Tungusic Historical Linguistics and the Buyla (a.k.a. Nagyszentmiklós) Inscription

Studia Etymologica Cracoviensiavol.20:17–46  Kraków2015doi:10.4467/20843836SE.15.002.2788

www.ejournals.eu/SEC

José Andrés ALONSO DE LA FUENTE (Vitoria/Barcelona)

TUNGUSIC HISTORICAL LINGUISTICS ANDTHE BUYLA (A.K.A. NAGYSZENTMIKLÓS) INSCRIPTION*

Abstract.ThispaperpresentsaproperlinguisticassessmentoftheTungusicreadingoftheBuylainscription,asproposedbythelateEugeneHelimski(1950–2007)whobe-lievedthatoneofthelanguagesspokenbytheEuropeanAvarswasTungusic.ThemainconclusionisthattheTungusicreadingshouldberejected.Thisoutcomepartlyagreeswith the communis opiniowherebytheBuylainscriptionhidesa(nunidentifiedsofar)Turkiclanguage.

Keywords:etymology,philology,medievalhistoryofAsia,historicalandcomparativelinguistics,Tungusic,Hungarian,Avars,migration

1. Introductory remarks

Inaseriesofarticles,thelateEugeneHelimski(2000a,2000b,2003,2004)arguedthatanaberrantformofTungusiccouldhaveenteredtheCarpathianba-sinduringtheAvarperiod,theonlyevidenceofwhichispreservedintheBuyla(orBoyla/Boila)inscriptionandahandfulofwordsfoundintheclassicalsourcesontheAvars.1Moreover,itispossibletoinferfromthewordingoftheauthor

* PapersupportedbytheResearchProjectDURSI2009SGR18(Spain).Iwouldliketoexpressmygratitudeandappreciationtotheanonymousreviewersfortheircor-rectionsandmanyvaluablesuggestions.Needlesstosay,Ialoneamresponsibleforallremainingerrorsandomissions.

1 FromtheviewpointofEuropeanhistory,theso-calledAsianAvarsaretraditionallyidentifiedastheRuanruan(402–555).ThetermAvarsreferstotheEuropeanAvars(567–822),i.e.theAsianAvarsthatenteredEuropein555AD(seei.a.Pohl2002).TheNagyszentmiklóstreasuretowhichtheBuylainscriptionbelongs(see§2 below) isassociatedwiththelastremnantsoftheEuropeanAvarculture,i.e.theonewhichspreadovertheCarpathianbasinduringthe8th–9thcenturies.Goodsummarieswithadditionalliteratureofthetwomajorcompetinginterpretationsregardingtheethno-linguisticaffinitiesoftheRuanruancanbefoundinGolden(1992:76–79),whopre-sentsthetraditionalpositionthattheRuanruanwereactuallyaMongoliclanguage

18 JOSÉ ANDRÉS ALONSO DE LA FUENTE 

thattheAvarconfederationcouldhavebeenconstituted,amongmanyotherun-knownnations,byasmallcontingentofTungusicindividuals(Helimski2000b:53fn.12).ItwastheTungusicreadingoftheBuylainscriptionthatledhimtothisconclusion.

TungusicisoneofthemanyindigenousethnolinguisticgroupsoftheAsiancontinent,itscurrenthabitatcoveringmostofEasternSiberiaandManchuria.Speakersof theNorthernTungusiclanguagescanbefoundinCentralandNortheasternChina,whereasthebulkoftheSouthernTungusicspeakerscon-centratesintheAmurianregionandtheNorthernmostpartoftheSakhalinIsland.2 Manchuricspeakersaside,aboutwhichweknowagreatdealthankstoChinesesources,theSiberianTungusicwerefirstreportedattheverybeginningofthe17thcentury.3ThetimedepthoftheTungusiclanguagefamilyisveryshallow,withManchuricbeingthemostaberrantgroup(specialistsconsiderthisconditiontobetheresultofMongolicandChineseinfluence).

population,andJanhunen(1996:190),whobelievesthatthelinguisticcoreoftheRuanruanwasTurkic.Beckwith(2009:390–391)pointsoutthat“[c]arefulstudyoftheJou-jan[=Ruanruan]namesintheChinesesourcescouldshedlightontheeth-nolinguisticaffinitiesoftheJou-jan;untilthatisdone,speculationonthesubjectispremature.”Inthesamevein,seeVovin’sremarks(2007:180,184–185).Incidentally,thehypotheticalconnectionbetweentheethnonymsruanruan and ju(r)cen‘Jurchen’echoedbyHelimski(2000b:137)ismostlikelyfalseandshouldbeabandoned(fortheetymologicalintricaciesofthetermju(r)cen,seeJanhunen2004).

2 AsiscustominrecentspecialistliteratureonTungusiclinguisticsandinagree-mentwithsomeoftheideasbyJanhunenonphonologicaltranscription(1987,1996:xiii–xiv),Helimski’s‹e›hasbeenreplacedwith‹ä›,‹j›with‹y›,‹ʒ&ǯ›and‹c&č›mergedin‹j›and‹c›,respectively,vowellengthiswrittenwithdouble-vowels.Otherconventions:NorthernTungusic(=Northwestern:Ewenki,Ewen,Solon,Negidal,Arman,Udihe),SouthernTungusic(=AmurianTungusic:Oroch,Nanay,Kilen,Kili,Ulcha,Orok),withUdiheandOrochservingasabridgebetweenonebranchandtheother,Manchuric(EarlyandLateJurchen,WrittenManchu[=WM],SpokenManchuandSibe),CommonTungusic[=CT](alllanguagesbutManchuric,i.e.NorthernTungusic+SouthernTungusic),andProto-Tungusic(=Pan-Tungusic=CommonTungusic+Manchuric).“Lit.”standsfor“Literary”,and‹-n›for(lightly)nasalizedfinalvowel.ThedifferencebetweenProto-TungusicandPan-Tungusicisthatthelatterdoesnotmakeanyclaimsregardingthe(genealogical)inheritanceofagivenword,i.e.itmayrefertobothinheritedandborrowedterms(seeforinstancethepresenceofEnglishloanwordsacrossentirelinguisticfamilies:theyarecommon,pan-elements,butnotproto-elements;theformeremphasizesthesynchronicdistribution,thelatteritsdiachronicdepth).

3 ItmaybeworthnotingthattheMiddleAmurregioniscommonlyidentifiedasthemostlikely UrheimatfortheparentallanguagefromwhichalltheTungusiclanguagesde-scend(seegeneraldiscussioninJanhunen1996:167–172,andalsoJanhunen1985,2012,2013:27–28;forfurtherdetailsontheNorthernTungusicexpansion,seeAtknine1997and,forthelargerAltaisticperspective,seeMiller1994).

TUNGUSIC AND THE BUYLA INSCRIPTION 19

TheAvar-Tungusictheoryisindeedaboldproposal.IfitturnsoutthatHelimskiisright,thentheBuylainscriptionwouldinstantlybecometheoldestlinguisticmonumentinanyTungusiclanguage,washingawayeventheearliestJurchenrecords.Inspiteoftheapparentrelevanceofsuchastatement,Helimski’sproposalwaspassedoverinsilenceintheTungusicspecialistliterature.NolesssurprisingistofindoutthatcriticsfromotherareasdisregardtheTungusicna-tureoftheBuylainscriptionwithoutdiscussingitssubstance.TheyareusuallyTurcologistsbelievingthattheonlypossiblereadingoftheinscriptionhastobeTurkic.ThemostexplicitstatementwasmadebyErdal:“[…]thehypothesisis,however,arrivedatbysomearbitrarystretchingofTungusdata,[it]isfar-fetchedbyitselfandisthereforeratherunlikely”(2007:79).4Erdaldidnotgointogreatdetailinordertoexplainthereaderwhatthe“stretchoftheTungusdata”involved.Therefore,thegeneralopinionisthattheTungusicreadingoftheBuylainscriptioniswrong,5butnoonecanexplainwhythatisso.

ThemaingoalofthispaperistoprovidethereaderwithanevaluationofHelimski’shypothesisbasedontheTungusicdata.NeitherthegeopoliticalscenariosetupbyHelimski(orbyanyotherauthorforthatmater)northepaleographicalanalysisoftheinscriptionshallbediscussedatlargeinthepresentcontribution.Theformerissueseeminglydependsinitsentiretyonthelinguistichypothesisthateachoftheauthorendorses.6Asforthelatter,thetopichasbeenapproachedbyspe-cialistsmuchmorequalifiedthanthepresentauthor(see i.a.Róna-Tas2001).

4 Berta/Róna-Tas(2011:1163)claimthatbothHelimski’shypothesisandFutaky’sHungaro-Tungusica(seeExcursusbelow)havebeenrefutedsomewhereelse(Kara2002andRóna-Tas2003).Unfortunately,thereferencesprovidedbytheauthorsconcernsonlyFutaky’swork,withnomentionwhatsoevertotheworkofHelimskiontheBuylainscription.Asshallbeshown,Helimski’sandFutaky’sworksrequiretheadoptionofdifferentapproaches,putanotherway,theconclusionsofcriticizingone,whilehavinganimpacttocertaindegree,cannotbeuncriticallyappliedtotheother.

5 ThisopinionisechoedinoneoftheobituariesaboutHelimski(Janhunen2009:368–369,seealsoJanhunen2013:55andSzalontai/Károly2013:367).Stachowski(2004)hasspokeninsupportofHelimski’shypothesis,atleastasfarasthepossi-bilityofidentifyingaTungusicsubstratuminEuropeisconcerned,andmadewhatseemsbestuseofitbyproposinganewetymologyofEnglishsabre(←French←German←Hungarian←Avar?←Manchuseleme‘adaggercarriedatthebelt’andotherTungusicrelatedterms).However,thereisnoharmwhatsoeverinremovingtheTungusicsubstratumelementoftheequation.Thisetymologycanbejudgedinitsownmeritsassuming,forinstance,thatsabreisaKulturwort/Wanderwort.

6 Thereisacausalrelationshipbetweenlinguisticsandgeopolitics.Thisfactapparentlyrendersthemincompatiblewitheachother:ifoneclaimsthattheBuylainscriptionreflectsaTurkiclanguage,thenitcannotbearguedthattheTungusicweretobecountedamongthe(Asian)Avars,andvice versa.ButsincetheAvarswereamultiethnicstate,theideathatbothTurkicandTungusicpopulationsmayhavebeenapartofitcannotberejectedoutofhand.Bethatasitmay,itisnecessarytohighlightthatHelimski’s

20 JOSÉ ANDRÉS ALONSO DE LA FUENTE 

Theorganizationofthepaperisthefollowing.AtabulatedsummaryoftheTungusicandTurkicreadingsbyHelimskiandErdal,respectively,isshownin §2.In§§3–7,IshalldiscusstheparticularsoftheetymologicalproposalsformostelementsfoundintheBuylainscription,includingfewgrammaticalele-ments.Theonlyquestiondealingwithmorphosyntax(i.e.thenominativeorzero-accusativeobjects)isanalyzedin§8.ThatsectionisfollowedbyabriefexcursusdiscussingsomeofFutaky’sHungarian–Tungusicetymologies.Conclusionsin§9 closethearticle.

2. The Buyla Inscription

TheNagyszentmiklóstreasureiscomprisedofseveralinscriptions,alloftheminscribedonvessels.Therearethreetypes:GreekinscriptionsinGreekletters,TurkicinscriptionsinGreekletters,andruniforminscriptions(writtenintheso-calledNagyszentmiklós-Szarvasalphabet,aruniform,Semitic-originscript;forfurtherdetails,seeRóna-Tas2001:121–127).TheBuylainscription–discoveredin1799andinitiallycataloguedas“objectnumberXXI(drinkingvessel)”–isusuallyincludedinthesecondgroup,despitethelackofaTurkicreadingforit(forafulllistcontainingalltheexistingproposals,seeGöbl/Róna-Tas1995:9–20,esp.18–19).ThishasencouragedsomeauthorstoproposealternativereadingsmovingawayfromthegeneralassumptionthattheinscriptioncontainaTurkictext.

