ∫なc加 がげ 翫rBθ πo'4畠 ワ〃 θrz麗 〃∬ ゴ

θわz'8∫'θ
〃Gθ わ腐π5'α8.Leiden:Bdl1,337-343.

一1988:`Z町DarsteHungderGutturaleindenindischenFlemdw6rtemdesU圭gurischen㌔CA/

3213-4:232-243.

一1996:`ZurSuf匠xklassifika丘onimAl仙rkischen'.ひ 〃 わ1V .F.141176-186.

Sh6gaito,M.1978:`KodaiUigurugoniokenlIndoraigenshakuy6-goinod6nyUkeironl

εsuile'。!螂σ41諺r伽8ε πgo∂ 配π㎞ 舵 π々ソ廊15:79-110.

一1991-3:∫ ∫
励 β∫∫
η 酌 θσ'g乃 〃'レセr5'oπ(ゾ 漉 εAう 雇4ん αr㎜ 々05αわぬδの 履繭∫
疏 α 勲"vδ 「∫
擁 ・Vo'∫ ・

1一〃,7セ 溜,7}rα 鰐'α ∫'oπα層Co〃3〃26π'α り1.Shokado.Hereref6πedtoasTA. .

Zieme,P.1969:σ η'6r∫κc勧 η8θηzμr∫cんr哲 麗η43ρrαcゐ θ4θr脚 αη'o層'5cぬ 一

漉rん'3c乃8η

乃4ゆ η嫉 惣.UnpロbLPh.D.dissertadon(Berlin).

一1984:`ZurVerwendungderBr盗hmi-SchriftbeidenUigu㈹n'.AoF11:331-346.

一2001:`AManichaean-TurklcdisputeinRunicscdpt' .P.Mirecki&」.BeDu㎞,η;θ 乙'g勉

α刀4∫ぬθDαr肋 θ∬.∫'μ4'θ5加 ルf伽'c加8∫5溺 απ4"5Wb〃4.Leiden:Brill,209-219.

(23)