Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
The morphosyntax of states: deriving aspect and event roles from argument structure
(USC Thesis Other)
The morphosyntax of states: deriving aspect and event roles from argument structure
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
DepartmentofLinguistics DornsifeSchoolofLetters,ArtsandSciences UniversityofSouthernCalifornia TheMorphosyntaxofStates DerivingAspectandEventRoles fromArgumentStructure AlfredoGarcía-Pardo AdissertationsubmittedtotheUSCGraduateSchool inpartialsatisfactionoftherequirementsforthedegreeof DoctorofPhilosophyinLinguistics August2018 ©CopyrightbyAlfredoGarcía-Pardo 2018 ThedissertationofAlfredoGarcía-Pardoisapproved: AudreyLi RoumyanaPancheva GabrielUzquiano MaríaLuisaZubizarreta(Chair) UniversityofSouthernCalifornia 2018 A quien me en señó a leer y e scribir, por darme el mayor de los regalos. TableofContents AbstractoftheDissertation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viii ListofTables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . x ListofAbbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xii 1 Introduction 1 1.1 Settingthestage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.2 Goalsandscopeofthisdissertation . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 1.3 Structureofthisdissertation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 2 Aktionsartandargumentstructure:Astateoftheart 8 2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 2.2 Theclassiceventtypes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 2.2.1 States . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 2.2.2 Activities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 2.2.3 Telics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 2.2.4 Thetests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 2.2.4.1 Eventsvs.States . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 2.2.4.2 Telicvs.Atelic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 2.2.5 Twotypesoftelicpredicates . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 2.2.5.1 Accomplishments . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 2.2.5.2 Achievements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 2.2.6 Againstsemelfactives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 2.2.7 Interimsummary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 2.3 ModelingAktionsartinthesyntax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 2.3.1 Originsofverbaldecomposition . . . . . . . . . . . . . . . . . . . . . . . . 29 2.3.2 L-syntax:Hale&Keyser(1993,2002) . . . . . . . . . . . . . . . . . . . . . . . 31 2.3.2.1 Anoverviewofthemodel . . . . . . . . . . . . . . . . . . . . . . 31 2.3.2.2 Criticalassessment . . . . . . . . . . . . . . . . . . . . . . . . . . 38 2.3.3 TheXS-model:Borer(2005b) . . . . . . . . . . . . . . . . . . . . . . . . . . 40 iv 2.3.3.1 Anoverviewofthemodel . . . . . . . . . . . . . . . . . . . . . . 40 2.3.3.2 Criticalassessment . . . . . . . . . . . . . . . . . . . . . . . . . . 47 2.3.4 Firstphasesyntax:Ramchand(2008) . . . . . . . . . . . . . . . . . . . . . . 49 2.3.4.1 Anoverviewofthemodel . . . . . . . . . . . . . . . . . . . . . . 49 2.3.4.2 Criticalassessment . . . . . . . . . . . . . . . . . . . . . . . . . . 56 2.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 3 Stativecausatives 61 3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 3.2 Non-dynamicevents:Fábregas&Marín(2017) . . . . . . . . . . . . . . . . . . . . . 62 3.2.1 AnoverviewofFábregas&Marín(2017) . . . . . . . . . . . . . . . . . . . . 62 3.2.2 ReviewofFábregas&Marín(2017) . . . . . . . . . . . . . . . . . . . . . . . 66 3.2.2.1 Eventivitytests? . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 3.3 Myproposal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 3.3.1 Definingtraitsofgobernar-verbs . . . . . . . . . . . . . . . . . . . . . . . . . 75 3.3.2 Thesyntaxofgobernar-verbs . . . . . . . . . . . . . . . . . . . . . . . . . . . 78 3.3.3 Derivingthepropertiesofgobernar-verbs . . . . . . . . . . . . . . . . . . . . 80 3.3.4 Comparisonwithatomicapproaches . . . . . . . . . . . . . . . . . . . . . . 83 3.4 Againstmaintain:causeandthetemporalrelationbetweenevents . . . . . . . . . . 85 3.4.1 causeandtemporalsequencing:Neeleman&vandeKoot(2012) . . . . . . . 85 3.4.2 ProblemswithNeeleman&vandeKoot(2012) . . . . . . . . . . . . . . . . . 87 3.4.3 Analternativeproposal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89 3.5 Compatibilitywithotheraccountsofcauseandtheexternalargument . . . . . . . . 91 3.5.1 VoiceP6=vP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 3.5.2 Unaccusativecause . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 3.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 4 Stativeresultatives 101 4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 4.2 AdjectivalpassivesinSpanish:astateoftheart . . . . . . . . . . . . . . . . . . . . . 105 4.3 Comparingpredicativeandattributiveparticiples . . . . . . . . . . . . . . . . . . . 110 4.3.1 Change-of-statestructure? . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 4.3.2 Externalarguments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114 4.3.3 Passives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 4.3.4 Perfectsemantics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119 v 4.3.5 Summaryoffindings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122 4.4 Myproposal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 4.4.1 Estar-PPrts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 4.4.2 Bare-PPrts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129 4.4.3 Perfectiveadjectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134 4.5 By-phrasesandagent/event-orientedmodifiersinAPass.BeyondSpanish . . . . . . 138 4.5.1 State-relevance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140 4.5.2 Problemswiththestate-relevanceapproach . . . . . . . . . . . . . . . . . . 141 4.5.3 Event-kindsandpseudo-incorporation . . . . . . . . . . . . . . . . . . . . . 144 4.5.4 Problemswiththeevent-kindsapproach . . . . . . . . . . . . . . . . . . . . 148 4.5.5 Wherearewe? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150 4.6 Crosslinguisticvariation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154 4.6.1 Permissivelanguages:anintroduction . . . . . . . . . . . . . . . . . . . . . 154 4.6.2 Previousaccounts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157 4.6.2.1 Kratzer(2000) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157 4.6.2.2 Alexiadouetal. (2015) . . . . . . . . . . . . . . . . . . . . . . . . . 159 4.6.2.3 ProblemswithKratzer(2000) . . . . . . . . . . . . . . . . . . . . 161 4.6.2.4 ProblemswithAlexiadouetal. (2015) . . . . . . . . . . . . . . . . 164 4.6.3 Analternativeaccount . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166 4.6.3.1 Option1.ParametrizationofAdj(tobediscarded) . . . . . . . . . 167 4.6.3.2 Option2.StatPandAdjP . . . . . . . . . . . . . . . . . . . . . . . 168 4.7 Towardsaunifiedtheoryofparticiples . . . . . . . . . . . . . . . . . . . . . . . . . . 171 4.8 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175 5 Stativepsychologicalandlocativeverbs 177 5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177 5.2 ‘Psych’verbs:stateoftheart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178 5.2.1 Experiencersandthelinkingproblem . . . . . . . . . . . . . . . . . . . . . 178 5.2.2 ‘Psych’effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181 5.3 Psychologicalandlocativeverbs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188 5.3.1 Spanishpsychverbsandtheireffects . . . . . . . . . . . . . . . . . . . . . . 188 5.3.1.1 StativeOEPVs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189 5.3.1.2 Psycheffects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190 5.3.2 Locativeverbsandtheir‘psych’effects . . . . . . . . . . . . . . . . . . . . . 192 5.3.2.1 Spanishlocativeverbs . . . . . . . . . . . . . . . . . . . . . . . . 192 vi 5.3.2.2 Psycheffectswithlocativeverbsinotherlanguages . . . . . . . . 195 5.3.2.3 Interimsummary . . . . . . . . . . . . . . . . . . . . . . . . . . . 200 5.4 Myproposal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201 5.4.1 APPforexperiencerobjects:Landau(2010) . . . . . . . . . . . . . . . . . . 201 5.4.2 Unifyingpsychandlocativeverbs . . . . . . . . . . . . . . . . . . . . . . . . 202 5.4.3 Stativepsychandlocativeverbsarenotcausative . . . . . . . . . . . . . . . 205 5.4.4 Stativepsychandlocativeverbsareunaccusative . . . . . . . . . . . . . . . 207 5.4.5 ThecommonthematicinterpretationofExperiencersandGrounds . . . . . 212 5.4.6 InterimConclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215 5.5 Agentivity? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216 5.5.1 Thereceivedviewonagentivity . . . . . . . . . . . . . . . . . . . . . . . . . 216 5.5.2 Agentivestativelocativeverbs . . . . . . . . . . . . . . . . . . . . . . . . . . 220 5.5.3 Agentivityandpsychverbs . . . . . . . . . . . . . . . . . . . . . . . . . . . 223 5.5.3.1 TheclassificationinMarín(2011) . . . . . . . . . . . . . . . . . . . 223 5.5.3.2 Agentivityand‘psych’effectsinOEPVs . . . . . . . . . . . . . . . 228 5.5.4 Agentivityasteleologicalcapability:Folli&Harley(2008) . . . . . . . . . . . 234 5.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238 6 Conclusions 241 6.1 Summaryoffindings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241 6.2 Maincontributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242 6.3 Futuredirections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243 6.3.1 Locativeinversion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244 6.3.2 Nominalizations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245 6.3.3 Stativityandgrammaticalcategory . . . . . . . . . . . . . . . . . . . . . . . 246 Bibliography 250 vii Th eMorphosyntaxofStates DerivingAspectandEventRoles fromArgumentStructure Adissertationby AlfredoGarcía-Pardo DoctorofPhilosophyinLinguistics UniversityofSouthernCalifornia,2018 ProfessorMaríaLuisaZubizarreta,Chair AbstractoftheDissertation W ithin the literature on event types, there is a well-established distinctionbetweendynamiceventualities,whichinvolvesome sort of change, and stative eventualities, which do not. Recent neoconstructionist theories seek to derive event types and the interpretation of event participants—thematic roles—from the syntax of the verbphrase—argumentstructure. Thesemodelsuniformlyassignaprivileged statustodynamiceventseitherinfocusorinthetheoreticaltoolstheyassume, relegatingstatestoasecondarystatuswithinthetheoryandasworthyobjects ofempiricalresearch. However,statesarenotasdifferentfromeventsasitmayseem:uponcloser inspection,aspectualandthematicnotionsgenerallyassumedtobeexclusiveto dynamicevents,likeagentivity,causativityandresultativity,canapplytostates viii aswell.Morestrikingly,arichvarietyofthematicrolescanbefoundwithinsta- tivepredicates(Experiencer/Stimulus,Figure/Ground,Initiator/Resultee...),which raisesthenon-trivialquestionofhowthisdiversitycanbederivedfromanim- poverishedstructure. Thisdissertationaddressesstativepredicateswithinageneraltheoryofevent typesfromaneoconstructionistprism.IanalyzeasetofSpanishverbs,thegob- ernar‘govern’-typeandarguethattheyarederivedbyabi-phrasalstructurethat isunambiguouslyinterpretedasastativecausativeeventuality,i.e.twostatesre- lated causally. In so doing, I enrich the typology of event types taxonomically and theoretically, the latter by integrating stative causatives within a compre- hensivesyntacticmodelofeventstructure. Ialsoexploreadjectivalpassivesasacasestudyofderivedstatives. Ishow howtheseconstructionsaretrulystative—andnotperfectiveorresultative,as is often argued. The underlying participle is fed by different kinds of stative structures—unaccusativeorcausative.Theformerislexicalizedbytypicaltelic verbs—i.e. verbs of change with an endpoint, e.g. break—whereas the latter is lexicalized by stative causative verbs—the gobernar-type. This explains many propertiesofadjectivalpassivescrosslinguistically,andreinforcestheideathat statesalsocomeindifferenttypescrosscategorially. Finally, I uncover a series of crosslinguistic grammatical parallelisms be- tween stative object-experiencer psychological verbs (e.g. worry, amaze...) and locative verbs (e.g. surround, cover...). I argue that they share a uniform struc- ture,articulatedbyasilentbirrelationalpreposition. Thisstructuredenotesan abstractrelationbetweentwoentitiesthatcanbeunderstoodaslocativeorpsy- chologicaldependingonthelexicaltypeofverbthatlexicalizesit. Ishowhow thisproposalcorrectlyderivesthethematicinterpretationofthesestativepred- icateswithouttheneedtoresorttothematicrolesasgrammaticalprimitives,in thespiritoftheneoconstructionistprogram. ix ListofTables 2.1 TheaspectualfeaturesofeventtypesinSmith(1991) . . . . . . . . . . . . . . . . . . 25 2.2 Classiceventtypesandtheirtests . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 3.1 TypologyofeventtypesinFábregas&Marín(2017) . . . . . . . . . . . . . . . . . . 66 3.2 Eventtypesandtheirtests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78 4.1 EnglishparticiplesinEmbick(2004) . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 4.2 Cut-shortandfull-fledgedparticiplesinSpanish . . . . . . . . . . . . . . . . . . . . 107 4.3 SyncreticparticiplesinSpanish . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108 4.4 Propertiesofestar-PPrts,bare-PPrtsandperfectE-PPrts . . . . . . . . . . . . . . . . 123 4.5 Greek-menosand-tosparticiples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156 4.6 Change-of-stateverbsinChichewawithstativeandpassive . . . . . . . . . . . . . . 169 4.7 Non-change-of-stateverbsinChichewawithstativeandpassive . . . . . . . . . . . 170 5.1 Psycheffectscrosslinguistically . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187 5.2 Psycheffectswithstativelocativeverbs . . . . . . . . . . . . . . . . . . . . . . . . . 200 5.3 Comparingstativecausativeswithstativepsychandlocativeverbs . . . . . . . . . . 210 5.4 Marín’s(2011)classificationofOEPVs . . . . . . . . . . . . . . . . . . . . . . . . . . . 229 5.5 Agentivity,aspectandpsycheffectswithOEPVs . . . . . . . . . . . . . . . . . . . . 234 6.1 Grimshaw’s(1990)typologyofnominalizations . . . . . . . . . . . . . . . . . . . . . 