Theinscriptionreadsasfollows(goodreproductionscanbefoundinGöbl/Róna-Tas1995:TafelXXIV,Kovács/Garam2001:41,Róna-Tas1999:127,GreekalphabetaccordingtoErdal1988:221):

✚ ΒΟΥΗΛΑ•ΖΟΑΠΑΝ•ΤΕCΗ•ΔΥΓΕΤΟΙΓΗ•ΒΟΥΤΑΟΥΛ•ΖΑΠΑΝ•ΤΑΓΡΟΓΗ•ΗΤΖΙΓΗ•ΤΑΙCΗ

Erdal’sproposalofaTurkicreadingforittranslates:“ItisBuylaŽoapanwhocarriesthebowl.Drinkingbowl,madebyButAulŽoapan”(1988:233).Helimski’salternativeTungusicreadingtranslates:“Thebasileusdeclaresthedisplacement

hypothesisiscoherentfromthegeopolitical,historicalandlinguisticviewpoints:itistheoreticallyplausiblethat(1)afractionoftheoriginalRuanruanpopulationwasTungusic,(2)thatfractionmighthavefledwestwardstoEuropeaftertheattackoftheTurkic-Chinesealliancein552,(3)alinguisticsubstratummayhavesurvivedinthelanguagesoftheCarpathianbasinregion.TheplausibilityofthisscenariosetupbyHelimskiiswhatmakesitunique:althougheverythingsoundsperfectlylogical,itistotallyindemonstrableunlessveryconvincinglinguisticevidenceispresented.Inthediscussionbelow,itwillbeshownthatthelinguisticevidenceactuallydoesnotstanduptoseriousscrutiny.

TUNGUSIC AND THE BUYLA INSCRIPTION 21

ofBuylaaszhupan.ThebasileusdeclareshisrecognitionandwatchoverButaulas(thenew)zhupan”(2000b:51–52).

TheTungusicreadingoftheBuylainscriptionproposedbyHelimskiisshowninthefollowingtable,whereA=transliterationoftheGreektext(Helimski2000:44),B=the“actualtext”inSlavicizedTungusicasreconstructedbyHelimski,C=theProto-Tungusicarchetypeunderliningthe“actualtext”,andD=abriefsummaryoftheTurkicreadingbyErdal(1988)7:

# A †B *C DI buila

(-yl-)buyla buyla personal name

[notTungusic]personal name / title

[cf.Buyla Baga Tarkan intheToñukukandBilgäKaganinscriptions]

II zoapan(ž-,j-)

župan župan title title

III tesi (-ī)

täsi tägä-si title [{sit.down-NLZ}]

†täwsi/täpsi{bowl-ACC?}

IV dügetügi(d- -y-,d- -y- -y-; -ī)

dügätägii jügä-t-rä.gii-Ø

{change-HAB-PRT.AOR-3SG}

†yǖd-δök-i{carry-PRT.PST-3SG.POSS}

V butaul butawul buta-wul personal name [{hunt-NLZ}]

personal name [†butogul,cf.But Qaya inUighursources]

VI zoapan(ž-,j-)

župan župan title title

VII tagrogi(-y-, -ī)

tagrogii taag-ra.gii-Ø

{recognize-PRT.AOR-3SG}

†tag-δök-i{make-PRT.PRF-3SG.POSS}

VIII icigi(ī-, -y-, -ī)

icigii icä-rä.gii-Ø

{see-PRT.AOR-3SG}

†iči-y.i<*iči-g-i{drink-DER-3SG.POSS}[izafetconstruction?]

IX tesi (-ī)

täsi tägä-si title [{sit.down-NLZ}]

†täwsi/täpsi{bowl-ACC?}

7 ThereisnospecialreasontochooseErdal’soverthereadingofotherspecialists.ItisrecommendabletosticktotheopinionexpressedbyRóna-TasregardingsomeTurkicreadings,namely“[these]interpretationsarebasedonVilhelmThomsen’swork–heldperhapsinoverlyhighesteem–aswellasaseriesofinconsistentandunfoundedassumptions,soforthetimebeingtheycanbeignored.Allthemoreso,sincewenowknowtheclaspwasattachedtothedrinkingvesselatalaterdate”(2001:129).ThiscommentcanbeextendedtoalltheTurkicreadings.However,thementionofatleastoneTurkicversionmaymakeagoodserviceasamodelagainstwhichtocontrasttheTungusicreadingbyHelimski.

22 JOSÉ ANDRÉS ALONSO DE LA FUENTE 

Ifthedataaretabulatedasintheprevioustable,itispossibletonoticeatonce–obviousdifferencesbetweentheTungusicandTurkicreadingsaside–thatoutoftheninewords,fouraretitles(onlytwo,twicerepeated)andtwopersonalnames.Helimskionlyhastoidentifythemeaningofwhat,inhisview,arethreeverbalforms(withthesameending!)andoneofthetitles.Additionally,heproposedalternativeetymologiesfor(II&VI)and(V)too,thoughhediditratherhesitantly.

3. ‹-ΓΗ›

ThecoreofHelimski’sanalysisliesintheverbalforms(IV,VII–VIII).Thesequence‹-ΓΗ›=†-gi(i)isidentifiedwithawellknownTungusicmorphemewhichplaysacentralroleintheTungusicverbalmorphology:theaoristparticipialending*-rA-gi(i).

Thereconstructionofmedial*-g- intheaoristparticiplemarker*-rA-gi(i) isaproposalusuallyascribedtoBenzing(TVSG128–129:§135[b]).Materialevidence,however,doesnotsupportit.Historicallanguagesshow -ray(i) or -raa,i.e.vowelcluster(hereafterdiphthongoid)orlongvowel.Benzingisawareofthefactthatdiphthongoidsandlongvowelsmaysometimesresultafterthelossofaconsonantbetweenvowels.Somelanguagespreservedthevowelsequenceassuchwiththepossibleinsertionofyod:*aCy> a(y)i,whereasotherlanguagesunderwentcrasis:*aCy>aa.Themostlikelycandidateforaconsonanttobelostinsuchacontextis*-g-,henceBenzing’sreconstruction*-ra-gi(i).

Itiscrucialtobearinmindthatthereisnohistoricaltestimonysupportingthereconstructionof*-g-.ThissegmentispreservedinNorthernTungusic,itslossbeingsystematicinSouthernTungusicandManchuric(seeTVSG29–31:§41).Thisleadstotheconclusionthat,atleastfromamethodologicalviewpoint,thereissomethingwrongwithBenzing’sreconstruction.Takingintoaccountthat(a)itissafetoassumethattheparentallanguagehaddiphthongoidswhichdonotrequirewhatsoevertheassumptionofapreviousstagewithmedialconsonants,and(b)diphthongoidsprovideaconvincingexplanationforsomehistoricallongvowels,i.e.*ay>aa(thisassumptionisalreadyimplicitinBenzing’sreason-ing!),itfollowsthatthereisactuallynoneedorjustificationtopostulatethelossofaconsonantsegmentintheancestralformsoftheendings -raa and -ray(i).Reconstructionendswhenallhistoricalformsareaccountedfor.Thereisnoneedtogoasbackas*-ra-Ci,because -C- doesnotsolveanyproblemregardingthesubstanceofthehistoricallanguages.Benzing’s*-ra-giisagoodexampleofpetitio principii:basedonit,onecouldargueforexamplethatalldiphthongoidssurfacedfromanoriginalsequence*VCV.Allinall,itseemsthatthemosthonest,

TUNGUSIC AND THE BUYLA INSCRIPTION 23

simplestandlogicalsolutionistoreconstruct -raa<*-ray> -ray(i).Benzing’s*-ra-giisaghost,putanotherway,anad hocreconstruction.8

IncaseofacceptingHelimski’sreading,wecouldlegitimatelytalkabouttheTungusic“laryngeal”,somehowechoingthefamedSaussure-Hrozný-KuryłowiczstoryinthefieldofIndo-Europeanstudies,inthesensethatreconstructionhadbeenprecludedbeforematerialevidencecameup.Unfortunately,unlessadditionalevidenceisbroughtup,thisinterpretationcannotbeaccepted.Furthermore,there-constructionof*ra-giposessomeadditionalproblems.Forexample,giventhepresenceof*-g-,itfollowsthattheBuylainscriptionmustreflectatrulyaberrant,archaicformofTungusic,perhapsevenaveryoldstageoftheTungusicparentallanguage.Thisconclusiongoesagainstotherlinguisticfeatureswhicharebetterdescribedasrecent(seediscussionbelow).

4. ‹ΒΟΥΤΑΟΥΛ›

Thesequence‹ΒΟΥΤΑΟΥΛ›hidesapersonalnameaccordingtobothTurkicandTungusicreadings. Helimskiinterpretsthat†Buta-wulcorrespondstoaProto-Tungusicwordmeaning‘tohunt,fish’(SS1.108b).Inreality,*buta- may have meantjust‘tofish’(EEW157–158[1714]),while‘tohunt’istheresultofasecondaryspecializationafterthespreadofNorthernTungusicandManchuricoutoftheAmurianregion.Thisscenariosuitsthesemanticdistributionof*buta-:fishingdominatesintheSouthernTungusiclanguages,e.g.Ulchabüta- ‘tohunt&fish’,butderivatesbüta+la‘fisherman’andbüta-nda- ‘togofishing’,LiteraryUdihebuta- ‘tolayin,store’,LiteraryNanay&Kili&Kilenbota- ‘tofish’.Themeaning‘tohunt’isrestrictedtoNorthernTungusic,seei.a.Ewenkibulta- ‘tohunt’orSolonbülüü- ‘id.’.AsforWMbuta- ‘tocatch(gameorfish)’andbuta-ra niyalma‘hunter’vs.nimaha buta-ra niyalma‘fisherman’,sinceitisnecessarytoaddtheelementnimaha‘fish’tothenounphrasebutara niyalma,itfollowsthatWMbuta- may haveoriginallyreferredonlytohunting(cf.,however,theambiguityinbutha-mbi ‘tohuntandfish’,butha-i niyalma‘hunter,fisherman,sportsman’).ThismaybeconfirmedinthePentaglotDictionary(Wǔtǐ2.663[3036–2]),wherebutambi (buta- plusinfinitivemarker -mbi)correspondstoChinese打牲 dǎshēng and Mongoliangörügele-müi,twotermsreferringexclusivelytohunting,cf.Chinese牲 shēng‘livestock’,Mongoliangörüge(n)‘antelope;game’.TheauthorsoftheSSrightlypointsoutthatalltheEwenkiformsalongwiththeSolonareYakutinorigin,cf.Yakutbultā-­‘tohunt’<Proto-Turkic(+Chuvash)*bul- ‘tosearch,look

8 Althoughaminorquestion,theverbalformations†taag-ra-gï and †ic(ä)-rä-gi are grammaticallyodd:itisalmostcustomarytofindoneormorevoiceormodalsuffixesbetweenthestemandtheparticipialending,asisthecaseoftheimperfective/habitualmarker*-t(i)- in †jügä-t-rä-gi.

24 JOSÉ ANDRÉS ALONSO DE LA FUENTE 

for’(Clauson1972:332a,Sevortjan1978:252–253).HelimskimentionstheEwenkiclan name Buta(n),butthismayreflectanarchaism(theoriginalwordcouldhavebeenentirelyreplacedbythenewbasebulta- ←Yakut)orbeofforeignorigin.

As for the onomastic element -wul,itisfoundinmalepersonalnamesonlyin(some)Ewenkidialects.ItspresenceinEwenisrare,andthereseemstobenotraceofitinNegidal,SolonorArman.9 Therefore -wulcannotbecalledNorthernTungusic,letalonePan-Tungusic,butonly“Ewenki”.AsVasilevičexplains(1974:299–300),theending -(w)ul ~ -(g)ulisalreadyattestedinMiddleMongolian(Secret History: boro’ul ~ boroγul from boro‘grey’),alsointheAmurianregionamongtheDagur(Mongolic).Mongolic*-b- >­-γ-­→Tungusic(Ewenki) -w- is aregularcorrespondenceinloanwords,ashasbeenalreadynotedbyPoppe(1966:189–192,cf.1972:97–98).ThesameelementcanbealsofoundinYukaghir(iso-lated),Nganasan(Samoyedic,Uralic),YakutandDolgan(bothTurkic),andeveninhistoricalrecordsofKott(Yeniseian).Thisdistribution,plusthepresenceof -(w)ul ~ -(g)ulinMongolicandotherlanguagesofNorthernEurasian,pointoutthatthisisratheraforeign,laterelementinEwenki.

InHelimski’sopinion,theBuylainscriptionissomehowclosertoSouthernTungusicandManchuric(see i.a.2000b:53).†Buta- fitsthebill(itcouldpassasatypicalSouthernTungusicelement),butthecombinationofthisbasewiththeonomasticsuffix -(w)ul,onlyattestedinEwenki,diminishesthepersuasivepowerofhisinitialproposal.

