Remarks on the History of the Indo-European Infinitive Dorothy Disterheft University of South Carolina - Columbia, Ddisterheft@Sc.Rr.Com

Home , Control (linguistics), Oblique case, Preverb

University of South Carolina Scholar Commons

Faculty Publications Linguistics, Program of

1981 Remarks on the History of the Indo-European Infinitive Dorothy Disterheft University of South Carolina - Columbia, [email protected]

Follow this and additional works at: https://scholarcommons.sc.edu/ling_facpub Part of the Linguistics Commons

Publication Info Published in Folia Linguistica Historica, Volume 2, Issue 1, 1981, pages 3-34. Disterheft, D. (1981). Remarks on the History of the Indo-European Infinitive. Folia Linguistica Historica, 2(1), 3-34. DOI: 10.1515/ flih.1981.2.1.3 © 1981 Societas Linguistica Europaea.

This Article is brought to you by the Linguistics, Program of at Scholar Commons. It has been accepted for inclusion in Faculty Publications by an authorized administrator of Scholar Commons. For more information, please contact [email protected]. Foli.lAfl//ui.tfcoFolia Linguistica HfltorlcGHistorica II/III/l 'P'P.pp. 8-U3—Z4 © 800£etruSocietae lAngu,.ticaLinguistica Etlropaea.Evropaea, 1981lt)8l

REMARKSBEMARKS ONON THETHE HISTORYHISTOKY OFOF THETHE INDO-EUROPEANINDO-EUROPEAN INFINITIVEINFINITIVE

DOROTHYDOROTHY DISTERHEFTDISTERHEFT I 1.1. INTRODUCTIONINTRODUCTION WithWith thethe exceptfonexception ofof Indo-IranianIndo-Iranian (Ur)(Hr) andand CelticCeltic allall historicalhistorical Indo-EuropeanIndo-European (IE)(IE) subgroupssubgroups havehave a morphologicalIymorphologically distinctdistinct II infinitive.Infinitive. However,However, nono singlesingle proto-formproto-form cancan bebe reconstructedreconstructed foforr themthem -— onlyonly a widewide arrayarray ofof actionaction nouns.nouns. A goodgood dealdeal hashas beenbeen writtenwritten aboutabout thethe developmentdevelopment ofof thethe infinitive,infinitive, almostalmost allall fromfrom a morphologicalmorphological perspective.perspective. ButBut thethe evolutionevolution ofof a singlesingle infinitiveinfinitive fromfrom a largelarge numbernumber ofof Proto-Indo-EuropeanProto-Indo-European (PIE)( ) actionaction nounnounss isis primarilyprimarily a syntactiosyntactic developmentdevelopment involvinginvolving notnot onlyonly anan expanexpan- sionsion inin thethe numbenumberr ofof clauseclause typestypes thatthat useuse infinitiveinfinitive asäs embeddedembedded predicatepredicate bubutt alsoalso changchangee inin ththee struoturestructure ofof thethe infinitiveinfinitive clausclausee itselfitself (i.e.(i.e. expansionexpansion ofof typestypes ooff subject/objecsubject/objectt occurrence),occurrence), changchangee inin objecobjectt casecase,, andand worwordd Ordeorderr shiftsshifts.. MMyy aiaimm herheree isis toto identifyidentify archaicarchaic featuresfeatures,, bothboth morphologicamorphologicall andand syntacticsyntactio,, ofof attesteattestedd IEIE languageslanguages whicwhiohh wilwilll givgivee a clearerclearer pioturepicture ooff whawhatt nonfinitenonfinite subsub- ordinatingordinating strategiestrategiess werweree useusedd iinn certaioertainn areaareass ofof ththee latlatee PDPIEE languaglanguage,e community.Community. This'discussionThis discussion differdifferss frofromm paspaatt syntaotiosyntactic studiestudiess (e.g(e.g.. thosthosee ooff HahHahnn 1950,1950, MilleMiUerr 1974,1974, ArArdd 1975)1975) whicwhiohh focusfocus oonn embeddeembeddedd structurestructuress ooff a further-developefurther-developedd typtypee (e.g(e.g.. thosethose wherwheree subject-raisinsubject-raisingg iiss involvedinvolved)) ttoo reconstrucreconstruott earlierearlier stagesstages ooff infinitivainfinitivall Subordinationsubordination.. I wilwilll arguarguee herheree thathatt undeunderr oertaincertain conditionconditionss ththee actioactionn nounounn wawass indeterminatelindeterminatelyy a nounounn oorr ththee prepre- dicatdicatee ooff a subordinatsubordinatee clauseolause.. BeanalysiReanalysiss äass aann infinitivinfinitivee resulteresultedd iinn a completcompletee shifshiftt iinn mosmostt language]an~agess ttoo parpartt ooff ththee verbaverball Systemsystem. .

*... A pre~preliminary sketchsketch ofof thisthis paperpapor waswas preSentedpresented atat thethe WinterWinter 19719788 LSLSAA meetingmeeting.. I wisWishh ttoo thanthankk CaroCaroll FF.. JustuJustuss anandd TheodorTheodoree MM.. LightneLightnerr foforr theitheirr comrnentcommentss oonn iit t andand alsoalso SandraSandra A.A. ThompsonThompson andand PatrickPatrick K.K. FordFord foforr heljhelp) duiinduringg previoupreviouss stagestagess ooff thithiss researehresearch.. OOff coursecourse,, I taktakee fülfuJIl rere- sponsibilitsponsibilityy foforr presentatiopresentationn anandd analysianalysiss ooff ththee datadata..

Brought to you by | University of South Carolina Libraries (University of South Carolina School of L) Authenticated | 172.16.1.226 Download Date | 5/31/12 9:18 PM 4

TheThe datadata presentedpresented herehere areare fromfrom thethe twotwo languagelanguage groups,groups, IIrIlr andand Celtic,lCeltic,1 thatthat preservepreserve severalseveral IEIE actionaction nounsnouns asäs infinitives.infinitives. I concludeconclude thatthat OldOld IrishIrish (OIr)(Olr) hashas subordinatesubordinate clausesclauses veryvery closeclose toto whatwhat PIE had,had, whilewhile thethe RigvedicRigvedic (RV)(RV) periodperiod ofof SanskritSanskrit hashas innovatedinnovated fromfrom purelypurely nominalnominal status;Status; AvestanAvestan (Av),(Av) , representingrepresenting OldOld Iranian,Iranian, alsoalso showsshows tracestraces ofof earlierearlier stagesstages eveneven thoughthough itit hashas standardizedstandardized objectobject casecase selectionselection whichwhich waswas variantvariant inin IIr.Ilr. OtherOther topicstopics likelike thethe extentextent toto whichwhich morphologymorphology andand syntaxsyntax areare concon- nectednected andand whatwhat conditionsconditions werewere necessarynecessary toto allowallow incipientincipient infinitiveinfinitive analysesanalyses willwill bebe explored.explored. I willwill alsoalso sketchsketch thethe morphomorpho- logicallogical andand syntacticsyntactic consequencesconsequences ofof thisthis reanalysisreanalysis andand comparecompare themthem withwith otherother IEIE languageslanguages whichwhich havehave moremore developeddeveloped infinitiveinfinitive clauses.clauses.

2.2. INDO-IRANIANINDO-IRANIAN AND'AND' CELTICCELTIC

TheThe infinitivesinfinitives ofof RVRV andand Av areare obliqueoblique casescases ofof productiveproductive actionaction nounnounss withwith thethe onlyonly exceptionsexceptions beingbeing RRVV -dhyai andand AAvv -dyai-dyäi («Il

1 1 MyMy descriptiondescription ofof CelticCeltic isis basedbased uponupon OldOld IrishIrish whichwhich isis thethe bestbest attestedattested ooff ththee oldeolderr languagelanguagess iinn thathatt groupgroup.. TraditionaTraditionall descriptiondescriptionss ooff thesthesee lanlan- guageguagess tentendd ttoo taktakee ththee positiopositionn thathatt ththee verbaverball nounounn iiss nonott aann infinitiveinfinitive (e.g(e.g.. ThurneyseThurneysenn 19461946:: 455455,, DilloDillonn 19551955:: 111122 f.)f.);; iinn thithiss discussiodiscussionn I shoshoww thathatt iinn certaicertainn clausclausee typetypess infinitivainfinitivall Statustatuss shoulshouldd bbee assigneassignedd ttoo somsomee OlOldd IrisIrishh structuresstructures.. 2 2 BuButt cfcf.. argumentargumentss liklikee thosthosee bbyy BartholomaBartholomaee 18901890,, SchwyzeSchwyzerr 1953/11953/1:: 808099 whicwhichh trtryy ttoo matcmatchh *-dhyäi*-dhyai witwithh ththee GreeGreekk middlmiddlee infinitivinfinitivee -sthai.-8thai. ThiThiss correspondenccorrespondencee iiss untenabluntenablee oonn phonologicaphonologicall groundsgrounds:: HIIrr *-dhy*-dhy- correspondcorrespondss ttoo GGkk -ss--88- (Benvenist(Benvenistee 1935b1935b:: 208)208).. I havhavee discussediscussedd elsewherelsewheree (Disterhef(Disterheftt 19801980:: 34—4034-40)) earlieearlierr elaimsclaims,, notablnotablyy bbyy BenvenistBenvenistee 1935a:1935a: ChapteChapterr 22,, thathatt -*dhyäi-*dhyai iiss a medio-passivmedio-passivee infinitiveinfinitive.. HiHiss twtwoo critericriteriaa areare:: 11.. iiff ththee infinitivinfinitivee hahass ththee samsamee meaninmeaningg äass ththee finitfinitee medio-passivemedio-passive,, iitt hahass ththee samsamee voicvoicee äass wellwell;; 22.. iiff ththee finitfinitee paradigmparadigmss araree predominantlpredominantlyy middlmiddlee oorr passiv:p~s~ivee ththee infinitiv~i~tiv~e iiss likewis~ikewisee middlemiddle.. MMyy examinatioexamir~ationn ooff ththee datdataa upouponn whichwhlCn* Benveniste'BenvenISte ss claiclaImm iISs basebasedd ddoo nonott supporsupportt hihISs argumentargument:: whilwhilee ououtt ooff a totatotall ooff 3333 verbaverball Sternss.tems,, 8 araree exclusivelexclusivelyy anandd 44 araree predominantlypredominantly middlemiddle,, 7 appeaappearr equallequallyy ofteoftenn iinn botbothh activactivee anandd middlemiddle,, 1111 araree infrequentlyinfrequently middlemiddle,, anandd 3 araree exclusivelexclusivelyy activeactive.. SincSincee middlmiddlee voicvoicee doedoess nonott changchangee tht~ee .marking.marking ooff agent'agent'ss ana~dd patient'pat~ent'ss relationshirel~tionshipp ttoo tht~ee predicatepredicate,, iitt iiss ofteoftenn difficuldlfflCul~t tt.oo infeInfe~r thathatt middlmIddlee iISs thth~e voicvOIcee ttoo bbee assigneassignedd ttoo ththee predicatpredicatee unlesunlesss iItt iISs obviouobVIOUSs frofromm ththee meaninmearungg ooff ththee verbaverball stestemm oorr frofromm ththee resrestt ooff ththee sentencesentence.. Brought to you by | University of South Carolina Libraries (University of South Carolina School of L) Authenticated | 172.16.1.226 Download Date | 5/31/12 9:18 PM 5 languages,languages, isis nnoo longerlonger useusedd nominallnominallyy inin Indic.Indic. -tavai (u)(u) is,is, accordaccord- inging ttoo ThurneysenThurneysen 1908,1908, derivedderived fromfrom -tave (dative(dative -tu- stem).3stem).3 RRVV andand AAvv mostmost commonlycommonly ususee dative,dative, accusative,accusative, andand locativelocative casescases ofof *-tu-,*-ta-, *-ti-,*-&*-, *-as-,*-os-, rootroot,, andand variouvariouss extendedextended heterocliteheteroclite stems.stems. OIrOlr verbalverbal nounnounss (the(the termterm forfor actionactioii nounsnouns inin traditionaltraditional CelticCeltic grammar)grammar) reconstrucreconstructt ttoo basicallybasically ththee samesame inventory,inventory, thoughthough somsomee restructuringrestructuring hahass taketakenn place.place. DativDativee andand accusativeaccusative OIrOlr verbalverbal nounnounss appearappear inin clausesclauses similarsimilar ttoo thosthosee ofof IIrIlr eveevenn thougthoughh thetheyy areare withouwithoutt exceptionexception membermemberss ofof nominanominall paradigms.paradigms. SinceSince mosmostt IIrIlr infinitivesinfinitives andand allall OIrOlr verbaverball nounsnouns areare morphomorpho- logicallylogically identicaidenticall ttoo nouns,nouns, iinn certaincertain subordinatesubordinate clausesclauses somesome dativedative andand accusativeaccusative actionaction nounnounss maymay bbee interpretedinterpreted asäs eithereither subordinatesubordinate predicatepre~~atess oror simplesimple NPsNPs.. ThisThis problemproblem ofof analysisanalysis hahass certainlycertainly beeilbeeh'recognized" recognized ; differentiationdifferentiation ofof substantivessubstantives fromfrom infinitivesinfinitives hashas beebeenn anan importantimportant themthemee inin previoupreviouss researchresearch witwithh emphasisemphasis placeplacedd onon morphologicamorphologicall factorsfactors oror objectobject casecase selectionselection (for(for literatureliterature seesee DisterhefDisterheftt 1980:1980r28-—34) '28-34).. HerHeree I shallshall focusfocus oonn ththee syntacticsyntactic structurstructuree ofof clauseclausess inin whicwhichh thesethese actioactionn nounsnouns appearappear inin orderorder ttoo determinedetermine ifif thetheyy areare simplesimple nounnounss oror predicatespredicates ofof subordinatesubordinate clauses.clauses. PastFast studiesstudies havhavee dependedependedd upouponn objecobjectt casecase selectionselection asäs thethe criterioncriterion forfor infinitivehood:infinitivehood: ifif objectobject isis thethe samesame casecase asäs thatthat ofof finitefinite verbverb (usually(usually accusative)accusative) ththee analysisanalysis ofof thethe passagpassagee inin questionquestion isis infinitival.infinitival. Conversely,Conversely, ifif ththee objecobjectt (or(or subject)subject) isis genitive,genitive, thethe formform isis nominal.nominal. ItIt isis indeedindeed thethe cascasee thatthat genitivegenitive agent/patientagent/patient markingmarking isis standardStandard withwith actionaction nounnounss asäs isis accusativeaccusative withwith "real""real" infinitives;infinitives; bubutt whenwhen wwee areare dealindealingg witwithh grammargrammarss thatthat havhavee transitionatransitionall formformss -— asäs ththee KRVV mostmost definiteldefinitelyy isis ttoo judgjudgee by·by itsits morphologmorphologyy -— ththee situationSituation isis nonott ssoo well-definedwell-defined.. ItIt cancan bebe arguedargued thathatt agent/patientagent/patient markinmarkingg isis nonott ththee onlyonly criterion.criterion. FoForr instance,instance, ththee infinitiveinfinitive isis predicatpredicatee inin sensen- tencetencess thatthat areare mildlymildly imperativalimperatival andand almostalmost alwaysalways havhavee accuaccu- sativesative object.object. ButBut inin (I)(1) (where(where thethe infinitiveinfinitive isis ththee locativelocative ooff anan extendedextended heterocliteheteroclite stem,stem, -tar-),-taro), ththee objectobject ofof ththee infinitiveinfinitive iiss genitivgenitivee eveneven thoughthough finitefinite formsforms ofof dty-dl~r- neveneverr seJectselect thatthat case.case.

3Thurneysen3 Thurneysen 19081908 suggestedsuggested thatthat .tavai-tavai withwith itsits anomalousanomalous secondsecond wordwprd accentaccent isis derivedderived fromfrom -tave (pre-Vedic(pre-Vedic *-tavai)*-tavai) followedfollowed bbyy ththee particleparticle vai:vdi: *-tavai*-tavai vai thenthen haplologizedhaplologized ttoo itsits historicahistoricall form.form. KleiKleinn 1978:1978: 164-7164-—7 agreedagreed onon mosmostt pointpointss withwith ThurneysenThurneysen,, addingadding thathatt 'Itu wawass addedadded ttoo ththee metricallmetricallyy incorrecincorrectt lineline thatthat resulteresultedd fromfrom ththee haplologyhaplology.. HeHe demonstl'l!'teddemonstrated nonott onlyonly thatthat -tavai isis useusedd iinn ththee samesame syntacticsyntactio andand metricalmetrical collocatIOnscollocations asäs •tave»-tave. butbut alsoalso thatthat itit appearsappears inin slightlyslightly youngeryounger textetexts,.