245 x ListofAbbreviations acc Accusativecase Adj-PPrt Adjectivalpastparticiple APass Adjectivalpassive Asp Q P AspectofquantityPhrase B-PPrt Boundpastparticiple cause Causalrelation caus Causativemorpheme CCF Crucialcontributingfactor CIP Clause-internalphase cl Clitic dat Dativecase DEPV Dative-experiencerpsychverb det Determiner dom Directobjectmarking EA Externalargument E-PPrt Eventivepastparticiple EP EventPhrase f Femininemarker gen Genitivecase IA Internalargument impf Imperfectiveaspect imp Imperativemood inf Infinitive irreg Irregular maintain Maintenancerelation m Masculinemarker neg Negation nom Nominativecase OEPV Object-experiencerpsychverb pass Passive pf Perfectiveaspect pl Plural poss Possesive PPrtP ParticiplePhrase PPrt Pastparticiple pres Presenttense pst Pasttense p Person refl Reflexive reg Regular R-PPrt Resultativepastparticiple RRC Reducedrelativeclause RUTAH RelativizedUniformityofThematic AssignmentHypothesis SEPV Subject-experiencerpsychverb sg Singular s-o-q subjectofquantity SP StatePhrase StC Stativecausative subj Subjunctivemood T/SM Target/Subjectmatter U-PPrt Unboundpastparticiple UTAH Uniformity of Thematic Assign- mentHypothesis VPass Verbalpassive V-PPrt Verbalpastparticiple xi Acknowledgments How do I even begin? As I am sitting at a table at the Nazarian Pavilion Courtyard giving the final touchestomydissertation,Ifeelakindofvertigolookingbacktoalltheyearsthatledtothismoment. ItseemslikeforeversinceIwasdoingmyundergradinMadridandtookmyfirstgenerativesyntax class,andyetIfeellikemylifeisbarelybeginning.IntheselinesIwouldliketoshowmyappreciation toallthewonderfulpeoplethatImetonthewayandhelpedthisdissertationbecomeareality. María Luisa Zubizarreta has been to me everything an advisor should be, and then some. She hasbeenagreatlistenerandsupportivewhenIneededherthemost,andhasalwaysgivenmecom- pletefreedomtopursuethetopicsIwantedsinceIstartedtheprogram. Heradvicehasalwaysbeen sharp and on point, her criticism always constructive, be it for research problems or for my future professionalpursuits. Heradviceduringthislastyearhasbeeninvaluableforthecompletionofthis dissertationintermsofbothcontentandform.Tothisday,Iadmireherclarityofmindandhervast knowledgeofthefield,andIfeelreallyfortunatethatsheagreedtofollowmeonthisjourneywhenI askedhertobemyadvisorduringmyfirstyearintheprogram.Muchísimasgracias. Myothercommitteemembershavealsobeenwonderful. RoumiPanchevahasbeentomeasec- ondadvisorinspirit,fromherintroductorysyntaxcourseco-taughtwithAudreyLitoherguidance andencouragementwithseveralprojectsthroughoutmystayhere.Ihavelearnedalotfromherrig- orous approach to linguistics and every meeting with her shed new light on my topics and kept me ontrack. ThesamegoesforAudreyLi. Iwouldneverleaveameetingwithherwithoutanon-point observationandamove-forwardquestion,andshewouldalwayshaveareassuringsmileandhelpful advicetoofferwheneverIneededit. ThankyoualsotoGabrielUzquiano,whoonceagainagreedto serveasanexternalmemberforaLinguisticsdissertationcommittee.Despitenotbeingalinguistby training,hetookthetimetoreadmydissertationandgavemesharpandthought-provokingremarks duringmydefense. IalsoappreciatetheclassesthatItookwiththefacultyattheLinguisticsdepartment,whichbroad- enedmyperspectiveofthefieldandhelpedmeprepareforconductingindependentresearch.Among themareAndrewSimpson,HajimeHoji,HagitBorer,ElenaGuerzoni,BarrySchein,LouisGoldstein, KhalilIskarous,ElsiKaiserand,ofcourse,myinternalcommitteemembers. Anhonorablemention within the Linguistics faculty goes to Rachel Walker. She never served as a member in any of my committees,yetItookseveralworkshopsonteachingandthejobmarketwithher.Theseworkshops provedveryhelpfulwhenIhadtoteachandwhenIwaslookingforwork.Moreover,herdedication, xii expertiseandkindnessasaprofessorandasahumanbeingarearealinspirationforme. Thankyou Rachel. IshouldalsothankmygreatcohortatUSC:HuilinFang,PeterAraGuekguezian,DashaHenderer, CaitlinSmithandUliSteindl. Theyweregreatclassmatesandfriends,aswellasanamazingsource ofsupportduringthefirststressfulyearsintheprogram. Studentsfromothercohortsalsoenriched mytimeinthedepartmentbothacademicallyandpersonally. Theveteranswerewelcomingandof- feredhelpfuladviceastohowtonavigatetheprogram—thankyouMythiliMenon,SarahOuwayda, Sergio Robles-Puente, Héctor Velásquez, Xiao He, Canan Ipek, Saurov Syed, Ellen O’Connor, Brian Hsu and Pryanka Biswas. To all the newcomers to the program, I hope I was half as welcoming to you as the veterans were to me: A big hug to Ana Besserman, Andrés Benítez, Sarah Lee, Charlie O’Hara,BhamatiDash,CynthiaLee,SamanthaGordon,JessicaHarmon,MairymLlorens,ReedBlay- lock,ThomasBorer,MonicaDo,andJesseBisogni.Lastbutnotleast,thankstoJoyePerez,whowent beyondherroleasastudentadvisortomakesureallofusgraduatestudentsfeltwelcomeinourbe- ginnings at USC, as well as Frankie Hayduk, Brandon Washington, Guillermo Pérez, Lisa Jo Keefer andthestudentworkersforalltheirhelpwithadministrativematters.Iwilltakeveryfondmemories ofyouallwithmewhenIleaveLosAngeles. TheSpanishdepartmentatUSCwasreallywhereIspentmostofmytimeasagraduatestudent. IamreallythankfultohavehadGabrielaZapataasateachingmentorandsupervisorwhenIstarted teachingSpanish. Sheiswell-versedinherfield, efficient, hard-workingandcompassionate. Atrue force of nature. Her love of teaching, her dedication and her high standards are contagious, and I wouldnotbeatenthoftheteacherIamtodayifitwasnotforher.Shehelpedmeenormouslyinthe beginningandsheputatrustinmethathumblesmetothisday.WhenGabrielaleftUSCforaposition elsewhere, Goretti Prieto came along, and I could not have been luckier. Her support for graduate studentsisrelentlessandheradviceandhelpwhenIwaslookingforworkwasinvaluable. Muchas graciasalasdos.Asanassistantlecturer,Ialsohadthepleasuretoworksidebysidewithveryinspiring professionals who were always willing to give advice or lend a helping hand during the numerous times I needed one: Leah Kemp, María Mercedes Fages Agudo, Gayle Vierma, Mar Gómez, Karen Pérez, Liana Stepanyan, Andrea Parra, David Zarazúa, Consuelo Sigüenza-Ortiz, Carolina Castillo, VianeyCabrera,IvetteGómez,LorenaGallego,CarolinaSitnisky-Cole,EllenOliveira,MartínOcón- Gamarraand,lastbutnotleast,thefourcolleaguesthatcametoteachintheSpanishlanguageprogram thesameyearasme:LoriMesrobian,JackieCohen-Steinberg,FaithHardenandJoséCésardelToro. I also leave behind other truly wonderful people at the Spanish Department as well. They are unsure of what I was doing for my dissertation, but the moments of sheer fun that I shared with themduringbreaksatourofficewereawelcomebreakfromwriting. AhugtoVincentCervantes— xiii Doesanyonehaveasix-foothoagie?—,JackieSheehan—Iwillmissouroperanights!—,SarahSkillen, VanessaOvalle,CésarPérez,NoraEdén,GuillermoRodríguez,ErinMizrahiandLacySchauwecker. Also in this office were Héctor Velázquez and Sergio Robles-Puente, whom I thank again for their compañerismo. ThanksalsotoMariannaChodorowska-Pilch, JuliánGutiérrez-Albilla, JamesFort- neyandGianMariaAnnovi,fortheirfriendshipandtheirwisepiecesofadvicethroughoutmytimeat thedepartment.ThanksalsotoAmeliaAcosta,LauraDelgado,BerthaArce,KatherineChanGuevarra andthestudentworkersforalltheirinvaluablehelpwithadministrativematters. IsurelywouldnothavemadeittograduateschoolinthefirstplaceifithadnotbeenforGema Chocano.Gemasparkedmyinterestingenerativesyntaxinmylastyearofundergraduatestudiesin MadridandencouragedmetopursegraduatestudiesintheUnitedStates. Herpassionforthefield andhercandidyetempathicattitudetoherstudents,aswellasherabilitytomotivatethem,remainan inspirationtomeuptothisday. Ihopetohavehonoredalittlebittheenormousconfidenceshehad inme.AmayaMendikoetxeaintroducedmetocomparativelinguisticsandfurtherencouragedmeto pursuegraduatestudies.AttheUniversidadComplutensedeMadrid,ItookacoursewithMaríaJesús FernándezLeboransonaspectandtherelationshipbetweenthesyntaxandthelexicon,aninterestI havemaintaineduptothisday.VioletaDemontegavemethefinalpush:sheencouragedmetoapply toUSCandshehaskeptherinterestinme,askingmeregularlyhowthingsweregoingandgivingme usefulsuggestionsintermsofprofessionaldevelopment.Hergenerositywithmehasbeenenormous andIonlywishIcouldrepayhersomeday.Muchísimasgraciasatodas. Throughoutthisjourney,Ihavemetmanylinguistsallovertheworldinconferences,workshops andseminars,someofwhomIhavetheprivilegetocallfriends. AntonioFábregasandRafaelMarín have always shown great interest for my work and their feedback has left this dissertation in much better shape. Similarly, Jonathan McDonald, Grant Armstrong have always given me constructive feedbackinthenumerousvenuesinwhichwehavecoincidedandhavealwaystreatedmeasapeer whenIwasanythingbut.ThesamegoesforJaumeMateu,whoseprolificworkonargumentstructure isaconstantsourceofinspiration. ThanksalsotoVíctorAcedo-Matellán,MaríaJesúsArche,David Basilico, Elena Benedicto, Josep María Brucart, Aniko Csirmaz, Olga Fernández-Soriano, Raffaella Folli, Itamar Francez, Maria LluÏsa Hernanz, Daniel Hole, Paula Kempchinsky, Gillian Ramchand, IsabelleRoy,CristinaSánchez,FlorianSchäfer,Anna-MariaDiSciullo,IdaToivonen,MyriamUribe- EtxebarriaandEvaWittenberg.Theyhaveallgivenmehelpfulfeedbackformydissertation.Thanks arealsoduetoSimoneAlex-Ruf,ElenaAnagnostopoulou,IdaLarssonandIreneRappforsendingme theirwork. Iamequallythankfultomygreatinformants: ItamarKastnerandIritMeirforHebrew, ThomasBorerandUlliSteindlforGerman,GiorgosSpathasforGreek,DeliaBentleyforItalianand EkaterinaChernovaforRussian. xiv Thanksalsotomylinguistfriendsfromalloverwhohavemademesmileduringthistime:Margot Vivanco—ay, nuestras charlas sobre aspecto—, Elías Gallardo—porque te se apetece un bocadillo—, Pablo Rico, Javier Fernández-Sánchez—pursing our lips—, Anna Pineda, Ager Gondra—we need to meet in person already!—, Itamar Kastner, Imanol Suárez-Palma, Julia Villa-García, Melania Masià and Jennifer Tan. To my non-linguist friends in California, Norman, José, Ariel, Alina, Baptiste, Brett—youwillbeagreatwritersomeday—andFilipe.TothefriendsIleftbackinMadrid,CarlosCre- spoandVincent,MiguelÁngel,Rodrigo—nowinCambodia—,CarlosBarea,NachoGalán‘Nachala Macha’,Chus,EvaandAngela—alwaysbetweenMadridandJFK.Thankstomybelovedfriendsfrom high school, María Ayllón, Sofía, Silvia—os confundiré siempre y lo sabéis—, María Barbero, Lara, Laura,Noe,EvayEneko. Tomyfriendsfromschool,Rodrigo,Ester,Eva,Sara‘Golfi’,César,Juancar, Juanjo, Tini, Santa, Chus, Sara Bacigalupe, Azucena, y muchos más que ahora no recuerdo. La vida noshallevadomuylejosatodos,perosiemprelosllevoenmicorazón. My time in Los Angeles would have been a lot less stimulating were it not for the incredible venues where I could satiate my thirst for classic movies and independent films. My thanks to the now-defunctCinefamilyattheSilentMovieTheater,totheAmericanCinemathequeattheEgyptian TheatreinHollywoodandtheAeroTheatreinSantaMonica,theSecretMovieClubandtheir35mm screeningsattheVistaTheatreinLosFeliz,withoutforgettingLaemmletheatersandtheiranniver- saryscreeningswithsomanyinterestingguests.Thanksarealsoduetomoviepass,formakinggoing tothemoviesalotcheaperinmylastmonthsofdissertationwriting. Ialsoackowledgeacoupleof WhatsappgroupsthatIamapartof, Cariñitos de amor and Syntaxdiscoursinyourface, whichcertainly addedasparkofhumortomyfinalyearsintheprogram. Ioweeverythingtomyfamily.ThevaluesthattheytransmittedmehavemademethepersonIam todayandwithouttheirconstantsupportandencouragementIwouldhaveneversetfootongraduate schoolinthefirstplace. Iknowmybeingsofarawayforsomanyyearshasnotbeeneasyforthem andIamgratefulforthemtohavemadeiteasyonme.Myfinalandbiggestacknowledgmentgoesto mylategrandmother. Beforeattendingpreschool,shemadesureIknewhowtoreadandwrite. She wouldtakemeforawalkaroundmyhometown,Burgos,andreadwithmethelabelsonthebenches alongthepromenades,allofwhichreadCajadeAhorrosdelCírculoCatólico.Fromthereshemovedto introducemetoclassicliteraryworksandfilms, whichhavebeenasourceofenrichmentinmylife andhaveprovidedsolaceduringlonelytimes.Thisdissertationisdedicatedtoher. ¡Muchas gracias! LosAngeles,25June2018 xv 1 Introduction 1.1 Settingthestage ThelastdecadeshavewitnessedsubstantialprogressinourunderstandingofAktionsart—i.e.thetem- poralcontourofanevent—anditsconnectiontothesyntaxoftheverbanditsargumentstructure—i.e. theprojectionoftheeventparticipantsinthesyntax. Forinstance, ithasbeennotedthattheprop- erties of the direct object determine the telicity of the whole predicate with verbs of consumption andcreation,asthecontrastin(1)shows(Verkuyl1972;Dowty1979;Krifka1992,a.o.). Ithasalsobeen observedthattheadditionofanadjective,asmallclauseoraPPtoanotherwiseatelicverbcreatesa resultativepredicate,asin(2)(Talmy1985;Hoekstra1992;Mateu2002;Wechsler2005;Zubizarreta&Oh 2007, a.o.). In addition, the different base-generation position of the single argument of intransitive verbs(i.e.theUnaccusativityHypothesis,seePerlmutter1978;Burzio1986;Levin&Rappaport-Hovav 1995, for specific proposals and unaccusativity tests in different languages) has been often argued to correlatewiththe(a)telicityofthewholepredicate(Dowty1991;vanHout2004;Borer2005b,a.o.)(e.g. (3)).Syntax,then,effectivelybuildsaspectually-relevanteventstructureconfigurations. (1) a. John{ate/drew}threeapples{*foranhour/inanhour}. (Quantizedobject,telic) b. John{ate/drew}apples{foranhour/*inanhour}. (Non-quantizedobject,atelic) (2) a. Maryhammeredthemetal(flat). (Adjectivalresultative) b. Maryran(hershoesragged). (Smallclause) c. Marypushedthecart(totheendofheroad). (TerminalPP) (3) a. Carl{arrived/died/fell}. (Unaccusative,telic) b. Harriet{laughed/sang/worked}. (Unergative,atelic) Theseconfigurationsnotonlydeterminethetemporalmeaningoftheverb,butalsotheinterpretation ofitsarguments,asithasoftenbeenobserved(Tenny1987,1992;Hale&Keyser1993,a.o.). In(2a),for 1 Section1.1. Settingthestage Chapter1 instance,thebareobjectthehammerdoesnotnecessarilyundergoachange,butonlyaprocessofbeing hammered:however,theadditionofflatforcesaresultativereadinginwhichthehammerundergoesa changeofstate.Similarly,in(2c),thecartundergoesachangeoflocationwithawell-definedendpoint whenitappearswithaPP;withoutthePP,thecartmaynotevenmovefromthepushing.Someobjects arenotevenlicitwithoutanaspectualdelimiter(cf. *Maryranhershoes,seeSimpson1983;Smith1983; Carrier&Randall1992fordiscussion). Crucially, Aktionsart effects can also beobserved within verbal morphology. Forinstance, dead- jectival verbs—roughly corresponding to Dowty’s (1979) ‘degree achievements’—are telic if the base adjective is closed-scale (e.g. (4a)) but atelic if the base adjective is open-scale (e.g. (4b)), unless we caninferacontextually-relevantendpoint(Hayet al. 1999). Withrespecttodeadjectivalanddenom- inalverbs,Hale&Keyser(2002)havearguedthattheyarederivedviarelationalsyntacticstructures withaspectualmeaning.Intheirproposal,thederivedtelicverbsaresyntacticallyisomorphictotheir underivedcounterpartswithlightverbsbecomeorput,whichheadatelicstructure(a‘terminalcoinci- dencerelation’,intheirterminology)(e.g.