5. ‹ΔΥΓΕΤΟΙΓΗ› and ‹ΖΟΑΠΑΗ›

InHelimski’sview,‹ΔΥΓΕΤΟΙΓΗ›=†dügätägi(i)isthecontinuationofPT*jügä(ä)ttägi<*jügä(ä)-t-rä-gi(thereisnoManchuriccognateforthisword).Theassimilationoftheinternalcluster*-tr- isregular,e.g.Ewenkijügäättä<*jügää-t-rä‘notchanging’(negativeparticiple),buttheinitialdepalatalization*j- >d- is clearlyananachronismfromtheviewpointofTungusichistoricalphonology,asdepalatalizationisonlysystematicinOrok.Helimski,whoisawareofthisdetail,mentionsthatthereexistexamplesinWMshowingthisirregularsoundchange.TheseexamplescomefromBenzing(TSVG36:§48).Ifmyreadingofthepassage

9 ClannamesamongtheEwenkiareusuallymarkedwiththesuffix -gir (PL -gil).Vasilevič’slist(1969:262–286)containsnoclannameswith -wul.Negidalalcakul (?),thenameoftheunidentifiedTungusicpeoplearoundtheČarariverbilyakur(?),Ewenki gayul (butcf.Ewenkimaugir ~ maul),Arman‹Gobdzur›= gobjur(?),Birarmalakul,Ewenkioceul,Ewenkitamtakul,Ewenkitonkul,Ewenkixängul and Ewenki ceernoul,allcontainaseriesofelements,namely -ul, -gul and -kul,thatcouldbehis-torically related to -wul.Note,however,thattheetymologyoftheclannamesabovearenotalwaysentirelyclear.

TUNGUSIC AND THE BUYLA INSCRIPTION 25

iscorrect,whatBenzingactuallysaysisthatWMj- >d- whenfollowedby/cjs/(“[…]*ǯ…>d…,wenninWortec,­ʒ,­sfolgen”,seefootnote2aboveonortho-graphicconventions).Thisisacommoncaseofconditionedfricativedissimilation.Thecontextfordepalatalizationdoesnotapplyin(IV).10

ThoughHelimskiseemstoacceptthecommonreadingof‹ΖΟΑΠΑΗ›asSlavic †župan,heoffersanalternativesolution:†jupan<*ju(w)an‘ten’.11 This is an inconsequentialdecision,becauseitcomesintoconflictwithpreviousproposal,e.g.*j- >†d- beforehigh(round)vowelsin(IV)dügetägi,and*-w- >Øin(V)Butaul.Ifindveryunlikelythattwodifferenthistoricaloutcomesoftwodifferentseg-mentsmaycohabitintheverysametext,especiallywhenthetextismadeupoftwosentences,andsupposedlycarvedbythesameperson.Furthermore,theresult*-w- >†-p- isunheardofinTungusic.Thealternation -p- ~ -w- is common only inNorthernTungusic,anditcanbealwaystracedbacktoanoriginal*-p- (TSVG32–34:§44).ThisfactaloneagaincontradictstheSouthernTungusicpedigreeoftheBuylainscriptionasassumedbyHelimski.

6. Slavicized Tungusic

AnotherveryimportantpillarofHelimski’shypothesisisthepresenceofSlavicindividualsintheverysamespotinhabitedbytheAvarsintheCarpathianbasinand,mostimportantly,thelinguisticinfluencetheymayhaveexertedontheAvar.ItisinthiscontextthattheallegedsoundchangeProto-Tungusic*ā&*ă>BuylaTungusica&omakessense,sincethesamedevelopmenthasbeende-scribedforCommonSlavic,e.g.Turkic*tavar[=ta var]→Proto-Slavic*tăwārъ>CommonSlavic*tovarъ‘good,commodity’(cf.Polishtowar,Russiantovár,Slovenian tóvor).Thesesoundchangeswouldaccountquiteelegantlyfor(VII),i.e.SlavicizedTungusic†tāgrăgī>ΤΑΓΡΟΓΗ(Helimski2000a:48)aswellas

10 NotethatWMduksi<*jüxi+ktä,theonlyexamplesuppliedbyBenzing(TSVG42§55,withaquestionmark!)andreproducedbyHelimski,maybeanintra-borrowingfromAmurianTungusic(itisnecessarytoassumemetathesisinWM,i.e. -ks- <*-sk- <*-xĭkt-),cf.Ulchajüstä,LiteraryNanayjusiktä,Orokdusiktä(SS1.256b).Thediagnosticfeaturepreciselyisthedepalatalizationofinitial*j(regular,non-conditionedonlyinOrok).Itmayalsobepossiblethatnoneofthesewordsarerelated.Thepreservationoftheconsonantcluster -ks- pointsoutthatthissequencemaybepartofthebase,ratherthantheremnantofasuffix,e.g.PT*tüksa‘housecovermadeofbirchbark’>Ewenkitiksa,WMtuksa in the collocation tuksa boo‘housemadeofbirchbark’(SS2.179a).Moreover,thenominalsuffix*ktAsystematicallyyieldsWM -hA(asinUdihe),e.g.Proto-NorthernTungusic*tii-lä- ‘tosearchforlice’>Ewenkitiilä-id.~Proto-SouthernTungusic*ti(i)+ktä‘louse’>Ulchatiktä,Nanayciktä,WMcihe(SS2.179a,181b).

11 Alemany(2009)hasrecentlyproposedthatžupan­mightbeablendformation,thefirstcomponentcorrespondingtoChinese州 zhōu‘regions’.

26 JOSÉ ANDRÉS ALONSO DE LA FUENTE 

anotherlexicalitemwhichdoesnotbelongtotheBuylainscription,namely†Boyar-inъ,PL†Boyare<SlavicizedTungusic*băyā.r(ъ),partlypreservedintheEwenkiclan names Boyar ~ Buyar(Helimski2000b:144–146).12 One wonders why this so regularlyappliedinthesetwowords,butitfailedin†Bută(w)ul>**Butowul(?)or †Apă.r>**opar(?)‘Avar’(intheory,apluralformationbuiltontheSouthernTungusicbase*apa- ‘toattack,assault’>Nanayapa- [→intra-borrowinginHailarSolon apa-ldi- id.],Kilen&WMafa- id.,theonlynominalderivateisWMafan ‘battle,fight’,cf.SS1.47a,seebelowforfurtherdetails).

Inthisconnection,HelimskisuggestedthatSlavicizedTungusic*băyā.r(ъ) is apluralformation:SG*baya.n‘rich,wealthy’vs.PL*baya.r(butWMbaya-sa).13 Theidentificationof*băyā.r(ъ)asTungusicsoundsnatural.14Note,however,thatHelimskihimselfmentionsthattheEwenkiclannamesBoyar ~ Buyar may be connectedwithWMbayara‘guard,troops(oftheEmperor)’(seeSS1.65b).Thelat-terhowevercontainsnopluralmarker,butthemarkeroftheso-calledaoristorimperfectparticiple -ra.InanattemptatsavingtheTungusiclink,onecouldargueasanalternativeexplanationthat*băyā.r(ъ)isnotapluralformation,butactu-allyaparticipialformation,thechangeSlavicizedTungusic*-ra→Proto-Slavic*-rъbeingtheresultofanalogicalreadjustmentsinspiredbyloanwordssuchasTurkic*tavar→Proto-Slavic*tăwārъ.Unfortunately,HelimskifailedtonoticethatalreadyVasilevič(1969:263–264)consideredthoseEwenkiclannamesasmerevariantsofthemoretransparentBayagir ~ Buyagir ~ Boyagir,withsporadic

12 TheoriginofthevocalisminBoyar ~ Buyarremainsunexplained(folketymologyandblendingwithRussianbogáč‘richman’,bogátyj‘rich’andperhapsbogatýr’ ‘hero (inRussianfolklore)’?).Theresemblancewiththevoweldevelopmentinotherwellknown“Altaic”loanwordsintheSlaviclanguagessuchasPolishkozak←Ukrainiankozák~Russiankazák‘Cossack’←CommonTurkickazak‘freeman,vagabond’cannotbedenied.However,thereisnounanimousaccountforthe/o/~/a/vowelalternationinthisword(thoughcommonlytreatedasaninternalSlavicprocess,itmayhavealreadybeenpresentinTurkic,seei.a.Pritsak2006:241fn.4paceDoerfer1967:462–468§1479;foradditionalexamplesandsomegeneralremarkson“Altaic”loanwordsintheSlaviclanguages,seeStachowski2005),thereforeitisunclearwhetherthesameexplanationcanapplytobothkozak and Boyar.Thisissuerequiresfurtherinvestiga-tion.IamindebtedtoDr.TomaszMajtczakforbringingthisfacttomyattention.

13 AsfarastheCommonTungusiclanguagesareconcerned(Manchurichasreducedtheoriginalsystemalmosttozero,cf.TSVG76–78§87),rulesregardingtheforma-tionofpluralsarestraightforward:basesendingin -n take -r,otherwisetheytake -l (thisincludesvowel,y-,l- and r-bases),e.g.Lit.Ewenkiurä‘mountain’⇉ urä-l,adil ‘net’⇉ adil.i-l&bur‘island’⇉ bur.i-l (i-epentheticvowelinsertion),gujäy‘pretty’⇉ gujäyl,oron‘deer’⇉ oro-r.Exceptionscoverkinshiptermsandcollectives.

14 IntheMongoliclanguages(theyarealsoinvolvedinthediscussionregardingtheeth-nolinguisticaffiliationsoftheAvars),thecorrespondingpluralformationis*baya.d (T-pluralsareonlyattestedinMongolic,SamoyedicandSogdian,thefewcasesinTurkicbeingmostlikelyofMongolicorigin,seeSinor1952,Poppe1977).

TUNGUSIC AND THE BUYLA INSCRIPTION 27

loss of -g- betweenvowels,perhapsundertheinfluenceofSouthernTungusicorevenManchu(ric).ItmaybeworthnotingthatWMbayara also refers to one of theoldestclansamongtheManchu(Shirokogoroff1924:20).

7. ‹ΤECΗ› & ‹ΤΑΙCΗ›

Paleographicdiscrepanciesaside,Helimskiproposesthat(III)&(IX),†täsi in hisreading,arethecontinuationof*tägä+si,withabasemeaning‘tosit’(thisisaveryregularbasewhichisusedeventocoinneologismsformodernutensils,e.g.Orok tääk(k)u‘chair’,seeOzoliņa1995:98,Ozolinja2001:364)andtheformant si correspondingtothewellknownnomina actorissuffixinWM.InManchugram-marsthissuffixisusuallylistedalongwith -msi, -ci and -(m)ji(seei.a.Zakharov20102:§42[10]),e.g.kumun‘music’⇉ kumusi‘musician’,adun‘herd,swarm’⇉ aduci‘herder’,boigon‘family’⇉ boigoji‘host,master’,butu‘dark,hidden’⇉ butumji‘cunning,deceitful’,taci- ‘tolearn’⇉ tacimsi‘student(oftheImperialAcademyofLearning)’.Theyshouldnotbecalledallomorphsbecausethedetailsconcerningthecriteriafortheirdistributionareunknown.Perhapsmoreobvious,however,isthefactthatthevariant -(m)jibelongswithLit.Ewenki -mdii,Negidal&Udihe -mni ~ -mdi,Ulcha -mdi ~ -mji,Oroch&Orok&Nanay -mji,allgoingbacktoCommonTungusic*mdi(i)(TSVG64:§75[d],65:§76[d]s.v.*-mgi,Sunik1982:92–100,Boldyrev1987:53–57).15AsforthevariantsWM -si and -ci,theyareofMongolianorigin.Theformervariantunderwentphoneticnaturalization,showinginconsequencethediagnosticsoundcorrespondenceWM/š/(/si/=[ši])vs.Mongolian/c/(Doerfer1985:177–179).ThelatterreflectstheMongolic -ci suffixassuch(seei.a.Poppe20062:40–41:§118;1987:274–275:§227).