Brought to you by | University of South Carolina Libraries (University of South Carolina School of L) Authenticated | 172.16.1.226 Download Date | 5/31/12 9:18 PM 66

(1)(1) saaa ...... brahma'J')a8patirbrahmanaspatir druMdruho hantahantä maMmaha '{tMyartasya dhartaridhartari thisthis Brahma~aspatiBrahmanaspati LieLie slayerslayer greatgreat TruthTruth upholduphold nomnom nomnom gengen nomnom gengen gengen infinf (II.(II. 23.1723.17)) 'May'May thisthis Brahma~aspati,Brahmanaspati, thethe slayerslayer ofof thethe Lie.Lie. upholduphold thethe greatgreat Truth:Truth/

ThusThus wewe seesee thatthat thethe singlesingle featurefeature ofof objectob ject casecase selectionselection isis inadeinade- quatequate forfor purposespurposes ofof definingdefining infinitives,infinitives, asäs RenouRenou 19541954 hashas alsoalso recognized.recognized. JI proposepropose herehere thatthat moremore propertiesproperties ofof thethe clauseclause mustmust bebe usedused toto definedefine infinitiveInfinitive status,Status, atat leastleast inin thethe transitionaltransitional infinitiveInfinitive grammarsgrammars ofof IIrIlr andand Celtic.Celtic. A moremore reliablereliable approachapproach isis toto includeinclude thethe treatmenttreatment ofof subordinatesubordinate subject:subject :44 ifif itit hashas thethe samesame propertiesproperties thatthat subordinatesubordinate infinitiveinfinitive subjectsubject doesdoes inin lanlan- guagesguages withwith morphologicallymorphologically separateseparate ~initives,infinitives, thethe actionaction nounounn inin questionquestion shouldshould bebe consideredconsidered infinitival.infinitival. MajorMajor processesprocesses whichwhich subjectssubjects ofof infinitivesinfinitives undergoundergo are:are: a.a. equivalentequivalent nounnoun phrasphrasee deletion;deletion ; bb.. raisingraising ttoo object;object; c.c. raisingraising ttoo subject.subject. WheWhenn subordinatesubordinate subjectsubject iiss nonott coreferencoreferentt witwithh a maimainn clausclausee NNPP wwee alsalsoo find:find: dd.. overovertt subjecsubjectt iinn dativdativee oorr accusativaccusativee case.case. \ InIn 2.1—2.2.1-2.33 ththee featurefeaturess ooff purpospurposee clauseclausess anandd ververbb complementcomplementss iinn RVRV,, AvAv,, anandd OlOJrr wilwilll bbee describedescribedd anandd ·thenthen comparecomparedd ttoo simplsimplee object/obliquobject/obliquee cascasee nounsnouns.. ActioActionn nounnounss whicwhichh araree indeterminatelindeterminatelyy infinitives/substantiveinfinitives/substantivess wilwilll alsalsoo bbee introduceintroducedd anandd contrastecontrastedd witwithh formformss knowknownn ttoo bbee infinitives.infinitives.

'- 2.2.11 INDIOINDIC.. ThThee nonfinitnonfinitee purpospurposee clausclausee iinn RRVV useusess aann infinitiveinfinitive äass predicatpredicatee witwithh nnoo conjunctioconjunctionn introducinintroducingg itit.. SubjecSubjectt iiss eithereither deletedeletedd bbyy coreferenccoreferencee witwithh aann NNPP iinn ththee maimainn clausclausee (subject(subject,, direcdirectt objectobject,'., indirecindirectt objectobject,, possessivepossessive,, prepositionaprepositionall phrasephrase)) oorr iiss aann overovertt dativedative.. IInn (2)(2),, ththee -dhyai infinitivinfinitivee hahass accusativaccusativee

4 4 SentenceSentencess witwithh aann infinitivinfinitivee subjecsubjectt liklikee LaLatt errare humanumhumanum esestt ddoo nonott occuoccurr iinn languagelangua.gess witwithh nominallnominallyy associateassocia.tedd formsforms.. SententiaSententiall subjectsubjectss witwithh aann infinitivinfinitivee typicalltypicallyy appeaappearr iinn grammargrammarss wherwheree ththee infinitivinfinitivee hahass beebeenn separate~eparB:tedd frofro~m th~hee nominanominall paradigparadigm.m fo forr ssoo lon~ongg thathatt iitt retainretainss nnoo paradigmatiparadIgmatICc associationaSSOCIatIOn,, sucsuchh äass cascasee markingmarking,, whicwhichh woulwouldd keekeepp obliquobliquee casecasess frofromm appearinappearingg iinn subjecsubjectt slotsslots.. . .

Brought to you by | University of South Carolina Libraries (University of South Carolina School of L) Authenticated | 172.16.1.226 Download Date | 5/31/12 9:18 PM 7 objectobject withwith subjectsubject deleteddeleted byby coreferencecoreference withwith mainmain clauseclause subsub- jectject..

(2)(2) tatd wgramvigram dkaitkedJiaithe ialMramjatharam 'Pr'(Uidkyaipfnadhyai (VI.(VI. 67.767.7)) DEMDEM strongstrong taketake stomachstomach fillfill nomnom dudu accacc 2 dudu imvimv accacc infinf 'You'You botbothh shouldshould taketake thethe strongstrong oneone inin orderorder toto fillfill youryour stomach.'stomach/

InIn (3)(3) thethe -tavai ('II.)(u) infinitiveinfinitive hashas subjectsubject deleteddeleted bbyy coreferencecoreference withwith thethe objectobject ofof thethe mainmain clauseclause verb.verb. ItsIts dativedative objectobject isis commoncommon inin thisthis clauseclause type.type. ((<' cx (3)(3) brrikmii'(l.abrahmdna iMramIndram maMyantomdhayanto arlctiirarkair o,vardkayannavardhayann BrahmansBrahmans IndraIndra magnifyingmagnifying songssongs they-have-strengthenedthey-have-strengthened nomnom accacc nomnom ppIl instrinstr anayeähaye Mnta'VitMAtava 'II.u (V.(V. 31.4)31.4) snakesnake killkill datdat ininff 'The'The magnifyingmagnifying BrahmansBrahmans havehave strengthenedstrengthened IndraIndra witwithh songssongs soso thatthat (he)(he) wouldwould slayslay thethe snake:snake/

EvenEven thoughthough RVKV hashas nono morphologicalmorphological voicevoice distinction,distinction, passivizapassiviza- tiontion isis indicatedindicated inin allall clauseclause typestypes byby markingmarking patienpatientt asäs subject.subject. (4)(4) hashas anan infinitiveinfinitive whosewhose subjectsubject isis deleteddeleted bbyy coreferencecoreference witwithh matrixmatrix indirectindirect object.object.

(4)(4) 8itsa nana ita VaMvdha ...•0. . rayim ..•..„ i,ayadkyaiisayadhyai (VI.(VI. 64.464.4)) DEMDEM usus P brinbringg wealthwealth strengthenstrengthen nonomm sgsg enclencl 2 sgsg imvaccimv acc infinf 'Bring'Bring wealthwealth toto usus ...... thatthat .(we)(we) maymay bbee strengthened:strengthened/

NominaNominall infinitivesinfinitives areare predicatespredicates inin clausesclauses identicalidentical toto thethe nonnon- nominalnominal onesones ofof (2}-(4).(2)—(4). (5)(5) iU6.yeütaye isis a membermember ofof a paradigm.paradigm. (dative(dative -ti- stem)stem) attestedattested inin mostmost casescases ofof botbothh singularSingular andand pluralplural ItIt isis listedlisted inin Grassmann'sGrassmann's WorlerbuchWörterbuch zumzum Rigveda (1873)(1873) onlyonly .underunder thethe nounnoun üti-uti- 'help'.'help*. However,However, thisthis dativdativee actionaction nou nounn iiss useusedd asäs anan infinitivalinfinitival predicatepredicate inin anan embeddedembedded clauseclause (subjec(subjectt deleteddeleted bbyy coreferencecoreference withwith matrixmatrix object,object, accusativeaccusative object):object):

Brought to you by | University of South Carolina Libraries (University of South Carolina School of L) Authenticated | 172.16.1.226 Download Date | 5/31/12 9:18 PM 8

(5)(5) eva vasifthavll8i~/ha Uulramindram ütayefUage nfn ...... gynätigr1]iiti (VII.(VII. 26.526.5)) thereforthereforee VasishthVasishthaa IndraIndra helhelpp memenn he-praisehe-praisess nonomm accacc ininff accacc 'Therefore,'Therefore, VasishthaVasishtha praisespraises IndrIndraa inin orderorder thatthat (he)(he) helhelpp ththee men:men/ vipfccharh 'inquire''inquire' (6)(6) iiss ththee accusativeaccusative ofof a rooroott nounounn iinn a purposepurpose clausclausee andand,, liklikee isayadhyaii~agQ,dhgai iinn (4),(4), isis ththee onlyonly constituentconstituent iinn itsits clauseclause ttoo appearappear inin thethe sentence.sentence. TheThe rooroott pfch-prch- iiss rarelrarelyy nominalnominal- izedized andand isis foundfound onlyonly inin ththee dativedative andand accusativeaccusative cases;cases ; inin thisthis passagpassagee preverpreverbb vi + accusativeaccusative actionaction nounounn cancan onlyonly bbee interinter- pretepretedd infinitivally.infinitivally.

(6)(6) upo emi cikit~ocikituso vipfccham (VII(VII.. 86.386.3)) towardstowards II-g -goo wise-onewise-oness inquirinquiree accacc infinf 'I approachapproach thethe wiswisee onesones iinn orderorder ttoo inquire:inquire/

LikewisLikewisee ththee dativdativee ofof a productivproductivee stemstem mamayy alsoalso bebe predicatpredicatee inin a clauseclause wherwheree itit isis thethe onlyonly constituent.constituent. subM'shine'oubhe 'shine' (7)(7) isis thethe dativedative ofof a rooroott nounnoun whichwhich isis attestedattested inin a fullfüll paradigmparadigm iinn thethe singular.singular.

(7)(7) samänamsamiinam angy anjateänjate subhesubM himkam samesame ointmentointment they^anoint-themselvethey-anoint-themselves'shines shine PTPT accacc accacc infinf (VII.(VII. 57.357.3)) 'They'They anointanoint themselvesthemselves inin ththee samesame ointmentointment iinn orderorder ttoo shine. ''

(8)(8) andand (9)(9) illustratillustratee dativedative subjectsubject witwithh dativedative infirdtiveinfinitivess formedformed onon rooroott stemsstems:: vrdM (8)(8) iiss intransitiveintransitive;; dfoedrse (9)(9) hahass accusativeaccusative object.object. LikeLike utayeütaye inin (5),(5), vrdMvfdM isis listedlisted in-thein"the dictionarydictionary underunder ththee nominanominall paradigm.paradigm. d?se,drse, howeverhowever,, isis foundfound asäs anan Infinitivinfinitivee underunder ththee verbalverbal roorootdts-t - andand hashas nnoo extantextant nominalnominal paradigm.paradigm.

(8)(8) dbhiabhi gandharvam atynadatr'l)ad abudhne§uahudhnösu rajaltsvrajdhsv toto GandharvaGandharva he-crossedhe-crossed bottomlesbottomlesss skiesskies accacc - locloc locloc aä indro brahmabhyabrahmtähya idid V'{dMvfdha (VIII.(VIII. 77.577.5)) P IndraIndra BrahmansBrahmans PTPT prosperprosper . nonomm datdat ininff

Brought to you by | University of South Carolina Libraries (University of South Carolina School of L) Authenticated | 172.16.1.226 Download Date | 5/31/12 9:18 PM 9

'Indra'Indra crossedcrossed ttoo ththee GandharvaGandharva inin thethe bottomlessbottomless skiesskies iinn orderorder thathatt thethe BrahmanBrahmanss mighmightt prosperprosper:/

(9)(9) ud-ad u tyarhtyam jätavedasamjatavedasarh devamdevarh vahaniivahanti therePTthitherePTthiss JatavedasJatavedasaa gogod.d they-carrythey-carry accacc accacc accacc ketavalbketavah dr8e vtäväyav£8vaya suryamsuryarh (I.(L 50.150.1)) rayrayss seesee allall sunsun nomnom infinf datdat acaccc 'The'The raysrays ·carrycarry thethe gogodd JatavedasaJatavedasa therethere iinn ordeorderr thathatt alalll mamayy seseee ththee sun:sun/

IInn ththee KVRV,, infinitivinfinitivee complementscomplements areare formedformed ttoo verbverbss whosewhose ./ 5 subjectssubjects controlcontrol equiequi deletiondeletion ofof infinitivinfinitivee subject.5 (10)(10) hahass thethe dativedative samdfsesarhdfSe 'see''see' witwithh accusativeaccusative object.object.

(10)(10) kavtnrkavtiir icchiimiicchämi samdföesarhdfse sumedhahsumedMl} (III.(III. 38.138.1)) poetpoetss I-wisI-wishh seseee wise-onwise-onee accacc ininff nonomm 'I,, thethe wisewise one,one, wishwish toto seesee thethe poets:poets/

IInn (11)(11) thethe complementcomplement infinitivinfinitivee iiss anan accusativeaccusative whichwhich alsoalso taketakess accusativaccusativee object:object:

(U)(11) oakemasakema vaj£noväjino yamamyamarh (II(II.. 5.1)5.1) may-we-be-ablemay-we-be-able war-horseswar-horses leadlead optopt accacc ininff 'May'May wwee bbee ableable toto lealeadd thethe wawarr horseshorses.'/

(12)(12) hahass accusativeaccusative ärdbhamarabham 'be'be alliealliedd withwith'^ äass aann intransitive:intransitive:

(12)(12) na pancabhirpaiu;ti,bhir duabhirdMhhir va§lyva$fy artibharhäraJbham (V(V.. 34.534.5)) NENEGG fivefive tetenn he-wishehe-wishess be-alliebe-alliedd instrinstr instrinstr ininff 'He'He doesdoes nonott wishwish ttoo bebe alliedallied withwith fivfivee (or)(or) withwith tenten (men):(men)/

II RV verbal roots in this category ara vaS-, i~- 'wish', ju~- 'like, deRire', * KV verba 4l roots in this categoiy are -, i$- 'wish', jut- 'like, desire', jnä-,jna-, cit-,cü-, vid- 'know',know', man- 'think''think',, ^Aak-, /c-, arh-,arn-, i~-$- *be'be,, able*able',, arthaya-arthaya- 'stive''st,iva',, am8-arhs- ·succeed'.*succeed*.