(5a-b)).Atelicdenominalverbsalsohavethesamestructure astheircounterpartswiththelightverbdo,whichheadsanatelicsyntacticconstellation(e.g.(5c)). (4) a. Theyfilledtheglass{*fortwoseconds/intwoseconds}. b. Thesealevellowered{fortwodays/*intwodays}. (5) a. Theskycleared.(cf.Theskybecameclear) b. Theycorraledthehorses.(cf.Theyputthehorsesinacorral) c. Peterdanced.(cf.Peterdidadance) The picture that emerges, then, is that verbal morphology directly determines the Aktionsart of the resultingpredicatepreciselybecauseitissyntax.Thatis,verbsarenottakenfromthelexiconwithan encodedmeaning,butrather,theyarebuiltinthesyntaxandassucharesubjecttosyntacticprinciples andrestrictions,andareinterpretedcompositionallyjustastheyarebuiltcompositionally.Thisview regardingeventandargumentstructurehasthrivedwithinrecentsyntactic-orientedapproachesto morphologysuchasDistributedMorphology(Halle&Marantz1993;Marantz1997),Hale&Keyser’s (2002)L-syntax,Nanosyntax(Starke2009)ortheExo-Skeletalmodel(Borer2005a,b,2013). Although theirtechnicalimplementationsandtheirviewsonthesyntax-lexiconinterfacediffer,theysharethe commonassumptionthatwords, justlikesentences, aresyntacticobjects. Thisresearchprogramis knownasneoconstructionism,whosemaintenetscanbestatedasin(6). (6) Maintenetsofneoconstructionism: a. Argumentstructuredeterminestheeventstructureoftheverb. 2 Section1.2. Goalsandscopeofthisdissertation Chapter1 b. Thematicinterpretationofargumentsisanentailmentfromeventstructure. c. Deverbalwordsarederivedsyntacticallyfromverbalstructure. 1.2 Goalsandscopeofthisdissertation Theoverarchinggoalofthepresentdissertationistostudythemorphosyntaxandsemanticsofsta- tive predicates in the light of the neoconstructionist research program outlined in Section 1.1 and integratetheminacomprehensivesyntacticmodel. Iexplorehowstativepredicatescomeabout,i.e. whatsyntacticprimitivesdeterminetheiraspectualandthematicmeaningandhowrootslexicalize suchstructure.Ifocusonthreedifferenttypesofstativeverbs—stativecausatives,object-experiencer psychologicalverbsandlocativeverbs—andstativepastparticiplesderivedfromverbs. Myguiding questionscanbestatedasin(7). (7) Researchquestionsofthisdissertation: a. Whatsyntactico-aspectualprimitivesderivethesubclassesofstativeverbs? b. Howdostativeeventstructuresassignthematicinterpretationtoverbalarguments? c. Whatisthestructuralrelationshipbetweenstativeparticiplesandtheirverbs? Stativepredicatesareanimportantareaofinquirybecausetheyaregenerallyassumedtosimplybe impoverished versions of eventive predicates, and have not received as much attention in the core neoconstructionistliteratureasdynamicpredicates.Notionslikeagency,causativityandresultativity are often assumed to belong to the eventive—as opposed to stative—domain. My goal with respect toquestion(7a)isshowthatthesenotionscanpertaintostatestoo, onceweadoptamoreinclusive view of stativity and a better understanding of what the properties of aspectual operators are and howtheyinteracttemporallywiththeeventstheymanipulate.Ishowhowmyproposalsimplifiesthe taxonomyofeventsandtheirsyntacticprimitivesandthustheoverallarchitectureoftheVP. Anotherchallengeforasimplifiedviewofstativepredicatesisthefactthattheirargumentsdisplay aratherricharrayofthematicroles—Holder/Possesor,Experiencer,Stimulus,Location,Agent,and soon.Ifthematicrolesarederivedfromeventstructure,andstativepredicateshaveanimpoverished event structure, then this thematic richness becomes a puzzle. I address question (7b) by looking at stativepsychologicalandlocativeverbs.Iproposeasyntacticaccountthatderivesthesepossiblethe- maticrolesfromtheaspectualstructureoftheverbalpredicate,byshowinghowsomethematicroles can be collapsed into a generalized one (e.g. Experiencer and Ground location into a more abstract relationalrole)whileothersaresimplyworld-knowledgeeffects(e.g.agentivity). Anessentialdomaintoexplorestativityfromaneoconstructionistperspectivearestativedeverbal words.SincetheassumptionisthatbothAktionsartandwordsarederivedinthesyntax,itiscrucialto 3 Section1.3. Structureofthisdissertation Chapter1 understandwhattheaspectualoperatorsatplayare,howtheyinteractwithverbaleventivestructure and, in short, how they deliver different types of stative predicates crosscategorially. This issue is addressedinquestion(7c),whichtacklestheeffectsthatthebaseAktionsarthasonbuildingthederived stateaswellasthecross-linguisticvariationwithinderivedstates. Asitprogresses,mydissertationweavesaneoconstructionistaccountofthesestativepredicates withwell-definedsyntacticprimitivesandaparticularconceptionoftheirinterfacewithsemantics and the lexicon. It not only derives the stative predicates discussed here, but it also makes testable predictions regarding the possible types of states crosslinguistically in terms of meaning and mor- phosyntactic complexity. The main language under focus is Spanish (Castilian variety), my mother tongue. However, since I work from a universalist perspective, I also undertake comparison with otherlanguagestofurthersupportmyproposal. 1.3 Structureofthisdissertation Chapter2. Aktionsartandargumentstructure: Astateoftheart Thischapterintroducesthecanonicalaspectualclassificationofeventsandtheirrelevanttestsasap- plied to Spanish, which I will adopt for the remainder of the dissertation. I also provide some his- torical background regarding the decomposition of verbs in the syntax in connection to Aktionsart, constrastingitwithatomicviewsthatencodeaspectualtypesintherootorindedicatedverbalheads. I provide an overview of the main current neoconstructionist models. I focus particularly on the first-phasesyntaxframeworkdevelopedbyRamchand(2008), whosecorepostulatesIadoptforthe remainderofthisdissertation. Chapter3. Stativecausatives Ifocusontheunderstudiedsetofverbsexemplifiedin(8).Theseverbsaresporadicallyandinconsis- tentlyclassifiedintheliteratureaseitheractivitiesorstates,sincetheyappeartoshowtraitsofboth aspectualtypes. (8) gobernar‘govern’,vigilar‘surveil’,controlar‘control’,proteger‘protect’,presidir‘preside’... I argue that these verbs are complex stative predicates that form a causative-resultative configura- tion.Ianalyzethemsyntacticallyashavingabi-phrasalsyntax,wheretheexternalargumentisintro- ducedbyinitPandtheinternalargumentbyresP,andwheretherootlexicalizesbothinitandresheads (e.g. (9b)). This configuration is unambiguously interpreted by the semantic component as a stative causative,i.e.twocausallyrelatedstates.Thetoysemanticsaregivenin(9c). 4 Section1.3. Structureofthisdissertation Chapter1 (9) a. Berta Berta gobierna governs el the país. country ‘Bertagovernsthecountry.’ b. initP DP Berta init’ init <gobierna> resP DP elpaís res’ res <gobierna> XP c. s9s 1 ,s 2 [s=(s 1 !s 2 )&gobierna(s 1 )&participant(Berta,s 1 )&gobierna(s 2 )&participant(el país,s 2 )] Icomparemyproposaltomorecomplexanalysesthatassumeeitheraricherontologyofeventtypes (Maienborn2005;Fábregas&Marín2017)oraricherarrayofaspectualoperators(Neeleman&vande Koot2012).Ishowhowitisbothfeasibleandmoredesirablebothfortheoreticalandempiricalreasons to keep a binary classification of eventualities (event and state) and a single relation cause between sucheventualities,whichmaybestativeoreventive.Myproposaldiffersfromrecentonesthatsever theexternalargumentfromtheintroductionofcause. Idiscusshowmyaccountiscompatiblewith suchproposals. Chapter4. Stativeresultatives ThischapterexplorestheformationofadjectivalparticiplesinSpanish.First,Idisentangleattributive and predicative participles and conclude that only the latter are uniformly adjectival. I argue that adjectivalparticiplesmaybederivedfromeitherstativecausativeverbs, discussedinChapter3(e.g. (10a)), or telic verbs (e.g. (11a)). I show that the former projects a full argument structure (e.g. (10b)), whereas the latter only projects resP, which in isolation is interpreted as a non-resultative state (e.g. (11b)). Thisisdue, Iargue, tothestativityrequirementthatparticipialadjectivizationimposesonits baseverbalpredicate. 5 Section1.3. Structureofthisdissertation Chapter1 (10) a. La the ciudad city está is protegida. protected ‘Thecityisprotected.’ b. AdjP Adj initP init resP (11) a. La the ciudad city está is destruida. destroyed ‘Thecityisdestroyed.’ b. AdjP Adj resP Myanalysisofadjectivalparticiplesderivedfromtelicverbsasbeingadjectivizedverbalstatesstands instarkcontrasttocurrentaccountsthatpositafulleventargumentstructureforthebaseverb,and an aspectual operator that delivers the stative meaning (e.g. Bosque 2014 for Spanish). I show how my proposal is theoretically simpler and fares better with the data. I conclude this chapter with a discussion of the crosslinguistic typology of adjectival passives and a proposal to account for their parametricdifferences. Chapter5. Stativepsychologicalandlocativeverbs This chapter focuses on stative object-experiencer psychological verbs (e.g. (12)) and challenges the received view that they assign an inherent Experiencer-role to their object. I start by reviewing psych effects, i.e. the special grammatical behavior observed for these verbs. I then provide evidence thattheclassofstativelocativeverbs(e.g.(13))alsoshowthesamepsycheffectscrosslinguistically.This provides negative evidence for the view that Experiencer objects are linked to special grammatical behavior,giventhattheobjectsoflocativeverbs,oftentimesinanimate,areclearlynotExperiencers. (12) preocupar‘worry’,asombrar‘amaze’,aterrar‘terrify’,aburrir‘bore’,molestar‘annoy’... (13) rodear‘surround’,cubrir‘cover’,flanquear‘flank’,bloquear‘block’,cercar‘fence’,obstruir‘obstruct’... I propose, following Landau (2010) for object-experiencer psychological verbs, that both verb types have a common structure (e.g. (14)) that involves a silent preposition P that assigns oblique case to itsobject. Iarguethatthisprepositiondoesnotassignaninherentthematicrole;rather,theabstract relationbetweenthetwoargumentsisinterpretedasspatialorpsychologicaldependingonthelexical classoftheverbthatlexicalizessuchstructure,asillustratedin(14b)and(15b). 6 Section1.3. Structureofthisdissertation Chapter1 (14) a. La the crisis crisis aterra frightens a dom las the familias. families ‘Thecrisisfrightensfamilies.’ b. VP V aterra PP Lacrisis Stimulus P’ P alasfamilias Experiencer (15) a. La the sábana sheet tapa hides la the entrada. entrance ‘Thesheethidestheentrance.’ b. VP Lasábana Figure V’ V tapa PP P laentrada Ground Ialsoargueagainstthereceivedviewthatpsycheffectsonlytakeplacewithnon-agentivepsychverbs, but disappear with agentive psych verbs. I show that the factor at play is aspectual, not thematic: stativepsychologicalandlocativeverbsshowpsycheffects,whereaseventive—i.e.dynamic—onesdo not;Agentivityhasnorelevantgrammaticalstatusintheseconstructions. Chapter6. Conclusions I present the overall conclusions of my dissertation. I synthesize my main findings and discuss re- mainingpuzzlesandfuturedirections. 7 2 Aktionsartandargumentstructure:Astateoftheart 2.1 Introduction ThischapterpresentsanoverviewoftheclassiceventtypesoriginallyproposedbyVendler(1957)and the tests that I adopt to diagnose each type, as applied to Spanish. I also introduce a state of the art oftheliteraturethathasendeavoredtomodelsucheventtypessyntactically,ultimatelycrystallizing intheneoconstructionistprogram. Fortheeventtypes,ItaketheseminalworkofVendler(1957)asa startingpoint. Ireviewcriticallyhisclassificationaswellasothervariationsthereof,andIconclude withaprovisionaltwo-waydistinctionbetweenstatesandeventsand,withinthelatter,afurthertwo- waydistinctionbetweenactivitiesandtelics.ThisclassificationwillbeaugmentedinChapter3,when Iaddstativecausativestothestatetypes. Theliteraturereviewonthesyntaxofeventtypespresentedhereisfarfromcomprehensive,given itsvastnessandthespacelimitationsofthisdissertation.Therefore,Ihavegivenprioritytothepieces of work that I consider seminal for the study of the syntax of event structure (e.g. Hale & Keyser’s 1993,2002workonlexicalsyntax)and/orarerepresentativeofcurrentmainstreamviewsregarding therelationshipofthelexiconandthesyntax(e.g. Borer’s2005bneoconstructionistapproach). Iwill evaluatetheirproposalsalongwithRamchand’s2008first-phasesyntaxmodel,thecurrentnanosyn- tacticapproachtoeventstructure. Iwillarguethatfirst-phasesyntaxcircunventsmanyoftheprob- lemsfoundinprevioussyntacticmodels,chieflythedissociationbetweeneventtypeandgrammatical categoryandthelackofeventdecompositiontoderivethematicstructureandscopefactsadequately. Thischapterisstructuredasfollows. Section2.2presentstheVendlerianeventtypologyandthe relevanttestsforSpanish.Section2.3presentsanoverviewofthecoresyntax-orientedapproachesto eventstructure,namelyL-syntaxinSection2.3.2,theXS-modelinSection2.3.3andfirst-phasesyntax inSection2.3.4.Section2.4synthesizesthetake-awaypointsofthechapter. 8 Section2.2. Theclassiceventtypes Chapter2 2.2 Theclassiceventtypes Ipresentbelowtheclassiceventtypesthataregenerallydiscussedintheliterature.Ashortterminol- ogynoteisinorder:followingBach(1986),Iwillusethetermeventualitytorefertoalleventsingeneral, regardlessofaspectualclassification. Also, Iwillusetheterm event, unlessotherwisenoted, torefer todynamiceventualities(telicsandactivities)andstatetostative(i.e.non-dynamic)eventualities. 1 Another important note on terminology: I use the terms event type, Aktionsart, lexical aspect and inner aspect interchangably, all referring to the temporal contour or constituency of the eventuality. Thetermeventstructure,inthiswork,issimilartothelattertermsbutitcommitstoAktionsarthaving acertainstructure,i.e.asetofprimitivesthathavemeaningfulrelationswitheachother. 