Allinall,itseemsmandatorytoassumethatthelanguageoftheBuylainscriptioniscertainlyveryclosetoManchuric,forthisistheonlyTungusiclanguageshowingtheresult si.However,thisisanachronisticiftakentogetherwithotherfeaturesaspresentedbyHelimski.Forinstance,itisremarkablyincon-sequenttoarguefortheretentionof*-g- in[1],and,atthesametime,toassume

15 InspiteofSunik’sefforts,Ewen­-mŋaamaynotbelonghereafterall.ContrarytowhatBoldyrevclaims(1987:56),Benzing(1953:113,TSVG56§69)neverproposedthat*-mdi(i)wasoriginallya“suffixednoun”meaning‘personconnectedtoanob-jectoranaction’(ifsomething,thatisBenzing’sprovisionaltranslationforasuffixwhichappearswithbothverbalandnominalbases!).Sunik,followingthetracksofG.Ramstedt,proposesthatthissuffixcanbesegmentintothedeverbalnoun*-m- + PT*gäy‘second;another’or*näri‘humanbeing’,whereasBoldyrev(1987:56–57),whilerejectingallthepreviousattempts,proposes*-mari~*-masior*-Buri (where B=/wbpm/)withnofurtherinsightsregardingtheexactvalueorlaterevolutionofeachofthecomponentshesetup.Thoughitisirrelevantforpresentpurposes,noneofthesehypotheseshasbeenacceptedsofar.

28 JOSÉ ANDRÉS ALONSO DE LA FUENTE 

itsloss.Needlesstosay,Helimskiiswellawareofthisfact.Inordertoame-liorateitsseriousness,headds:“[i]tcanbeonlyremarkedthatproblemsofthatkindfrequentlyaccompanytheattemptstogiveanaccuratedescriptionofthereflexesof“weak”consonantsinTungus-Manchurian(andinotherAltaic,aswellasnon-Altaic)languages.Justoneexample:T[ungus-]M[anchurian]*daga‘root’isattestedinNegidalasdā,thoughnormally -g- ispreservedhere(TM*daga ‘near’>Negidaldaγa)”(2000b:50fn.9).Fromsuchapicture,onemaythinkthatthesituationinTungusicischaoticandthereforeoneisatlibertytochoosewhicheverscenariomaybemostconvenient.

NotwithstandingHelimski’sagileresponse,thesituationisactuallynotcha-otic.Asamatteroffact,theexamplebroughtupbyHelimskiisproblematic.OneofthemostsalientisoglossessettingapartUpperandLowerNegidalisthefateoftheparentallanguagevelarplosives:PT*-k- yieldsUpperNegidal -k-,butLowerNegidal -x- (Myl’nikova/Cincius1931:133),andPT*-g- is retained in the former,butlostinthelatter(accordingtoXasanova&Pevnov2003:6–7,thelossof -g- alongwith -y-, -w- and­-ŋ-,ismostnoticeableintheUst’-AmgunsubdialectofLowerNegidal).16Negidaldaga‘near’isattestedinbothUpperandLowerdialects,thereforethereisroomtospeculatethatspeakersofLowerNegidalmayhavesecondarilyadoptedthecorrespondingUpperNegidalform(forexample,throughmixedmarriages).PT*daga‘root’isonlyattestedinLowerNegidalanditshowstheregularlossof*-g- betweenvowelsasexpected.

AgainstthepreconceivedchaoticsituationarguedforbyHelimski,itmayalsobementionedthatthedistributionofthisisoglossinNegidalisnotfortuitous.Asiswellknown,UpperNegidalislinguisticallyandculturallylinkedtotheEwen,i.e.aNorthernTungusicpeople,whereasLowerNegidalisverycloselytiedtotherealmoftheAmurian(=Southern)Tungusicpeoples(see i.a.Xasanova&Pevnov2003:228–229).Generallyspeaking,PT*-g- continuesintactinNorthernTungusic,butitissystematicallylostinSouthernTungusic(TSVG29–31:§41),thereforetheisoglossintheNegidaldialectshasanarealbasis.Allinall,Helimski’s

16 Schmidt(1923:8),followingtheadviceofcertainMr.K.D.Loginovskij,dividedtheNegidallanguageterritoryintofourmajordialectalareas:Amgun’suppercurrent,middlecurrent,lowercurrent,andthatofthevillageTyr(virtuallyidenticaltosomeEwenkivarieties).Tothispicture,somelaterauthorsaddedSamagir(see i.a.Doerfer1978:10),avarietyspokeninthevalleyoftheGorin,thelefttributaryriveroftheAmur,veryclosetotheAmgun,andhasbeensincecharacterizedas“nanaizedNegidal”(see i.a.Ligeti1953;anaberrantNanaydialectisspokenintheverysamevalley,seePutinceva1954).SamagirandTyrare,inoverall,transitionaldialects,oneleaningtowardsSouthernTungusic,theothertowardsNorthernTungusic.Currentviews,however,givemuchcredittothebinaryclassification,especiallyaftertheworkofMyl’nikovaandCincius(1931,seealsoCincius1982:17–19),whichweadopthereforthesakeofclarityandsimplicity.

TUNGUSIC AND THE BUYLA INSCRIPTION 29

appreciationaboutthechaotichistoryofthe“weak”consonantsintheTungusiclanguagesisanoversimplification.

As for the fact that †täsi>†tesiisactuallynotdocumentedinthe(very)extensiveManchucorpus,Helimskiqualifiesitas“[…]afactthatcanhardlybesurprising,becauseallpoliticalterminologyofthislanguageconsistsofChineseandMongolianloan-words.”Onceagain,thisstatementisnotentirelyaccurate:althoughitistruethattheadministrativeandpoliticalterminologycomesfromChineseandMongolian,eachtermhasitscorrespondingManchutranslation.Interestinglyenough,theBeijingpentaglotdictionary(Wǔtǐ­qīngwénjiàn) contains around580termsbelongingtothisverysemanticfield(Wǔtǐ1.107–138)andnoneofthemresemblestheoneofwhichHelimskiisinneed.17

8. Some thoughts on morphosyntax

TheBuylainscription,accordingalwaystoHelimski’shypothesis,containstwoprototypicaltransitivesentences.ThetypologicalprofileoftheTungusiclanguagesdeservesnospecialtreatment:theyareagglutinative,SOV,andac-cusative.Ifthesecharacteristicsareborneinmind,theonlystrikingfactintheBuylainscriptionisthelackofdirectobjectmarkersin†taagragii,†jügäträgii and †butawul.However,Helimski(2000b:51)mentionsthattherearenumerouscasesofzero-accusatives(=nominativeobjectsinHelimski’swording)inNanay,especiallywhenthepatientisundetermined(=indefiniteobjects).Regretfully,Helimskiagainsimplifiesahighlycomplicatedissue.Inreality,therearetwodifferentissueswhichneedtobedealtwith:(a)actualzero-accusativesand(b)theso-calledindefiniteaccusativeordestinative.

a) Nominative objects or zero-accusativesTheexistenceofnominativeobjectsorzero-accusativescanbesupported

withextensivedocumentation.HelimskiprofitsfromAvrorin’sresearchonNanaysyntax.Avrorin(1981:155–158)explainsthattherearenominativeobjectsal-readydocumentedin19thandearly20thcenturysources(e.g.Protodiakonov,Dobrolovskij).ThesesourcesactuallyreflectKilen(=SungariNanay),alanguagewhosegrammaticalstructurestandsveryclosetoSouthernTungusic(see i.a.

17 TheManchuworddasan‘rule,government,control’,whichlooselyremindsof†täsi,istraditionallyexplainedasa(recent)Mongolism,cf.Manchujasak‘chiefofaMongolbanner’afterdissimilation(Rozycki1994:121;seesomeremarksonthesoundchangein §5 above).SSeditors(1.201a)mentionChinese治 zhì ~ chí ‘towork,make,regulate,govern>well-governed,ingoodorder’,whichtheyderivefrom*dai(Pulleyblank1991:56,408;Schuessler2007:619).

30 JOSÉ ANDRÉS ALONSO DE LA FUENTE 

Sem1976:14–24,Janhunen1996:61–61,Doerfer1977:57,60:§2).18Whythisisrelevant,Ishallexplaininthefollowingparagraphs.Interestinglyenough,nei-therAvrorinnor,consequently,Helimskiprovideexamplesofthisphenomenon.ThetwofragmentsbelowweregatheredbyI.A.DobrolovskijandpublishedbyKotvičъ(=WładysławKotwicz).Thefirstexample(simplesentence)showstheuseoftheovertaccusativemarker.Thesecond,however,showsinthefirstplacetheovertaccusativemarkerinacontextwhereitisexpected,butafterwards,anominativeobjectisfoundinstead19:

Kilen (1) ‹Aбкаендурiбаwöоwоĭ·дунiджолодзiоwо·ханi.› heavenspiritsky-ACCset-PRT.AOR-LOC-3SG.POSSstone-INSTset-PST-3SG ‘Whenthecelestialspiritmadetheheavens,hemade(it)withstone(s).’ (Kotvičъ1909:217,TextIII,lines1–2) (2) ‹“неубiемубаду,емубансiджефахаi·,хесумi·ачарсiнсакдi”.–синидже-

фахасиукесакдi?–“Мiнiджефахаi·тенi·сакдi:бунiнгунбöjеемубансiwöджефемiилан·iненгiилан·долбоджефехаĭкоптомöнiхафоlдi·ларцiн.[…]”›

young.brotherIoneplace-LOConepelmeneat-PST-1SGsay-CV.SGnot.possiblebigyou.GENeat-PST-2SGsobigI.GENeat-CV.PLthisbigwesixpersononepelmen-ACCeat-CV.SGthreedaythreenighteat-CV.PLjacket-3SG.POSS-EMPHpierce-NEG-PST.3SG

‘“(Young)brother,atacertainplaceIateonepelmen[akindofravioli],sobig,onewouldsayitcannotbe(that)big”,–“Thatwhatyouate,howbigwasit?”,“WhatIatewasthatbig!We(were)sixmen,wehadonepelmenforthreedaysandthreenights,andthejacketdidnotget(even)pierced!”’

(Kotvičъ1909:218,TextIV,lines4–10)20

18 Foralongtime,itwascustomarytotreatsomeaberrantvarietiesofAmurianTungusicasNanaydialects.ItwasafterDoerfer’sworkthatNajxin(=Literary)Nanay,Kilen(=BikinNanay)andKili(=Kur-UrmiNanay),togetherwithHezhe(n),areconsideredautonomousvariants(seeJanhunen2012:16,items[13–15]inAppendix,andAlonsodelaFuente2011b,esp.pp.15–17forsomelinguisticfeaturesdistinguishingLiteraryNanay[Janhunen’s“properNanai”],KilenandKili).

19 BothexampleshavebeenLatinized(theywereoriginallywrittenintheCyrillicscript)andtranslatedintoGerman(Walravens1992:1–12,esp.10and11,respectively).

20 Thisfragmentcontainssomeobscurewordsdeservingclarification.BasedonthetranslationprovidedbyKotwicz,itispossibletospeculatethatthemodifier†uke correspondstoKilenäkä=LiteraryNanay äyä‘this(close)’(seeAvrorin1959:269–270,Sem1976:61–62;oneistemptedtomentionManchuuhe‘(comm)unity,unified;atone,inconcert;ingeneral,onthewhole’ortheemphaticweke‘heyyou!’).ApparentlyisolatedwithinTungusic,theverbalbase†foldi.la- may be related to Manchufolo-mbi‘tocarve,engrave’orfolko-mbi‘toleaveapace,makeaninterval,makeapause’.†täniicorrespondstotheemphaticparticle=tAni which is commonly attachedtothedesiderativeverbalmood(Avrorin1961:267–268).†koptomo-ni-ha (with‹h›=[γ]<-k-?)containstheemphaticparticle=kA(Avrorin1961:268),thoughonewouldnotexpectithere.

TUNGUSIC AND THE BUYLA INSCRIPTION 31

Ononehand,example(2)showsthetypicalcontextwheretheACCmarkerisdropped:directobjectsreferringtogenericentitiesordirectobjectswhicharementionedforthefirsttime(‘onepelmen’).Ontheotherhand,examples(1)andpartly(2)makeclearthatdefinitenessisthefeaturethattriggerstheobligatorypresenceoftheACCmarker(‘theheavens’,‘thepelmen(thatone,thebigone)’).AsfarasManchuricisconcerned,thepresenceoftheACCmarkerbe isobliga-toryinClassicalManchuonlywhenthedirectobjectofatransitiveverboccursinadistantpositionorinnegativeutterances.However,indeclarative,affirmativesentences,iftheobjectoccursimmediatelybeforethegoverningtransitiveverb,theACCmarkercanbeleftout.ItisalsoclaimedthatthepresenceoftheACCmarkergreatlydependsonwhetherthedirectobjectreferstospecificentities(seei.a.Tamura1990,Gorelova2002:170–172,Larsen2007).