Brought to you by | University of South Carolina Libraries (University of South Carolina School of L) Authenticated | 172.16.1.226 Download Date | 5/31/12 9:18 PM 1100

CompleraentComplementss ttoo mähwi8h mamayy alsalsoo havhavee overtovert,, noncoreferennoncoreferentt subjectsubject iinn ththee dativdativee caseCMe::

(13(13)) yathäyatha ta u&maeiuSmasi istayeil}lUye ((1.L 30.1230.12)) whewhenn yoyouu we-wiswe-wishh hastehastenn dadatt ininff 'whe'whenn wwee wiswishh thathatt yoyouu hastenhasten''

IInn ththee abovabovee exampleexampless illustratinillustratingg purpospurposee clauseclausess anandd ververbb complementscomplements,, a combinatiocombinationn ooff featurefeaturess leadleadss ttoo ththee determinatiodeterminationn thathatt thesthesee araree Infinitivinfinitivee clausesclauses.. EitheEitherr ththee overovertt presencpresencee ooff a subjecsubjectt (8)(8),, (9) (9), , (13(13)) oorr a dativdativee (3(3)) oorr accusativaccusativee (2)(2),, (5) (5),, (9)(9),, 10)10),, (11(11)) objecol)jectt givegivess syntactisyntacticc confirmatioconfirmationn ttoo analysianalysiss äass infinitives.infinitives. LikewiseLikewise,, a nonnominanonnominall infinitivinfinitivee sucsuchh äass (4)(4) isayadhyaii~ayadhyai guaranteeguaranteess ththee samsamee Interpretatiointerpretationn eveevenn thougthoughh iitt ,isis ththee onlonlyy surfacesurface concon- stituenstituentt ooff ththee clauseclause.. ThThee accusativeaccusativess vipfcchamviptccham (6)(6) anandd arabhamärdbham (12)(12) araree ththee onlonlyy membermemberss ofof clauseclausess afterafter approach andand wish,wi8h, respectively,respectively, anandd thuthuss coulcouldd conceivablconceivablyy bbee analyzeanalyzedd äass direcdirectt objectsobjects.. However,However, thetheyy belonbelongg ttoo prefixeprefixedd stemsstems whicwhichh araree attesteattestedd onlonlyy inin infinitivainfinitivall casecasess (dative,(dative, accusative)accusative).. 8ubMsubM (7(7)) iiss ththee dativdativee ofof a well-establishewell-establishedd paradigparadigmm iinn a sentencesentence whicwhichh woulwouldd nonott makmakee sensesense ifif analyzedanalyzed noninfinitivally.nomnfinitivally. ParallelParallel toto thethe infinitivalinfinitival accusativesaccusatives ofof (6),(6), (11),(11), (12)(12) areare formsforms thatthat cancan onlyonly bbee interpretedinterpreted asäs directdirect objectsobjects ofop accusativesaccusatives ofof goagoall afterafter verbsverbs ofof motion.motion. ForFor example,example, inin (14)(14) pitim 'drink'' drink' (accusative(accusative -ti- stem)stem) expressesexpresses ththee goalgoal ofof yahi 'come' :

(14)(14) ugre~vugrasv in nunu...... trf,kadruke~utrikadrukesu pahi 86mamsomam iiulraindra •...•• powerfulpowerful PTPT nownow TrikadrukasTrikadrukas drinkdrink somasoma IndraIndra locloc loc..loc 2 sgsg imvimv accacc vocvoc yahi haribhyarhharibhyäm 8utasya.sutasya . pitimpiUm (II.(II. 11.17)11.17) comecome golden-onesgolden-ones extractextract drinkdrink 2 sgsg imvimv instrinstr gengen accacc 'Now,'Now, 0o Indra,Indra,...... drinkdrink thethe somasoma amongamong thethe powerfulpowerful TriTri- kadrukaskadrukas ...... comecome withwith thethe goldengolden onesones toto thethe drinkingdrinking ofof ththee extract.'extract/

HereHere thethe accusativeaccusative actionaction nounnoun isis certainlycertainly substantival:substantival: therethere isis nono evidenceevidence thatthat a coreferentialcoreferential infinitiveinfinitive subjectsubject hashas beenbeen deleteddeleted andand thethe meaningmeaning ofof thethe entireentire stanzastanza mi~igatesmitigates againstagainst interpretinginterpreting Brought to you by | University of South Carolina Libraries (University of South Carolina School of L) Authenticated | 172.16.1.226 Download Date | 5/31/12 9:18 PM I

1111 pttimpitim, ass anan accusativeaccusative infinitiveinfinitive.. yyahi hi ••.. «. ppiti1n im isis struoturallystruoturally paralleparallell toto pahipii,hi •••... somam. somam. Thus,Thus , inin 'drink'drink thethe somasoina ...... come coine toto thetho drinking'drinking* thethe twtwoo aoousativesaccusatives (the(the seoonsecondd derivedderived frofromm a verbaverball rootroot,, ththee firstfirst oneone notnot)) areare paralleparallell ttoo eacheaoh other.other. NotNotee thatthat thethe oonstituentsconstituents ofof (14)(14) (acousative(accusative actionaction nounnoun followingfollowing a ververbb ooff motionmotion)) areare ververyy closeclose toto (6)(6) upo emi&mi...... 'IlipfcchamvipfccJiam 'Iº approaohapproaoh inin orderorder ttoo inquire'.inquire'. InIn (I5)(15) thethe actionaction nounounn (utim.( tim, accusativaccusativee -Bi-ti- sternstem)) isis ththee directdirect objectobject ofof 'ask'*ask' :

(15(15)) vf~antamasyavfsantamasya Tihumahe mahe 'l'Ui.rhtim (I.(L 10.1010.10)) most-manlmost-manlyy we-askwe-ask helhelpp gellgen accacc 'We'We askask thethe helhelpp ofof ththee mostmost manlmanlyy one.'one/ '. < I. " DativeDative substantivessubstantives likewiselikewise appearappear inin sentencessentences similarsiinilar toto thosethose wherwheree thethe dativedative hahass beebeenn analyzedanalyzed ass a purposivpurposivee infinitiveinfinitive (5)(5),, (7),(7), (8),(8), (9).(9). FoForr example,example, thethe dativedative rootroot nounounn έηΜιέ8ubM (16)(16) isis identicalidentical toto thatthat ofof (7)(7)..

(16)(16) vak§altsuvaksahsu rukmdfirulcmii,ii, adhi yetire oubhoSubM (1.(L 64.464.4)) breastbreastss goldgold uponupon they-binthey-bindd adornment'adornment0 100loc acoacc pIpl datdat 'They'They binbindd thethe goldgold uponupon theirtheir breastbreastss ass anan adornment.'adornment/

InIn contrastcontrast ttoo ththee actionaction nounnounss whicwhichh areare infinitivalinfinitival (2)-(13)(2)—(13) oror nominanominall (14)-(16),(14)—(16), therethere iiss a larglargee groupgroup ofof dativesdatives andand accuaccu- sativessatives whosewhose statusStatus iiss difficultdifficult ttoo establish.establish. InIn additionaddition ttoo aann infinitivinfinitivee interpretation,Interpretation, ththee dativdativee formsforms maymay bebe analyzeanalyzedd ass nounnounss expressinexpressingg ththee goalgoal ooff aann actioactionn whilwhilee ththee accusativeaccusativess mamayy bbee eithereither ththee directdirect objectobject ofof a transitivtransitivee ververbb oror ththee goalgoal ofof a motiomotionn verb.verb. FoForr example,example, 800MsubM (17),(17), unlikunlikee thethe infinitiveinfinitive ofof (7(7)) oror thethe substantiveSubstantive ofof (16),(16), hashas anan analysisanalysis whicwhichh isis indeterminate.indeterminate.

(17)(17) ένύΗέAu1JM kdrhkam yyanti nti rathat1trbhirrathatarbhir aa8vai(/, vaih (1.(I. 88.288.2)) glorify/glorglorify/gloryy PPTT they-gothey-go wagon-pullinwagon-pullingg horsehorsess infinf datdat NNPP instrinstr ppIl instrinstr

•β TheThe differencedifference betweenbetween Lhetho gl08f\gloas ofof aubhoBubhG herahere andand inin (7)(7) iHw tho,tho rlll'ult ofof changeschanges inin transitivity;transitivity: έυώΗ-Aubh· isi translatetra.nsla.tedd a.sa 'shine,'ehino, glow'glow* whonwhen IOt.rull·intmn· sitive,itive, butbut 'adore,*adore, glorify'glorify* whewhenn transitivetra.nsitivG..

Brought to you by | University of South Carolina Libraries (University of South Carolina School of L) Authenticated | 172.16.1.226 Download Date | 5/31/12 9:18 PM 1212

dativedative ofof goal:goal: 'They'They gogo withwith thethe wagon-pullingwagon-pulling horseshorses forfor gloryglory.'/ PurposivePurposive infinitive:Infinitive: 'They'They gogo withwith thethe wagon-pullingwagon-pulling horseshorses inin orderorder toto bebe glorified.'glorified/

Ut4yeütaye (18)(18) isis thethe samesame actionaction nounnoun foundfound inin thethe purpospurposee clauseclause ofof (5).(5). HereHere itit hashas nono objectobject andand appearsappears afterafter invoke, asäs doesdoes thethe substantiveSubstantive ofof (15)(15)..

(18)(18) iiulrdmindram Iia huvaTiuva utayeütaye (I.(L 111.4)111.4) IndraIndra P I-invokeI-invoke helphelp accacc datdat NP/infNP/inf

ItItss dativedative cascasee andand thethe meaningmeaning ooff ththee entireentire sentencesentence yielyieldd thethe followingfollowing interpretations:interpretations: noun:noun: 'I invokeinvoke IndraIndra forfor help.'help/ infinitive:infinitive: 'I invokeinvoke IndraIndra inin orderorder thatthat (he)(he) mightmight help.'help/ ActionAction nounsnouns afterafter wish, desire likewiselikewise maymay bebe ambiguousambiguous sincesince ththee RRVV allowsallows botbothh nominalnominal andand sententialsentential objectsobjects withwith it.it. InIn (19(19)) thethe meaningmeaning ofof thethe actionaction nouI!.noun gatUmgätum cancan bebe extendedextended fromfrom 'a'a going'going' toto 'path''path' andand cooccurscooccurs withwith thethe dativedative namename manave:

(19)(19) 6;traaträ dasasyadäsasya ndm'UC~namuceh 8irooiro yadydd avartaypavartayp thenthen demondemon Namuci.Namuci headhead whenwhen you-caused-to-turnyou-caused-to-turn gengen gengen accacc manave gatumgätum icchdnicchan (V.(V. 30.730.7)) ManuManu go/pathgo/path desiringdesiring datdat . infinf accacc NPNP nomnom sgsg ptcptc

SinceSince gätumgatum isis accusative,accusative, itit cancan bebe interpretedinterpreted eithereither asäs thethe objectobject ofof desire ('at('at thethe timetime whenwhen youyou causedcaused thethe headhead ofof thethe demondemon NamucNamucii toto turn,turn, desiringdesiring a patpathh forfor Manu')Manu') oror itsits complementcomplement withwith dativedative subjectsubject (' ...... desirindesiring g thatthat ManuManu go').go'). TheThe similarsimilar clauseclause iinn (13)(13) differsdiffers fromfrom (19)(19) bbyy havinghaving a dativedative actionaction nounnoun whichwhich allowsallows onlyonly a nonobjective,nonobjective, infinitivalinfinitival interpretation.Interpretation. TheThe discussiondiscussion soso farfär hashas shownshown thatthat inin thethe RV : a.a. InfinitivesInfinitives inin verbverb complementscomplements andand purposepurpose clausesclauses appearappear witwithh botbothh nonnominal(-dhyai,nonnominal (-dhyai, -sani, -tavai (u)}(u)) andand nominalnominal (datives,(datives, accusativesaccusatives ofof productivproductivee a~tionaction nouns)nouns) forms;forms;

Brought to you by | University of South Carolina Libraries (University of South Carolina School of L) Authenticated | 172.16.1.226 Download Date | 5/31/12 9:18 PM 1313

bb.. AllAll nonnominalnonnominal formsforms (2)-(4)(2)-—(4) havehave exclusivelyexclusively infinit,ivalinfinitival interpretations ;; c.c. AccusativesAccusatives ofof productivproductivee paradigmparadigmss areare mostmost oftenoften usedused asäs complementscomplements toto subject-equisubject-equi verbsverbs (10.)-(12)(10)—(12) andand maymay bebe analyzeanalyzedd indeterminatelyindeterminately asäs complementscomplements toto oror directdirect objectsobjects ofof a finitefinite verbverb (19);(19); d.d. DativesDatives ofof productivproductivee paradigmparadigmSs areare predominantlypredominantly purposivepurposive infinitivesinfinitives (5),(5), (7),(7) , (8),(8), (9)(9) butbut SOmesome areare iildeterminatelyindeterminately predicatespredicates inin purpospurposee clausesclauses oror dativedative nounsnouns (17),(17), (18) ; e.e. ProductiveProductive accusativesaccusatives andand dativesdatives whichwhich areare onlyonly nominanominall (direct(direct objectobject (14),(14), (15)(15) oror dativedative ofof goalgoal (16),(16), respectively)respectively) preservepreserve thethe earliestearliest stagestage inin whichwhich thesethese hadhad onlyonly nominalnominal properties;properties ; f.f. DativesDatives whic:Q:(arewhicly

2.2.2.2. IRANIAN.IRANIAN. TheThe clausclausee typestypes andand syntacticsyntactic detailsdetails ooff thethe AvAv infinitivesinfinitives areare almostalmost identicalidentical toto thethe RV.BV. TheThe mostmost strikingstriking differencedifference isis thatthat accusativeaccusative objectobject hashas beebeenn generalizedgeneralized bbyy thethe timtimee thatthat ththee textstexts werewere recordedrecorded (20)(20).. (20)-(22)(20)—(22) illustrateillustrate purposepurpose clausesclauses whichwhich regularlyregularly havehave nnoo subordinatingsubordinating marker.marker. TheThe subjectssubjects ofof nonnominalnonnominal -dyai-dyäi (IIr(Ilr *-dhyai)*-dhyäi) anandd ththee dativedativess davoi, aniUeanäSe araree deleteddeleted bbyy coreferenc~coreference withwith mainmain clauseclause subject ':':

(20)(20) !caBO,kaßä aSaiaSäi drufJmdruföm dyqm zastayo ninl himhlm howhow TruthTruth LieLie I-would-placeI-would-place handhand downdown itit datdat accacc locloc 'accacc m~€f:Zdyaim&rqzdyäi ()wahyadwahyä mq()raiSmqdräis s§nghahyä8~ngkahya vmavaiilm~mavaitim sinqm davoidävoi cast-downcast-down youryour wordswords teachingteaching powerfulpowerful chasmchasm givgivee infinf instrinstr gengen aceacc aceacc ininff dragvasii,drdgvasü aä isIS dva/8angdväßzng mazdamazdä aniile qstqs- co,cä(Y44.14 (Y 44.14)) LiarsLiars toto themthem sufferingsuffering WiseWise brinbringg hostilityhostility andand 100loc aceacc aeeacc voevoc infinf aceacc 'How wouldwould I placplacee thethe LieLie inin thethe handhand ofof TrutTruthh inin orderorder toto castcast itit downdown bbyy thethe wordswords ofof youryour teaching,teaching, inin orderorder toto putput

Brought to you by | University of South Carolina Libraries (University of South Carolina School of L) Authenticated | 172.16.1.226 Download Date | 5/31/12 9:18 PM 1144

a powerfupowerfull chaschasmm betweebetweenn ththee LiarsLiars,, iinn Ordeorderr ttoo brinbringg sufferingsuffering anandd hostilithostilityy ttoo themthem,, o0 WisWisee One?One~''

ThThee onlonlyy exceptioexceptionn ttoo ththee accusativaccusativee objecobjectt rulrulee ooff ththee purpospurposee clauseclausess iiss ononee dativedative::

(21(21)) mqOrwn'11I48rml. spdntwnspantml. maraeta ...... uxineuxsne xraQwexra{)we (V(Vdd 4.454.45)) worwordd holholyy they-speathey-speakk increasincreasee wisdowisdomm acaccc ac accc ssgg ininff datdat 'The'Theyy speaspeakk ththee holholyy worwordd iinn ordeorderr ttoo increasincreasee wisdomwisdom.'/

ThThee Infinitivinfinitivee iiss passivizepassivizedd bbyy promotinpromotingg patienpatientt ttoo subjectsubject.. IInn (22 (22)) ththee dativdativee infinitivinfinitivee witwithh passivpassivee readingisformedfroreadingisformedfromm aann extendextend- eedd heteroclitheteroclitee stestemm anandd hahass subjecsubjectt deletedeletedd bbyy coreferenccoreferencee witwithh maimainn clausclausee object:object:

(22)(22) aoi '11I4mmqm staomaine stüiöistuilJi yadaya()a P meme praisepraise praisepraise in-order-that .. accacc infinf 2 sgsg imvimv mämii aparacitaparaci! saofyantösaosyanto stavqn ((YY 9.29.2)) meme futurefuture saviorssaviors they-praisethey-praise accacc nomnom pIpl nonomm 'Praise'Praise meme thatthat I maymay bebe praised,praised, thatthat futurefuture saviorssaviors maymay praisepraise me.'me/ *