2.2.1 States Statesarepredicatesthatholdintime,butdonottaketime,inthewordsofTaylor(1977:206).Inother words,theydonotencodeanykindofinternalchangeordynamicity,butarereallystaticthroughout the timespan where they hold. Dowty (1979) neatly illustrates this defining trait of states with his subinterval property, defined as follows: if a non-dynamic (i.e. stative) predicate is true at an interval I, it will also be true at every subinterval of I, down to instants. For instance, in (16), if it is true at a certaintimeintervalthatAlbertohatespollutionorthatVictoriaownsfiveestates,itwillalsobetrue atanyinstantwithinthatinterval,i.e.theyhavethesubintervalproperty. (16) a. Alberto Alberto detesta detests la the contaminación. contamination ‘Albertodetestspollution.’ b. Victoria Victoria posee owns cinco five fincas. estates ‘Victoriaownsfivestates.’ 2.2.2 Activities Unlike states, activities are predicates that encode internal change: they are dynamic predicates. As such, they do not have the subinterval property: they need a minimal interval to be evaluated. For instance,adrivingorarocking(e.g.17)mustbeevaluatedataminimalintervaloftimefortheaction tocountasadrivingorarocking,asopposedtoameresittinginsideacarorabeingincontactwith acradle. 1 InthisIdepartfromBach(1986), forwhomeventualitiesotherthanstateswere, simply, ‘non-states’. Atelicnon-states wouldbe‘processes’andtelicnon-stateswouldbe‘events’.Seealsofootnote2. 9 Section2.2. Theclassiceventtypes Chapter2 (17) a. Berta Berta condujo drove el the coche. car ‘Bertadrovethecar.’ b. Carlos Carlos meció rocked la the cuna. cradle ‘Carlosrockedthecradle.’ 2.2.3 Telics Telicpredicatesarethosethatdenoteanendpointortelos. 2 Takethesentencesin(18)asanillustration: theverbalpredicatesdibujaruncírculo‘drawacircle’andcerrarlapuerta‘closethedoor’haveendpoints, namely,thecompletionofthecircleandthecompleteclosureofthedoor. (18) a. Juana Juana dibujó wrote un the círculo. article ‘Juanawrotethearticle.’ b. Pedro Pedro cerró closed la the puerta. door ‘Pedroclosedthedoor.’ 2.2.4 Thetests Throughoutthedecadesofstudyofverbalaspect, manytestshavebeenadducedfordiscerningbe- tween the different event types. In this section I will present the ones I deem relevant and reliable forSpanishandleaveoutthosethatareeithernotapplicabletoSpanishorareinfactunreliable. In- deed,someoftheclassictestsarenotconclusive:suchisthecaseoftheprogressive,which,although generallyassumedtotellapartbetweenstatesandevents(Lakoff1970), canactuallybefoundacross alleventtypes(Leech1970;Comrie1976;Verkuyl1989;Levin&Rappaport-Hovav1995:70,andseedis- cussioninSections3.2.2.1and5.5.2). Thebatteryofagentivitytests(complementsof persuadeor force, possibilitytoappearintheimperative,etcetera)belongalsointhisgroup; originallyassumedtotell aparteventsandstates(Lakoff1970),theyareunreliablesincetherearemanyeventivepredicatesthat arenotagentive(Dowty1979;Verkuyl1989;Filip1999;Levin&Rappaport-Hovav2005,a.o.) 3 andalso 2 TelicscorrespondtowhatKenny(1963)calledPerformancesandMourelatos(1978);Bach(1986);Verkuyl(1989);Parsons (1990)calledEvents. 3 InLevinandRappaport-Hovav’swords,agentivityseems‘orthogonaltoaspectualclassification,withagentiveandnona- gentivepredicatesbeingfoundineveryaspectualclass’(2005:89). 10 Section2.2. Theclassiceventtypes Chapter2 manystativepredicatesthatcanbeagentiveaswell(Dowty’s(1979)agentivestates,andseeSection5.5 forfurtherdiscussion). 4 Ihavedividedthetestsintosmallersectionsaccordingtowhateventtypes theytellapart. 2.2.4.1 Eventsvs. States Here I present the tests that can tell apart events from states. First, events and states have different readingsinthepresenttense,asnotedbyKenny(1963). Eventshaveahabitualreadinginthepresent tense(e.g.(19a-b)),whereasstateshaveanongoing,non-habitualinterpretation(e.g.(19c)). (19) a. David David diseña designs su his propia own ropa. clothes ‘Daviddesignshisownclothes.’ b. Berta Berta baila dances flamenco. flamenco ‘Bertadancesflamenco.’ c. María María sabe knows geografía. Geography ‘MaríaknowsGeography.’ Anothertestthatdifferentiateseventsfromstatescomesfromtheirdifferentreadingsintheperfect. Assomeauthorshavenoted(eg.Iatridouetal. 2001;Portner2003),statesgiverisetoauniversalreading inthepresenceofatemporaladverbialheadedbysince:suchmodifierspecifiesaleftboundaryforthe eventualityandassertsthattheeventualityholdsbetweenthatleftboundaryandthereferencetime, which is the right boundary. However, events in the perfect cannot give rise to a universal reading andtheyareoutinthepresenceofasince-adverbial. 5 4 Forinstance,Iwillnotadmittheprogressiveonitsownasareliabletest,sinceitdoesnotgiveconsistentresultsacross theboard(seealsomydiscussiononSection3.2.2.1 5 However,wecanhaveauniversalreadingforeventsiftheyappearintheprogressive: compare(ia)and(ib)with(20a) and(20b),respectively. (i) a. Juan Juan ha has estado been pintando painting su his casa house desde since ayer. yesterday ‘Juanhasbeenpaintinghishousesinceyesterday.’ b. Fernando Fernando ha has estado been paseando walking al acc.the perro dog desde since las the seis six de of la the tarde. evening ‘Fernandohasbeenwalkingthedogsince6pm.’ 11 Section2.2. Theclassiceventtypes Chapter2 (20) a. Juan Juan ha has pintado painted su his casa house (*desde (*since ayer). yesterday ‘Juanhaspaintedhishouse(*sinceyesterday).’ b. Fernando Fernando ha has paseado walked al dom.the perro dog (*desde (*since las the seis six de of la the tarde). evening ‘Fernandohaswalkedthedog(*since6pm).’ c. Laura Laura ha has vivido lived en in Londres London desde since 2009. 2009 ‘LaurahaslivedinLondonsince2009.’ DeMiguel(1999)notesthattheSpanishaspectualperiphrasispararde+V.inf‘stopV-ing’doesnotallow forstates(e.g. (21a)): anotherperiphrasis, dejar de + V.inf‘stopV-ing’mustbeusedinsteadtoexpress cessative aspect (e.g. (21b)). Events, on the other hand, readily accept the parar de + V.inf periphrasis (e.g.(21c)). (21) a. *Julia *Julia paró stopped de of {saber {know inglés/ English conocer know Roma/ Rome odiar hate a dom Luis/ Luis ser be alta/ tall tener be sed}. thirsty intended: ‘Juliastopped{speakingEnglish/knowingRome/hatingLuis/beingtall/being thirsty}.’ b. Julia Julia dejó stopped de of {tener {have sed/ thirst odiar hate a dom su her primo}. cousin ‘Juliastopped{beingthirsty/hatinghercousin}.’ c. Julia Julia paró stopped de of {andar {walk un a momento/ moment construir build la the casa house cuando when llegaron arrived las the lluvias}. rains ‘Juliastopped{walkingforamoment/buildingthehousewhentherainscame}.’ (deMiguel1999:3012) ThesyntheticfutureinSpanishcanhaveanepistemicreadinginadditiontoitstemporalreading(Gili Gaya1961callsitthe‘futureofprobability’). Itseems,however,thatthisreadingisonlylicensedwith stativepredicatesasin(22a):bothactivitiesandtelicsdisallowtheepistemicreading,as(22b)and(22c) show,respectively. (22) a. Un a kilo kilo de of manzanas apples costará cost.fut cinco five euros. euros 3 Temporalreading:akilogramofappleswillcostfivekilosatsomepointinthefuture. 3 Epistemicreading:Itismyguessthatakilogramofapplescostsfivekilos. 12 Section2.2. Theclassiceventtypes Chapter2 b. Juan Juan mecerá rock.fut la the cuna. cradle ‘Juanwillrockthecradle.’ 3 Temporalreading:Juanwillrockthecradleatsomepointinthefuture. 7 Epistemicreading:ItismyguessthatJuanrocksthecradle. c. Esther Esther pintará paint.fut un a retrato. portrait ‘Estherwillpaintaportrait.’ 3 Temporalreading:Estherwillpaintaportraitatsomepointinthefuture. 7 Epistemicreading:ItismyguessthatEstherpaintsaportrait. Conditionalsentenceswiththemainclauseinthefuturetenseandthedependentclauseinthepresent tense(i.e.theSpanishcounterpartofEnglish’s‘type-1conditionals’)constituteanothertesttotellapart statesandevents(GómezVázquez&GarcíaFernández2013;Jaque2013). Whenthedependentclause hasastateasitsmainpredicate,cangetbothaprospectiveandpresentreadingforsuchpredicate(e.g. (23a)).However,onlytheprospectivereadingisavailablewithbothactivities(e.g.(23b))andtelics(e.g. (23c)). (23) a. Si if Sara Sara tiene has un a amante, lover su her novio boyfriend se refl pondrá put celoso. jealous ‘IfSarahasalover,herboyfriendwillgetjealous.’ 3 Prospectivereadingofthedependentclause:ifSarahweretohavealoverinthefuture, herboyfriendwillgetjealous. 3 Presentreadingofthedependentclause:IfSarahcurrentlyhasalover,herboyfriend willgetjealous. b. Si if David David baila dances con with María, María Daniela Daniela se refl enfadará. get.angry.fut ‘IfDaviddanceswithMaría,Danielawillgetangry.’ 3 Prospective reading of the dependent clause: David will dance with María at some pointinthefuture,andthenDanielawillgetangry. 7 Presentreadingofthedependentclause:DavidcurrentlydanceswithMaría,andthen Danielawillgetangry. c. Si if María María construye builds una the casa, house Pedro Pedro se refl alegrará. get.happy.fut 13 Section2.2. Theclassiceventtypes Chapter2 ‘IfMaríabuildsthehouse,Pedrowillgethappy.’ 3 Prospectivereadingofthedependentclause:IfMaríabuildsahouseatsomepointin thefuture,thenPedrowillgethappywhensheisdone. 7 Presentreadingofthedependentclause:IfMaríaiscurrentlybuildingahouse,Pedro willgethappywhensheisdone. Alongsimilarlines,Jaque(2013)arguesthattheexpression<esperoque‘Ihopethat’+presentsubjunc- tive>mayhaveaprospectiveandpresentreadingwithstativeverbs(e.g.(24a)).Thisisnotthecasewith activitiesandtelicverbs,whichonlyallowforaprospectivereading(e.g.(24a)and(24b),respectively). (24) a. Espero hope.1p.sg que that vivas live.subj.2p.sg en in Madrid. Madrid ‘IhopethatyouliveinMadrid.’ 3 Prospectivereading:IhopethatyouliveinMadridatsomepointinthefuture. 3 Presentreading:IhopethatyoucurrentlyliveinMadrid. b. Espero hope.1p.sg que that corras run.subj.2p.sg en in la the maratón. marathon ‘Ihopethatyouruninthemarathon.’ 3 Prospectivereading:Ihopethatyourunthemarathonatsomepointinthefuture. 7 Presentreading:Ihopethatyouarecurrentlyrunningthemarathon. c. Espero hope.1p.sg que that repares repair.subj.2p.sg el the ordenador. computer ‘Ihopethatyourepairthecomputer.’ 3 Prospectivereading:Ihopethatyourepairthecomputeratsomepointinthefuture. 7 Presentreading:Ihopethatyouarerepairingthecomputer. Aremarkneedstobemadeabouttheselastthreetests,soastoavoidconfusion:asJaque(2013)notes, when eventive verbs appear in the progressive form or are interpreted habitually, they pattern as states. Ishowthatforthesyntheticfuturein(25),conditionalsentencesin(26)and espero que‘Ihope that’constructionsin(27). (25) a. Pedro Pedro estará be.fut comiendo eating una a ensalada. salad 3 Temporalreading:Pedrowillbeeatingasaladatsomepointinthefuture. 3 Epistemicreading:Pedroisprobablyeatingasaladrightnow. 14 Section2.2. Theclassiceventtypes Chapter2 b. Pedro Pedro conducirá drive.fut un a taxi. taxi 3 Temporalreading:Pedrowilldriveataxiatsomepointinthefuture. 3 Epistemicreading:ItismyguessthatPedrodrivesataxi. (26) a. Si if Pedro Pedro está is fregando washing los the platos, dishes su his anfitrión host se refl enfadará. get.mad.fut ‘IfPedroiswashingthedishes,hishostwillgetmad.’ 3 Prospective reading of the dependent clause: if Pedro happens to be washing the dishes at some point in the future (e.g. when his hosts arrives home), his host will getmad. 3 Presentreadingofthedependentclause:IfPedroiswashingthedishesatthemoment, hishostwillgetmad. b. Si if Pedro Pedro escribe writes novelas, novels ganará win.fut.3p.sg mucho much dinero. money ‘IfPedrowritesnovels,hewillmakealotofmoney.’ 3 Prospectivereadingofthedependentclause: ifPedrowritesnovelsinthefuture,he willmakealotofmoney. 3 Presentreadingofthedependentclause:IfPedrocurrentlywritesnovels,hewillmake alotofmoney. (27) a. Espero hope.1p.sg que that estés are.subj trabajando. working ‘Ihopethatyouareworking.’ 3 Prospectivereading:Ihopethatyouareworkingatsomepointinthefuture. 3 Presentreading:Ihopethatyouarecurrentlyworking. b. Espero hope.1p.sg que that comas eat.subj.2p.sg verduras. vegetables ‘Ihopethatyoueatvegetables.’ 3 Prospectivereading:Ihopethatyoueatvegetabesatsomepointinthefuture. 3 Presentreading:Ihopethatyoucurrentlyeatvegetables. 2.2.4.2 Telicvs. Atelic TheclassictestfortelicitycomesfromVendler’soriginalwork.Henotesthattelicpredicatescantake in x timemodifiers,expressingthetimeittakesfortheeventtobeaccommplished,butnotfor x time 15 Section2.2. Theclassiceventtypes Chapter2 modifiers,sincethesemodifiersdonotentailacompletion(e.g.(28a)).Atelicpredicates,conversely,do notacceptin x timemodifiersbecausetheydonottaketime,butmerelygoonorholdforsometime period,andthusacceptforxtimemodifiers.Iprovideexamplesforactivities(28b)andstates(28c). (28) a. Pedro Pedro construyó built una a carretilla wheelbarrow {en {in dos two horas/ hours *durante *for dos two horas}. hours ‘Pedrobuiltawheelbarrow{intwohours/*fortwohours}.’ b. Carlos Carlos empujó pushed la the carretilla wheelbarrow {*en {*in dos two horas/ hours durante for dos two horas}. hours ‘Carlospushedthewheelbarrow{*intwohours/fortwohours}.’ c. Ester Ester tuvo had un a coche car rojo red {*en {*in dos two años/ years durante for dos two años}. years ‘Esterhadaredcar{*intwoyears/fortwoyears}.’ A caveat is in order about the for x time test. It may actually be possible with telic predicates in two cases. First,ifthetelicpredicatehasanassociatedresultstatethatcanbeconceivedastemporary,a forxtimephraseislicitinthesentence. Iprovideanexamplein(29),adaptedtoSpanishfromDowty (1979:250). 6 Thesecondcaseiswhentheteliceventisinterpretedasbeingrepeatednumeroustimes throughoutthetimespandenotedbythe for x timephrase(whatMcDonald2008callsthe sequence of identical events interpretation). This is illustrated in (29b), under the interpretation that Carlos kept throwing the ball up to the roof repeatedly for two hours (say, the ball kept rolling back down and Carlosgrabbeditandthrewitupagain). 7 (29) a. El The sheriff sheriff de of Nottingham Nottingham encarceló jailed a acc Robin Robin Hood Hood durante for cuatro four años. years ‘ThesheriffofNottinghamjailedRobinHoodforfouryears.’ b. Carlos Carlos lanzó threw el the balón ball al to.the tejado roof durante for dos two horas. hours ‘Carlosthrewtheballtotherooffortwohours.’ 6 DowtytakesthisexamplefromMorgan(1969)andMcCawley(1973),whointurnattributeittoRobertI.Binnick. 7 Imustaddthat,atleastinmydialect,thesequenceofidenticaleventsinterpretationcomesaboutmorenaturallywith aprogressive,asin(ii). (ii) Carlos Carlos estuvo was lanzando throwing el the balón ball al to.the tejado roof durante for dos two horas. hours ‘Carlosthrewtheballtotherooffortwohours.’ 