ThetwosentencesoftheBuylainscriptionseemtofulfilltherequirementsfordefiniteness,butwehavenoACCmarker.Itsabsencethereforemayhaveadifferentmotivation.InUdihe(see i.a.Nikolaeva&Tolskaya2001:120–123),forinstance,thecircumstancesunderwhichtheaccusativemaysurfaceasthenominative,namelywiththezero-marker,are(1)phonetic,e.g.allegropronuncia-tionoftheaccusativemarkerafterbasesendingin/o/,/u/,/wa/,/fa/,e.g.au(-wa) ‘cup’,iŋofo(-wo)‘birdcherrytree’mäwa(-wa)‘heart’,mafa(-wa)‘oldman;bear’,orbeforewordsbeginningwith/wa/(especiallytransitiveverbs),e.g.wa- ‘tokill’,or(2)semantic,e.g.withnon-specificpatients(massorgenericnouns)orifthecorrespondingparticipantisbeingintroducedinthetextforthefirsttime(seeKilenexample[2]above).CoulditbethattherearenoACCmarkersintheBuylainscriptionbecausethedirectobjectsrefertoentitiesmentionedforthefirsttime?Thisisveryunlikely,because“thefirsttime”contextrequiresthatthedirectob-jecthastobementionedonseveraloccasions,asintheKilenexample(2)above,where‘pelmen’appearstwice:thefirsttimewithouttheACCmarker,butwithitattachedthesecondtime.

Fromtheperspectiveoftextualtypology,theclosestparallelandmodelfortheBuylainscriptionaretheJurchenmemorials.Jurchen,togetherwithManchuandSibe,constitutestheManchuric(orJurchenicinJanhunen’sterminology)branchoftheTungusiclanguagefamily.TheselanguagesareveryclosetotheSouthernTungusicbranch,hencethepertinenceofthecomparisoninregardstoHelimski’shypothesis.TheJurchenmemorialshaveaveryrigidformulaiclanguageasshowninthefollowingexamplebelongingtotheTōyōBunkocollection(Kiyose1977:208–209[MemorialXVII];Ifollowhistransliterationandtranslation):

Jurchen †haisigitanweidujihuwijaligijejimei/jaulamaiahaicinhojuwasunjaaniyajuwajuwe/biyajuwauyuninengibahabidiejeheiweilebe/teeaniyasaladawajirbiyejuwii/cuyanhadigunširaruahaibahabidi/ejeheiweilebejaulamaibahabi/aciburuhagannisahi.

32 JOSÉ ANDRÉS ALONSO DE LA FUENTE 

HaisiGitanguardregionalcommissionerJaligibe.awed-CV.AORmemorial-DER-CV.AORslaveCh’eng-huatenfiveyeartentwomonthtenninedayget-CV.PRF“of”21postthing-ACCnowyearagedfinish-DERexistsonCuyanhacome-DERinherit-HORslaveget-CV.PRF“of”postthing-ACCmemorial-DER-CV.AORget-CV.PRFdivineemperorknow-PRT.PRF

I,Jaligi,RegionalMilitaryCommissioneroftheGitanGuardofHai-hsi,respect-fullymemorializeconcerningthepositionwhichIobtainedonthe19th day of the12thmonthinthe15thyearoftheCh’eng-huaperiod.Ihaveagednowandhaveason,Cuyanga.Lethimcomeandassume(my)position.WouldthattheDivineEmperorbutacknowledgemypetition.

Notethatthedirectobjectejehei weile‘position’appearstwicecarryingtheACCmarker,irrespectiveofwhetheritoccursimmediatelybeforethegoverningtransitiveverboritsreferenthasbeenmentionedforthefirsttime.ThisfactcastssomedoubtsonHelimski’sstatement.

b) Indefinite accusativeTheindefiniteaccusative(IND.ACC,Russianвинительныйнеопределен-

ныйпадеж)couldshedsomelightonthequestionraisedbyHelimskiregardingzero-accusatives.SomeauthorsnowarguethattheIND.ACCisconnectedtothedestinativeordesignativecase(DES,Russianназначительныйпадеж;notethatinhiscomparativetreatise,Benzinglocatesitsub“Partitiv”,cf.TSVG81–83:§92,Menges1952,Kazama2012).

IntheNorthernTungusiclanguages,theso-calledindefiniteaccusativesignalsanunknownorindefinitedirectobjectanditisobligatoryinnegativetransitivesentences(oneofthetwoobligatorycontextsinwhichtheACCmarkerisobliga-toryinManchuricorNanay).Thedestinative,ontheotherhand,“[…]designatesathingwhichispreparedforthefutureuseofthepersonspecifiedbythepersonsuffix,sothispossessiverelationiscalled‘futurepossession’.Syntacticallythedesignativecasefunctionsliketheaccusative[…]”(Kazama2012:124).TheDESappearsalwayswithpossessiveendings.Incontrast,iftheIND.ACCisfollowedbypossessiveendings,itcarriesthesamefunctionastheDES.WhenbothIND.ACCorDESandACCco-occurinthesamesentence,thelatersignalsmoreab-stract,secondaryobjects,e.g.

Ewenki (1) jäwgää-yä gamii, aya bi-mcä. food-ACC.INDtakegoodbe-COND ‘Itwouldbegoodtogetsome(any)food.’ (LiteraryEwenki;Bulatova&Grenoble1999:9)

21 ThepresenceofthiselementisusuallyexplainedastheproductofChineseinfluence(calques),e.g.†bakjumeidi‘tobehostile’istranslatedinChinese對敵 duìdí ‘to con-front,facetheenemy’,lit.‘opponent’+‘tobehostile’,cf.Manchubakcin‘opponent,

TUNGUSIC AND THE BUYLA INSCRIPTION 33

(2) Kïldïnakaan mata Mängunkaan ahaatkaan-ma ahï-ya-wï ga.da-n. KïldïnakaanheroMängunkaangirl-ACCwife-IND.ACC-REF.SGtake.

AOR-3SG ‘TheherocalledKïldïnakaantookaswifeagirlcalledMängunkaan.’ (SakhalinEwenki;Vasilevič1936:106[line47])

Nanay Si ami-si bu.ji-i juliälä-ni täy äktä-wä asi-go-a-si ga-ci-ni. youfather-2SG.POSSdie-PRT.AOR-POSS.REFbefore-3SG.POSShere

woman-ACCwife-DES-OCM-2SG.POSSbuy-PST-3SG ‘Beforehepassedaway,yourfatherboughthereawomanaswife.’ (Avrorin1981:159)

Generallyspeaking,thedistributionoftheIND.ACCandDESmarkerswith-intheTungusiclanguagesiscomplementary(cf.Kazama2012:143):NorthernTungusiclanguagesexhibitIND.ACC,22whereasSouthernTungusichaveDES.Manchurichasnoneofthem,thelossofthismarkerbeingtraditionallyascribedtoMongolicandChineseinfluence.ItfollowsnaturallythattheIND.ACCandtheDESmayhavesharedthesamesourceinthedistantpast.ThefollowingtablesummarizesthedistributionoftheIND.ACCandDES:

Language Ending Description ReferenceEwenki Cº-a ~ Vº-ya Indefinite

accusativeKonstantinova(1964:49)

Negidal ClassI(Vº) -ya,e.g.joo‘house’⇉ joo-yaClassII(/gylmŋ/º) -ya ~ -ña,e.g.laaŋ

‘trap’⇉ laaŋ-ñaClassIII(/kxpts/º) -ya,e.g.es‘larch’⇉

ees-ya ClassIV(nº) -a,e.goyon‘deer’⇉ oyon-oPL -la,e.g.joo-l-la

Indefiniteaccusative

Cincius(1982:27,Table I)

Solon Cº-a~Vº-ya IndefiniteaccusativePartitivus

Poppe(1931:113:§31)Hú/Cháokè(1986:25)

Ewen ClassI(Vº) -ga-ClassIIa(C[-strong]º) -ga- ClassIIb(C[+strong]º) -ka- ClassIII(nº)­-ŋa-

+POSS.REF Destinative Novikova(1960:188–195)

oppositeside’.TheChinesetranscriptionof†bakjumei di has 的 de for the last element (Kiyose1977:141[797],fn.326and327).ItisworthnotingthatChinese的 de is placedbetweentheobjectanditsgoverningverbwhenusedasemphaticparticle.

22 NotethatinLiteraryEwenki,theIND.ACCmarkeris -(y)a,andtheregularACCis -wa.TheformerandtheManchuricACCmarker(-)bearecognates(seei.a.TSVG80–81§91).

34 JOSÉ ANDRÉS ALONSO DE LA FUENTE 

Language Ending Description ReferenceOroch ClassI(Vº) -yaa- ~ -laa-

ClassII(nº) -naa-ClassIII(gº) -laa-

+POSS.REF Destinative Avrorin/Boldyrev(2001:115–118)

Udihe -na- +POSS.REF Destinative Nikolaeva/Tolskaya(2001:126–127)

Kili -na- +POSS.REF Destinative Sunik(1958:72)Orok -ddoo- +POSS.REF Partitive Petrova(1967:51–52)Ulcha POSS -ju-

SG.REFL -ju-ïPL.REFL -ju-wa-r(i)

+POSS.REF Destinative Sunik(1985:34–35)

Kilen ClassI(Vº) -go-ClassII(nº) -(ŋ)go-

+POSS.REF Destinative Sem(1976:40–41)

Nanay 1&2SG&PL.POSS -go-a-1SG.POSS -go-i-wa1PL.POSS -go-po-SG.REFL -go-i PL.REFL -go-a-ri

+POSS.REF Destinative Avrorin(1959:179)

Thehistoryofthiscaseisfairlycomplicated.TheEwenkimarker -(y)agoesalongwithSolon&Negidal&Oroch -(y)a.BenzingexplainsthatUdihe(&Kili) -(n)a (POSSpersonalendings)istheresultofreinterpretingthefinalsegmentofn-basesaspartofthemarker,i.e.CVCVn-a>CVCV-na.23 Since deaffrication is regularinOrok,itispossibletolinkUlchaandOrok.Benzing,however,providesnoreconstruction,mostcertainlybecausehefindsverydifficulttoreconcilethey-endingswithEwen -ga (POSSpersonalendings),Nanay -go- andUlcha-ju-.HadtheoriginalCTendinghave*-g-,SouthernTungusiclanguageswouldhaveyieldedØ.Moreover,thesoundchange*g>/j/isirregularintheTungusiclan-guages,thereforethereislittletorecommendintheexplanationthatUlcha&Orok/j/isadevelopmentofEwen&Nanay/g/.ThepluralformationinNegidalpointsoutthatthereisaconsonantassimilatedtothepluralmarker/l/,e.g.joo-l-la<*joo-l-Ca,butthisisnotreflectedinEwenki,e.g.oro-r-o{deer-PL-DES},withregularCº-aending.ItissafetoassumethatNegidalelusive“C”is/y/.

23 Kazama’sskepticismaboutthevalidityofthisexplanation(2012:144)isunfounded.Suchaprocess,i.e.reinterpretationofmorphemeboundaries,cross-linguisticallyisverycommon.TheoriginalorprimarystageispreservedintheallomorphyofOroch&Ewen.ItispossibletospeculatethatUdihe&Kili,duetoexternalinfluences(Mongolic,Chinese),generalizedthen-allomorph(thegeneralization,orsimplifica-tion,ofallomorphyisanotherfairlycommonprocess,seetheIND.ACCinEwenki&Solon,or,withmoreprofoundconsequences,thehistoryofManchuric,wherethereisalmostnoconsonantalallomorphy).