InIn additionaddition toto thethe purposepurpose clauses,clauses, thethe infinitiveinfinitive complementscomplements a listlist ofof subject-equisubject-equi verbsverbs whichwhich areare similarsimilar toto thosethose ofof thethe RVRV (23),(23), (24).7(24)7

(23)(23) visaiti dim /rayrarayofräyrärayo (N(N 19)19) he-knowshe-knows himhim enlivenenliven accacc infinf 'He knowsknows howhow toto enlivenenliven him.'him/

AsAs inin thethe RV,RV, wish admitsadmits a complementcomplement withwith independentindependent subjectsubject (accusative)(accusative) ::

7 7 AvAv rootsroots inin thisthis groupgroup are:are: aes-,ae$-, vas-vas- 'wish,'wish, desire',desire', iz-- 'be'be eager'eager' , manman- 'think','think', vaed-vaed- 'know','know', Xa8(Y)-xa§(y)- 'be'be able',able', Bpar3z-spam- 'strive','strive', upa-drarj-upa-drag- 'finish'finish , succeed'.succeed*. . ,

Brought to you by | University of South Carolina Libraries (University of South Carolina School of L) Authenticated | 172.16.1.226 Download Date | 5/31/12 9:18 PM 1515

(24)(24) aa!t toitoi atrm,ätrSm ahuraahurä •.•«.. u.mnahius&maM •..... stoi thethenn youyourr firefire LordLord we-wiswe-wishh bbee accacc vovocc ininff

rapante ciOrä.avarfidmciOra.ava'Y}ham (Y(Y 34.434.4)) supportersupporter brilliant-helbrilliant-helpp datdat acaccc 'Then'Then wewe wishwish,, 0o LordLord,, thatthat youryour firefire ...... mamayy bbee ofof brillianbrilliantt helphelp toto youyourr supporter.'supporter/ '

A morphologicalmorphological differencedifference betweenbetween RVBV andand AvAv isis thatthat thethe nominanominall paradigmsparadigms whichwhich yielyieldd ththee AAvv infinitivesinfinitives havhavee undergonundergonee a certaincertain amount(9famount,-of disintegration;disintegration; oftenoften onlyonly a fewfew casecase formsforms areare attestedattested forfor anyany actionaction nounnoun.. SyntacticSyntactic differencesdifferences lielie mainlmainlyy inin ththee confinementconfinement ofof objectobject casecase toto accusative.accusative. StillStill thertheree isis somesome evidenceevidence thathatt AvAv actionaction nounsnouns areare ap.timesat times indeterminatelyindeterminately nominomi- nal/infinitival.nal/infinitival. (25),(25) , forfor example,example, hashas a dativedative verbaverball abstractabstract fräxSndne/raun'fJne 'instruct''instruct' whicwhichh isis ambiguouslyambiguously a nominalnominal dativedative ofof goagoall oorr a passivpassivee infinitive.infinitive.

(25)(25) hyat-cähyal-Cii moi mraoSmraos aAamaSvm, jasöJaso fräxswdnefraun'fJne ((YY 43.12)43.12) thithiss andand mmee you-sayou-sayy TruthTruth you-goyou-go instructinstruct accacc dadatt acaccc dadatt infjNPinf/NP nounnoun:: 'This'This yoyouu saysay ttoo me : "You"You gogo toto TruthTruth forfor instruction":instruction'Y infinitive:infinitive: ' ...... iin n orderOrder ttoo bbee instructed.'instructed/

InIn contrast,contrast, nono ambiguityambiguity cancan bbee detecteddetected withwith verbverb complements.complements.

2.2.33 OLDOLD IRISIRISHH usesuses botbothh accusativeaccusative andand dativedative verbaverball nounnounss inin subordinatesubordinate clausesclauses similasimilarr toto thethe aboveabove examples.examples. DativeDative verbalverbal nounsnouns (preceded(preceded byby prepositionpreposition do 'to'S)'to'8) areare useusedd inin purpospurposee clauseclausess whilwhilee accusativeaccusativess araree predicatespredicates iinn complementcomplementss whoswhosee subjectssubjects areare eithereither coreferentialcoreferential oror independentindependent (and(and thusthus overt).overt). TheThe OIrOlr verbaverball nounounn appearsappears inin ththee restrest ofof ththee grammargrammar inin manymany uses,uses, mostmost oftenoften asaäs a simplesimple nounnoun:: subject/objectsubject/object ofof ththee finitefinite verb,verb, obliqueoblique prepositionaprepositionall phrases,phrases, etc.etc. ThisThis isis duedue ttoo ththee facfactt thatthat thethe verbaverball nounnoun isis completelycompletely withiwithinn ththee nominanominall paradigparadigmm andand nominanominall

8 8 OtherOther pre:positionsprepositions ma.ymay a.lsoalso bebe usedused withwith thethe verbalverbal nounnoun inin purposepurpose clausesclauses (e.g.(e,g. iINn witwithh accusative),accusative), butthesbut thesee areare nonott asäs numerousnumerous;; forfor exampleexampless seesee DisterhefDisterheftt 150150 Lf.

Brought to you by | University of South Carolina Libraries (University of South Carolina School of L) Authenticated | 172.16.1.226 Download Date | 5/31/12 9:18 PM 1616 versusversus infinitivalinfinitival analysisanalysis isis onlyonly determiueddetermined syntactically.syntactically. IfIf a fullfüll noun,noun, objectobject isis genitive;genitive; ifif pronominal,pronominal, itit isis a procliticproclitic pospos- sessivesessive pronoun.pronoun. TheThe dativedative verbalverbal nounnoun inin purposepurpose clausesclauses hashas nono subordinatorsubordinator butbut isis simplysimply placedplaced toto thethe rightright ofof thethe mainmain clause.clause. InIn (26)(26) subjectsubject isis deleteddeleted byby coreferencecoreference withwith a mainmain clauseclause NPNP andand objectobject iiss a procliticproclitic pronounpronoun suffixedsuffixed toto thethe prepositionpreposition do.da.

(26)(26) dandersaig dia dia m-brith huan chadraigchadraig he-stirred-himhe-stirred-him GodGod for-theirfor-their carrycarry from-thefrom-thecit cityy nomnom vn-datvn-dat datdat (M166(Ml 66 c 1414)) 'God'God stirredstirred himhim inin orderOrder toto carrycarry themthem fromfrom thethe city:city/

IndependentIndependent subjectsubject takestakes thethe formform ofof agentalagental dodo plupluss nounnoun (27(27)) oror pronounpronoun (28).9(28).9

(27)(27) dodo atrob dodo diadia andand (Wb(Wb 2121 c 7a)7a) ttoo dwelldwell ttoo GodGodin-i in-itt vn-davn-datt datdat 'tha'thatt GodGod mamayy dwelldwell inin it'it'

(28(28)) dduu buith dait anandd feinfein secech talmain (M(M184l 84 a 22)) ttoo bbee to-yoto-youu in-iin-itt itselitselff beyond-everybeyond-everjMan landd vn-davn-datt 'that'that yoyouu shoulshouldd bbee iinn iitt itselfitself beyonbeyondd evereveryy landland'"

IfIf ththee verbaverball nounounn iiss intransitiveintransitive,, a proclitiprocliticc possessivpossessivee pronoupronounn iiss optionalloptionallyy substitutesubstitutedd foforr ththee agentaagentall ddoo phrasphrase;e ; iinn sucsuchh examplesexamples,, ththee possessivpossessivee pronoupronounn subjecsubjectt isis,, liklikee ththee possessivpossessivee objecobjectt pronoupronounn iinn (26)(26), , suffixedsuffixed to·the to the dodo which which precedesprecedes thethe verbalverbal noun:noun :

(29(29)) dum fortachtfortacht huait- siu (M(MIl 8888 a 66)) to-mto-myy aiaidd by-yoby-youu PTPT . vn-davn-datt empemphh 'i'inn ordeorderr thathatt I be,aideb~ aidedd bbyy youyou''

9 9 ThiThiss ddoo phrasphrasee woulwouldd bbee difficuldifficultt ttoo consideconsiderr a trutruee subjecsubjectt sincsincee iitt iiss alsalsoo useusedd äass agenagentt iinn passivpassivee clauseclausess anandd iinn nominanoininall constructionsconstructions.. Brought to you by | University of South Carolina Libraries (University of South Carolina School of L) Authenticated | 172.16.1.226 Download Date | 5/31/12 9:18 PM 1717

10 ComplementsComplements toto subject-equisubject-equi verbs10 taktakee thethe formform ofof accusativaccusativee verbaverball nounnoup.ss (30),(30), (31).(31). ObjectObject markingmarking isis thethe samesame ass inin purposepurpose clauses:clauses: genitivegenitive nounounn objectobject (30)(30) oror proclitiprocliticc possessivepossessive propro- nounnoun (31)(31)..

30)30) conicimm digail 7 C08Ccosc neich (Wb(Wb 2200 d 66)) I-80m-ableI-am-able punishpunish anandd reprimanreprimandd anyonanyonee vn-acvn-accc vn-acvn-accc gegenn 'Iº amam ableable toto punishpunish andand reprimandreprimand anyone.'anyone/

(31)(31) aá n-ru-n-aninraccaigestar dia a ditin- 80m$ora(Ml62b21 (M162 b 21)) wbewhenn he-refusehe-refusedd GodGod hihiss protecprotectt PTPT .x,ã nonomm vn-accvn-aec emphemph 'when'when GodGod refusedrefused ttoo protectprotect himhim''

UnlikeUnlike ththee similarsimilar IIrIlr clauses·clauses wherwheree coreferencecoreference triggerstriggers manman- datorydatory subjectsubject deletion,deletion, itit isis optionaloptional witWithh activeactive butbut neveneverr applieappliedd ttoo passivpassivee verbaverball nounsnouns (cf.(cf. alsoalso (29»(29)) :

(32)(32) aiairr ni tormenatar- 80msom etir a forfor NENEGG they-thoughthey-thoughtt PPTT at-allat-aUthei theirr emphemph n-ditin 7 aann icc (M!(Ml 106106 d 1111)) protecprotectt andand theirtheir savsavee vn-acvn-accc vn-acvn-accc 'for'for theythey diddid nonott atat allall thinthinkk thathatt thetheyy wouldwould bebe protecteprotectedd anandd saved'saved*

ThatThat a n- 'their''their' isis indeedindeed ththee subjectsubject ofof a passivpassivee verbaverball nounounn anandd nonott thethe objectobject ofof anan activeactive oneone isis demonstrateddemonstrated byby thethe factfact thatthat thethe verbalverbal nounounn ditin paraphraseparaphrasess ththee LatinLatin protecti 8untsunt 'the'theyy werweree protectedprotected'' inin ththee passagpassagee whichwhich thethe OIrOlr glosses:glosses: contra ergo omnem 8pemspem DeDeii auxilio protecti 8untsunt (Ps(Ps 86)86) 'therefore'therefore aga.instagainst everyevery hopehope thathatt thetheyy woulwouldd bebe protecteprotectedd bbyy ththee helphelp ofof God'.God'. ComplementsComplements toto 8ay,say, declare, think, decide, believe, de8iredesire maymay havehave independentindependent subjectsubject inin thethe formform ofof thethe do agentagent phrasephrase::

10 10 TheseThese include:inolude: do-moinethar 'thinks','thinks', midithir 'decides','decides', ad-cobra 'desires','desiree', con-icc 'is*is able',able', dligid'deserves',dligid *d^erves', ro-tecro-uccc diriug 'succeeds','succeeds', ro-Zaimetharro-laimethar 'dares',*dares% intinnscanaintinnecana 'begins'.'begins'.

Brought to you by | University of South Carolina Libraries (University of South Carolina School of L) Authenticated | 172.16.1.226 Download Date | 5/31/12 9:18 PM 1188

(33(33)) ...• .is.is airiat" asbeir- som a epert doib (M(MI31l 31 b b 17)17) it-iit-iss on-accounton-account-- he-sayhe-sayss PPTT itssaits sayy to-theto-themm of-iof-itt empemphh vn-acvn-accc 'I'Itt iiss oonn accounaccountt ooff iitt hhee saysayss thathatt thetheyy easayy itit:/

(34(34)) do futhractarfuthractar forlor n-imdibe- eisi (W(Wbb 2200 c 2323)) they-desirthey-desiree youyourr circumciscircumcisee PTPT vn-acvn-accc empemphh 'The'Theyy desirdesiree thathatt yoyouu bbee circumcisedcircumcised.'/

ThThee complementcomplementss illustrateillustratedd iinn (30)—(34(30)-(34)) indicatindicatee thathatt eveevenn thougthoughh ththee accusativaccusativee verbaverball nounounn iiss fullfullyy a membememberr ooff a nominanominall paradigmparadigm,, ititss ususee correspondcorrespondss ttoo Infinitivinfinitivee constructionconstructionss iinn otherother IE languageslanguages.. OIr,Olr, liklikee IlrIIr,, hahass accusativaccusativee ververbb complementcomplementss whicwhichh areare indeterminate:indeterminate: thetheyy mamayy bebe eithereither embeddeembeddedd complecomple- mentmentss oror simplesimple nounsnouns.. ThiThiss happenhappenss iinn manmanyy casecasess whewhenn a genitivgenitivee isis useusedd whoswhosee relationshirelationshipp ttoo ththee verbaverball nounounn iiss unspecifiedunspecified.. FoForr instance,instance, inin (35)(35) ththee pronoupronounn a n- 'their''their' appe~rsappears beforbeforee ththee accuaccu- sativesative icic 'save':'save':

(35)(35) ni ru frescachtarfrescachtar an-ica n-ic (M(MIl 2266 b 2525)) NENEGG they-expectethey-expectedd theitheirr savsavee vn-accvn-acc

InIn thisthis sentence,sentence, a n-n- isis identicalidentical toto thethe subjectsubject ofof expect andand maymay bebe interpretedinterpreted asäs a possessivepossessive modifyingmodifying a substantivalsubstantival verbalverbal noun:noun: 'They'They diddid notnot expectexpect their·their salvation:.salvation/ JustJust asäs likely,likely, itit couldcould bebe a subjectsubject pronounpronoun coreferentcoreferent withwith thethe mainmain clauseclause subjectsubject,, followingfollowing thethe patternpattern ofof (32):(32): 'They'They diddid notnot expectexpect toto bebe saved:saved/ ThisThis ambivalenceambivalence isis reflected_reflected^i inn previousprevious translations,translations, suchsuch asäs Stokes'Stokes' andand Strachan'sStrachan's (1901:(1901: 51)51) asäs a nounnoun versusversus Gagnepain'sGagnepain's (1963:(1963 : 84)84) asäs a passivepassive infinitive.infinitive. CompareCompare (36)(36) withoutwithout thetheambiv ambiv- alentalent pronoun:pronoun:

(36)(36) conn- gestais huili taidchortaidchor so-thatso-that they-wo~ld-praythey-would-pray allall releaserelease nomnom vn-acvn-accc asaa indoiri ...... (MI(M l 131131 d d 13) 13 ) fromfrom captivitycaptivity datdat

Brought to you by | University of South Carolina Libraries (University of South Carolina School of L) Authenticated | 172.16.1.226 Download Date | 5/31/12 9:18 PM 1919 wherewhere ththee verbalverbal nounnoun maymay bebe analyzedanalyzed asäs directdirect objecobjectt ofof 'pray''pray* ('so('so thatthat thetheyy alalll shouldshould praprayy forfor releaserelease fromfrom captivity')captivity') oror asäs itsits complementeomplement ('('. ..•. .tto o bebe releasereleasedd fromfrom captivity').captivity'). AgaiAgainn paspastt transtrans- lationslations havehave beebeenn botbothh nominalnominal (Stokes(Stokes andand StrachanStrachan 451)451) andand passivpassivee infinitivalinfinitival (Gagnepain(Gagnepain 83)83).. PurposePurpose clausesclauses areare notnot liablliablee ttoo soso mucmuchh indeterminacyindeterminacy asäs thethe aboveabove accusativeaccusative complements,complements, bubutt doubledouble readingsreadings dodo sometimessometimes obtain.obtain. InIn (37)(37) ththee encliticenclitic genitivegenitive pronounpronoun precedinprecedingg ththee verbalverbal nounounn mamayy bebe reareadd asäs eithereither ththee subjectsubject ofof a passivpassivee infinitivainfinitivall verbalverbal nounnoun oror asäs ththee objecobjectt ofof a nominanominall one.one.

