RootsIIIWorkshop,Jerusalem,June2011 1

Thecausative‐passivealternationinFrench

MarieLabelle,Départementdelinguistique,UniversitéduQuébecàMontréallabelle.marie@uqam.ca

1. ProblemCausative‐passivealternation:(1) a. Thecausativeconstruction: Evefait[offrirunposteàPaul]. (faire=‘make’) EveCAUSE[offerapositiontoPaul] ʻEvecauses(pro)toofferapositiontoPaul’ b. The‘passivesefaire’construction(PSF): PauliSEfait[offrirunposteei]. PaulSECAUSE(?)[offeraposition] ʻPaulisofferedaposition’ (mayalsohaveareflexivecausativereading:Paulcauses…tohimself.)Causative‐anticausativealternation:(2) a. Thelexicalcausative: Pierrecassalabranche ʻPierrebrokethebranch’ a’. [[PeterACT]CAUSE[vaseBECOME<broken>]] b. Theresultanticausativeconstruction: LabrancheiSEcassaei ʻThebranchbroke’ b’. [vaseBECOME<broken>] Inbothalternations,acomplementin(a)surfacesassubjectin(b); Inbothcases,SEiscruciallyinvolvedinthealternation; InbothcasesaCAUSEpredicateisinvolved:presentin(a),absentin(b).ThissuggeststhatSEin(1b)istheanticausative(ordecausative)morphemefoundin(2b)(e.g.Koenig&Pedersen,1992,Labelle,2002).

RootsIIIWorkshop,Jerusalem,June2011 2

However,suchananalysisraisessomeintriguingquestionsconcerning(1b):a. Starting with the embedded event (offrir un poste à Paul ‘offer a position to

Paul’), why would one want to add to it a causative verb (faire), only todecausativizethepredicateimmediatelyafter,usingSE?

b. Whywould thatoperationof causativization‐decausativizationyieldapassivereading?

c. Whatisthesemanticimportofthehigherclauseinthestructure?QuestionsraisedbyMarantz(2007):“Howmuchofthesemanticeventstructurefindsrepresentationinsyntacticheadsandrelations?i.isthereaCAUSEheadincausativevPs?"

2. SEinResultAnticausativeclauses(3) Marantz(2007):

Labelle&Doron(2010),Doron&Labelle(2011)’sanalysisofresultanticausatives.Accomplishementsinvolvealayered[vP[VP]]structurewhere:

o Vheadsachangeofstatesubeventandlicensesaresultphrase,o vheadsanactivitysubevent.

No meaning such as ‘causative’ attached to the higher v : ‘cause’ is the normalinterpretationoftheconfiguration(Hale&Keyer2002:10&174ss.;Marantz2005).[+Active] Voice: introduces the external argument and assigns it a thematic rolecompositionally determined by the complex verb phrase (Marantz 1984, Kratzer1996):

RootsIIIWorkshop,Jerusalem,June2011 3

o Ifarootmergeswiththeheadofanactivitysubevent,theexternalargumentisanAgent;

o If a root merges with the head of a change of state subevent, the externalargumentisaCause.

[–Active]Voice:doesnotintroduceanexternalargumentinsyntax.In Result Anticausatives, SE heads [–Active] Voice and combines directly with achange‐of‐statesubevent(VP).SE selects for an open predicate(λPλxλe[P(e,x)])& creates a property‐denotingpredicatewhosesubjectcorrespondstotheVPinternalmissingargument.(4)a.Transitiveofaccomplishmentverb

Pierrecassalabranche.‘Pierrebrokethebranch.’

b.Anticausativeofsameverb

Labranchesecassa.‘Thebranchbroke.’

3. SEinthecausative‐passivealternation3.1. CausativefaireDoesnotpassivize:(5) a. Marcafait[chasserPierredel’organisation].(Active) MarcAUXCAUSdismissPierrefromtheorganization ʻMarchadPierredismissedfromtheorganizationʼ b. *Pierreaétéfait[chasseredel’organisation].(Passive) ↑__________X____________|>> suggests that there is no restructuration with formation of a complex verb[faire+infinitive]:

RootsIIIWorkshop,Jerusalem,June2011 4

(6) Marcafait­chasserPierreACCdel’organisation.o (6)shouldallowpassivizationofaccusativecomplement.