16 Section2.2. Theclassiceventtypes Chapter2 A test that also features in Vendler (1957) is the How long did it take? test. According to Vendler, that question,alongwithitsassociatedanswer,isonlypossiblewithtelics. IillustratethetestforEnglish in(30)and(31),adaptedfromVendler(1957:145). (30) a. Howlongdidittakehimtodrawthecircle? b. Ittookhim20secondstodrawthecircle. (31) a. #Howlongdidittakehimtopushthecart? b. #Ittookhim20secondstopushthecart. Interestingly, I observe that this test translates to (Castilian) Spanish with two different verbs, llevar and tardar‘take’,whichprovidetwodifferentdiagnostics. Thefirstone, llevar,issimilartoVendler’s viewofEnglishtakeinthatitdiagnosestelicity: withtelics,theverbllevarislicit(e.g. (32)),butnotso withactivitiesandstates(e.g.(33)). (32) a. ¿Cuánto ¿How.much le him llevó took a dat Imanol Imanol reparar to.repair el the ordenador?. computer ‘HowlongdidittakeImanoltorepairthecomputer?’ b. A dat Imanol Imanol le him llevó took dos two horas hours reparar to.repair el the ordenador. computer ‘IttookImanoltwohourstorepairthecomputer.’ (33) a. *A *dat Juan Juan le him llevó took dos two horas hours conducir to.drive el the coche. car ‘IttookJuantwohourstodrivethecar.’ b. *A *dat César César le him llevó took dos two años years vivir to.live en in Londres. London ‘IttookCésartwoyearstoliveinLondon.’ Theverbtardar,althoughverysimilarinmeaning,cruciallydiagnosesdifferentaspectualproperties. Unlike llevar, tardar ispossible withall eventtypes. This isso becausepredicateswith tardar donot necessarilydenotethetimethatgoesbyinthedevelopmentofateliceventualityfromitsinception until its culmination, but rather, the time that it passes from a contextually relevant point until the eventuality comes about—the result state in the case of telic eventualities, and the activity or state itself in the case of atelic eventualities. For instance, in (34), it is not necessarily the case that Pablo tooktwohourstodrawacirclesincehetookthepencilandstarteddrawing,butrather,thathedid notgettocompletethecircleuntiltwohoursafteracontextuallyrelevanttime(e.g.whenhisteacher toldhimthatheshoulddrawacircle). 17 Section2.2. Theclassiceventtypes Chapter2 (34) a. ¿Cuánto ¿How.much tardó took Pablo Pablo en in dibujar to.draw el the círculo?. circle ‘HowlongdidittakePablotodrawthecircle?’ b. Pablo Pablo tardó took dos two horas hours en in dibujar to.draw el the círculo. circle ‘IttookPablotwohourstodrawthecircle.’ Withateliceventualities,crucially,thereadingisnotoneofculmination,butratheroneofinception. For instance, (35a) means that after the contextually relevant time point (e.g. when we all gathered together),halfanhourelapseduntilCarlosbeganhistalking. Thesentencedoesnotmean,crucially, thatthetalkingwentonforhalfanhouruntilaculminationpointwasreached. Thesameinceptive readinghappenswithstativepredicates(e.g.(35b)). (35) a. Carlos Carlos tardó took media half hora hour en in hablar. to.talk ‘IttookCarloshalfanhourtostarttalking.’ b. Juan Juan tardó took varios several años years en in confiar to.trust en in Diego. Diego ‘IttookJuanseveralyearstotrustinDiego.’ Anothertelicitytest,takenfromKenny(1963),isthedifferententailmentsfromtheprogressivetothe perfect: telicpredicatesintheprogressivedonotentailtheirperfectcounterpart(e.g. (36a))whereas atelicpredicatesdoentailit(e.g.(36b)). (36) a. Pedro Pedro está is dibujando drawing un a círculo circle 9 9 Pedro Pedro ha has dibujado drawn un a círculo. circle ‘Pedroisdrawingacircle9Pedrohasdrawnacircle.’ b. Carlos Carlos está is acariciando petting al acc.the perro dog ! ! Carlos Carlos ha has acariciado petted al acc.the perro. dog ‘Carlosispettingthedog!Carloshaspettedthedog.’ Dowty(1979)notedthattheadverbalmosthastworeadingswithtelicpredicatesbutonlyonewithac- tivities. 8 Forinstance,withthetelicpredicatein(37a),thereisthereadingwhereEvahadtheintention of painting a picture but ended up not doing anything at all, but also the reading where Eva started 8 Forprecision’ssake,Dowty(1979)restrictedthisclaimregardingscopeambiguityofalmosttotelicaccomplishmentverbs (discussed in Section 2.2.5.1). Achievement verbs do not show the ambiguity, according to him. However, I show this is incorrectinSection2.2.5.2. 18 Section2.2. Theclassiceventtypes Chapter2 painting the picture but did not finish it completely. However, activities only allow for the former reading(e.g. In(37b),thereisonlythereadinginwhichPedrothoughtaboutwhistlingbutdidn’tdo it).Statesdonotseemtoallowalmostmodificationatall(e.g.(37c)). (37) a. Eva Eva casi almost pinta paints un a cuadro. picture ‘Evaalmostpaintedapicture.’ b. Pedro Pedro casi almost silba. whistles ‘Pedroalmostwhistles.’ c. *Laura *Laura casi almost sabe knows geografía. Geography intended:LauraalmostknowsGeography.’ Morgan(1969);McCawley(1973);Dowty(1979)notesimilareffectswiththeadverbagain:itisambigu- ouswithtelicsbutunambiguouswithatelics.In(38a),forinstance,wehavetworeadings:theexternal reading, where Juan has performed the action of closing the door before, and the internal reading, which merely says that Juan brings the door back to a closed state, without him necessarily having performed the action of closing it previously (Dowty 1979:252). Atelic eventualities as in (38) do not showthatambiguity. (38) a. Juan Juan cerró closed la the puerta door de of nuevo. new ‘Juanclosedthedooragain.’ b. Carmen Carmen acarició swam al of gato new denuevo. ‘Carmenswamagain.’ c. Pedro Pedro cree believes en in Dios God de of nuevo. new ‘PedrobelievesinGodagain.’ 2.2.5 Twotypesoftelicpredicates Withintelicpredicates,wecandistinguishtwosubtypes: Accomplishmentsandachievements. Iwill discussthetwosubtypesbelow,withaparticularattentiontoachievements,giventhatitisanevent typewhosedefiningproperties,inmyopinion,havebeenandstillareoftenmisunderstood. 19 Section2.2. Theclassiceventtypes Chapter2 2.2.5.1 Accomplishments In Vendler’s classic work, accomplishments are defined as predicates that develop in time up to an endpoint. Thisiscontrastedwithactivities,whichdevelopintimeinahomogeneouswaysuchthat anysubpartthereofislikethewhole.Iprovideexamplesin(39). (39) a. Juana Juana escribió wrote el the artículo. article ‘Juanawrotethearticle.’ b. Carlos Carlos paseó walked al acc.the perro dog hasta until el the final end del of.the puente. bridge ‘Carloswalkedthedoguptotheendofthebridge.’ AsnotedinDowty(1979),accomplishmentverbsbehavelikeactivitiesintheabsenceofadirectobject oraprepositionalcomplement(e.g.(40)). (40) a. Juana Juana escribió wrote {durante {for horas/ hours *en *in una one hora} hour ‘Juanawrote{forhours/*inonehour}.’ b. Carlos Carlos paseó walked al acc.the perro dog {durante {for horas/ hours *en *in una one hora} hour ‘Juanawrote{forhours/*inonehour}.’ Also, as noted by many authors (Verkuyl 1972; Bach 1986; Krifka 1989, a.o.), the direct object or the prepositionalcomplementhastohavecertainpropertiesfortheaccomplishmentreadingtoemerge: thedirectobjectmustspecifysomequantitythereof(41)andtheprepositionalcomplementneedsto specifyanendpoint(42). (41) a. Juana Juana escribió wrote el the artículo article {*durante {*for horas/ hours en in una one hora}. hour ‘Juanawrotethearticle{*forhours/inonehour}.’ b. Juana Juana escribió wrote artículos articles {durante {for horas/ hours *en *in una one hora} hour ‘Juanawrotearticles{forhours/*inonehour}.’ (42) a. Carlos Carlos paseó walked al acc.the perro dog hasta until el the final end del of.the puente bridge {*durante {*for horas/ hours en in una one hora}. hour ‘Carloswalkedthedoguptotheendofthebridge{*forhours/inonehour}.’ 20 Section2.2. Theclassiceventtypes Chapter2 b. Carlos Carlos paseó walked al acc.the perro dog por around el the parque park {durante {for horas/ hours *en *in una one hora} hour ‘Carloswalkedthedogaroundthepark{forhours/*inonehour}.’ Accomplishments,then,areactivityverbswithaquantizeddirectobjectoraterminalprepositional phraseasacomplement.Inthissense,thetelicityofaccomplishmentpredicatesiscompositional,since itisconstruedbythesyntax,andtheseverbsarethereforeaspectuallyambiguousbetweenactivities andaccomplishments(seealsoMittwoch1991,whoparticularlyemphasizesthispoint). 2.2.5.2 Achievements In his classic work, Vendler contrasts accomplishments with achievements. While this author takes accomplishmentstobepredicatesthattaketimetodevelopbeforereachingtheirendpoint(remember Section 2.2.5.1), he views achievements as punctual predicates that cannot be said to denote ongoing processes(e.g.intheexample(43b),someonecanbesaidtobeengagedinrunningaraceforacertain timeperiod,butawinningcannotdevelopintime:ithastobepunctual). (43) a. Johnwrotealetter. (Accomplishment) b. Johnwontherace. (Achievement) Vendlertreatsachievementsonaparwithstates,onthebasisthatneitheroneofthemareprocesses that go on in time: the former happen at a single point in time and the latter merely hold during a certaintimeperiod.Moreradically,Verkuyl(1989)arguesthatachievementsandstatesformanatural class.Theviewofachievementsasadistinctaspectualclassdefinedintermsoftemporalpunctuality hasbeenveryinfluentialsinceVendler(1957)(Mourelatos1978;Bach1986;Smith1991;Ramchand2008; Travis2010,tonameafew:twonotableexceptionstothistrendarePustejovsky1991andBorer2005b). 9 However,itturnsoutthatwedonothavereliabletestsfordistinguishingaccomplishmentsfrom achievementsintermsofpunctuality.AclassictestfromVendler(1957)wastheunacceptabilityofthe progressivewithachievements,buthasbeensincechallengedbymany(Leech1970;Comrie1976;Vlach 1981;Parsons1990;Verkuyl1989;Borer2005b)whonotedthatbothaccomplishmentsandachievements canappearwiththeprogressive.ThesamesituationgoesforSpanish,andIillustrateitbelowin(44). Aswecansee,theparadigmisfarfromexceptional. 9 Pustejovsky(1991)arguedthattheaspectualstructureofaccomplishmentsandachievementsisidenticalandtheyonly differ in terms of agentivity, accomplishments being agentive and achievements being non-agentive. Borer (2005b) also claims that the temporal structure of these two types is not distinct: for her, achievements are merely verbs that are lexicallyspecifiedfortelicstructure. 21 Section2.2. Theclassiceventtypes Chapter2 (44) a. Pedro Pedro está is llegando arriving a at casa. home ‘Pedroisarrivinghome.’ b. María María está is muriéndose. dying.refl ‘Maríaisdying.’ c. El the árbol tree se refl está is cayendo. falling ‘Thetreeisfallingdown.’ d. La the empresa company está is logrando achieving sus its objetivos. goals ‘Thecompanyisachievingitsgoals.’ e. Carmen Carmen está is reparando repairing el the ordenador. computer ‘Carmenisrepairingthecomputer.’ f. Víctor Víctor está is ganando winning la the maratón. marathon ‘Víctoriswinningthemarathon.’ Thefactthattheachievementsin(44)accepttheprogressiveshouldnotbesurprising,however,once wetakeacloserlook. Take(44e)forinstance. Thetransitionfromacomputerbeingbrokentobeing repaired is intuitively instantaneous, but there is nonetheless a previous, preparatory process that leadstothecomputerbeingfullyfunctionalagain. 10 Itseemsthatsuchpreparatoryprocessshouldbe grammaticallyrepresentedsomehow,giventhattheprogressivecanaccessit. SeealsoBorer(2005b), foracritiquealongtheselines. Anothertestthatpurportedlydiagnosesthepunctualnatureofachievementsisthelackofscope ambiguitieswithalmost(Dowty1979). (45) Johnalmostnoticedthepainting. However, it turns out that many achievement verbs do in fact show a scope ambiguity with almost, as shown in (46) for its Spanish counterpart casi. It needs to be noted that the different scenarios in 10 Rothstein(2012)actuallyclassifiesrepairasanaccomplishment,andrightlyso,ifoneadscribestotheviewthataccom- plishmentstaketimebutachievementsareinstantaneous.Butnotewellthatrepairdoesnotbehavelikeanaccomplishment inthatitstelicityisnotcompositional,i.e.thatverbisinherentlytelic. 22 Section2.2. Theclassiceventtypes Chapter2 (46)aregrammaticallyrelevant,i.e.theydiagnoseatrueambiguity.Evidenceforthisisthatwhenthe predicateisintheprogressive(e.g.(47)),itisonlycompatiblewiththesecondscenarioin(46),i.e.that inwhichcasimodifiestheresultstate—entailingthattheprocesspartoftheeventhasindeedbegun, henceitscompatibilitywiththeprogressive. (46) a. Pedro Pedro casi almost repara repairs mi my ordenador. computer ‘Pedroalmostrepairedmycomputer.’ 1. It was almost the case that Pedro got to repair my computer (but in the end another technicianwasassignedtorepairmycomputer) 2. Pedro worked on my computer and he almost repaired it, but stopped working on it withoutreturningittoafunctionalstate. b. Sara Sara casi almost llega arrives a in Londres London ayer. yesterday ‘SaraalmostarrivedinLondonyesterday.’ 1. ItwasalmostthecasethatSarastartedherjourneytoLondon, which, hadshestarted, wouldhavetakenhertoLondonyesterday. 2. SarawasonherwaytoLondonandclosetoarriving,butsomethingcameupthatpre- ventedherfromarrivingyesterday. (47) a. Pedro Pedro está is reparando repairing mi my ordenador. computer ‘Pedroisrepairingmycomputer.’ Onlycompatiblewithscenario2in(46a). b. Sara Sara está is llegando arriving a in Londres. London ‘SaraisarrivinginLondon.’ Onlycompatiblewithscenario2in(46b). Theambiguityofalmost,then,cannotbeadefiningtraitofachievementsasaninstantaneouseventu- ality.Intuitively,thereasonthatitishardtoseeascopeambiguitywithnoticein(45)isthatitdoesseem tohappeninstantaneouslyor,atleast,inanextremelyshortamountoftime(i.e. whenonerunsinto apainting,onetypicallynoticesitrightaway). Fordiscussionofotherclassictestsforachievements thatalsocollapseunderfurtherscrutiny,suchastheittookxtimetestorthestart/stopVingtest(Dowty 1979),thereaderisreferedtoBorer(2005b)fordiscussion. 23 Section2.2. Theclassiceventtypes Chapter2 Instead, I take the defining trait of achievements to be that their telicity is not compositional, unlike accomplishments. That is, the verb does not require syntactic satellites for a telic reading to arise: achievement verbs denote a change of state, and as such they have an inherent endpoint. In other words, they are lexical resultatives. In fact, I take so-called instantaneous achievements to not beaspectuallydifferentfromotherchange-of-stateofverbslike destroyor open, whichareoftennot classifiedasachievementsbutnonethelessarenotdistinctwithrespecttogrammaticaltests. 