(37)(37) an- asäs torbe ddoo sochudi dia n-icc (Wb(Wb 1111 c 66)) whawhatt it-isit-is prpfiprpfitt ttoo multitudemultitude for-theirfor-their savesave nonomm datdat vn-davn-datt nounnoun:: 'What'What isis profiprofitt ttoo ththee multitudmultitudee foforr savingsaving them?them~'' infinitiveinfinitive:: 'What'What isis profiprofitt ttoo ththee multitudmultitudee thatthat theythey bbee saved?saved~''

(38)(38) iiss analyzableanalyzable asäs eithereither a passivpassivee complementeomplement witwithh nonappl(nonappli" cationcation ofof equi,equi, asäs anan activeactive complementeomplement withwith a procliticproclitic objectobject,, oror asäs a dativedative ofof goagoall witwithh possessiv~ossessivee pronounpronoun:: ..

(38)(38) ...... condarbaisconcIarbais frecndarcus dudu fortachtae dunaib so-that-they-so-that-they- presencpresencee youryourhel helpp to-theto-the may-shomay-showw gegenn trebaib ssoo dia soirad ....•• (M(MI.101l 101 c 77)) tribetribess PPTT to-theirto-their deliverdeliver datdat emphemph vn-datvn-dat nounnoun:: 'in'in ordeorderr thathatt yoyouu mamayy showshow ththee presencpresencee ofof youyourr helhelpp ttoo thesethese trihestribes for theirtheir deliverance'deliverance> infinitiveinfinitive:: ''...... iin n orderorder thatthat thetheyy bebe delivered';delivered'; 'in'in orderorder ttoo deliverdeliver themthem''

2.2.44 AAss mentionedmentioned inin thethe IntroductionIntroduction,, ththee morphologicalmorphological simisimi- laritlarityy betweebetweenn IIrIlr andand CelticCeltic isis preservepreservedd nowherenowhere elseeise iinn IIEE anandd cacann onlonlyy bbee considereconsideredd anan archaismarchaism.. WWee havhavee seeseenn thatthat purpospurposee clausesclauses areare formedformed inin ththee samesame way:way: thetheyy areare introducedintroduced bbyy nnoo conjunctionconjunction andand subjectsubject iiss deleteddeleted byby coreferencecoreference withwith a mainmain clauseclause NPNP.. TJsUsee ofof independent,independent, overtovert subjectsubject iiss founfoundd onlonlyy iinn RRVV withwith nnoo tractracee inin AvAv;; OIrOlr usesuses a prepositionaprepositionall agentagent phrasphrasee whicwhiohh

Brought to you by | University of South Carolina Libraries (University of South Carolina School of L) Authenticated | 172.16.1.226 Download Date | 5/31/12 9:18 PM 2020 isis probablyprobably a laterlater innovation.Innovation. InIn RV,RV, allall formsforme ofof infinitivesinfinitives .ma.mayy bebe used,used, butbut dativesdatives dodo predominate.predominate. InIn Av,Av, assignmentassignment ofof infinitiveinfinitive formform accordingaccording toto clauseclause typetype nono longerlonger obtains.obtains. InIn OIr,Olr, dativedative verbalverbal nounsnouns (always(always withwith prepositionpreposition do)do) areare predicatepredicate inin nonnon- finitefinite purposepurpose clauses.clauses. ObjectObject casecase ofof a purposivepurposive infinitiveinfinitive needneed notnot bebe thethe samesame asäs thatthat ofof finitefinite verbs:verbs: RVUV objectobject isis mostmost oftenoften accusative,accusative, butbut dativedative andand genitivegenitive areare foundfound inin substantialsubstantial numnum- bers.bers. Av,Av, whichwhich ha.~has leveledleveled infinitiveinfinitive objectobject toto accusative,accusative, preservespreserves oneone oror twotwo datives.datives. OIrOlr maintainsmaintains itsits strictstrict adherenceadherence toto genitivegenitive objectobject ofof verbalverbal nouns.nouns. I suggestsuggest thatthat onon thethe basisbasis oftheseof these languages,languages, latelate stagesstages ofof PIEPIE usedused thethe infinitiveinfinitive asäs predicatpredicatee inin purpospurposee clausesclauses withwith thethe followingfollowing featuresfeatures:: a.a. clauseclause hashas nono subordinator;subordinator; bb.. subjectsubject isis optionallyoptionally deleteddeleted bbyy coreferencecoreference witwithh a mainmain clauseclause NPNP;; c.c. objectobject isis usuallyusually notnot markedmarked likelike thatthat ofof thethe finitefinite verb,verb, butbut isis dativedative oror genitive;genitive ; d.d. infinitiveinfinitive hashas dativedative casecase (at(at leastleast inin earlie:r;earlier stages).stages). ThThee discussiondiscussion ooff ververbb complementscomplements soso farfar hahass beebeenn limitelimitedd toto complementscomplements ofof verbsverbs whosewhose subjectssubjects controcontroll coreferencecoreference witwithh lowerlower subjectsubject andand ttoo complementscomplements thathatt havehave anan independent,independent, overtovert subject.subject. OtherOther classesclasses ofof verbverb complementscomplements (object-equi(object-equi andand subject-raisingsubject-raising,, ttoo bbee discussediscussedd iinn §3§3)) simplysimply ddoo nonott showshow syntacticsyntactic parallelparallelse nonorr dodo thetheyy havhavee indeterminatindeterminatee analyses.analyses. VerbVerbss whoswhosee subjectsubjectss controlcontrol coreferenccoreferencee iinn IlIIrr anandd CeltiCelticc araree wish, like, know, think,fhinky decide, bbee able, succeed, strive, finish,finish, begin, etcetc.. WheWhenn anan objecobjectt appearsappears,, iitt iiss alwayalwayss accusativaccusativee iinn A v anandd mosmostt ofteoftenn ssoo iinn ththee RVRV;; OlOIrr adhereadheress ttoo ititss genitivgenitivee objecobjectt (fo(forr accusativeaccusative,, seseee (48(48)) (49(49)) below).below) .While While Av Av does does not not showshow a a preferencepreference for for anyany spespe- cificificc infinitivinfinitivee clasclasss oorr casecase,, RRVV hahass a predominancpredominancee ooff accusativesaccusatives herheree anandd OlOIrr limitlimitss thithiss clausclausee typtypee ttoo accusativaccusativee verbaverball nounsnouns.. DuDuee ttoo ththee sharesharedd featurefeaturess ooff aa.. accusativaccusativee infinitive,infinitive, bb.. genitivgenitivee oorr accusativaccusativee objectobject,, cc.. similaritsimilarityy iinn rängrangee ooff maimainn clausclausee verbverbss iitt appearappearss thathatt complemencomplementt ttoo aatt leasleastt severaseverall oofthef the just-mentionejust-mentionedd lislistt ooff subject-equsubject-equii verbverbss mamayy havhavee hahadd sententiasententiall objectobjectss witwithh infinitivinfinitivee iinn latlatee PIE.PIE. ThThee twtwoo clausclausee typetypess foforr whicwhichh I havhavee suggestesuggestedd a PIPIEE datdatee sharsharee aann additionaadditionall featurefeature:: thetheyy arare.e subjecsubjectt ttoo a certaicertain~mountn amount

Brought to you by | University of South Carolina Libraries (University of South Carolina School of L) Authenticated | 172.16.1.226 Download Date | 5/31/12 9:18 PM 2211 ofof indeterminacyindetermiiiacy inin eacheach ofof thethe languageslanguages bbyy virtuevirtue ofof ththee factfact thathatt theythey areare morphologicallmorphologicallyy identicalidentical ttoo nounnounss whicwhichh occuroccur iinn ththee samesame slotslot inin thethe sentence.sentence. TheyThey areare oftenoften toto bbee analyzedanalyzed asäs eithereither accusativeaccusative directdirect objects,objects, accusativesaccusatives afterafter motiomotionn verbs,verbs, oror dativesdatives ofof goal.goal. OneOne additionaladditional typetype doesdoes seemseem toto bebe reconstructreconstruct- ibleible forfor IIrIlr butbut nonott forfor otherother branchesbranches:: complementscomplements ttoo wish witwithh noncoreferent,noncoreferent, overtovert subject.subject. TheThe factfact thatthat thithiss typetype isis alsoalso subjectsubject toto indeterminacyindeterminacy indicatesindicates anan earlyearly development,development, butbut probablyprobably onlyonly withiwithinn thethe IIrIlr period.period. InIn concontrast tra.st, , purpospurposee clauseclausess witwithh overtovert subjectsubject (and(and othersothers discusseddiscussed inin §3§3 belowbelow)) areare neveneverr indeterminate.indeterminate. InIn OIr,Olr, ththee peculiarpeculiar patterninpatterningg ofof subjectsubject inin bothboth ververbb complementscomplements andand purpospurposee clausesclauses lendslends itselfitself ttoo multiplmultiplee analyses.analyses. ,/ .\ 33.. BEYONDBEYOND INDETERMINACYINDETERMINACY

InIn additionaddition ttoo ththee structuresstructtires disc'Q&.seddiscussed above,above, therethere areare complecomple- menmentt typestypes inin eacheach ofof thesethese languageslanguages thathatt undeunderr nono circumstancescircumstances cancan bbee confusedconfused witwithh simplesimple NPsNPs.. Furthermore,Furthermore, a lacklack ofof similaritysimilarity inin theirtheir syntacticsyntactic featuresfeatures indicatesindicates thathatt theythey havhavee developeddeveloped independently.independently. Object-equiObject-equi clausesclauses areare oneone suchsuch typetype:: bothboth KRVV (39)(39) andand AvAv (40)(40) havhavee suchsuch structures.structures.

(39)(39) sa i'lhIm mamiidamamäda muhimähi karma kartave (II.(II. 22.122.1)) DEMDEM hihimm he-hahe-hass greatgreat worworkk ddoo nomnom accacc inspiredinspired aceacc accacc ininff 'He'He hahass inspiredinspired hihimm ttoo dodo greatgreat work:work/

(40)(40) vav& naecisnaecls dar'iJstdarfst ite (Y(Y 43.1343.13)) yoyouu no-onno-onee he-compellehe-compelledd ggoo aceacc ppIl nonomm ininff 'No'No oneone compelledcompelled yoyouu ttoo go:go/

ItIt is,is, however,however, unlikelunlikelyy thathatt thithiss structurestructure developeddeveloped beforbeforee thethe 11 lIrIlr period:period: thethe inventoryinventory ofof maimainn clauseclause verbsll doesdoes nonott matcmatch;h ; unambiguouunambiguouse dativesdatives areare usedused asäs infinitivesinfinitives herehere;; OIrOlr doedoess nonott havehave thithiss strategy.strategy.

11 Verbs whose objects control equi deiet.ion are: RV i· 'clItreat', mand· 'inspire', mah· 'grant', vr- 'choose', Itu· 'call, invite', cit· 'perceive', avo 'help';*help'; AAvv kav·kav- 'promise',*pr nwtii~ 'command', daxA-, zAa·, 8qlt- 'teach', hak· 'incite','incite', dar£~darA· 'compel.^ompel^ dare',dare', yam-yam· 'allow'.'allow*.

Brought to you by | University of South Carolina Libraries (University of South Carolina School of L) Authenticated | 172.16.1.226 Download Date | 5/31/12 9:18 PM 2222

ÁAvν raiseraisess Infinitivinfinitivee subjecsubjectt ttoo objecobjectt ooff man- 'think'think'' anandd m rav-rav 'say'say'' bbyy markinmarkingg iitt ass maimainn clausclausee objecobjectt witwithh accusaaccusatt ivivee cascase:e :

(41(41)) afal QwOwa mvnghimlJnghi paourvim mazdmazda yazytizum m anandd yoyouu I-thoughI-thoughtt firsfirstt WisWisee laslastt acaccc acaccc vovocc acaccc atoi8toi τηαηαηΗάmanafJM ((YΥ 31.831.8)) bbee minmindd ininff instinstrr 4 'And And witwithh mmyy mindmind,, o0 WisWisee OneOne,, I havhavee thoughthoughtt yoyouu ttoo bbee ththee firsfirstt (and(and)) ththee lastlast.'/

RRVV hahass nnoo strategiestrategiess whicwhichh matcmatchh AAvv subject-to-objecsubject-to-objectt raisingraising.. BByy ththee timtimee ooff ClassicaClassicall SanskriSanskritt ththee infinitivinfinitivee onlonlyy hahass ononee form,form, -turn,-tum, whicwhichh hahass becombecomee disassociatedisassociatedd frofromm ththee nominanominall paradigparadigmm sincsincee ththee VediVedicc perioperiodd whewhenn ititss d&iive'dative' (-tave), alonalongg witwithh otheotherr stems,stems, wawass useusedd predominantlpredominantlyy iinn ththee samsamee functionfunction.. ThuThuss verbverb complementcemplementss (42(42)) ass welwelll ass purpospurposee clauseclausess (43(43)) iinn ththee latelaterr periodperiod regularlyregularly havhavee thethe samesame distincdistinctt infinitivinfinitivee formform (examples(examples fromfrom (Cardona(Cardona 1976:1976: 149)149)..

42)42) lcartumkartum icchati devadattalJ,-devadattah makmakee he-wanthe-wantss DevadattDevadattaa in!inf nomnom 'Devadatta'Devadatta wantswants toto makemake ... . .'/

(43)(43) bholctumbhoktum vrajati devadattalJ,devadattah eateat he-goeshe-goes DevadattaDevadatta iufinf nomnom 'Devadatta'Devadatta isis goinggoing inin orderorder toto eat.'eat/

However,However, ClassicalClassical SanskritSanskrit cancan stillstill useuse thethe dativedative actionaction nounnoun inin a phrasephrase equivalentequivalent toto (43):(43):

(44)(44) bhojanayabhojan ya vrajati devadattalJ,devadattah eateat he-goeshe-goes Devadatta-Devadatta datdat -- nonomm

UseUse ofof thethe samesame dativedative thematicthematic actionaction nounnoun hadhad earlierearlier beebeenn admittedadmitted inin thethe RVRV ass anan infinitive,infinitive, ass thethe accusativeaccusative objectobject inin (45)(45) -makesmakes clear:clear: _ -

Brought to you by | University of South Carolina Libraries (University of South Carolina School of L) Authenticated | 172.16.1.226 Download Date | 5/31/12 9:18 PM 2233

(45)(45) abodhi hotäMta yajathäyayajdthaya devdndevän (V.(V. 1.21.2)) he-has-awakenedhe~has-awakened sacrificersacrificer worshiworshipp godgodss nonomm ininff accacc 'The'The sacrificersacrificer hahass awakenedawakened iinn orderorder ttoo worshipworship thethe godsgods.'/

-aya-äya hahadd obviouslyobviously beenbeen parpartt ofof ththee patterpatternn ofof subordinationSubordination whicwhichh allowedallowed infinitivalinfinitival purpospurpose.e clausesclauses toto ususee anyany dativedative actionaction noun.noun. However,However, onconcee -turn-tum becamebecame dominantdominant,, -äya-aya thenthen resumedresumed itsits exclusiveexclusive functionfunction asäs actionaction nounounn strippedstripped ofof formerformer infinitivalinfinitival propertiesproperties.. IInn OIrOlr ththee subjectsubject ofof thethe dativedative verbalverbal nounounn mamayy bbee raisedraised toto objectobject ofofthink, think, know, andand hear (46).(46). HerHeree thethe lowerIpwer subjectsubject hashas notnot onlyonly accusativeaccusative casecase markinmarkingg bubutt hahass beebeenn movedmoved ttoo matrixmatrix objecobjectt \' Positiopositionn precedinprecedingg verbaverball noun..noun.