Proposal:(1) a.Evefait[offrirunposteàPaul]. EveCAUSE[offerapositiontoPaul] 'Evecauses(pro)toofferapositiontoPaul'(7)

o Crucially: √faire is not a causative head, but a root that lexicalizes V,licensingthemergeofaresultclause(=thecausedevent).

o Thecaused(embedded)event isavP (notaVoicephrase=noposition forthesubject=realizedasaby‐phraseinPSF).

o Active Voice introduces the external argument = Cause (because the rootmergeswithV)

Thecompletestructurestatesthattheexternalargumentcausedthecoming‐aboutofthecausedevent,viewedasasituationresultingfromsomeunspecifiedactionoftheexternalargument.

3.2. Passivesefaire(8) b. PaulSEfait[offrirunposte]. PaulSECAUSE(?)[offeraposition] ʻPaulisofferedapositionʼ

RootsIIIWorkshop,Jerusalem,June2011 5

(10)

o SErealizes[–Active]Voiceo SEselectsforanopenpredicate(Labelle2008forreflexiveclauses).o TheDPmergedaboveVoicecombineswith theopenpredicate toyieldasentence.

Theconstructiondenotesasituationwhereanentitycomestobeinaresultstateexpressedbythelowerclause,hencethepassivereading.

3.3. FaireasmodifyingV

3.3.1. Inchoativereadingofsefaire:(9) Lefromagesefait. thecheesecomes‐to‐be.(10)Lucsefaitvieux.(McIntyre,2006) LucSEFAIREold ‘Lucisbecomingold’(11)Lemondenes’estpasfaitenunjour. theworldNEG‐PRTSEAUXNEGmadeinoneday ‘Theworlddidnotcometobe(/wasnotmade)inoneday.’= √FAIRE merged with V adds no semantic modification to it: by definition V,expressesachangeofstatesubevent,isinterpretedasmeaning«come‐to‐be»=>faire=atypeofexpletivehead:lexicalizestheheadwithoutmodifyingit.=>When √FAIRE merges with V, it lexicalizes the result subevent and licenses aresultativecomplement.

RootsIIIWorkshop,Jerusalem,June2011 6

(12)

3.3.2. Re‐prefixationMarantz (2007): re‐prefixation creates the restitutive of the verb (not therepetitive).(13)a.Thewallsinthehouseweregreenlongbeforetheypurchasedit;theyplan

tore‐paintthem(white)assoonastheycan. b.*Johnre‐smiled./(vsJohnsmiledagain.)InEnglish,re‐takesscopeoverthechangeofstateevent:(14)*Theyre‐putthebookonthetable.(15)*Theyre‐gaveJohntheaward.(fromMarantz‐‐Restitutivere‐andthefirstphasesyntax/semanticsoftheVP)(16)and(17)areOKinFrench:(16)Ilsontremislelivresurlatable(aprèsqu’ilsoittombé). theyput[AGAIN[thebookonthetable]](17)IlsontredonnéleprixàJohn. theygavetheawardtoJohnback/asecondtime/againre‐ may scope higher than the change of state event. But, crucially, it does notrequire the subject to have done the first action (i.e. not repetitive). Holsinger(2008)introducesathirdinterpretation,reiterativescope,thatmaybeadequate:

RootsIIIWorkshop,Jerusalem,June2011 7

(18)

Similarly,incausatives:(19)compatiblewithasituationwhereEvedidnotcausethefirstpaintingact(e.g.shejustboughtthehouse):(19)a.Evearefaitpeindrelamaisonenrouge. b.EveACTCAUSEDAGAIN[cometobepaintedthehouseinred]=Thehousewasalreadyred.(20)a.Eveafaitrepeindrelamaisonenrouge. b.EveACTCAUSED[cometobe[[AGAINpaintedthehouse]inred]]]=Thehousecouldhavebeenadifferentcolour.(21)a.EvearefaitoffrirunposteàPaul b.EveACTCAUSEDAGAIN[cometobeofferedapositiontoPaul]=Eveisnotnecessarilyresponsibleforthefirstoffer.RE‐scopesoverthechangeofstatepredicate.Itdoesnotscopeoveracausepredicate.Theroot√FAIREheadsthechangeofstatepredicate.(notacausepredicate)