11 2.2.6 Againstsemelfactives Anothereventtypethathasbeenproposedintheliteratureisthatofsemelfactives,suchas knockor cough. Originally proposedby Smith (1991), semelfactivesare defined as atelic, instantaneous events. Smithnotesthattheseverbscanbecoercedintoaderivedactivity,wheretheeventualityisinterpreted repetitively (i.e. in (48), the for-phrase covers a time interval with multiple instances of a coughing, and similarly, the progressive in (49) focuses in part of a composite process with repeated events of coughingorknocking). (48) Johncoughedfor5minutes. (49) a. Johnwascoughing. b. Billwasknockingatthedoor. (Smith1991:56) Smithproposesasystemofbinaryaspectualfeaturesformodelingeventtypes(seeTable2.1),inwhich activitiesareonlydistinguishedfromsemelfactivesintermsofthevalueofthe[DURATION]fea- ture.Bothare[–STATE]and[–TELIC](sincetheyareneitherstativenortelic)butactivitiesare[+DU- RATION]andsemelfactivesare[–DURATION]. Rothstein(2004)adoptsadifferentview,andclaimsthatsemelfactivesareactuallyinstantaneoustelic events,essentiallyachievements:thedurativereadingisderivedbyS-summing,anoperationthatputs together a series of punctual events. Ramchand (2008) takes a similar position and proposes that semelfactives are special in that they are ambiguous between an achievement and an activity read- ing. If Rothstein and Ramchand were correct, however, and semelfactives were achievements, we would expect theprogressive version of these sentencesto be aspectually ambiguous as well. How- 11 My view on achievements departs from Pustejovsky’s (1991) and Borer’s (2005) in that they assume the same aspectual structureforbothaccomplishmentsandachievements. IadoptaviewsimilartoRamchand’s(2008)inthatthetelicityof accomplishmentsandchange-of-stateverbsisbuiltindifferentways,butIdepartfromRamchandinthatshealsoposits distinctstructureforchange-of-stateverbsandachievementsinordertoaccountfortheperceivedtemporalpunctuality ofthelatter. 24 Section2.2. Theclassiceventtypes Chapter2 Table2.1:TheaspectualfeaturesofeventtypesinSmith(1991) [STATE] [TELIC] [DURATION] States + – + Activities – – + Accomplishments – + + Achievements – + – Semelfactives – – – ever,itturnsoutnottobethecase.In(50a),thereadingisnotoneinwhichJuanisabouttojump,but one in which he is already jumping, i.e. he is either up on the air or he is in the midst of a series of jumps.Crucially,wefindtheoppositesituationwithachievements:in(50b),theonlyreadingisonein whichJuanisabouttoarrivehome:heiswalkingtowardsitandheisonlyashortdistanceaway.The momenthegetshome,wecannolongertalkabouthimarrivinghome,muchlesscanwetalkabout himbeinginthemiddleofarrivinghomeagainandagain. (50) a. Juan Juan está is saltando. jumping ‘Juanisjumping.’ b. Javi Javi está is llegando arriving a at casa. home ‘Johnisarrivinghome.’ Ramchand(2008)alsoarguesthat,forsemelfactivesofmotion,theremustbeanassociatedresultthat denotes the endpoint of the distance traveled by virtue of that motion, which can furthermore be expressedexplicitlybyalocativePP(e.g.(51)). (51) Katherinejumpedinthelake/onthetable. (Ramchand2008:79) However,itiseasytoconstruescenariosfortheseverbsinwhichnodistanceistraveled,aswecansee intheexamplesin(52)forEnglishandSpanish.Theideathatsemelfactivesinherentlyencodetelicity, then,isfurtherweakened. (52) a. Johnjumpedinplace. b. Elías Elías saltó jumped sin without moverse moving.refl del of-the sitio. place ‘Elíasjumpedinplace.’ 25 Section2.3. ModelingAktionsartinthesyntax Chapter2 Iconcludethatsemelfactivesareinfactpureactivityverbs,andthattheirperceivedtemporalpunc- tualityis,aswasthecasewithachievements,amatterofworldknowledgeratherthanagrammatical difference. 2.2.7 Interimsummary AsummaryofmydiscussioninthissectionisprovidedinTable2.2. Table2.2:Classiceventtypesandtheirtests Tests Telics Activities States Inxtime-phrases 3 7 7 Llevarxtiempo‘totakextime’ 3 7 7 Terminativereadingwithtardar‘totake’ 3 7 7 Scopeambiguitywithalmost 3 7 7 Scopeambiguitywithagain 3 7 7 Forxtime-phrases 7 3 3 Progressiveentailsperfect 7 3 3 Inceptivereadingwithtardar‘totake’ 7 3 3 pararde+V.inf‘stopV-ing’ 3 3 7 Habitualreadinginthepresenttense 3 3 7 Universalreadingintheperfect 7 7 3 Ambiguityinthefuture 7 7 3 Ambiguitywithconditionals 7 7 3 Ambiguitywithesperoque‘Ihopethat’ 7 7 3 2.3 ModelingAktionsartinthesyntax In classic Government and Binding Theory (Chomsky 1981, 1986a,b), the syntax of a transitive VP is typicallyrepresentedasin(53),withthedirectobjectascomplementofVandthesubjectin(Spec,IP), or,alternatively,asin(54),withthesubjectin(Spec,VP)undertheVP-internalsubjecthypothesis,see Fukui&Speas1986;Tikagawa1986;Kuroda1988;Koopman&Sportiche1985,1988,1991,a.o.). 26 Section2.3. ModelingAktionsartinthesyntax Chapter2 (53) IP Subject I’ I VP V Object (54) VP Subject V’ V Object The standard model assumes that verbs are stored in the lexicon with syntactico-semantic relevant informationwithrespecttotheircategorialclassification(i.e.verb),thesyntacticargumentsitcantake (c-selection),thekindofentitythattheargumentsneedtodenote(s-selection),thethematicrolesthat it assigns and the arguments that it assigns them to (the-grid). All this information constitutes the lexicalentryoftheverb(e.g.(55))anditdeterminestheD-structureoftheVP,insofarastherearewell- formednessconditionsthatsyntaxmustobeyforasyntacticrepresentationtobelicitwhichmustbe readoffthelexicalentryoftheverb(e.g.theProjectionPrincipleandtheTheta-criterion). 12 (55) Lexicalentryofeat: category c-selection s-selection -grid -roles V NP entity x Agent NP entity y Theme Asisimmediatelyapparent,thismodelremainssilentregardingAktionsart,andsodoes,forthemost part,currentstandardMinimalisttheory(Chomsky1995,2001,2008).Tobesure,alltypesofAktionsart 12 Chomsky(1981)definestheProjectionPrincipleandtheTheta-criterionasin(iii)and(iv),respectively. (iii) Projection Principle: Representations at each syntactic level (i.e., LF, and D- and S-structure) are projected from thelexicon,inthattheyobservethesubcategorizationpropertiesoflexicalitems. (FromChomsky1981:29) (iv) Thetacriterion: (i) Eachargumentbearsoneandonlyone-role (ii) Each-roleisassignedtooneandonlyoneargument. (FromChomsky1981:36) 27 Section2.3. ModelingAktionsartinthesyntax Chapter2 canbefoundintransitiveconfigurations(e.g. (56)),orinintransitiveonesforthatmatter(e.g. (57)). 13 Thesyntacticstructuresin(53)or(54),however,donotdifferentiatebetweeneventtypes,butassumea commonsyntaxforsentenceslikethosein(56).Nordolexicalentrieslikethatin(55)specifyormodel inanywaywhattheaspectualpropertiesoftheresultingpredicatewillturnouttobe. (56) a. Peterrespectedthelaw. (State) b. Marypushedthecart. (Activity) c. Charlesatethecake. (Accomplishment) d. Sarahtouchedthewall. (Achievement) (57) a. Dinosaursexisted. (State) b. Marysang. (Activity) c. Charlesarrived. (Achievement) The strategies of theoreticians who have done research on the grammatical properties of Aktionsart couldbedividedintofourmaingroupsasin(58),dependingonwhethertheytakethelexiconorthe syntax to be the locus of aspectual meaning (a) and b) vs. c) and d)) and whether such calculation is done via meaningful grammatical features or in a compositional, structural fashion (a) and c) vs. b) andd)). (58) MaintypesofaccountsforAktionsart. a. Aspectual features in the lexicon: verbs carry aspectual features as part of their lexical entries(Smith1991;Olsen1994,1997;Kearns2000,a.o.) b. Lexical structure: the lexical entry of a verb contains aspectually-meaningful structural configurations(theresemblanceofthosestructurestosyntax-propervaryingwithinpro- posals)(Grimshaw1990;Pustejovsky1991;Hale&Keyser1993;Travis2010,a.o.) c. Aspectual features in the syntax, i.e. in functional heads with specific aspectual values (Harley1995;Marantz1999;Cuervo2003;Lin2004;Folli&Harley2005,a.o.) d. Syntaxproper:syntaxbuildsaspectually-meaningfulsyntacticconfigurations(Borer2005b; Ramchand2008;Acedo-Matellán2016,a.o.). The classification of the works in (58) is not meant to be absolute and categorical, as many of these Aktionsartareactuallyhybridtosomeextent,buttheyarenonethelessrepresentativeofthesetypesof approachestoeventstructure.Sincemymainconcernispresentingthecaseforastructuralapproach 13 Accomplishmentscannotbefoundinintransitiveconfigurationssince,asIdiscussedinSection2.2.5.1,theyareparasitic onadirectobject:Anintransitive‘accomplishment’verbisanactivity. 28 Section2.3. ModelingAktionsartinthesyntax Chapter2 toAktionsart,inparticularoneframedinsyntacticterms,Iwillfocusmyliteratureoverviewonthose kindsofapproaches,fromthelexicalistonestothecurrentsyntacticones. 2.3.1 Originsofverbaldecomposition Theideathatverbsarenotgrammaticalatoms,butrather,thattheyarecomposedofstructure,how- ever characterized, has been present since the beginnings of the research on event structure. The earliest work on verbal decomposition, as far as I am aware of, comes from Generative Semantics, whichproposedthatthedeepsyntacticstructureofaverbwascomposedbysmallerabstractseman- tic components. Particularly famous is McCawley’s decomposition of the verb to kill. According to him,thedeepstructureofxkillsywouldbeasin(59).Grammaticaltransformationswouldthenrear- rangethiselementsintoawell-formedsyntacticconstituentthatwouldlicensethelexicalinsertion ofkillattherelevantpointofthederivation. (59) S CAUSE x S BECOME S NOT S ALIVE y InspiredbytheworkonGenerativeSemantics(aswellasinMontague’slogicsystem), Dowty(1979) proposesanaspectcalculussystemtoderivetheclassiceventtypes.Hearguesthatstatesarethebasic type of predicate, and different aspectual operators derive the more complex event types. For him, the aspectual operators operate over propositions, and so the resulting verbs are actually complex sentencesindisguise. Dowty’s operators are presented in (60), simplifying somewhat for expository purposes. BE- COMEtakesapropositionanddenotesatransition(thepropositiondoesnotholdimmediatelybefore timetbutholdsattimet).TheoperatorCAUSEintroducesacausativerelationbetweentwoproposi- tions,followingacounterfactualanalysisofcausation(Lewis1973).Finally,DOtakesanindividualand apropositionandsaysthatsuchpropositionisdirectlycontrolledbytheagent.Thelogicalstructures oftheclassiceventtypesinDowty’ssystemareprovidedin(61). (60) a. BECOMEistrueattiffistrueattandfalseatt–1. (Dowty1979:76) 29 Section2.3. ModelingAktionsartinthesyntax Chapter2 b. [CAUSE ]istrueiffisacausalfactorfor (AdaptedfromDowty1979:108) c. [DO(,)$^undertheunmediatedcontroloftheagent() ] (whereisanindividualandisasentence) (AdaptedfromDowty1979:118) (61) a. States:Nooperators b. Activities: S1 [DOxS2] c. Accomplishments: S1 [ S2 [DOxS3]CAUSE S4 [BECOMES5]] d. Achievements: S1 [BECOMES2] Pustejovsky(1991)continuesthetraditionofrepresentingaspectualmeaninginsemanticstructures. He assumes a model of the lexicon-syntax interface where verbs are semantically specified with re- spect not only to event structure, but also to subcategorization frames and argument structure (see also Grimshaw 1990; Williams 1981; Chomsky 1981). Like Dowty, he takes states to be the most basic eventtype.Foractivities(hecallsthem‘processes’),heisinspiredinDowty’ssubintervalaccount(see sections2.2.1and2.2.2)andmodelsthemasatemporalsequenceofevents.Fortelics,heisalsoinspired inDowtyinrepresentingthemas‘transitions’fromthenon-existenceofaneventualitytotheexis- tenceofsucheventuality.Iillustratehislexical-conceptualstructuresin(62),takenfromPustejovsky (1991:40). (62) a. State(S):asingleevent,whichisevaluatedrelativetonootherevent Examples: besick,love,know Structuralrepresentation: S e b. Process(P):asequenceofeventsidentifyingthesamesemanticexpression Examples: run,push,drag Structuralrepresentation: P e 1 ...e n c. Transition (T): an event identifying a semantic expression, which is evaluated relative to itsopposition Examples: give,open,build,destroy Structuralrepresentation(whereEisavariableforanyeventtype): T :E 1 E 2 30 Section2.3. ModelingAktionsartinthesyntax Chapter2 2.3.2 L-syntax: Hale&Keyser(1993,2002) 2.3.2.1 Anoverviewofthemodel Intheirseminalwork(startingin1993andcrystallizingintheir2002book),Hale&Keyseraputforth thehypothesisthatargumentstructure,derivationalmorphologyandthethematicinterpretationof argumentsobeyandcanbederivedbysyntacticprinciples. Althoughtheirmainconcernisderiving patternsofargumentstructure,andnoteventtypes per se(especiallyintheir2002book),theirwork nonetheless lays the foundation for a study of verbal syntax as the determining source of aspectual andthematicinterpretation. Hale&KeyserassumeaL(exical)-syntaxwiththelexicalcategoriesV,N,A,Pasheadsinitsinven- tory. Such heads, they argue, project distinct argument structure configurations that are defined in syntacticterms,asin(63). (63) Thefundamentalrelationsofargumentstructure(Hale&Keyser2002:12) a. Head-complement: If X isthe complementofahead H,then X istheuniquesisterof H (X andHc-commandoneanother). b. Specifier-head: IfX isthespecifierofaheadH,andifP 1 isthefirstprojectionofH(i.e. H’, necessarilynon-vacuous),thenX istheuniquesisterofP 1 . Giventheserelations,HaleandKeyserproposethattherearefourpossiblelexicalstructures,shown in(64). Wecanhaveheadsthatonlytakeacomplement(64a),headsthattakebothacomplementand aspecifier(64b),headsthatrequireaspecifierbutcannotprojectit,soitneedstobethecomplement ofanotherheadthatcanprojectaspecifier(Head*in(64c)).Afinaloptionisaheadthattakesneithera complementnoraspecifier(64d).Languagesmayvarycrosslinguisticallywithrespecttowhatlexical categoryprojectstheseconfigurations, butthetypicalpatterninEnglishis, fortheexamplesin(64): a)V;b)P;c)A;d)N. (64) Thestructuraltypesoflexicalargumentstructure(Hale&Keyser2002:13) a. Head Head Comp b. Head Spec Head Head Comp 31 Section2.3. ModelingAktionsartinthesyntax Chapter2 c. Head* Spec Head* Head* Comp d. Head Hale&Keyser(2002)showhowmanyargumentstructurepatternscanbeexplainedresortingtotheir structuraltypes.Theytakeanintransitivepairofverbslikeclearandlaughin(65)forillustration. (65) a. Thescreencleared. b. Thechildrenlaughed. Hale and Keyser argue that these verbs have a different l-syntax. The first one is derived from an adjectiveA(clear),whichrequiresaspecifierviaaseparatehead,inthiscaseaverbalhost.Thesubject inthespecifierofVisthescreen.Theadjectiveisverbalizedbyconflation,amorphologicalprocessthat copiesthephonologicalmatrixofaheadcomplementintotheheadthatgovernsit(V,inthiscase).The secondone, illustratedin(66b), isderivedfromalexicalN(laugh), whichhasthemonadicstructure of (64d), and is selected by a specifierless V. Note that V does not project a specifier in itself (unless requiredbyitscomplement,asin(64c)andexample(66a)),i.e.ithasthel-structurein(64a). (66) a. V DP thescreen V V A clear (Hale&Keyser2002:16) b. V V N laugh (Hale&Keyser2002:15) Thesedistinctstructuresexplaintheasymmetriesintransitivizationpatterns: asshownin(67), clear maytransitivize(e.g. (67a))butlaughcannot(e.g. (67b)). Transitivization,Hale&Keyser(2002)argue, is derived by a matrix V (V 1 ) that takes another V (V 2 ) as a complement, introduces a subject (truly externaltol-syntax,intherelevantsense)andassignsaccusativecasetothesubjectofitscomplement V 2 (i.e.theinternalargument)(e.g.