(46)(46) ma-ni fessed/e8sed comdidnadwmdidnad dodo ifif NENEGG it-kneit-kneww consolationconsolation toto'' *. accacc thiarmorru'ktthiarmoracht inaind uiteuilc (Ml(Ml 8787 d 44)) followfollow thethe evilevil vn-davn-datt gegenn 'if*if itit diddid nonott knoknoww thathatt consolationconsolation followsfollows evil'eviT

VerbalVerbal nounsnouns iinn subject-raisingsubject-raising environmentsenvironments (46)(46) differdiffer inin ononee importanimportantt aspectaspect fromfrom thethe complementscomplements toto subject-equisubject-equi verbsverbs (30)-(32):(30)—(32): dativedative casecase markingmarking precludeprecludess Interpretatiointerpretationn asäs direcdirectt object.object. This,This, coupledcoupled withwith ththee strongstrong infinitivalinfinitival characteristicscharacteristics ooff subject-raising,subject-raising, iiss anan importantimportant stepstep awayaway fromfrom nominanominall statusstatus.. TheThe dativedative verbaverball noun,noun, asäs aann unambiguousunambiguous embeddedembedded predicatepredicate,, appearsappears toto bebe expandingexpanding itsits useuse duringduring thethe OIrOlr periodperiod:: fromfrom onlonlyy subject-raisingsubject-raising verbsverbs toto includincludee thosethose withwith coreferentiacoreferentiall subject.subject. A fewfew examplesexamples areare foundfound inin whichwhich ththee dativedative verbalverbal nounnoun complecomple- mentsments subject-eqwsubject-equi verbs:verbs: bbee able, desire, love. NotNotee thatthat forfor thethe firstfirst timetime,, objectobject ooff thethe verbalverbal nounounn iiss a preposepreposedd accusativeaccusative::

(47)(47) ...... nini cumcat aithirgiaifhirgi ndonao denum (Ml(Ml 2323 a 55)) NENEGG they-are-ablthey-are-ablee repentancetrepentance too ddoo accacc vn-datvn-dat 'They'They areare notnot ableable toto dodo repentance:repentance/

Brought to you by | University of South Carolina Libraries (University of South Carolina School of L) Authenticated | 172.16.1.226 Download Date | 5/31/12 9:18 PM 2424

TheThe dativedative verbalverbal nounnoun inin anyany typetype ofof complementcomplement isis rarerare outsideoutside ofof thethe MilaMilann glosses.glosses. InIn thethe glossesglosses onon OIrOlr verse,verse, slightlyslightly laterlater iinn thethe periodperiod thanthan thethe biblicalbiblical glosses,glosses, restrictionsrestrictions onon subjectsubject corefcoref- erenceerence loosenloosen andand useuee ofof thethe dativedative verbalverbal nounnoun spreadsspreads toto as-heiras-beir inin thethe meaningmeaning ofof 'tell,'teil, order'order' (48).(48). HereHere thethe embeddedembedded subjectsubject isis deleteddeleted byby coreferencecoreference withwith thethe indirectindirect objectobject ofof 'tell'.'teil'. VerbaVerball nounnoun objectobject is,is, asäs inin (47),(47) , a preposepreposedd accusative.accusative.

(48)(48) atr'llbairlatrubairt Brenaind friaIria gilla a ehockullchochull he-toldhe-told BrendanBrendan to-theto-the servantservant hishis cloakcloak nomnom datdat acaccc dodo en'llrchur lorroforro (LH(LH 335.27)335.27) toto putput upon-theupon-themm vn-datvn-dat 'Brendan'Brendan toldtold hishis servantservant toto puputt hishis cloakcloak uponupon them.'them/

(48)(48) hashas replacedreplaced anan olderolder patterpatternn where,where, afterafter tell,teil, order, allow followedfollowed byby prepositionalprepositional phrase,phrase, thethe verbalverbal nounnoun waswas accusativeaccusative withwith postposepostposedd genitivgenitivee object:object:

(49)(49) intan asnindet dia 7 /orcongairforcongair dudu d6inihdoinib whenwhen he-speakshe-speaks GodGod andand he-ordershe-orders toto peoplepeople nonomm datdat eomalladcomattad a firinneI'rinne ...... (M}(M l 9494 b 33}) fulfillfulfill hishis righteousnessrighteousness , VD-accvn-acc gegenn 'when'when GodGod speaksspeaks andand ordersOrders thethe peoplepeople toto fulfillfulfill hishis righteous-righteous- ness'ness' "

(46)-(48),(46)—(48), then,then, showshow thathatt inin thethe laterlater OIrOlr perioperiodd a tendencytendency toto avoidavoid 'indeterminacyindetermiiiacy waswas starting:starting: dativedative verbalverbal nounsnouns couldcould havehave subjectsubject raisedraised toto mainmain clauseclause objectobject (46),(46) , couldcould complemencomplementt subject-equisubject-equi verbs'verbs (47),(47) , oror tellteil (48).(48) . RVRV andand AvAv havehave developeddeveloped awayaway fromfrom widespreadwidespread indeterminacyindeterminacy alsoalso withwith object-equiobject-equi strucstruc- turestures (39), (40)(40) andand AvAv hashas raising-to-objectraising-to-object (41).(41). WhileWhile RVRV stillstill preferpreferss dativedative infinitivesinfinitives inin purpospurposee clauses,clauses, otherother formsforms cancan appearappear inin thesethese clauseclause typestypes,, e.g.e.g. -dhyai (2),(2) , (4),(4) , -tavai (u)(u) (3),(3), andand accuaccu- sativessatives (6)(6) inin purposepurpiose clauses,clauses, dativesdatives (10)(10) inin complements.complements. ButBut byby thethe classicalclassical period,period, SanskritSanskrit hashas a morphologicallymorphologically separateseparate form,form, -tum.-turn. A v hashas nono tracetrace ofof thethe earlierearlier infinitiveInfinitive casecase assignmentassignment andand allowsallows anyany infinitiveinfinitive formform toto bebe usedused inin allall clauses.clauses.

Brought to you by | University of South Carolina Libraries (University of South Carolina School of L) Authenticated | 172.16.1.226 Download Date | 5/31/12 9:18 PM 2525

44.. OTHEROTHER INDO-EUROPEAINDO·EUROPEANN LANGUAGESLANGUAGES

OtherOther Indo-EuropeanIndo-Europeanlanguagespresumablywentthroug languages presumably went throughh a stagestage similarsimilar toto thatthat describeddescribed aboveabove forfor IIrIlr andand Celtic.Celtic. ByBy thethe timetime theythey areare attested,attested, however,however, mostmost havehave onlyonly oneone infinitiveInfinitive withwith "verbal""verbal" characteristics.characteristics. ThisThis is,is, ofof course,course, withwith thethe exceptionexception ooff modernmodern Irish,Irish, whichwhich hashas notnot developeddeveloped anan infinitiveinfinitive asäs a separateseparate category.category. HittiteHittite hashas twotwo forms,forme, -anna andand -(u)wanzi, whichwhich havehave longlong beenbeen separatedseparated fromfrom thethe dativesdatives thatthat theythey werewere formedformed fromfrom (Benveniste(Benveniste 1935b:1935b: 3030 f.,f., KammenhuberKammenhuber 1954:1954: 4444 f.).f.). -anna wawass originallyoriginally assignedassigned toto verbsverbs withwith ablautingablauting atems,sterns, -(-(u)wanziu )wanzi toto ththee others.others. GermanicGermanic usesuses *-onom (accusative(accusative n-stem;Tastern; cf.cf. GothiGothicc -an), BalticBaltic andand Sla~icSlayic havehave a *-ti- formform andand alsoalso a supinesupine (*-tum)(*-tum) toto expressexpress purposepurpose ~terafter verbsverbs ofof motion;motion; cf.cf. thethe similarsimilar LatinLatin supinesupine inin -um. InIn separateseparate developments,developments, thethe LatinLatin andand GreeGreekk infinitivesinfinitives havehave enteredentered thethe verbalverbal systemSystem andand eveneven acquiredacquired tensetense andand voicevoice marking.marking. MorphologioalMorphological voicevoice distinctionsdistinctions werewere apparentlyapparently notnot mademade inin Pro.to-Proto-ItalicItalic:: ,LatinLatin hashas activeactive ore-re (locativ(locativee 8-stem)5-stem) versusversus passivepassive -i-l (dative(dative rootroot noun)!-rinoun)/-n (dative(dative 8-stem)s-stem) whilewhile OscanOscan andand UmbrianUmbrian havehave oneone formform fromfrom accusativeaccusative *-um. TensTensee markingmarking waswas developeddeveloped inin thatthat periodperiod byby periphrasiperiphrasiss ofof thethe approappro- priatelypriately tensedtensed participleparticiple withwith toto bebe (Leumann(Leumann etet 0.1.al. 1972:1972: 342)342).. GreekGreek onon thethe otherother handhand showsshows evidenceevidence ofof a moremore unifiedunified morphomorpho- logicallogical development:development: -ein (or(or itsits dialectaldialectal reflexes;reflexes; Attic-IonicAttic-Ionic -un-een < *-esen; cf.cf. RVKV -8ani)-sani) waswas assignedassigned toto presenpresentt thematicthematic stemstemss veryvery early.early. LikewiseLikewise middlemiddle meaningmeaning waswas acquiredacquired atat anan earlyearly datedate byby -8t'hai-sthai (a(a disputeddisputed formation;formation; seesee notenote 2);2); butbut inin HomerHomer,, activeactive formsforme cancan stillstill be'be usedused wherewhere medio-passivemedio-passive isis expecteexpectedd (Schwyzer(Schwyzer 1953,1953, I:I: 805).805). illIn contrast,contrast, thethe appearanceappearance ofof -nai (dative(dative n-stem),n-stem), -menai (dative(dative extendedextended heteroclite),heteroclite), andand -men (locative(locative extendedextended heteroclite)heteroclite) forfor athematicathematic infinitivesinfinitives indicatesindicates thatthat thisthis waswas notnot standardizedstandardized untiluntil muchmuch laterlater inin thethe individualindividual dialects.dialects. ReturningReturning toto syntax,syntax, parallelsparallele toto IIrIlr dativedative objectobject casecase inin purpur- posepose clausesclauses areare Beenseen inin HittiteHittite eveneven thoughthough accusativeaccusative isis thethe standardStandard case :

(50)(50) nu SAL.MESSAL.MES ukturiya OaitiyaijiaStiyaS ld8uwanzileSSuwanzi PTPT womenwomen bonesbones collectcollect nomnom datdat ininff

Brought to you by | University of South Carolina Libraries (University of South Carolina School of L) Authenticated | 172.16.1.226 Download Date | 5/31/12 9:18 PM 2266

pänzipii,nzi (KU(KUBB XXXXXX 1155 I 1-21-2)) they-gthey-goo

'Th'Thee womewomenn ggoo ttoo ththee ukturiyaulcturiya ttoo colleccollectt bonesbones:/ Hittite'Hittite'ss dativdativee objecobjectt iiss undoubtedlundoubtedlyy inheriteinheritedd frofromm ththee samsamee sourcesourcess äass HrlIr.. IItt iiss reasonablreasonablee ttoo assumassumee thathatt ththee latlatee PIPIEE trantran- sitionasitionall infinitiveinfinitivess coulcouldd marmarkk theitheirr objectobjectss witwithh aa cascasee thathatt wawass neitheneitherr nominanominall (genitive(genitive)) nonorr whollwhollyy verbaverball (accusative)(accusative).. ThThee evolutioevolutionn frofromm fewefewerr ttoo mormoree clausclausee typetypess anandd ththee increasincreasee iinn verbverbss whicwhichh araree admitteadmittedd ttoo thesthesee structurestructuress araree confirmeconfirmedd bbyy otheotherr IIEE languageslanguages.. HittitHittitee preservepreservess a syntasyntaxx thathatt reflectreflectss ththee verbaverball affiliatioaffiliationn ooff -annal(u)wanzi.-anna/{u)wanzi. HerHeree mormoree ververbb complemencomplementt typetypess araree attestedattested.. NoNott onlonlyy ddoo a larglargee inventorinventoryy ooff subject-equisubject-equi verbverbss unambiguouslunambiguouslyy taktakee sententiasententiall objects,objects,1212 bubutt threthreee havhavee structurestructuress whoswhosee objectobjectss controcontroll equi-deletioequi-deletion:n : tarn- 'allow'allow,, let'let',, watarnalJ,lJ,-l·' Order''order',, fyalzäilJ,alziii-- 'calT'call',, äass iinn , , (51)(51) nu namma kiel SSAA KUR.KUR.T/JiKUR.KUR.PIMf LÜ.MELU.M:ESS PEMITEMI PPTT moreovemoreoverr thithiss PPTT landlandss memenn messagmessagee gegenn ge genn gengen MAffAEMAQAR nUTU.SIDUTU.&T uwauanziULuwauanzi ÜL tarniiitarnäi (KUB(KUB XIVXIV II frontfront Majesty-mMajesty-myy come.come NENEGG he-allowshe-allows Rs.Rs. 3030 f.)f.) 'Moreover,'Moreover, hehe doesnotdoes not letlet thethe messengersmessengers ofof thesethese landslands comecome beforebefore mymy Majesty:Majesty/

TheThe infinitiveinfinitive complementscomplements threethree impersonalimpersonal v~rbsverbs (UL(ÜL ara-ära- 'it'it isis notnot rightright (to)"(to)' , ULUL lciS-ki$- 'it*it isis notnot possiblepossible (to)"(to)' , frandäilJ,andai-- andand itsits SumerogramSumerogram SI+SA-SI+SÄ- 'be'be ordained')ordained') withwith whichwhich thethe lowerlower subjectsubject maymay eithereither remainremain inin thethe obliqueoblique case:case :

(52)(52) nu-mu-lcannu-mu- kau apiya-ya nISKURBISKUE URUlJATTIUHUgATTI EN.Y. A PTPTm mee PTPT thenthen andand StormgodStormgod HattiHatti Lord-myLord-my enclencl piran tiyauwanzitiyauwanzi lJandaittathandäittat (KUB(KUB VV 66 IVIV 1313)) forthforth gogp it-was-ordainedit-was-ordained infinf

12 4 III AmongAmongthe themm are:are: sanlJ-,wek~,ilaliya-$arih-,wek·,ilaliya- 'desire',desire', ep-ep- (reflex),zikk-(reflex),zikk-,$ABÄTSABATUU 'take,'take, begin',begin', a8sanu,a$$anu, irhai:-'irhäi-, tarup-,tarup-, zinna-zinna- 'be'be ready,ready, stop,stop, finish'finish' , handalliyahandalliva· 'dare','dare't J i , mat-mat-__« 4 'endure',*endiare'1^*»_, tarlJ-tarfy-T C. 1 'be'be able',able'* , waqqar-waqqar-i i» 1·»'fall 4fal1 l shortshor. <«..'"'·t (of)"(of)'. ,., -kars-"kars-. -V . _ 'fail*fail , neglectt,neglect', ZI-ZI- 'intend','intend', ling-ling- 'be'be boundbound byby oath',oath', tarkumlii-'tarkumäi- 'announce';'announce', KARABUKARÄBU 'promise','promise', memma-memma- 'refuse''refuse'.. . .

Brought to you by | University of South Carolina Libraries (University of South Carolina School of L) Authenticated | 172.16.1.226 Download Date | 5/31/12 9:18 PM 2727

'And'And thenthen itit waswas ordainedordained thatthat I gogo forthforth toto mymy Lord,Lord, thethe StormgodStormgod ofof Ratti.'Hatti/ oror bbee raisedraised toto subject:subject:

(53)(53) '11/1.£nu T*INGTR.LUMDINGffi.LUM piran tiyanna SIXSA.-atSIxSÄ- * (KUBV6IV8)(KUBV6IV8) PTPT godgod forthforth gogo it-was-ordainedit-was-ordained nomnom infinf 'The'The godgod waswas ordainedordained toto gogo forth'.forth'.