3.4. Koenig&Pedersen'sargumentsagainstananticausativeanalysis

3.4.1. Cross‐linguisticdifferencesTheequivalentof(10)ispossibleinSpanish,butthepassivereadingofsehacerisnot(samewithCatalan,Italian).Why?

RootsIIIWorkshop,Jerusalem,June2011 8

•Differencesbetweenpossibleandrealizedconstructions•Microparametricsyntax:

E.g.inItalian,thecausativemaypassivize:(22) Lamacchinasaràfattaripararedomani. Thecarwill‐bemaderepairtomorrow ‘Thecarwillbehadrepairedtomorrow.’ (Radford1978:37,fromLoCascio1970:187) ImpersonalSEconstructionsinItalian,Spanish…

3.4.2. AffectednatureofthesubjectSemantic/pragmaticspecializationofPSF.K&Pclaimthat

thesubjectmustbeanimate(23) theeventmustaffectthesubjectreferent(27),and For speakers who accept an inanimate subjects, “the inanimate subject is

affectedinsuchawaythatinitschangedstateitaffectsarelevantananimateentity.”(p.153)

PSFwithinanimatesubject:(23)Monsandwichs’estfaitmangerparcesalauddeJean. Mysandwichrefl‐be‐prmake‐prtbythisbastardofJean ‘MysandwichwaseatenbyJohn,thebastard.’ ==starredbyK&P(ex.32,p.151),butOKforPaul&me.Counter‐examples:(24)monmessages'estfaitmanger mymessagewaseatenup (http://blog.aufeminin.com/blog/seeone_20478_2806008)(25)[maplante]s'estfaitmangerlesfeuillessolide! myplantgotherleaveseatensolid!(=alot) (http://jardinfamilial.bbactif.com/t67134‐une‐horticultrice‐dans‐la‐salle)(26)Vous pourriez perdre votre connexion […]à cause d'un câble qui s'est fait

mangerparvotrechien. Youcouldloseyourconnexionbecauseofacablethathasbeeneatenbyyour

dog (http://onfaitduweb.com/programmation/nouveau‐drm‐pour‐ubisoft/)

RootsIIIWorkshop,Jerusalem,June2011 9

PSFwithexperiencerverb:(27)Marcs’estfaitcraindrepar/detoutlemonde. Marcrefl‐be‐prmake‐prtbyalltheworld ‘Marcwasfearedbyeverybody’ ==starredbyK&P(ex.42p.153) (possiblewithareflexiveinterpretation;moredifficult,butnotimpossibleasa PSF)Counter‐examples:(28)Instaureruneautoritéquisefaitcraindre. Establishanauthoritythatisfeared.(29)Unbeauvisageestuntraîtrequisefaitcraindreetqu'onregardeavecplaisir.

(Plutarque.) AbeautifulfaceisatraitorthatisfearedandthatonelooksatwithpleasurePossiblytruethatifthesubjectisinanimatethesentenceisunderstoodasaffectingan animate entity.Why? The apparent semantic/pragmatic specialization of PSFmaystemfrom:1)Twodistinctrealizationsofnon‐activeVoice:passivevsSE. Passive:Paulaétéfrappé. PSF:Pauls’estfaitfrapper.2)Inthepassive,onlyoneeventisinvolved. InPSF,thehighereventintroducesa“cometobe”/affectedeventontopofthe

embeddedevent.Couldthattriggeramore“affective”interpretation?Inanycase,thisshouldnotpreventusfromassigningasyntacticstructuretotheconstruction.

3.5. Extensions:

3.5.1. Reflexivecausatives(fromLabelle2008)Intruereflexives,SEismergedunder[+ActiveVoice].1‐ combineswithanopenVP;2‐ introducesanexternalargument;3‐ identifiestheexternalargumentwiththefreevariablewithintheVP.(30)Lucseparle.