TUNGUSIC AND THE BUYLA INSCRIPTION 35

Fromafunctionalviewpoint,Kazama(2012:146–147)convincinglyarguesthatIND.ACCisoriginal,theDESbeingalaterdevelopment.Unfortunately,hedoesnotelaboratefurtheronthedetailsofthisfunctionalevolution,nordoesheex-plaintheformaldiversityofthemarkerinthehistoricallanguages.ToaccountforEwen -ga,Malchukov/Nedjalkov(2010:347–349)proposedagrammaticalizationscenariowherebythesourcestructure,inoriginaserial-verbconstruction,wouldrequirethepresenceofanon-finiteformoftheverbga- ‘totake’withasubordinateobject(Malchukov’s‹h›=ours‹x›):

Hinturki-ga-semu-re-m<

*[Hinturki(-w)ga-ga-s]emu-re-myoursledge-DES-2SGbring-AOR-1SG

yoursledge(-ACC)take-CV-2SGbring-AOR-1SG

Thisingenioussolutionhassomeadvantages:Ewenga- isthecontinuationofaProto-Tungusicverbalbase*ga- ‘totake’(SS1.133–134),thereforeitispossibletoproposethatthesourceconstructionwasalreadyusedintheparentallanguage.IwouldaddthatthegrammaticalizationdescribedbyMalchukov/Nedjalkovmayhaveinvolvenotonly*ga-,butalso*gaju- ‘tobringortakesomethingback’.Thiswordwhoseetymologyisnotentirelyclear(traditionalaccountsrelatedthelastsegmenttoSouthernTungusic+Manchuric*ji- <*di- ‘tocome’,seeSS1.255a)hasalsoProto-Tungusicpedigree:

PT(SS 1.133–134,

EEW [3932, 3940])

*ga- ‘to take’ *ga.ju- ‘to (go to) bring or take something back’

Ewenki ga- ga.ju-Ewen&Arman ga- ga.jï- Negidal ga- gajï-Solon ga- gajüü-Oroch ga- gay- ‘tobring’Udihe ga- gaji- ~ gaju-,cf. gagi- ‘togather,takeback’Nanay ga- gajo-,cf. gago- ‘togather,takeback’Kili ga-Kilen ga- gajï-Ulcha ga- gajü-Orok ga- gasü-WM gai-,imperative

gaisu24gaji- ‘tobring’,imperative†gaju ~ gaji (cf.Sibe gaju- ‘tobring’)

24 IshallelsewheredealwiththeoriginsofManchu-i in gai-aswellastheimperativegaisu.

36 JOSÉ ANDRÉS ALONSO DE LA FUENTE 

Thereisnothingludicrousinproposingthat*gaju- couldaccountfortheformaldiversityoftheIND.ACCandDESmarkers.Thishypothesisshouldbeabletoovercometwophoneticobstacles:(1)theanomalousdistributionof/g/ in theonset(thesoundchange*-g- >ØisregularonlyinSouthernTungusicandManchuric),and(2)thebackarticulationofthevowelinSouthernTungusic.Theinitialautonomyof*ga- wouldexplainwhy*ghasnotbeenregularlylostinSouthernTungusic(asamatteroffact,thishappenedonlyinOrokandUlcha,e.g.*…gaju-­>*…(a)jU-;afterthegrammaticalizationwascompleted,idiosyncraticchangesinOroktookplace,i.e.*-j- > -s-).ItisreasonabletoassumethatNorthernTungusic(+Oroch,Udihe)*-(y)acouldhaveoriginatedafterthe(irregular)completelossof*ga- and the lenition of -j- to -y-,asinOrochga- vs.gay- (while admittedlyirregular,suchascenarioisnotunheardof,seeTVSG36–37:§48).25 ThereductioninNanay&Kilen -go- <*-gaju- isalsounexpected.Inthiscase,however,thepointofdeparturemightbejust*ga-: labialization is fairly com-mon,especially,thoughnotexclusively,afterlabialandvelarconsonants,whenadoptingregularharmonicvowelpatterns(e.g.instrumentnounsfromverbs:Lit.Ewenki­-ŋkii­vs.Nanay­-ŋko,oralienablepossession:Lit.Ewenki ŋii vs.Nanay ŋgo,etc.,seeBoldyrev1987:32–39and1976:130–142,respectively).Anothersolutionwouldinvolvethepresenceoftwo“competing”IND.ACCmarkers:*-(y)avs.*-ga( ju),theformerisoriginalandpartiallypreservedinNorthernTungusic,Udihe(&Kili),whereasthelatter,asecondaryproduct,developedinEwenandSouthernTungusic.

HowcouldHelimski’shypothesisbenefitfromtheforegoingdiscussion?IftheBuylainscriptionreflectsalanguageclosertoSouthernTungusic,itislegitimatetospeculatethatitcouldreflectasortofintermediatestageinwhichtheIND.ACCisundergoingthefunctionalchangetowardstheDES:†taag- ‘torecognize’and †icä- ‘tosee(>watchover)’canactuallytargetobjectswiththeDESmarker,whichismostfrequentlyattachedtotheobjectsofverbsbelongingtoaveryspecificsemanticclass(‘discovering,pursuing,making,achievement,appear-ance’).Forexample,Kazama(2012:126–127)mentionsNanaybaogo- ‘tosee’,ta- ‘todo,make’,gaajo- ‘tobring’(thelexicalsourcefortheveryDESending)or baa- ‘toget,obtain’(itscognateispresentintheJurchenmemorial).However,ifthechangeIND.ACC>DESobligatorilyrequirestheuseofPOSSmarkersorthegrammaticalizationof*ga( ju)-,thenitisverydifficulttoreconcilethisrequirementwiththematerialevidenceintheBuylainscription,wherenothingcanbetracedbacktopossessivesor*ga( ju-).

25 Contractionandphoneticsimplificationinvolvingirregularsoundchangesarecommonstagesintheprocessofgrammaticalization(foranin-depthdiscussionwithadditionalexamplesinTungusic,seeAlonsodelaFuente2011a,esp.18–24,105–110).

TUNGUSIC AND THE BUYLA INSCRIPTION 37

EXCURSUS. Hungaro-Tungusica.

HelimskiagreeswithFutakyinthatthereareafewHungarianwords,theetymologyofwhich,beingimpossibletobediscernedthroughtheFinno-Ugricbackgroundofthelanguage,couldbesolvedbyinvokingTungusicsubstratum.26 SinceFutakyalsomentionstheAvarquestion,HelimskimayfindsupportforhisownhypothesisbyacceptingFutaky’s.Thereisnoplaceinthisbriefcontri-butionforanexhaustivereviewofFutaky’sproposal,whichhasbeenreceivedverynegatively.Amongothers,theHungarianMongolistGy.Kara(Kara2002;seenowKnüppel2013:194–197)explainsthatmostetymologiessufferoflamesemantics,inconsistentsoundcorrespondences,anachronisms,etc.Inwhatfol-lows,thecoreofthediscussionwillrevolvearoundfourwellknownetymologiesthat,accordingtoHelimski(2000:53),arefaultless.ThesefouretymologiesallegedlyaccountfortheTungusicbackgroundofHungarianbeteg‘ill’,oldal ‘side’,hamar‘quickly’,andkanál‘spoon’.ItmustbehighlightedthatthepresentauthorisnottrainedinFinno-UgriclinguisticsandthereforecannotdiscussHungarianmattersinitspropercontext,thereforecommentswillberestrictedtoTungusic.

# Hungarian(Futaky 2001 /

Kara 2002)

Tungusic(reconstruction [adapted] = EEW / materials = SS)

I beteg‘ill’(35–37/492)

PT*bö.dä- (PRT.AOR)~*bö.cä- (PRT.PST)‘todie’(147[1562])>Lit.Ewenkibu.dä- & bu.rä- (bothPRT.AOR,thelatterisanalogical),Lit.Ewen(&Arman)bu-ni‘deceased’,butään ‘pain,indisposition’⇉ butääk ‘sickly,unhealthy’(cf.Armanbutääkñä),Solonbusé ~ buc’á(PRT.PST),Negidalbu.dä- ~ buldä-,Nanaybu(y)- & bur-,Orok&Ulchabu(l)-,Kili&Kilen&Oroch&Lit.Udihebu.dä-,WMbude- ~ buce- (Sibe becë-)(I.98–99),allmeaning‘todie’.

II oldal‘side’(65/495)

CT*xoldaa.n‘side’(642[8521])>Lit.Ewenkioldoon,Lit.Ewenoldaan,HailarSolonoldon,Negidaloldon,Armanoldaanji ‘around’(<CT*xoldaa.n+ji),Lit.Udihe&Orochogdo(n-),Kilioldon,Ulchaxoldo(n),Nanayxoldon,Orokxoldo(n-),Kilenxoldon(II.13),allmeaning‘side’.

26 NotinterferingwithUralicandFinno-Ugriccomparativelinguisticsisthemaindiffer-encebetweenFutaky’sandpreviousattemptsatlinkingHungarianandManchu(ric)orTungusic,atraditionwhichapparentlybegunwithCononvonGabelentz’s46Manchu–Hungarian(lexical)comparisons(ConondelaGabelentz1832:6–8).

38 JOSÉ ANDRÉS ALONSO DE LA FUENTE 

# Hungarian(Futaky 2001 /

Kara 2002)

Tungusic(reconstruction [adapted] = EEW / materials = SS)

III hamar‘quickly’(51–52/493)

PT*(x)ama ‘quick(ly),fast’(69[466,467])>Lit.Ewenki ama,Dial.Ewenkikamaña ~ kamakaan / gamakaan,Solonamarii,Lit.Ewenaamrǝk,Dial.Ewenamdak/aamnak,Armanaasmak(I.34),allmeaning‘quick(ly),fast’.Cf.Lit.Ewenkiamiltaan- ‘tobeintime,catchup’,Lit.Ewenaamǝltǝn-­id.,Arman aamaltan- id.(I.38a).

IV kanál ~ kalán ‘spoon’(60–61/494)

CT*kala.n‘kettle’(448[5748])>Ewenkikalan (⇉ kala+kaan {DIM}‘littlekettle’&verbalderivatesfromkala-ruu-,cf.Arman kalaa-was ~ kalaa-wüs‘bag’),Nanay&Kilekalan,Ulcha&Lit.Udihekala(n-) (I.364–365),allmeaning‘kettle’.

EverythingwouldbealrightifthesoundcorrespondenceswhichFutakyestablishedonthebasisoftheseetymologieswouldapplyintheremainingcaseswiththesameregularity.Justtomentiononeillustrativeexample:Hungarian‹e›and‹é›maycorrespondtoTungusic*ö,*e,*ior*ya(thisholdstruealsoforFutaky’sHungarian-Mongoliccomparisons,e.g.Hungarianbeze‘really,truly’>bezzek‘ofcourse’,bíz(ik)‘toentrust,confide’,deber[>deberke]‘akindofdump-ling’andkebel‘abdomen,lap’arecomparedwithMongolicbiz,bisira,debure and kebeli,respectively,seeFutaky2001:23–32s.vv.).Thereisnoproblemwiththisscenarioaslongasthediversityofresultsisexplained.Unfortunately,Futakyprovidesthereaderwithnoexplanationsatthisregard.

ThereisalsoastrikinglackofconsistencywiththerelativechronologyofsomesoundchangesoccurringinthehistoryofTungusicphonology.Forexample,Hungarianoldal(II)reflectsthevowelsequenceo…a,certainlyveryarchaicinTungusic(ithasbeenpartiallypreservedonlyinLiteraryEwen),butthereisnotraceoftheinitial*x,anotherveryarchaicfeatureinTungusicwhichiscommonlytakentoco-occurwiththevowelsequence*CoCa.However,Futakyassumesthat(III)goesbackto*xamar,inspiteofthelackofSouthernTungusiccognatessupportingthereconstructionofinitial*x.27NotethatdialectalformsinEwenkiwith initial g- (East:Kacug,Nercin,Tokmin;South:StonyTungus)andk- (East: Aldan,Tokmin,Ucur;North:Erbogocen;South:StonyTungus)maybetheresultofsecondaryprocesses,e.g.contaminationwithCT*kama- ‘tooppress,prohibit;beataloss’(SS1.369).Asforg-,itcouldbeaprotheticconsonant,perhapstriggered

27 Futakywasapparentlyawareofthisfact,e.g.hajdan‘intimespast,informertimes’←*xaal(ï)daa{when-LOC}‘sometime’(2001:50–51),thederivateisonlyattestedinSouthernTungusicandOroch,beingabsentinNorthernTungusic;FutakydoesnotexplainHungarian -n,whichcannotbeofTungusicoriginbecauseTungusic -n never followsgrammaticalendings(SS1.32a,Kara2001:493).

TUNGUSIC AND THE BUYLA INSCRIPTION 39

byanalogywithother(postpositional)formssuchasxamari‘down’.28Whywas*x leftuntouchedinoneword,butlostintheother?