OtherOther mIE languageslanguages whichwhich areare attestedattested muchmuch laterlater alsoalso exhibitexhibit anan expansionexpansion inin numbernumber ofof clausesclauses withwith infinitive.infinitive, ArdArd 19751975 describesdescribes a shiftshift fromfrom finitefinite toto nonfinitenonfinite complementationcomplementation inin OldOld English.English. InIn thethe earlierearlier periodperiod thethe numbernumber ofof suchsuch complementscomplements waswas smallsmall butbut bbyy thethe laterlater parpartt ofof thatthat periodperiod,, mostmost verbverb typestypes werewere reprerepre- sentedsented (p.(p. 1616 f.).f.):"~Thus' 'Thus finitefinite complementscomplements toto verbsverbs likelike order,Order, accuse, permit, expect, intend asäs inin

(54)(54) pa heht hehe hiBhis geferan, paetjJaet hio 80htonsohton sumne earme pear/anpearfan (AElfred,(AElfred, Bede,Bede, 388,388, 1010)) 'Then'Then hehe orderedordered hishis companionscompanions thatthat theythey seekseek a certaincertain poorpoor beggar.'beggar/ usuallyusually antedateantedate anan infinitiveinfinitive complement.complement. ArdArd arguedargued thatthat thethe ModemModern EnglishEnglish constructionsconstructions whichwhich areare synchronicallysynchronically derivedderived viavia raisinraisingg cannotcannot bbee soso diachronically.diachronically. ReHe wentwent onon ttoo showshow thatthat thethe nonfinitenonfinite clauseclause typetype cannotcannot bbee derivedderived solelysolely fromfrom applyingapplying equiequi- NPNP deletiondeletion toto a finitefinite structurestructure sincesince thethe twotwo typestypes occuroccur sideside bbyy sideside inin earlyearly OldOld EnglishEnglish (p.(p. 23).23). TheThe datadata suggestsuggest ratherrather thatthat thethe embeddedembedded subjectsubject isis copiedcopied asäs objectobject ofof thethe mainmain clauseclause verb,verb, pronominalized,pronominalized, andand subsequentlysubsequently deleteddeleted (p.(p. 27).27) . RaisingBaising strucstruc- turestures withwith typicaltypical NP-infinitive-NPNP-infinitive-NP sequencessequences likelike

(55)(55) John believes Fred toto bebe a plumber. synchronicallysynchronically cancan undergoundergo to-be deletion,deletion, yieldingyielding twotwo derivedderived objectsobjects NPNPss ("predicate("predicate adjunctadjunct construction",construction", pp.. 32) :

(56)(56) John believesbelieves Fred a plumber.

ButBut likelike complementscomplements toto object-equiobject-eqüi clauses,clauses, NP-NNP-NPP sequencesequencess cannotcannot bebe diachronicallydiachronically derivedderived fromfrom NP-infinitive-NP:NP-infinitive-NP: theythey

Brought to you by | University of South Carolina Libraries (University of South Carolina School of L) Authenticated | 172.16.1.226 Download Date | 5/31/12 9:18 PM 2288 predatepredate raisedraised constructionsconstructions andand werewere oneone ofof thethe sourcessources ofof NPNP- infinitive-NPinfinitive-NP complementscomplements toto eay/Jay andand think.ihink. ThusThus withwith NP-NNP-NPP complements,complements, itit isis notnot thethe casecase thatthat oneone elementelement hashas beebeenn movemovedd fromfrom thethe embeddedembedded clauseclause toto thethe mainmain clauseclause withwith thethe restrest ofof thethe complementcomplement remainingremaining inin anotheranother clause.clause. RatherRather thethe entireentire concon- tenttent ofof thethe embeddedembedded complementcomplement whichwhich remainsremains onon thethe surfacesurface hashas beebeenn movedmoved toto thethe mainmain clauseclause andand hashas becombecomee thethe surfacesurface objectsobjects ofof itsits verbverb (p.(p. 40).40). ArdArd claimedclaimed thatthat thethe NP-infinitive-NPNP-infinitive-NP patternpattern maymay possiblpossiblyy havehave arisenarisen underunder LatinLatin influence,influence, bubutt hadhad difficultydifficulty explainingexplaining itsits spreadspread iinn laterlater EnglishEnglish periodperiodss asäs duedue toto LatinLatin style.style. HeHe notednoted (p.(p. 42)42) thatthat thisthis patternpattern becambecamee increasinglyincreasingly commoneommon afterafter 8ay,eay, think asäs casecase distinctionsdistinctions werewere lostlost andand thatthat casecase markingmarking maymay actuallyactually havehave hinderedhindered theirtheir acceptanceacceptance inin OldOld English.English. Raising-to-subjectRaising-to-subject likewiselikewise cannotcannot bbee ascribedascribed toto anyany singlesingle changechange thatthat cancan bebe describeddescribed asäs raisingraising inin thethe historyhistory ooff English.English. ArArdd stated:stated:

StructuresStructures whichwhich ar~ee derivedderived viavia Raising-to-SubjectRaising-to-Subject inin ModernModern EnglishEnglish developedevelopedd diachronically,diachronically, inin thethe main,main, fromfrom oneone oror bothboth ofof thethe followingfollowing sources:sources: 1..1. impexsonalimpersonal verbsverbs withwith twotwo arguments.arguments -.— anan obliqueoblique NNPP andand a sententialsentential complementcomplement forfor whichwhich thethe obliqueoblique NPNP becamebecame subjectified;subjectified; 2.2. personapersonall verbsverbs whichwhich OCCUlTedoccurred inin structuresstructures ofof thethe correctcorrect syntacticsyntactic shapeshape NPNP V nonfinitenonfinite V X

butbut withwith semanticsemantic interpretationInterpretation 'incompatibleincompatible withwith a raisingraising analysis,analysis, forfor whichwhich thethe semanticsemantic interpretationInterpretation changedchanged toto aJ,lowallpw a raisingraising analysisanalysis (p.(p. 53)53)..

TheThe historyhistory ofof LatinLatin alsoalso indiQatesindicates anan expansionexpansion ofof complementcomplement structures:structures: subje~t-equisubject-equi verbsverbs taketake infinitivesinfinitives inin thethe earliestearliest textstexts bubutt otherother typestypes ofof nonfinitenonfinite complementationcomplementation onlyonly appearappear later.later. MostMost ClassicistsClassicists agreeagree thatthat accusativeaccusative subjectsubject ·ofof thethe infinitiveinfinitive inin LatinLatin andand GreekGreek originatedoriginated inin clausesclauses afterafter order (Lat(Lat iubeo, GkGk keleo) (Schwyzer(Schwyzer II:II: 373373 f.,f., LeumannLeumann etet a1.al. II:II: 353353 f.;f.; alalll LatinLatin examplesexamples areare takentaken byby HahnHahn 19501950 fromfrom PlautusPlautus andandTerence) Terence) :

(57)(57) iube hunc abire (HT(ET 585-6)585—6) . orderorder himhim gogo imimvv accacc ininff 'OrderOrder himhim toto go:go/

Brought to you by | University of South Carolina Libraries (University of South Carolina School of L) Authenticated | 172.16.1.226 Download Date | 5/31/12 9:18 PM 2929

(58)(58) tlWreksaitt reksai hehe Mleu.eMleue ...... AchaioUsAchaious (Iliad(Iliad 2.11)2.11) armarm DEDEMM he-orderehe-orderedd AohaeanAchaeanss ininff nonomm accacc ¹'HeÈ ordereorderedd ththee AchaeansAchaeans ttoo armarm themselvesthemselves.'/

HahnHahn 19501950 showedshowed hohoww verbsverbs ofof perceptionperception (=(= object-equiobject-equi verbsverbs)) ververyy earlearlyy tootookk a sententialsentential objecobjectt consistingconsisting ooff a participlparticiplee modifyingmodifying andand agreeingagreeing witwithh ththee accusativeaccusative objectobject noun ;j thuthuss NP(acc)NP(acc)- participle(acc)participle(acc) obtainedobtained inin sentencessentences liklikee

(59)(59) vidi eett illam etet hospitemMspitem complexamcomplexam I-saI-saww andand heherr anandd guestguest huhugg accacc accacc ptc-accptc-acc osculantem (¥iI533-4)(Mil 533—4) kiss .• ( ptc-acptc-accc º'I sawsaw herher andand thethe guestguest hugginghugging andand kissing.'kissing/

OthertimesOthertimes ththee participlparticiplee isis useusedd ambiguously:ambiguously:

(60)(60) meminestin me gravidam? (HT(ET 626626)) you-remembeyou-rememberr meme pregnanpregnantt accacc ptc-acptc-accc yielrungyielding interpretatiointerpretationn ass attributiveattributive (NP-participle):(NP-participle): 'Do'Do youyou rememberremember mmee beinbeingg pregnant?pregnant~'' oror predicativpredicativee (NP-to(NP-to be-participle):6e-partieiple): 'Do'Do yoyouu rememberememberr thathatt I wawass pregnant?pregnant r ' SometimesSometimes hear cancan onlyonly havhavee a sententialsentential complementcomplement iinn whicwhichh casecase a predicativpredicativee relationshiprelationship isis establishedestablished betweebetweenn thethe accusativeaccusative NNPP andand itsits modimodifier:f ier:

(61)(61) faenus creditum audio (Most 629629)) interestinterest creditcredit I-heaI-hearr aceacc ptc-accptc-acc 'Iº hearhear thatthat itit hashas beenbeen creditedcredited forfor thethe interest.'interest/

HahnHahn thethenn demonstrateddemonstrated howhow sino 'allow'allow'' andand patiar 'suffer''suffer' areare useusedd botbothh withwith infinitive:Infinitive:

Brought to you by | University of South Carolina Libraries (University of South Carolina School of L) Authenticated | 172.16.1.226 Download Date | 5/31/12 9:18 PM 3300

(62(62)) siguidemsiquidem potespotu esseuse ttee patipati iinn lepidolepido lociloci (Poen 696696)) ijff you-areyou-are-- bbee yoyouu suffesufferr iinn pleaeanpleasantt placplacee ablablee ininff dadatt ininff lolocc lo locc 'i'iff yoyouu cacann suffesufferr ttoo bbee iinn a pleasanpleasantt placeplace'' oorr witwithh ththee NP-participlNP-participlee constructionconstruction::

(63(63)) vosneVOSM egegoo patiar cum mendicismerulicis you-Pyou-PTT I I-will-alloI-will-alloww witwithh poopoorr acaccc ppIl dadatt ppIl nuptasnwptas ...... viris? viris? (Stich(Stich 132)132) marriemarriedd memenn ptc-acptc-accc dadatt ppIl 'S'Soo shalshalll I alloalloww yoyouu (t(too bebe)) marriemarriedd ttoo poopoorr memenn t'V

SinceSince nnoo examplesexamples cacann bbee foundfound witwithh presenpresentt participlparticiplee oorr withwith nonparticipianonparticipiall adjectives,adjectives, HahHahnn surmised'surmised (p(p.. 121)121) thathatt iinn eacheach ooff ththee examplesexamples citedcited,, wewe havehave notnot a participlparticiplee bubutt a paspastt tenstensee infininfin- itiveitive (which(which isis formedformed frofromm paspastt participlparticiplee plupluss toto bebe)) withwith copulcopulaa deleted.deleted. ThusThus iinn

(64)(64) abductam illam aegre pati (Merc 251251)) abductedabducted herher scarcelyscarcely suffersuffer ptc-accptc-acc accacc infinf 'to'to scarcelyscarcely suffersuffer herher (to(to be)be) abducted'abducted'

l abductam isis underlyinglyunderlyingly abductam useesse withwith copulacopula (inf(inf u8e)esse) deleted,deleted, asäs isis regularregulär inin thethe restrest ofof Latin'Latin grammar.grammar. OnOn analogyanalogy toto suchsuch constructionsconstructions withwith pastpast infinitive,infinitive, presentpresent infinitives,infinitives, whicwhichh areare notnot formedformed periphrastically,periphrastically, areare possible:possible:

(65)(65) 8edsed tutu enumquam pi8cat0r..empiscatorem vidi8tividisti . ... . piscem butbut youyou everever fishermanfisherman you-sawyou-saw fishfish accacc accacc cepisse? (Rud(Rud 987~8)987—8) catchcatch infinf 'But'But diddid youyou everever seesee aa fishermanfisherman catchcatch aa fishfish? " '

OtherOther LatinLatin clauseclause typestypes alsoalso exhibitexhibit aa gradualgradual spreadspread throughthrough thethe rangeränge ofof possiblepossible structures.structures. SaltarelliSaltarelli 19791979 demonstrateddemonstrated thatthat

Brought to you by | University of South Carolina Libraries (University of South Carolina School of L) Authenticated | 172.16.1.226 Download Date | 5/31/12 9:18 PM 3131

ththee followingfollowing clauseclause typestypes areare sequentiallysequeiitially developeddeveloped bbyy causativescausatives witwithh facere 'make':'make*: a.a. finitefinite complementcomplement witwithh subordinatorsubordinator ut;ut; bb.. finitefinite complementcomplement withoutwithout ut;ut; c.c. subject-to-objectsubject-to-object raisinraisingg witwithh finitefinite complement;complement; d.d. subject-to-objecsubject-to-objectt raisingraising witwithh activeactive andand passivpassivee infinitives.infinitives. InIn RomanceRomance,, clauseclause uniounionn competescompetes withwith thethe reflexesreflexes ofof facerefacere andand isis currentlcurrentlyy ththee onlonlyy possibilitpossibilityy witwithh ItaliaItaliann farefare anandd SpanishSpanish faire. TheThe materiamateriall presentepresentedd herehere frofromm otherother IEIE languageslanguages indicatesindicates a continuingcontinuing developmentdevelopment inin eacheach fromfrom fewerfewer toto moremore clauseclause types.types. LikLikee purpospurposee clausesclauses anandd subject-equisubject-equi complements,complements, raisinraisingg andand object-equiobject-equi structuresstructures iinn OldOld BnglisEnglishh andand LatiLatinn evolveevolve fromfrom matemate- riariall inin ththee maimainn clauseclause whicwhichh developsdevelops propertiepropertiess ofof a predicatepredicate.. AlAlll thesthesee clausalflmovationsclausal innovations adaddd furtherfurther evidenceevidence ttoo Ard'Ard'ss claimclaim thatthat diachronicdiachronic processeprocessess cannotcannot bbee assumedassumed toto recapitulatrecapitulatee a _ synchronicsynchronic description.description. WhilWhilee ththee riserise ofof thethe fourfour clauseclause typestypes involvesinvolves a reanalysisreanalysis ofof maimainn clause.clause. surfacesurface strings,strings, thethe processeprocessess areare different.different. PurposePurpose clausesclauses andand subject-equisubject-equi complementscomplements develdevel- opedoped atat a timtimee whewhenn thethe infinitiveinfinitive wawass stillstill anan actionaction nounounn anandd partpart ofof nominalnominal paradigms.paradigms. InIn itsits capacitycapacity asäs ththee objectobject ofof a finitefinite ververbb oror asäs a dativedative addingadding materiamateriall ttoo ththee sentence,sentence, ththee actionaction nounnoun movedmoved intointo clausalclausal statusstatus asäs objectobject complementcomplement oror asäs a purpur- posepose clause,clause, respectivelyrespectively,, upouponn developingdeveloping subjectsubject coreference.coreference. OOnn thethe otherother hand,hand, wwee findfind object-equiobject-equi andand raisingraising developingdeveloping atat a timetime whewhenn thethe infinitiveinfinitive wawass completelycompletely separatedseparated fromfrom nominalnominal parapara- digmsdigms andand hahadd acquiredacquired moremore propertiepropertiess ofof a predicatepredicate.. IInn LatinLatin wewe seesee NP-accusativeNP-accusative'" participlparticiplee sequencessequences reanalyzereanalyzedd asäs NPNP- infinitiveinfinitive afteafterr object-equiobject-equi verbsverbs.. IInn OldOld EnglishEnglish,, raisinraisingg ttoo objectobject evolvedevolved frofromm NP-NNP-NPP sequencessequences beinbeingg reanalyzereanalyzedd äass NP-infinitiveNP-infinitive- NPNP.. EveEvenn thougthoughh ththee OldOld EnglisEnglishh andand ththee LatiLatinn studiesstudies araree lanlan- guage-specificguage-specific descriptionsdescriptions,, ththee mechanismmechanismss involvedinvolved areare generagenerall tendencietendenciess thathatt cancan probablyprobably bbee observedobserved forfor otherother languagelanguagess developingdeveloping suchsuch structures.structures.