Lucspeaks‐tohimself.

RootsIIIWorkshop,Jerusalem,June2011 10

(31) λe[speak‐to(e,Luc)∧Agent(e,Luc)] Luc Voice'λxλe[speak‐to(e,x)∧Agent(e,x)]

Voice VP λyλe[speak‐to(e,y)] se | λPλxλe[P(e,x)∧Agent(e,x)] V parle λyλe[speak‐to(e,y)]Similarlyinreflexivecausativeconstructions:(32)Lescitoyensse sont tous très souvent laissé bêtement

The citizens SE AUX-3P all very often let-PP stupidly [VP berner <θ> par le maire]. deceive __ by the mayor ‘The citizens very often let themselves all stupidly be deceived by the mayor.’

o Noneof the verbs is semantically reflexive: there is no coreferencebetween

theirthematicsubjectandtheirthematicobject.o Therelevantverbsaremorphologicallyandsyntacticallyindependent—they

areseparatedbyadverbs:nocomplexpredicateformation.The translation of the causative verb allows it to combine with VP’s, i.e. withrelationsbetweenindividualsandevents(33fromLabelle2008):(33)

o Bothintruereflexivesandinanticausatives,SEselectsforanopenpredicate.

Intruereflexives,SEismergedunderActiveVoice Inanticausatives,SEismergedunderNon‐ActiveVoice.

RootsIIIWorkshop,Jerusalem,June2011 11

o Inbothcases,thesubjectisassociatedwithanemptypositionwithinthelowerevent.

o In both cases, the positionwithin the embedded eventmay be an unselecteddativecomplement:

(34)Marcs’estfaitcouperlescheveux. Marcrefl‐be‐prmake‐prtcutthehair (a)Marcgothimselfahaircut (b)Marcgothishaircut(bysomeone)(possiblyunwillingly)Strongpointsofthepresentanalysis:

- Inreflexiveandnon‐reflexivesentences,SEcombineswithanopenpredicate;theprojectionof the external argument thematic role isdeterminedby thenatureoftheVoiceheadunderwhichSEisrealized.

- Noobject‐to‐subjectmovemento DativeargumentsdonotraisetosubjectinFrench,buttheymayformthesubjectinSE‐FAIREconstructions.

- Nolexicaloperationonacomplexpredicate:o EvenunselecteddativesmaysurfaceasthesubjectofSE‐FAIRE.

- Noneedtopostulatetheconstructionofacomplexfaire+infinitiveverb.o The two verbs are distinct, the embedded verb may has its ownexternalargumentwhichmayberealizedasaby‐phrase.

o Thetwoverbsmaybeseparatedbyadverbs:

(35)Pauls’estfait[brutalementagresserdanslemétro]. PaulSEAUXFAIRE[brutallyagressinthemetro] ‘Paulwasbrutallyagressedinthemetro’

3.5.2. Nullpronoun/operatormovement(NOP)Labelle (2008): “A purely syntactic alternative to the derivation of (47) [=(32)]would be to follow Bruening (2006) in assuming that the verb’s object is a nullpronounthatmovestoadjoinabovetheVPdominatedbylaissé.Movementofthenullpronounleavesatrace,andabstractsovertheadjoined‐tostructure,creatingaconstituentoftype<e,st>.Thiswouldbeasin(50)[=36)]:”

RootsIIIWorkshop,Jerusalem,June2011 12

(36)

Supposethat,alternatively,thenullproadjoinstotheembeddedvP:(37)Paul[VoiceSE[VP[√faire‐V][vPpro[vPoffrirunpostepro]]]] ↑______________________|This analysismakes the construction similar to the Chinese passive constructioninvolvingnulloperatormovementdiscussedbyHuang(1999):(40)Zhangsanibei[IPNOPLisidaleNOP] |___predication_____||____movement____|- NOPmovementtotheleftperipheryofembeddedIP- NOP movement has the effect of creating a property‐denoting predicate =lambdaabstraction

- mainpredicatemeaning‘get,acquire,orendupwiththepropertyof....’- Predicationofthepropertyonthemainpredicate’ssubject- InPSF:(37)' Paul[VoiceSE[VP[√faire‐V][vPpro[vPoffrirunpostepro]]]] |_________predication______________||_______movement________|‐ NOPmovementpossiblewithaccusative&dativeobjects.