Inthisconnection,thereseemstobeapreferenceforNorthernTungusicmateri-alsinFutaky’sworks,althoughthereisnoexplicitstatementtothiseffect.ItremainsunclearwhyFutakychoseLiteraryEwenkiiiy(ä)‘horn’andrelatedforms(seeSS1.298–299)asthepointofdeparturefortheetymologyofHungarianíj ‘bow’,whenitiswellknownthatthecorrespondingformintheparentallanguageis*xüyä‘id.’(>Orok&Ulchaxuyä,LiteraryNanayxuyin,Manchuweihe ~ uihe,etc.,seei.a.EEW386[4973]).EvenlesscomprehensibleishisdecisiontorelateHungarianbuk(ik)‘tofall,collapse’(Futaky2001:37)toLiteraryEwenkibuk- ‘tothrowon,upon;torush’(SS1.103b).Thepresenceofthiswordinonlyahandfulof(Northern)EwenkidialectsanditsremarkableresemblancewithYakutbok kïn- ‘to do some-thingimmediately,atonce’(Pekarskij1959:489)cannotbefortuitousandbegsforacommonexplanation.29

InpreviousparagraphsitwasshownthatsomeHungarianwordsapparentlyretaincertainsalientarchaicfeaturesofTungusic,e.g.initial*x,butFutaky’sprac-ticeoftargetingonlyNorthernTungusicmaterialsgoesveryoftenagainstthistrait.Again,thescenariomaybeplausibleaslongasFutakyexplainswhytheHungarianlexiconhasNorthernTungusicitems,inspiteoftheimplications,namely,amuchrecentchronologicallayerthatmakesthedescriptionthejourneyofthoseTungusicwordsintotheCarpathianbasinevenharder.ThedispersionofEwenkispeakersoverNorthernSiberiaisaratherrecentevent(itiscommonlytakentobenoear-lierthanthe12thc.,seei.a.Janhunen2013:35)whichcannotbelinkedunderanycircumstancewithpopulationmovementsheadingtowardssoutheasternregions.

28 Castrén’svocabularyisoneoftheearliestwitnessestothealternation‹h›~Ø,e.g.hokto(Urulgindialect)~okto(Manikovadialect)‘path,track’,halgan ~ algan‘foot’,or häŋä ~ äŋä‘friend,companion’(1856:82–83),etc.ThisisaregularalternationamongEwenkidialects,thepointofdepartureofwhichisPT*p- (preservedinUlcha&Orok p-,againstLit.Ewen&Negidal&Udihe&Orochh- [h~x],thereforeitcan-notbeequatedwithamar,seeTSVG32–34§44).Could†[γamar]beanexampleofhypercorrection?Secondary/h/,whichappearsasaprotheticsegmentinwordsbegin-ningwithhighbackvowels,isaverycommonfeatureinEwenkidialects,e.g.Sym(Southern)Ewenki(h)utä ~ ’utävs.Lit.Ewenkihutä‘son’,withputativePT*p-,buthoroktovs.oro(o)kto‘grass’,fromPT*oraa+kta‘grass,hay’,cf.Lit.Ewenoraat (SS 2.24,Vasilevič1948:64).

29 TheYakutconnection,ignoredbyFutakyeventhoughitisalreadypointedbytheSSeditors,isnotentirelyclear,though:thedistributionofbok seems restricted also in Yakut(Pekarskijprovidesonlyonesource)andthesoundcorrespondenceEwenki/u/:Yakut/o/,thoughnotuncommon,isambiguous,cf.Yakutüüt turaan‘vollständigeStille’←Ewenkitoron‘Stille’orYakutnoxto ~ nolto‘Ader,Vene’←Ewenkiun-guktaid.(Kałużyński1982:266,268;forfurtherdetailsonthephonologyofthevowelcorrespondences,seeRomanova/Myreeva/Baraškov1975:34–42).Allinall,thiswouldratherpointtoaTungusicloaninYakut,andnottheotherwayaround.

40 JOSÉ ANDRÉS ALONSO DE LA FUENTE 

OnewouldexpectthatFutaky’sandHelimski’slinguisticevidences30wouldshowsomekindofhomogeneity.Forexample,theHungarianwordsdiscussedbyFutakycouldreflectsomeoftheSlavicfeaturesdescribedbyHelimski.Butthisisnotthecase,evenwhenthecontextwouldfavorsuchascenario.Whatismore,theobviousNorthernTungusicbiasaside,itisimpossiblebyanystretchoftheimaginationtoprovideaspecificsetoffeaturescharacterizingtheTungusiclanguagebehindFutaky’sHungarianetymologies.Anotherexampleofthenon-complementaryrelationshipbetweenHelimski’sandFutaky’shy-potheses:onseveraloccasionsHelimskisuggestedthattheTungusiclanguagewhicharrivedtotheCarpathianbasicwasverysimilartoSouthernTungusicandManchuric.Onceagain,noonewillfindanindependentconfirmationofthisinFutaky’shypothesis.

OnecouldarguethatthedifferencesbetweenFutaky’sandHelimski’shy-potheseslieinthefactthattheyexploitdifferentsources.Helimskiworkedwithatextualsamplewhichconsequentlydemandssomekindoflinguistichomogeneity.However,Futakypreferredtodealwithindividualitems,scatteredthroughtheHungarianlexicon,forwhichthereisalackofthemostbasicinformationabouttheiroriginalculturalandlinguisticcontexts.Therefore,itwouldbelegitimatetoconcludethattheapparenthomogeneityproposedbyHelimskicannotbecalledforinFutaky’smaterials,becauseeachitemanalyzedbytheHungarianscholarmaywellbelongtodifferentchronologicallayers.Butevenifthisisso,thefactremainsthatbothhypothesesmustberejectedduetotheirmanyinconsistenciesandfaultymethodology.

9. Conclusions

Intheforegoingdiscussion(see§§3–8),ithasbeenshownthattheTungusicreadingoftheBuylainscriptionbyE.Helimskiposessomeinsurmountableprob-lems:thereconstructionof*-g- intheaoristparticiplemarker,theghostword†täsi,theinconsistencyoftheso-called“Slavic”features(=SlavicizedTungusic),theratherarbitrarypresenceofNorthernelementswhenitisclaimedthattheinscriptionstandsclosertoSouthernTungusicandManchuric,theuncertaintiessurroundingthemorphosyntacticanalysis,etc.MoreproblematicisFutaky’sidearegardingthepresenceofTungusicwordsinHungarian(seeExcursus).AlthoughbothHelimski’sandFutaky’shypothesescouldgreatlyprofitfromeachotherowing

30 Incidentally,itcouldbementionedthatculturaltraitslinkingAsianAvarsandTungusianshavebeenalsoproposedinthespecialistliterature.Kőhalmi-Uray(2004:119–120)hasrecentlysuggested,“[a]lthoughnotwithoutdoubtsandreluctance”,thattheremightbeaconnectionbetweenahistoricalepisodeoftheAsianAvars(=theRuanruan,seefootnote1above)andthetaleoftheNišanshamaness.

TUNGUSIC AND THE BUYLA INSCRIPTION 41

tothemoreorlesscommonbackgroundtheyshare,inrealitythetwoproposalsaremutuallyexclusive.

Inanaudaciousleapofimagination,Helimskitriedtofindthesolutiontoalong-standingproblembyadoptinganentirelynewperspective.AsS.J.Gouldarguedinamemorableessay(1980:59–68),thesuccessofanyscientificendeavordependsonthebalancebetweeninductivism(“[a]newandsignificanttheory[…]canarisefromafirmfoundationoffacts.[…]eachfactisabrickinastructurebuiltwithoutblueprints.Anytalkorthoughtabouttheory(thecompletedbuild-ing)isfatuousandprematurebeforethebricksareset”)andeurekaism(“[g]reatscientists[…]aredistinguishedmorebytheirpowersofhunchandsynthesis,thantheirskillinexperimentorobservation”).GouldillustratedthispointinvokingDarwin’sachievements.Helimskididnotfindthemiddleroadbetweeninductiv-ismandeurekaism,butratherstrayedoffandgotlostinthelatter.

JoséAndrésAlonsodelaFuenteul.Okólna12/32PL–30-684Kraków[[email protected]]

A b b r e v i a t i o n s

1,2,3=person;ACC=accusative;AOR=aorist;COND=conditional;CT=CommonTungusic;CV=converb;DER=derivative(suffix);DES=destinative;EMPH=emphatic(particle);GEN=genitive;HAB=habitual;HOR=horta-tive;IND=indefinite;INST=instrumental;LOC=locative;NEG=negative;NLZ=nominalizer;OCM=obliquecasemarker;PL=plural;POSS=possessive;PRF=perfect;PRT=participle;PST=past;REF=reflexive;SG=singular;WM=WrittenManchu.

R e f e r e n c e s

AlemanyA.2009.FromCentralAsiatotheBalkans:theTitle*ču(b)-pān.–AllisonCh./Joisten-PruschkeA./WendtlandA.(eds.),From­Daēnā­to­Dîn:­Religion, Kultur und Sprache in der iranischen Welt. Festschrift für Philip Kreyenbroek­zum­60.­Geburtstag.Wiesbaden:3–12.

AlonsodelaFuenteJ.A.2011a.Tense, Voice and Aktionsart in Tungusic. Another Case of “Analysis to Synthesis”?.Wiesbaden.

AlonsodelaFuenteJ.A.2011b.Venjukov’s[1858]1862/1868Nanaimaterials.–Journal de la Société Finno-Ougrienne93:11–35.

42 JOSÉ ANDRÉS ALONSO DE LA FUENTE 

AtknineV.1997.TheEvenkiLanguagefromtheYeniseitoSakhalin.–Northern Minority Languages: Problems of Survival. Senri Ethnological Studies44:109–121.

AvrorinB.A./BoldyrevB.V.2001.Grammatika­oročskogo­jazyka.Novosibirsk.AvrorinV.A.1959.Grammatika­nanajskogo­jazyka,vol.1:Fonetičeskoe­vvedenie­

i­morfologija­imennyx­častej­reči.Moskva/Leningrad.AvrorinV.A.1961.Grammatika­nanajskogo­jazyka,vol.2:Morfologija glagol’nyx

i narečnyx­častej­reči,­meždometij,­služebnyx­slov­i­častic.Moskva/Leningrad.AvrorinV.A.1981.Sintaksičeskie­issledovanija­po­nanajskomu­jazyku.Leningrad.BeckwithC.I.2009.Empires of the Silk Road. A History of Central Eurasia from

the­Bronze­Age­to­the­Present.Princeton/Oxford.BenzingJ.1953.Remarquessurleslanguestongousesetleursrelationsavecles

autreslanguesdites“altaiques”.–Ural-altaische Jahrbücher25.1–2:109–118.BoldyrevB.V.1976.Kategorija­kosvennoj­prinadležnosti­v­tunguso-man’čžurskix­

jazykax.Moskva.BoldyrevB.V.1987.Slovoobrazovanie­imen­suščestvitel’nyx­v­tunguso-man’čžurskix­

jazykax­v­sravnitel’no-istoričeskom­osveščenii.Novosibirsk.BulatovaN./GrenobleL.1999.Evenki.München.CastrénM.A.1856.Grundzüge­einer­tungusischen­Sprachlehre­nebst­kurzem­

Wörterverzeichniss.St.Petersburg.CinciusV.I.1982.Negidal’skij­jazyk.­Issledovanija­i­materialy.Leningrad.ClausonG.1972.An Etymological Dictionary of Pre-Thirteenth-Century Turkish.

Oxford.ConondelaGabelentzH.1832.Élémens de la grammaire mandchoue.Altenbourg.DoerferG.1967.Türkische und mongolische Elemente im Neupersischen,Band3:

Türkische Elemente im Neupersischen (gīm bis kāf).Wiesbaden.DoerferG.1977.IstKur-UrmiischeinnanaischerDialekt?.–DecsyGy./Dimov-

BogoevCh.D.(eds.),Eurasia nostratica. Festschrift für K.H. Menges.Wies-baden,vol.1:51–63.

DoerferG.1978.ClassificationProblemsofTungus.–WeiersM.(ed.),Tungusica. Beiträge­zur­nordasiatischen­Kulturgeschichte.Wiesbaden:1–26.

DoerferG.1985.Mongolo-Tungusica.Wiesbaden.EEW=DoerferG.2004.Etymologisch-ethnologisches Wörterbuch tungusischer

Dialekte.UnterMitwirkungvonM.Knüppel.Hildesheim.ErdalM.1988.TheTurkicNagy-Szent-MiklósInscriptioninGreekLetters.–Acta

Orientalia Academiae Scientiarum Hungaricae42.2–3:221–234.ErdalM.2007.TheKhazarLanguage.–GoldenP./Ben-ShammaiH./Róna-

TasA.(eds.),The­World­of­the­Khazars.Leiden/Boston:75–108.FutakyI.2001.Nyelvörténeti­vizsgálatok­a­Kárpát-medencei­avar-magyar­kapcso-

latok­kérdéséhez.Budapest.GöblR./Róna-TasA.1995.Die­Inschriften­des­Schatzes­von­Nagy-szentmiklós.­

Eine paläographische Dokumentation.Wien.