5,5. CONCLUSIONSCONCLUSIONS

InIn comparingcomparing ththee syntaxsyntax ofof ththee ververbb complementscomplements andand purpospurposee clausesclauses inin KVRV,, AAvv andand OIr,Olr, itit isis clearclear thatthat ththee OIrOlr verbaverball noun'noun'ss completelycompletely nominanominall statusstatus isis morphologicallymorphologically closercloser ttoo PIPIEE thathann

Brought to you by | University of South Carolina Libraries (University of South Carolina School of L) Authenticated | 172.16.1.226 Download Date | 5/31/12 9:18 PM 3232 thetho otherother languageslanguages whichwhich havehave somesome infinitivesinfinitives notnot partpart ofof nounounn paradigms.paradigme. OIrOlr syntax,syntax, however,however, containscontains botbothh highlyhighly archaicarchaic featuresfeatures andand somesome innovations.innovations. ArchaismsArchaisms areare mostmost strikingstriking withwith subjeotsubject ofof acousativeaccusative complements:complements: equi-deletionequi-deletion isis optionaloptional andand overtovert subjectsubject isis eitheeitherr aann agentagent phrasphrasee (with(with do) oror a genitivegenitive whosewhose exactexact relationshiprelationskip toto thethe verbalverbal nounnoun isis unspecifiedunspecified.. AdmitAdmit- tanoetanee ofof dativedative verbalverbal nounnoun complementscomplements representsrepresents a developmentdevelopment withinwithin thethe laterlater OIrOlr period:period: itit startsstarte toto appearappear inin raising-to-objectraising-to-object clausesclauses whilewhile subject-equisubject-equi complementscomplements havehave aocusativeaccusative object.object. RVBV hashas developeddeveloped further,further, botbothh morphologicallymorphologically (having(having nonnomnonnom- inalinal infinitives)infinitives) andand syntacticallysyntactically (object(object beingbeing mostmost oftenoffcen dativedative oror accusative;accusative; lessless restrictionsrestrictions onon coreferentcoreferent andand independentindependent subsub- ject)ject).. AvAv hashas gonegone thethe furthestfurthest byby havinghaving nono restrictionsrestrictions onon assignassign- mentment ofof infinitiveinfinitive formform andand bbyy standardizationstandardization ofof aocusativeaccusative object.object. AllAll threethree languageslanguages areare subjectsubject toto indeterminacy,indeterminacy, butbut onlyonly inin thethe olderolder clauseclause types.types. OnOn thethe basibasiss ofof thisthis evidence,evidence, I concluaeconcluäe thatthat thesethese featuresfeatures ddoo indicateindicate thathatt PIEPIE useusedd subordinationSubordination withwith nominalizations.nominalizations. ThThee infinitiveinfinitive was,was, however,however, notnot morphologicallymorphologically separateseparate asäs itit isis inin mostmost ofof thethe historicallyhistorically attestedattested daughterdaughter languages.languages. SyntacticSyntactic featuresfeatures alonealone decideddecided whenwhen thethe actionaction nounnoun shouldshould bebe interpretedinterpreted asäs a nounnoun oror asäs anan infinitive.infinitive. A comparisoncpmparison ofof IlrIlr andand OIrOlr yieldsyields featuresfeatures soso similarsimilar thatthat theythey mustmust havehave developeddeveloped duringduring thethe periodperiod ofof unity.unity. Equi-deletionEqui-deletion waswas a PIEPIE developmentdevelopment withwith verbverbss 13 whosewhose subjectssubjects controlcontrol coreferencecoreference inin purpospurposee clauses., lS FromFrom OIrOlr 14 (32)(32) andand scatteredscattered RVKV evidence14 equi-deletionequi-deletion maymay havehave beebeenn optiOIl,aloptional inin PIE.PIE. Latin'sLatin's admittanceadmittance ofof coreferentcoreferent infinitiveinfinitive subsub- jecjectt afterafter wish maymay alsoalso bebe a holdoverholdover fromfrom optionaloptional applicationapplication inin PIE.PIE. DifferencesDifferences inin object-equiobject-equi constructionsconstructions andand thethe absenceabsence ofof raisingraising inin RVRV mitigatesmitigates againstagainst theirtheir reconstructionreconstruction forfor PIE.PIE. GenitiveGenitive wawass probablprobablyy thethe majoritymajority objectobject casecase whilewhile dativedative -— toto judgjudgee fromfrom IlrIlr andand HittiteHittite -— waswas probablprobablyy a transitiontransition casecase iinn subordinatesubordinate clauses.clauses. AccusativeAccusative waswas aann independentindependent developmentdevelopment

13 13 MyMy evidenceevidence fromfrom TIr,Ilr, Celtic,Celtic, andand HittiteHittite (Disterheft(Disterheft 182-4)182—4) contra;contra· dictsdicts M;iller'sMüler's 1974:1974: 230230 argumentargument fromfrom LatinLatin andand GreekGreek (and(and thethe inferenceinference ~hatthat thisthis isis truetrue forfor allall IE)IE) thatthat equiequi NPNP deletiondeletion developeddeveloped independentlyindependently min thesethese languages.languages. *~~ 14 14 RVadmitsRV admits a.a coreferentcoreferent subjectsubject pronoupronounn inin thethe folIofollowi wigg purposepurpose clauses:clauses: tari~anitcvrl$ani VI.VI. 37.7,37.7, ~lay6itfaye I.I. 113.5,6,113.5, 6, utaY6ütaye VIIIVIII.. 21.9,21.9, pltayepitay6 I.I. 16.3,16.3, pra-yai X.X. 104.3,104.3, tariiyataräya VIII.VIII. 96.1,96.1, mMiiyamadäya IX.IX. 109.20.109.20.

Brought to you by | University of South Carolina Libraries (University of South Carolina School of L) Authenticated | 172.16.1.226 Download Date | 5/31/12 9:18 PM 3333 inin eacheach subgroupsubgroup (as(äs OlrOlr shows)shows) asäs a resulresultt ofof thethe developmendevelopmentt ooff mormoree "verbal""verbal" infinitiveinfinitive syntax.syntax. AlAlll IEIE languagelanguagess surveyedsuryeyed herheree uniformluniformlyy displaydisplay a trendtrend towartowardd anan expansionexpansion iinn typestypes ofof clausesclauses thathatt admitadmit infinitivinfinitivee andand iinn thethe numbenumberr ofof verbverbss thathatt useuse .infinitiveinfinitiv e complementscomplements withiwithinn eacheach ofof ththee semanticsemantic classes.classes. AfteAfterr incipienincipientt purpospurposee clauseclause andand subject-equisubject-equi comple~entscomplements ooff latlatee PIPIEE havehave becombecomee establishedestablished asäs productivproductivee complements,complements, object-equiobject-equi clausesclauses startstart toto develodevelopp asäs earlyearly asäs HittitHittitee andand areare alreadyalready proliferatinproliferatingg inAvin Av andand RV.RV. OlrOlr doesdoes nonott ususee object-equiobject-equi peperr se,se, bubutt a semanticallysemantically paralleparallell typetype whicwhichh deletesdeletes a verbalverbal noun'snoun's subjectsubject whewhenn coreferentcoreferent witwithh thethe objectobject ofof a prepositionaprepositionall phrasphrasee (literally:(literally: HHee ordersOrders ttoo ththee peoplpeoplee (=(= preprepp phretsephrase)) ttoo dodo thisthis;; forfor examplesexamples seesee DisterhefDisterheftt 145145 f.).f.). ByBy thethe timtimee ofof GreekGreek andand LatiLatinn thisthis typetype isis stronglystrongly attested.attested. Raising-to-objecRaising-to-object\·e.-t £- nonott foundfound inRVin RV oror HittitHittitee -— iiss startingstarting inin AvAv andand isis quitquitee developeddeveloped inin Olr.Olr. Raising-to-subjectRaising-to-subject iiss a distinctlydistinctly latelate typetype.. ItItss presencepresence hahaSs beebeenn notenotedd forfor threthreee verbsverbs inin HittitHittitee (52),(52), (53),(53), butbut itit doesdoes nonott appearappear iinn other.other languageslanguages discusseddiscussed herehere untiluntil ILatin .. atin,, Greek,Greek, andand OlrOlr (Disterheft(Disterheft 152-5).152—5). WeWe havhavee seeseenn infinitiveinfinitive complementationcomplementation startstart iinn ththee IEIE perioperiodd witwithh subject-equsubject-equii andand purpospurposee clauses,clauses, thethenn spreadspread independentlindependentlyy iinn ththee subgroupsubgroupss fromfrom object-equiobject-equi ttoo raising-to-objecraising-to-objectt thethenn raising-to-subjectraising-to-subject,, alalll asäs ththee resulresultt ofof reanalysireanalysiss ofof surfacesurface strings·strings withiwithinn thethe maimamn 15 clause.I5 InIn thisthis discussiondiscussion I havhavee dealtdealt onlyonly witwithh a portioportion·n ofof thethe IEIE 1l1nguagelanguage area;area; mosmostt ofof thethe languagelanguagess omittedomitted areare attestedattested latelaterr andand areare arguablyarguably nonott asäs valuablevaluable for reconstruction.reconstruction. FurtherFurther investigationinvestigationss intintoo thethe developmentdevelopment ofof structuresstructures inin otherother languaglanguagee groupsgroups whicwhichh earlierearlier werewere infinitivelessinfinitiveless willwill ascertainascertain hohoww mucmuchh ofof ththee successivsuccessivee spreadspread ofof clauseclause typestypes cancan bbee attributedattributed toto crosscross- linguistilinguisticc tendencies.tendencies.

DOROTHDOROTHYY DISTERHEFTDISTERHEFT DepartmentDepartment ofof EnglishEnglish UniversityUniversity ofof SouthSouth CarolinCarolinaa COLUMBIA,COLUMBIA, SCSO 2920292088 USUSAA

15 IS TheThe chronologychronology poff otheotherr constructionsconstructions notnot dealdealtt withwith herheree (e.g.(e.g. imperaimpera- tivtivee infinitives,infinitives, predicatepredicate infInitives,infinitives, nounounn anandd adjectivadjectivee complementscomplements)) aroaro discussediscussedd inin DisterhefDisterheftt 181-92.181—92.

Brought to you by | University of South Carolina Libraries (University of South Carolina School of L) Authenticated | 172.16.1.226 Download Date | 5/31/12 9:18 PM 3434

REFERENCESREFERENCES

ARD,ABD, WILLIAMWILLIAM .Tosu:JOSH (1976).(1975). Raising andand Word Order inin DiachronicDiachronie Syntax.Syntax. UOLAUCLA dissertationdissertation reproducedreprodueed andand distributeddietributed byby IndianaIndiana LinguisticsLinguistics Club,Club, 1977.1977. . BARTHOLOMAE,BABTHOLOMAE, C:a:RISTJANCHBISTIAN (1890).(1890). "Das"Das griechischegriechische InfinitivsuffixInfinitivsuffix -afia.£",-cr#at", Rheini8che8Rheinisches MU8eurnMuseum 45:45: 151-3.151—3. BENVENISTE,BENVENISTE, EMILEEMILE (1935a).(1935a). LuLee in/initi/sinjinitifs avutiquu.avestiques. Paris:Paris: AdrianAdrien MaisonMaison- neuve.neuve. -— (1935b)(193öb) OriginesOriginee dede la formation desdes nOmIJnoms enen indo-europeen. Paris:Paris: AdrienAdrien Maisonneuve.Maisonneuve. CARDONA,CABDONA, GEORGEGEOBGE (1976).(1976). "Some"Some FeaturesFeatures ofof Pii.r.llnianPäninian Derivations",Derivations", History o/linguisticof linguistic thought andand contemporary linguistics (ed.(ed. H.H. Parret),Parret), Berlin:Berlin: dede Gruyter,Gruyter, 137-58.137—58. DILLON,DrtxoN, MYLESMYLES (1955).(1955). "On"On thethe SyntaxSyntax ofof thethe IrishIrish Verb",Verb", TPhS 1955:1955: 104-16.104-16. DISTER:a:EFT,DISTEBHEFT, DOROTHYDOBOTHY (1980).(1980). TheThe Syntactic Development ofof thethe InfinitiveInfinitive inin Indo-European. Columbus,Columbus, Ohio:Ohio: Slavica.Slavica. GAGNEPAlN,GAGNEPAIN, JEANJEAN (1963).(1963). LaLa syntaxe dudu nom verbal danalulangUe8dans les langues celtiqUeB.ceUiques. 1.I. Irlandais.Irlandais. Paris:Paris: Klinoksieok.Klineksieek. GRASSMANN,GBASSMANN, HERMANNHEBMANN (1873).(1873). WOrterbuchWörterbuch zumzum Rig-Veda. [Reprinted[Reprinted byby WissensohaftlicheWissenschaftliche Buohgesellschaft,Buchgesellschaft, Darmstadt,Darmstadt, 1964. J] HAHN,HAHN, E.E. ADELAIDEADELAIDE (1950).(1950). "The"The InfinitiveInfinitive withwith SubjeotSubject Acousative",Accusative", TAPA 81:81: 117-29.117-29. KAMMENHUBER,KAMMENBTTBEB, ANNELIESANNELIES (1954).(1954). StudienStudien zumzum hethitisohenhethitischen Infinitivsystem.Infinitivsystem. 1.I. DieDie MorphologieMorphologie derder InfinitiveInfinitive aufauf -anna (-atar) undund -(u)wanzi (-uwar)(-uwar),, MIOMIO 2:2: 44-77.44-77. KLEIN,KLEIN, JAREDJABED (1978).(1978). TheThe PartideParticle u inin thethe Rigveda: A Synchronic anandd DiachronieDiachronic Study (Ergänzungshefb(Erganzungsheftee zurzur ZeitschriftZeitsohrift·- fiirfür vergleiohendevergleichende SpraohforschungSprachforschung 27).27). GöttingenGottingen:: VandenhoecVandenhoeokk & Rupreoht.Ruprecht. LEUMANN,LEUMANN, MANUMANu,, J.J. B.B. HOFMANN,HOFMANN, ANTOANTONN SZANTYRSZANT^B (1972)(1972).. Lateinische Grammatik.Grammatik, MunichMunioh:: BeckBeok.. MUXEBMILLER,, DD.. GARYGABY (1974)(1974).. "On"On thethe HistoryHistory ooff IrifinitiveInfinitive ComplementatioComplementationn inin LatinLatin anandd Greek"Greek",, JIES 2:2: 223—46223-46.. RENOU,RENOÜ, LOUISLouis (1954)(1954).. LLee passagepas8age dudu nonomm d'aceiond'action a Vinfinitifl'infinitij danaledans le Rigveda, Sprachgeschichte unundd Wortbedeutung (Fs(Fs.. AlberAlbertt Debrunner).Debrunner).·* BernBern:: FranckeFrancke,, 383—7383~7.. . . SALTABEIXISALTARELLI,, MABIMARIoO (1979)(1979).. "Syntacti'~Syntaotioc Diffusion"Diffusion".. PapePaperr reareadd atat ththee FourtFourthh InternationaInternationall ConferencConferenoee oonn HistoricaHistorioall LinguisticsLinguistics,, Stanford.Stanford. 2 SCRWYZER,SCHWYZEB, EDUABEDUARDD (1953)(1953).. Griechische Grammatik.2 MunichMunich:: BeckBeck.. STOKESSTOKES,, WHTELEWHITLEYY anandd JOHJOHNN STBACHASTRACHANN (1901)(1901).. TheBaurU8Thesaurus PalaeohibernicuePalaeohibernicu8 I.I. [Reprinte[Reprintedd bbyy ththee DubliDublinn InstitutInstitutee foforr AdvanceAdvanoedd StudiesStudies,, 1975.1975.]] THUBNEYSENTRURNEYSEN,, RUDOLPRuooLl'lIH (1908)(1908).. "Altindisc"Altindisoh.etavai",h etavai", Melanges ddee linguistique offertsojfert8 a MM.. Ferdinand ddee Saussure.SaU88ure. ParisParis:: ChampionChampion,, 225—7225-7.. —- (1946(946)) A grammar ooff Old Irish. DublinDublin:: InstitutInstitutee foforr AdvanceAdvanoedd StudiesStudies..

Brought to you by | University of South Carolina Libraries (University of South Carolina School of L) Authenticated | 172.16.1.226 Download Date | 5/31/12 9:18 PM