RootsIIIWorkshop,Jerusalem,June2011 13

ApotentialextensionoftheanalysisisindeedEastAsiancausativeswhichhaveapassivereading,likeKorean,MongolianandTaiwanese.

4. ConclusionQuestionsinsection1:

a. Startingwiththeembeddedevent(offrirunposteàPaul‘offerapositiontoPaul’), why would one want to add to it a causative verb (faire), only todecausativizethepredicateimmediatelyafter,usingSE?

b. Why would that operation of causativization‐decausativization yield apassivereading?

c. Whatisthesemanticimportofthehigherclauseinthestructure?Answers:

PSF permits the passivization of unpassivizable predicates (namely:predicateswithadativecomplement).

Thepassivereadingstemsfromtheuseofnon‐activeSEwitharootheadingachangeofstatesubevent.

IsthereaCAUSEheadincausativevPs?Perhapsnotevenincausativessentences…universally?

RootsIIIWorkshop,Jerusalem,June2011 14

References.Bruening, Benjamin. 2006. The Morphosyntax and Semantics of Verbal Reciprocals. Ms. University of Delaware.

[availableat:http://udel.edu/~bruening/downloads.html]Doron, E. &M. Labelle. 2011. “An ergative analysis of the French valency alternation.” To appear in a volume of

proceedingsofLSRL2010.Hale,Ken&SamuelJayKeyser.2002.Prolegomenontoatheoryofargumentstructure.Cambridge,MA:MITPress.Holsinger, Edward (2008) Re‐, Again A semantic investigation of restitutive again and re‐ in English. LSA 2008

SummerMeeting,OhioStateUniversity,March25th,2008. Availableonline:http://www.ripelacunae.net/wp‐content/uploads/2010/08/LSA_Again.pdfHuang, C.‐T. J. 1999. “Chinese passives in comparative perspective.” Tsing Hua Journal of Chinese Studies.

(downloadablefrom:http://aris.ss.uci.edu/ling/personnel/jhuang.html)Koenig,Jean‐PierreandEricPederson(1992)‘‘SemanticUnityandConstructionalParticularity:TheFrenchSeFaire

Construction”,BerkeleyLinguisticsSociety18,pp.143–156.Kratzer, A. 1996. “Severing the external argument from its verb”. In JohanRooryck&Laurie Zaring (eds.)Phrase

structureandthelexicon,109–137.Dordrecht:Kluwer.Labelle,M.2002. “TheFrenchnon‐canonicalpassive in ‘se faire’.”Haraguchi, Shosuke,BohumilPalekandOsamu

Fujimura (eds.) Proceedings of Linguistics and Philosophy 2002. Tokyo : Charles University Press and MeikaiUniversity.

_____.2008.“TheFrenchreflexiveandreciprocalse.”NaturalLanguageandLinguisticTheory26(4):833‐876._____&EditDoron.2010.“Anticausativederivations(andothervalencyalternations)inFrench.”Probus22‐2,303‐

316.Marantz,A.1984.Onthenatureofgrammaticalrelations.MITPress.Marantz,A.2005.Objectsoutofthelexicon:objectsasevents.Marantz, A. 2007. Restitutive re‐ and the first phase syntax/semantics of the VP. Handout. Workshop "(Lexical)

Decomposition" Organized by Artemis Alexiadou & Florian Schäfer (Project B1/C1 SFB 732 “IncrementalSpecification in Context”), University of Stuttgart, March 2‐3, 2007. (available on‐line at:http://web.mit.edu/~marantz/Public/Handouts/MarylandRe.pdf)

Radford, Andrew. 1978. Agentive Causatives in Romance: Accessibility versus Passivation. Journal of Linguistics(14:1):35‐58.StableURL:http://www.jstor.org/stable/4175423.