TUNGUSIC AND THE BUYLA INSCRIPTION 43

GoldenP.1992.An Introduction to the History of the Turkic Peoples.Wiesbaden.GorelovaL.M.2002.Manchu­Grammar.Leiden/Boston/Köln.GouldS.J.1980.The­Panda’s­Thumb.­More­Reflections­in­Natural­History.

NewYork/London.HelimskiE.=XelimskijE.2000a.Jazyk(i)Avarov:Tunguso-man’čžurskijas-

pekt.–Folia Orientalia36(=StudiainhonoremStanislaiStachowskidicata):135–148.

HelimskiE.2000b.OnProbableTungus-ManchurianOriginof theBuylaInscriptionfromNagy-Szentmiklós(PreliminaryCommunication).–Studia Etymologica Cracoviensia5:43–56.

HelimskiE.=XelimskijE.2003.Tunguso-man’čžurskij­jazykovoj­komponent­v av-arskom kaganate i slavjanskaja ètimologija (Materialy k dokladu na XIII Mež-du­narodnom­sъezde­slavistov,­Ljubljana,­15–21­avgusta­2003).Hamburg.

HelimskiE.2004.DieSprache(n)derAwaren:Diemandschu-tungusischeAlternative.–NaeherC. (ed.),Proceedings of the First International Conference­on­Manchu-Tungus­Studies­(Bonn,­August­28­–­September­1,­2000),vol.2:Trends in Tungusic and Siberian Linguistics.Wiesbaden:59–72.

HúZēngyì,Cháokè.1986.Èwēnkèyǔ­jiǎn­zhì.Běijīng.JanhunenJ.1985.TheTungusPeopleandtheConquestofSiberia.–JarringG./

RosénS.(eds.),Altaistic­Studies.­Papers­Presented­at­the­25th­Meeting­of­the­Permanent­International­Altaistic­Conference­(Uppsala­1983). Stockholm: 73–77.

JanhunenJ.1987.TowardsaunifiedphonologicaltranscriptionoftheSiberianlanguages.–Journal de la Société Finno-Ougrienne81:151–174.

JanhunenJ.1996.Manchuria. An Ethnic History.Helsinki.JanhunenJ.2004.FromChosontoJucher:OnthePossibilitiesofEthnonymic

ContinuityinGreaterManchuria.–Studia Etymologica Cracoviensia 9:67–76.

JanhunenJ.2009.EugeneHelimski(1950–2007).–Finnisch-Ugrische Forschungen 60.1–3:364–371.

JanhunenJ.2012.TheExpansionofTungusicasanEthnicandLinguisticProcess.–MalchukovA./WhaleyL.(eds.),Recent Advances in Tungusic Linguistics. Wiesbaden:5–16.

JanhunenJ.2013.TheTungusicLanguages:AHistoryofContacts.–KimJ./KoD.(eds.),Current Trends in Altaic Linguistics. A festschrift for Professor Emeritus­Seong­Baeg-in­on­his­80th Birthday.Seoul:18–60.

KałużyńskiS.1982.EinigetungusischeLehnwörterimJakutischen.–Acta Orientalia Academiae Scientiarum Hungaricae36.1–3:261–269.

KaraGy.2002.ReviewofFutaky2001.–Magyar Nyelv98:491–496.KazamaS.2012.DesignativeinTungusiclanguages.–MalchukovA./Whaley

L.(eds.),Recent Advances in Tungusic Linguistics.Wiesbaden:123–152.KiyoseG.N.1977.Study of the Jurchen Language and Script.Kyoto.

44 JOSÉ ANDRÉS ALONSO DE LA FUENTE 

KnüppelM.2013.ProfessorIstvánFutaky(12.5.1926–21.1.2013)alsTungusologe.–Studia Linguistica Universitatis Iagellonicae Cracoviensis130:191–198.

Kőhalmi-UrayC.2004.TheMythofNishanShaman.–HoppálM. (ed.),Rediscovery of Shamanic Heritage.Budapest:113–124.

KonstantinovaO.A.1964.Evenkijskij­jazyk.­Fonetika.­Morfologija.Moskva/Leningrad.

KotvičъW.1909.Materialydljaizučeniatungusskixъnarěčij.I.Obrazcyjazykasungarijskixъgol’dovъ.–Živaja­Starina18.2–3:206–218.

KovácsT./GaramÉ.(eds.)2001.The­Gold­of­the­Avars.­The­Nagyszentmiklós­Treasure. Catalogue of the Exhibition.Budapest

LarsenT.W.2007.InstrumentalNounPhrasesandUnmarkedObjectsinClassicalManchu.–WadleyS./NaeherC.(eds.),Proceedings of the First North American Conference on Manchu Studies, vol. 2:Studies in Manchu Linguistics.Wiesbaden:21–41.

LigetiL.1953.Leproblèmedel’ethnogenèsedessamaghirs.–Acta Orientalia Academiae Scientiarum Hungaricae3.1–2:33–44.

MalchukovA. /NedjalkovI.2010.DitransitiveConstructions inTungusicLanguages.–MalchukovA./HaspelmathM./ComrieB.(eds.),Studies in Ditransitive Constructions. A Comparative Handbook.Berlin/NewYork:318–351.

MengesK.H.1952.ZureinigenProblemendertungusischenGrammatik.–Ural-Altaische Jahrbücher24.1:112–121.

MillerR.A.1994.TheOriginalGeographicDistributionoftheTungusLanguages.–AronsonH.I.(ed.),Non-Slavic languages of the USSR. Papers from the Fourth Conference.Columbus:272–297.

Myl’nikovaK.M./CinciusV.I.1931.Materialypoissledevanijunegidal’skogojazyka.–Tunguskij sbornik1:107–218.

NikolaevaI./TolskayaM.2001.A­Grammar­of­Udihe.Berlin.NovikovaK.A.1960.Očerki­dialektov­èvenskogo­jazyka.­Ol’skij­govor,­čast’­1.

Moskva/Leningrad.OzoliņaL.1995.PresentdaylinguisticsituationamongtheValOroks(Uilta)of

EasternSakhalin.–Linguistic­and­Oriental­Studies­from­Poznań2:91–104.OzolinjaL.B.2001.Oroksko-russkij slovar’.Novosibirsk.PekarskijE.K.19592.Slovar’­jakutskogo­jazyka.Jakutsk.PetrovaT.I.1967.Jazyk­orokov­(ul’ta).Leningrad.PohlW.2002.Die­Awaren.­Ein­Steppenvolk­in­Mitteleuropa­567–822­n.Chr.

München.PoppeN.N.1931.Materialy­po­solonskomu­jazyku.Leningrad.PoppeN.1966.OnSomeAncientMongolianLoan-wordsinTungus.–Central

Asiatic Journal11:187–198.PoppeN.1972.OnSomeMongolianLoanWordsinEvenki.–Central Asiatic

Journal16:95–103.

TUNGUSIC AND THE BUYLA INSCRIPTION 45

PoppeN.1977.TheAltaicPluralSuffix -*t.–Studia Orientalia47:165–174.PoppeN.1987.Introduction to Mongolian Comparative Studies. Helsinki.PoppeN.20062.Grammar­of­Written­Mongolian.Wiesbaden.PritsakO.2006.TheTurkicEtymologyoftheWordQazaq‘Cossack’.–Harvard

Ukrainian Studies28.1–4:237–243.PulleyblankE.G.1991. Lexicon of Reconstructed Pronunciation in Early Middle

Chinese, Late Middle Chinese, and Early Mandarin.Vancouver.PutincevaA.P.1954.Morfologija govora gorinskix nanaj.Leningrad.RomanovaA.V./MyreevaA.N./BaraškovP.P.1975.Vzaimovlijanie­èvenkijskogo­

i­jakutskogo­jazykov.Leningrad.Róna-TasA./BertaÁ.2011.West Old Turkic. Turkic Loanwords in Hungarian,

2vols.[withtheassistanceofL.Károly].Wiesbaden.Róna-TasA.1999.Hungarians and Europe in the Early Middle Ages.Budapest.Róna-TasA.2001.TheInscriptionsoftheNagyszentmiklósTreasure.–Ko-

vácsT./GaramÉ.(eds.),The­Gold­of­the­Avars.­The­Nagyszentmiklós­Treas-ure. Catalogue of the Exhibition.Budapest:120–129.

Róna-TasA.2003.Börtönszavunkésazavar-magyarkapcsolatok.–Nyelvtudományi Közlemények100:219–225.

RozyckiW.1994.Mongol Elements in Manchu.Bloomington.SchmidtP.P.1923.TheLanguageoftheNegidals.–Acta Universitatis Latviensis5:

3–38.SchuesslerA.2007.ABC Etymological Dictionary of Old Chinese.HonoluluSemL.I.1976.Očerk­dialektov­nanajskogo­jazyka.­Bikinskij­(Ussurijskij)­dialekt.

Leningrad.SevortjanÈ.,ed.1978.Ètimologičeskij­slovar’­tjurkskix­jazykov,vol.2“B”.

Moskva.ShirokogoroffS.M.1924.Social­Organization­of­the­Manchus.­A­Study­of­the­

Manchu­Clan­Organization.Shanghai.SinorD.1952.OnSomeUral-AltaicPluralSuffixes.–Asia Major (NS)2:203–230.SS=CinciusV.I.,ed.1975–1977.Sravnitel’nyj­slovar’­tunguso-man’čžurskix­

jazykov,vol.2.Leningrad.StachowskiM.2004.TheoriginoftheEuropeanwordforsabre.–Studia Ety-

mologica Cracoviensia9:133–141.StachowskiM.2005.Uwagiozapożyczeniachałtajskichw- językuprasłowiań-

skimikwestiepokrewne.–Siemieniec-GołaśE./PomorskaM.(eds.),Turks and Non-Turks Studies on the History of Linguistic and Cultural Contacts. Special­Issue­Presented­to­Professor­Stanisław­Stachowski­to­His­Seventy­Fifth Birthday(=Studia Turcologica Cracoviensia10).­Kraków:437–454.

SunikO.P.1958.Kur-Urmijskij dialekt. Issledovanija i materialy po nanajskomu jazyku.Leningrad.

SunikO.P.1982.Suščestvitel’noe­v­tunguso-man’čžurskix­jazykax.Leningrad.SunikO.P.1985.Ul’čskij­jazyk.­Issledovanija­i­materialy.Leningrad.

46 JOSÉ ANDRÉS ALONSO DE LA FUENTE 

SzalontaiCs./KárolyL.2013.RuniformFragmentsoftheLateAvarPeriodfromHungary.–Acta Orientalia Academiae Scientiarum Hungaricae66.4:365–396.

TamuraK.1990.Manshū-gonotaikakuniokerumokuteki-gonogentei-shō.–Chūkyō­daigaku­kyōyō­ronsō31.2:579–601.

TSVG=BenzingJ.1956.Die tungusischen Sprachen. Versuch einer vergleichenden Grammatik.Wiesbaden.

VasilevičG.M.,ed.1936.Sbornik materialov po èvenkijskomu (tungusskomu) fol’kloru.Leningrad.

VasilevičG.M.1948.Očerki­dialektov­èvenkijskogo­ /tungusskogo/­ jazyka. Leningrad.

VasilevičG.M.1969.Èvenki.­Istoriko-ètnografičeskie­očerki­(XVIII–načalo­XX v.).Leningrad.

VasilevičG.M.1974.Antroponimyunarodovural’skojialtajskojsemej,rasse-lennyxvSibiri(opytkartografirovanija).–BrukS.I./BorodinaM.A.(eds.),Problemy­kartografirovanija­v­jazykoznanii­i­ètnografii.Leningrad:296–302.

VovinA.2007.Onceagainon theEtymologyof theTitleqaγan.–Studia Etymologica Cracoviensia12:177–187.

WalravensH.1992.Die frühesten goldischen Texte. Faksimile-Nachdruck einiger Übersetzungen­von­Prokopij­Protodiakonov.­Mit­einem­Essay­von­Władysław­Kotwicz­über­die­Erforschung­der­goldischen­(Nanai-)Sprache.Berlin.

Wǔtǐ=CorffO.et alii.2013.Auf kaiserlichen Befehl erstelltes Wörterbuch des Manjurischen in fünf Sprachen,2vols.Wiesbaden.

XasanovaM.M./PevnovA.M.2003.Mify­i­skazki­negidal’cev.Suita.ZakharovI.20102.A­Grammar­of­Manchu­/­Grammatika­man’čžurskago­jazyka.

Folkestone[Sanktpeterburgъ].