(68)). 32 Section2.3. ModelingAktionsartinthesyntax Chapter2 (67) a. Iclearedthescreen. b. *Theclownlaughedthechildren. (68) a. V 1 V 1 V 2 DP thescreen V 2 V 2 A clear (Hale&Keyser2002:17) b. V 1 V 1 V 2 V 2 N laugh (Hale&Keyser2002:15) HaleandKeyserarguethatdenominallocativeverbslikeshelvehaveasimilarstructureasdeadjectival verbs,onlythatinthiscasethespecifierisprovidedbyalocativePofterminalcoincidence,denotinga changeoflocationofaFigure(inthespecifierofP)toaGround(theNcomplementofP).Thelexical headisverbalizedthankstoaVthattakesPasacomplement,withwhichthelexicalNheadconflates, alongwithP.Iprovideanexamplein(69). (69) a. Petershelvedthebooks. b. V V P DP thebook P P N shelf (Hale&Keyser2002:31) Hale & Keyser (1993) make the crucial claim that the structural configurations that lexical heads are infacteventstructureconfigurations,andthattheinterpretationoftheargumentsoftheseheadsare 33 Section2.3. ModelingAktionsartinthesyntax Chapter2 entailmentsfromtheeventstructuretheyappearinandnothingmore.Eachoftheheadsinthelexical syntaxdenoteaseparatesubevent,andwhentwoheadsarecombinedinthesyntaxtheinterpretation betweenthetwosubeventsisoneof implication. Takeasentencesuchas(70a), withthestructurein (70b)(fromHale&Keyser1993:72). (70) a. Thecookthinnedthegravy. b. VP NP (Thecook) V’ V VP NP (Thegravy) V’ V AP thin Giventhat thinisadeadjectivalverb, thelexicalheadisAanditisacomplementofV.Thecomple- mentationstructureisdepictedin(71a). Hale&Keyser(1993)assumethatlexicalheadsareassociated withadistinctnotionaltype: thelexicalheadAdenotesastateandthelexicalheadVanevent. The resultingsemanticrelationisthatofaneventimplicatingastate,i.e.achangeofstate(see(71b),where ‘!’denotestheimplicationrelation,eaneventandsastate). (71) a. V’ V AP thin b. e!s (Hale&Keyser1993:72) Theinternalargument,the gravy,isinterpretedasatheme(anentityundergoingchange),theauthors argue,byvirtueofhavingasyntacticallyunambiguousrelationwiththecomplementationstructure in(71a)(itisinthespecifierofthehigherprojection,asyoucanseein(72a)).Thethematicrelationofthe internalargumentwithrespecttothecomplex(semantic)eventstructuredenotedbythe(syntactic) complementationstructureisrepresentedas‘>’,nbeingthetypeofnouns(see(72b)). 34 Section2.3. ModelingAktionsartinthesyntax Chapter2 (72) a. VP NP (Thegravy) V’ V AP thin b. n>(e!s) (Hale&Keyser1993:73) The complementation structure of the higher and lower VPs (represented in (73a)), in turn, has the denotation in (73b), where the relationship is again one of implication, understood in this case as a causalrelationbetweentwoevents. (73) a. V’ V VP b. e 1 !e 2 (Hale&Keyser1993:69) Theexternalargumentin(Spec,VP),byitsunambiguoussyntacticrelationwiththelowerstructure, isinterpretedasanagent:itistheentityimplicatedinthecausingeventuality. (74) a. VP NP (Thecook) V’ V VP b. n>e 1 !e 2 Withrespecttostativeverbs,Hale&Keyser(1997,2002)hypothesizethattheyfallintwotypes:thefirst typeiscomprisedofdenominalverbslikerespect(e.g. (75)),whosestativityisderivedfromaterminal 35 Section2.3. ModelingAktionsartinthesyntax Chapter2 coincidence P (eg. (75e)). 14 The fact that sometimes these verbs may have event-like properties, for them,stemsfromthefactthatthereisverbalstructureontopofP. 15 (75) a. Johnfearedthetruth. b. Johnknewthetruth. c. Johnadmiredthetruth. d. Johnlikedthetruth. e. Johnrespectedthetruth. (Hale&Keyser2002:208) 14 ThesamestativestructurecanbefoundwithanovertPinsentenceslike(v),whichhasthestructurein(vi). (v) With[AnnaninBaghdad],wecanrelax. (vi) P DP Annan P P in N Baghdad (Hale&Keyser2002:220) 15 However,theteststhatHale&Keyser(2002)adduceinfavoroftheirevent-likepropertiesarefarfromconclusive.They argue that they may appear in the imperative (e.g. (viia)), in the progressive (e.g. (viib)) and accept in x time phrases (e.g. (viic)). IalreadymentionedinSection2.2.4thatneithertheimperativenortheprogressivearevalideventivitytests(see alsoSections3.2.2.1and5.5ofthisdissertation).Theinxtimephrasein(viic),moreover,makesreferencetothebeginningof thestate,ratherachange-of-state. Notethatifitwerereallyachange-of-state/telicreading,itshouldbepossibletohave a reading in the progressive in which that change-of-state were about to take place. But that is not the case: a sentence like(viii)canonlymeanthatthetroopsalreadyrespectthecommander,notthattheyareaboutto. (vii) a. Respectyourparents. b. Heislikinghisnewjob. c. Thetroopsrespectedtheirnewcommanderinminutes. (Hale&Keyser2002:210) (viii) Thetroopsarerespectingtheirnewcommander. 36 Section2.3. ModelingAktionsartinthesyntax Chapter2 (76) V V P DP thetruth P P N respect (Hale&Keyser2002:210) Thesecondtypearecopula-likeverbswithameasurephrasecomplement(e.g.(77)).Withoutgivingan explicitsyntacticanalysis,theauthorsclaimthattheseverbsareinherentlystative,beingcopula-like (infact,alltheverbsin(77)canbesubstitutedwiththecopula be)areinherentlystative,i.e. stativity inthesecasesisnotderivedbyacentralcoincidenceP,butbyasubsetoftheVcategory. (77) a. Thathousecostsfiftythousanddollars. b. Thisbullweightsoneton. c. Twoandtwoequalsfour. d. Threebookscomprisetheentirecollection. (Hale&Keyser2002:214) Stativity, then, can be derived in three different ways: either a P with a central coincidence value, a subsetoftheVcategory(thecopula-likeVheads)andtheextendedprojectionoftheAinitsstative use,asshownin(78). (78) a. Wefound[theskyclear]. b. DP thesky A clear (Hale&Keyser2005:24) Observingthatthesameargumentstructureconfigurations,asdefinedin(63),mayhavedifferentas- pectualmeanings(forinstance),Hale&Keyser(2002,2005)reachtheconclusionthatargumentstruc- tureandaspectarebylargeindependentofeachother.Intheirownwords, aspectisorthogonaltoargumentstructure. Wheneverwedealwithquestionsofinter- faceandinteractioninthisdomain,weobservethatargumentstructureisforthemost part autonomous. Its properties and characteristics are strictly local, being defined in 37 Section2.3. ModelingAktionsartinthesyntax Chapter2 termsofthestructuralrelationsofcomplementandspecifier. Tobesure,anyargument structureconfigurationassociatedwithanactualpredicateinsententialsyntaxwillbein- terpretedintermsofoneoranotheraspectualtype(achievement,accomplishment,etc.) anditsargumentswillbeassociatedwithoneoranotheraspectualrole(measure, path, terminus,etc.(Tenny1994)).Butargumentstructureisadistinctandseparatecomponent ofgrammar. (Hale&Keyser2002:225) 2.3.2.2 Criticalassessment Hale&Keyser(1993,2002)isaseminalworkinthestudyofargumentstructure,andagiantstepto- wards neo-constructionism. Although their model is assumed to be one of lexical structure within thelexicon,andhencepriortosyntaxproper,theynonethelessshowhowphenomenasuchasgram- maticalcategory,derivationalmorphology,aspectualinterpretationandthematicinterpretationscan be derived by syntactically-constrained argument structure configurations. This is the crucial dif- ferencewithotherlexicalistdecompositionalmodels(Dowty1979;Grimshaw1990;Pustejovsky1991) whoseproposedstructuresdonotconformtothesamecombinatorialandwell-formednessprinci- plesasthoseofsyntaxproper,anditisinthissensethattheirworkisadoortoneo-constructionist accounts. AsIreadtheirseriesofwork,however,therearetwoHaleandKeyserswithaltogetherdifferent views regarding the aspectual and thematic interpretation of the structures. Their first article from 1993, as I commented in the previous section, attacked-theory and claimed that it has no place in thetheory. Whatwecall‘-roles’areinterpretiveentailmentsfromtheeventstructureofthepredi- cate,syntacticallybuilt,anditisalsothesyntacticpositionoftheargumentswhichdeterminestheir relationwithrespecttothateventstructure. Doingawaywith-theoryandconstraintssuchasthe -criterion, and deriving the notional-roles from independently motivated principles, is indeed a welcomeconsequenceforthetheory. 16 Hale & Keyser (1993), as we also saw, associated grammatical category with the semantic types from which articulated event structure is built (A denotes states, V denotes events, and so on). This canhardlybecorrect,becauseaspectualtypescutacrosscategories,astheythemselvesacknowledge in their 2002 book: 17 the V category may denote events and states (e.g. (79) and (80)), but also the 16 A precursor of the idea that-roles are not relevant for the syntax and that they should be associated to aspectual meaning (in the lexicon) can be found in Tenny (1992). In her words, ‘the syntax proper does not need to "see" thematic roles.Itonly"sees"certainsyntactic/aspectualstructuresthethematicrolesareassociatedwith.’(1992:2) 17 ThisobservationactuallygoesasfarasDowty(1979),whonotedthat‘inadditiontoverbs,adjectivesandnounsalsosplit intostativeandnon-stativecategories,accordingtowhethertheprogressivecanbeusedwhentheyappearaspredicative 38 Section2.3. ModelingAktionsartinthesyntax Chapter2 othercorepredicationalcategories,i.e. adjectives(e.g. (81)and(82))andnouns(e.g. (83)and(84)). This makestheauthorsreachtheconclusion(contratheiroriginal1993work,althoughtheydonotmake theconnectionexplicit)thataspectandargumentstructurearemostlyindependentofoneanother. (79) Stativeverbs a. Thebookcostsfivedollars. b. Peterfearshisparents. (80) Eventiveverbs a. Maryjumpedacrosstheriver. b. Johnchurnedthebutter. (81) Stativeadjectives a. Hansishandsome. b. Aweeklynewspaper. (82) Eventiveadjectives a. Marywasbeingunkind. b. Carmenisusuallyquitegenerous. (83) Stativenouns a. Thelovetowardsyourneighbor. b. Yourfearofgunsiswarranted. (84) Eventivenouns a. Theviolentfightbetweenthegangs. b. Mycousin’slongweddingparty. Ifindthistobeamostunfortunateconclusion.AsIseeit,makingtheclaimthataspectandargument structureareindependentofeachotherseriouslyjeopardizestheiroriginalproposalregardingevent structureconstrualandwithittheirargumentagainst-theory. Thatis,ifargumentstructuredoes not correlate with aspectual meaning, where does aspect come from? If it is not in the syntax and notinthelexicon,whereisit? Iseventstructure(andwithit‘thematicstructure’,intheirsense)con- struedinanindependentgrammaticalmodule? Ifso,thestructuralpropertiesofAktionsartbuilding, whichever they turn out to be, would no longer correlate, presumably, with those of syntax (e.g. a feature-based system à la Smith 1991). This is the worst theory, not only because it would assume a separategrammaticalmodule,undesirablebyOccam’srazor,becauseittakesusbacktomodelssuch asGenerativeSemantics,whosehighlysemanticallydecomposedstructureswerenotgrammatically justified. Another option is that aspect were just part of the conceptual meaning of the root. If so, then wewouldnotneedaseparategrammaticalmoduleandaspectwouldbeinterpretedmerelyfromthe lexicalsemanticsoftheverb.Butargumentstructureconfigurationsdoshowimportantcorrelations withaspect.IalreadydiscussedthedeterminingeffectsofsyntacticcomplementssuchasDPsandPPs inSection2.2.5.1,butevenwiththeargumentstructuretypesthatHale&Keyser(2002)discussthere adjectives and predicate nominals. Cf. John is being careful vs. *John is being tall, John is being a hero vs. John is being a grandfather.’(130,ft.6) 39 Section2.3. ModelingAktionsartinthesyntax Chapter2 areaspectualpatternsthatcanbedrawnfromthem.Lookattheirpatternsin(85)and(86),fromHale &Keyser(2002:1)(andseealso(65)and(67)). (85) a. Thepotbroke. b. Ibrokethepot. (Unaccusative-Transitive) (86) a. Theenginecoughed. b. *Icoughedtheengine. (Unergative) As Hale & Keyser (2002) discussed, the verbs break in (85) and cough in (85) pattern differently with respecttocausativizationbecausecoughisanunergativeverbanditalreadyhasanexternalargument, whereasbreakisunaccusativesoanexternalargumentcanbeaddedontopoftheinternalargument. Butnotethat,inaddition,theunergativeverbsthatHaleandKeyserdiscussseemtobeconsistently activitiesaspectually(i.e.atelic,e.g.(87a)),whereastheunaccusativeverbstheydiscussarechange-of- stateverbs(i.e.telic,e.g.(87b)). 18 (87) a. cough,dance,sneeze,laugh... (Atelic) b. break,thin,clear,open... (Telic) Hale&Keyser’s(2002)claimthataspectandargumentstructureareindependentmissesthesecrucial generalizations. Their original 1993 work is much more insightful in this respect, as I have argued. Someoftheir1993claims,however,needtobereconsideredandrectified,suchastheassociationof lexicalcategorywithadistinctsemantictype,whichwesawmakesthewrongpredictions. 2.3.3 TheXS-model: Borer(2005b) 2.3.3.1 Anoverviewofthemodel Borer (2005b) puts forth a grammatical framework that posits the syntax as the sole source of as- pectual meaning. 19 The model is neo-constructionist in that the grammatically-relevant properties of theverbandtheverbphrase(i.e. argumentstructure,eventstructure,thematicinterpretation...) are notbuiltfromthepropertiesofthelexicalroot,aslexicalistapproaches(alsoknownasprojectionistin Rappaport-Hovav&Levin1998orendo-skeletalinBorer2005a)wouldhaveit,butrather,theyemerge from the syntax, which builds structure independently from the lexical root. Borer metaphorically 18 SeeBorer(2005b)onthispoint,whoarguesindetailthatunaccusativitycorrelateswithtelicity. 19 Borer(2005b)ispartofatrilogythatdevelopsaneo-constructionistmodel(theXS-model)forgrammaticalstructure. Thetrilogyalsostudiesthenominaldomain(Borer2005a)aswellasword-formation(Borer2013). 40 Section2.3. ModelingAktionsartinthesyntax Chapter2 names her approach as the Exo-Skeletal (XS) model, capturing the idea that, as with exo-skeletal or- ganisms, itisthe‘propertiesofthe‘outside’, largerstructurewhichultimatelydeterminetheoverall ‘shape’ofwhatiswithin,ratherthantheotherwayaround’(Borer2005a:15). Borerassumesanextremelyimpoverishedlexicon—orencyclopedia,astheauthorcallsit—composed ofitemsknownas listemeswhichcontainphonologicalinformationandconceptualmeaningthatis notgrammaticallyrelevant. Listemes,then,donothaveanyinformationregardingevent,argument or thematic structure, or grammatical category for that matter. Borer gives support for her theory with data as in (88): although the word siren is usually a noun, it can nonetheless function as a verb inamultiplicityofargumentstructureconfigurations. Alexicalistmodelwouldhavetoassumethat theverbsirenhasmultipleentriesspecifyingtheselegitimateargumentstructureconfigurations,and there would also need to be a separate entry for the nominal use of siren. The same would apply to allthelexicalitemsthatshowedthesealternations,whichresultsinaplethoraoflexicalentriesinthe lexicon—andtheredundancyandlearnabilityissuesassociatedwithit. (88) a. Thefirestationssirenedthroughouttheraid. b. Thefactorysirenedmiddayandeveryonestoppedforlunch. c. ThepolicesirenedthePorschetoastop. d. Thepolicecarsireneduptotheaccident. e. Thepolicecarsirenedthedaylightsoutofme. (FromClark&Clark1979,apudBorer 2005a:8) Inaneo-constructionistmodel,instead,thepatternin(88)isderivedbyresortingtotheverylimited number of possible argument structure configurations that the combinatorial properties of syntax allows. A listeme like siren, given that it does not have any grammatical information that imposes restrictions on its syntactic realization, can, in principle, be inserted under any of those structures, effectively deriving patterns like those in (88). It can also be inserted under nominal structure, in whichincasewewouldhavethenounsiren(asintheannoyingpolicesiren,forinstance). IntheXS-model,syntaxiscomposedoffunctionalprojectionswhoseheadcontaincategorialin- formationaswellasopenvaluesthatneedtobeassignedrange(theopenvalueisrepresentedashei andthecategorylabelasasubscripttoitsright). RangeassignmentcanbedoneindirectlybySpec- Head agreement (89a), provided the syntactic phrase in the specifier has the relevant grammatical properties,ordirectlybyeitherafunctionalmorpheme(f-morph)(89b)oranabstractfeaturethatcan begivenaphonologicalrepresentation(89c).Rangeassignmentisrepresentedbyasuperscriptin(89). Theburdenofthesyntacticderivation,then,isplacedinfunctionalmaterialandthestructuralrela- 41 Section2.3. ModelingAktionsartinthesyntax Chapter2 tionstheyenterwitheachother:listemesgiveconceptualcontenttotheresultingsyntacticstructures, butdonotgiverisetothemnordotheyimposeconditionsonthem,contralexicalistaccounts. (89) a. [ XP YP 1 [ X’ he 1 i X ]] b. [ XP Spec[ X’ f-morph 2 he 2 i X ]] c. [ XP Spec[ X’ hheadfeature 3 ihe 3 i X ]] Borerproposesthattelicstructureislicensed(inEnglishandsimilarlanguages)viaaquantityDP.A quantity DP, in Borer’s terms, is a DP that projects a quantity phrase #P, whose job in the nominal structureistoassignquantitytomassnounsorcountnouns,whoseheadisassignedrangebyfunc- tionalitemssuchasthree,muchorfew.IfaDPlacksa#P(essentiallybarepluralsandbaremassnouns), thenominalpredicatewillbehomogeneous,i.e. cumulativeanddivisive. Forinstance,thehomoge- neousnominalpredicatesalt(90a)iscumulativeinthatsaltplussaltisstillsalt,anddivisiveinthata subset of salt is still salt as well. Quantity predicates (e.g. (90b)) are not homogeneous because they donothavethoseproperties: threeflowersplusthreeflowersisnotthreeflowers,butsix,justlikea subsetofmanybooks(inaparticularcontext)maynolongercountasmanybooks. 20 (90) a. salt,vodka,apples,books... Non-quantityDPs b. Threeflowers,theglass,manybooks,somewater... QuantityDPs Establishingaparallelismbetweenthenominalandverbaldomain,Borerarguesthattelicityarisesvia aprojectionshelabelsAsp Q P,whichisassignedrangebyaquantityDP.Justlike#Pdoesinthenominal domain, Asp Q P imposes quantity structure, which for Borer is the hallmark of telicity. Given that quantityisapropertyofaneventargument,BorerpositsaprojectionshelabelsEP(EventPhrase)that mapsthequantitypredicatetoanevent.Theargumentin(Spec,Asp Q P)isinterpretedasasubject-of- quantitys-o-q(roughly,subjectofstructuredchange).BoreralsoassumesthatTPislocatedbetween EPandAsp Q P. Ifthetelicpredicateistransitive,therewillbeanargumentin(Spec,EP)distinctfromtheonein (Spec,Asp Q P) which will be interpreted as anoriginator by default. It will also assign range to the 20 ThesenotionsareformalizedinBorer(2005a)asfollows: (ix) a. Quantity PisquantityiffPisnothomogeneous b. PishomogeneousiffPiscumulativeanddivisive i. Pisdivisiveiff8x[P(X)!9y(P(y)^y<x)]^8x,y[P(x)^P(y)^y<x!P(x-y)] ii. Piscumulativeiff8x[P(X)^P(y)!P(x[y)] (FromBorer2005a:127) 42 Section2.3. ModelingAktionsartinthesyntax Chapter2 open value of EP. 21 If the telic predicate is intransitive, the subject in (Spec,Asp Q P) will also raise to (Spec,EP) but it will not receive an additional role given that it already has one (s-o-q). TP, in this model,issandwichedbetweenEPandTP,anditmustbefilledfornominativecaseassignmentpur- poses. I provide examples of transitive and intransitive telic structures in (91) and (92), respectively, alongwiththeirNeo-Davidsonianeventsemanticsrepresentations. 22 (91) Transitivetelic a. Annareadthebook. 21 BorerarguesthatassigningrangetotheopenvalueofEPamountstobindingtheeventargumentexistentially,andas suchitcanonlybedonebyanelementwithexistentialforce,likealocativeoraspecificDP. 22 IntheXS-model,achievementverbsareassumedtobespecialfromothertelicverbsinthattheyrequiretheprojection Asp Q P, i.e. they arise from listemes that are specified in the lexicon for a Asp Q P, unlike the rest of listemes who are not (rememberthatthissystemclaimsthatlistemeshavenogrammaticalinformation).Borerclaimsthattherearetwotypes of achievements: one type is presentational, and for this one the open value of Asp Q P is an open locative value, which forcestheDPinthespecifiertobeweak:thesecanthenbeunderstoodasexistentialconstructionsofsorts(xa).Theopen locativevalueisassigneddirectlybyaheadfeaturehloc/9i.Thelexicalentryisgivenin(xia),whereL-Dstandsforlexical domain.Theothertypeofachieventsisliketheothertelicstructureshithertodiscussed:Asp Q needstobeassignedrange via a quantity DP in (Spec, Asp Q P) and hence we cannot have a non-quantity argument (e.g. (xb), and see the relevant listemespecificationin(xib)). (x) a. Marynoticed(*the)inkonhersleeve. b. Thebabyfinished*(the)milk. (xi) a. [ AspQ he loc/9 i # [ L-D notice]] b. [ AspQ hei # [ L-D finish]] (FromBorer2005b:337) 43 Section2.3. ModelingAktionsartinthesyntax Chapter2 b. EP Anna E E’ TP Anna T T Asp Q P thebook Asp Q Asp Q VP read c. 9e[quantity(e)&originator(Anna,e)&subject-of-quantity(thebook,e)&read(e)](From Borer2005b:85) (92) Intransitivetelic(unaccusative) a. Theflowerwilted. b. EP theflower E E’ TP theflower T’ T Asp Q P theflower Asp Q Asp Q VP wilt (Borer2005b:84) c. 9e[quantity(e)&subject-of-quantity(theflowers,e)&wilt(e)] (FromBorer2005b:84) 44 Section2.3. ModelingAktionsartinthesyntax Chapter2 With respect to atelic predicates, Borer argues that they are characterized by the absence of quan- tity structure in the syntax (i.e. they are interpreted as homogeneous by default). If the predicate is intransitive, the single argument is generated in (Spec,TP) and then raised to (Spec,EP) where it is interpreted as an originator. If transitive, the internal argument is licensed at by a functional shell shelabelsF s Pthathasnosemanticproperties,butmerelylicensescase. Assuch,itcanhostquantity and non-quantity DPs in its specifier with no aspectual effects. The argument in (Spec,F s P) is inter- pretedasadefaultparticipant. Theatelictransitiveandintransitivestructuresandtheirsemantic interpretationaregivenin(93)and(94). (93) Transitiveatelic a. Kimpushedthecart. b. EP Kim E’ E TP Kim T’ T F s P thecart F’ F VP push c. 9e[activity(e)&originator(Kim,e)&participant(thecart,e)&push(e)] (FromBorer 2005b:109) (94) Intransitiveatelic(unergative) a. Peterworked. 45 Section2.3. ModelingAktionsartinthesyntax Chapter2 b. EP Peter E E’ TP Peter T’ T VP work c. 9e[activity(e)&originator(Peter,e)&walk(e)] Forstatives,Borerremainsmorespeculative.Shesuggeststhatstativepredicatesarebuiltbyaprojec- tionshecallsSP(StatePhrase). UnlikeAsp Q PandEP,whichsheassumesverbalizewhateverlisteme isundertheirscope(rememberthatlistemesareuncategorizedinhersystem), SPpreemptsverbal- ization.Thiscapturesthefactthatstativepredicatescanalsobeadjectival.Theargumentin(Spec,SP) isinterpretedasasubject-of-state.BelowSP,Borerargues,theremustbeanargumentintroducing projectionsuchasF s PorAsp Q P,insofarastherearetransitivestates(shedoesnotmentionwhatrole wouldtheinternalargumentofstateshave:IassumesomethinglikethedefaultparticipantofF s Ps). Herproposedstructureisasin(95b),andthesemanticinterpretationwouldbesomethinglike(95c): Borerdoesnotgiveaprecisesemanticsforstatives,so(95c)ispartlyaguessonmypart. (95) Transitiveatelic a. JohnknowsPaul. 46 Section2.3. ModelingAktionsartinthesyntax Chapter2 b. EP John E E’ TP John T’ T SP John S’ S F s P/Asp Q P Paul F’/Asp Q ’ F/Asp Q VP know c. 9e[state(e)&subject-of-state(John,e)&Defaultparticipant(Paul,e)&know(e)] 2.3.3.2 Criticalassessment As we have seen, the XS-model departs radically from lexicalist models in that it contends that the lexicon does not impose well-formedness conditions on syntactic structure, nor does it determine thestructureinanyway: rather,thecomputationalsystembuildsstructureindependentlyoflexical items, which merely act as modifiers of that structure. A welcome theoretical consequence of this modelisthatthelexiconiseliminatedasacomputationalmodule,therebysimplifyingthemodelof grammarsignificantly,withnon-trivialimplicationsforlanguageacquisition. Themainobviousproblemofthisviewofgrammar,asmanyhaveobseved,Borerherselfincluded (Borer1994,2005b;Sorace2004;Folli&Harley2005,a.o.),isthatthesystemovergeneratesmassively. Forinstance,assumingalexicaldomainwiththelistemesman,eat,andcakeasinputsforatransitive verbal structure, all the sentences in (96) are in principle equally well-formed outputs. Syntactically theyarewell-formedindeed,butwhereas(96a)isperfectlynatural,(96b)and(96c)clearlysoundde- viant.Borer’ssuggestionisthatsentencescannotbeeasilyinterpretedbecauseofaclashbetweenthe 47 Section2.3. ModelingAktionsartinthesyntax Chapter2 semanticinterpretationofthestructureandtheconceptualknowledgeassociatedwiththelistemes. 23 Afurthercomplicationisthat,asBoreralsoacknowledges(2005b:350–1),otherlanguagesdonotshow themorphosyntacticelasticityofEnglish,e.g.Romancelanguages. (96) a. Themanatethecake. b. ??Theeatmannedthecake. c. ??Thecakemannedtheeat. More problematically for Borer, in my view, is the existence of argument structures that are well- formedandperfectlycompatiblewiththeencyclopediccontentofthelistemestheyappearwith,yet they are completely ungrammatical. That is, why should sentences like (97) be ungrammatical? It doesnotseemhardtointerpret(97a)asalexicalcausativeinwhichMarytookPetertoBerlin,or(97b) as an unaccusative structure in which John got into a state of fright, and yet both sentences are not well-formed. (97) a. *MaryarrivedPeterinBerlin. b. *Johnfrightened.(cf.ThestormfrightenedPeter) Also,theXS-modeldoesnotassumeaone-to-onemappingfromsyntacticstructuretoeventstruc- ture, inthesensethatsyntacticcomplexitydoesnottranslateintoeventcomplexity: rather, theex- tendedverbphrase,howeverarticulated,alwaysdenotesasingleeventuality(dynamicorstative).This semanticimpoverishmentmakesithardattimestounderstandthecorrelationbetweenthesyntactic structure and the semantic interpretation. Take the originator role, for instance. In the XS-model, as discussed, the originator role is an entailment that arises for an argument in (Spec,EP) when it is nototherwiseassignedarole. Whatisnotclearishowthatentailmentcomesabout. Toillustrate, I provideanexampleofanatelicsentencein(98a)anditsXS-syntaxin(98b). Forthesemanticrepre- sentation in (98c), I have purposedly (98c), but where the originator role would be represented (as a relationbetweenanentityandtheevent),Isimplywriteeventrole. (98) a. Huilindrovethecar. b. [ EP Huilin[ F s P thecar[ VP drive]]] c. 9e[activity(e)&Eventrole(Huilin,e)&participant(thecar,e)&drive(e)] Asitturnsout,thereisnothinginthesyntacticstructureorinitssemanticrepresentationthatentails thatsucheventroleshouldbeanoriginator(andnot, say, aGoaloraLocation). Thisproblemdoes 23 Remember,tothisrespect,Chomsky’s(1957)famousexampleof Colorless green ideas sleep furiously,asemanticallynon- sensicalsentenceyetperfectlyformedinsyntacticterms. 48 Section2.3. ModelingAktionsartinthesyntax Chapter2 notariseinthemodelputforthinHale&Keyser(1993)(orRamchand2008,aswewillseeinSection 2.3.4), which posits an event decomposition where the subevent denoted by the higher projection is interpretedascausingthesubeventdenotedbythelowerprojection.Assuch,thesubjectofthehigher projectionisinterpretedasanoriginatorbyentailment.Unlessoneassumesaminimallevelofevent decomposition,then,itishardtoseehow(atleastcertain)eventrolescanbederivedfromtheaspec- tualstructureofthepredicate.Itissimilarlynotclearwhattheroleoftheobjectofstativeverbscould be,orwhattheinterpretationofthefunctionalstructuretheyappearincouldbe. Finally,BorerrunsintosimilarproblemsasHale&Keyser(1993)byassociatingstativitywithnon- verbal predication. If her stativizing SP projection preempts verbalization, how do verbal statives comeabout?Inthefollowingsection,Ipresentthefirst-phasesyntaxmodeldevelopedinRamchand (2008)who,althoughdoesnotelaborateatlengthonstativepredicates—likeBorer(2005b)—doesin- tegratebothstativeandeventiveverbalheadsinhersystem. 2.3.4 Firstphasesyntax: Ramchand(2008) 2.3.4.1 Anoverviewofthemodel Ramchand(2008)hasrecentlyproposedadecompositionalsyntacticmodelfortheVPthatshelabels firstphasesyntax.Hersystemassumesthreeargument-introducingverbalheads,universallyordered: init(itation), proc(ess)and res(ult). Eachoftheseverbalheadsdenoteasinglesub-event(uality): initand resdenoteastateandprocdenotesadynamicevent.Thecombinationoftheseverbalprojectionsinthe syntaxbuildsupcomplexeventualitiesfollowinginessencetheaccountofHale&Keyser(1993): the subeventdenotedbythehigherprojectioninahead-complementrelationissemanticallyinterpreted ascausingthesubeventdenotedbythelowerprojection,andthesubjectsofthoseprojectionsreceive aneventroleasanentailmentfromtheaspectualstructureinwhichtheyappear. ThemaximalstructureoftheVPisthenasin(99a), withthesemanticsin(99b). In(99b), eisthe macro-eventthatresultsfromthecompositionofthesubevents,‘!’isthecausingrelation,asinHale &Keyser(1993),and init’, proc’ and res’ standforthepossiblelexicalitemsthatlexicalizethoseverbal heads. 24 24 Thesemanticsin(99b)areslightlymodifiedfromRamchand(2008).Theauthorcallstherelationbetweenthesyntactic subjectoftheverbalprojectionandtheeventitdenotes‘Subject’.However,Iopttocallsuchrelationparticipant,which showstransparentlythatwearedealingwithasemanticnotion. ThankstoRoumiPanchevaforbringingupthisissueto me. 49 Section2.3. ModelingAktionsartinthesyntax Chapter2 (99) a. initP initiator init’ init procP undergoer proc’ proc resP resultee res’ res XP b. e9s 1 ,e 1 ,s 2 [e=s 1 !(e 1 !s 2 )& init’(s 1 )&participant(x,s 1 )& proc’(e 1 )&participant(x,e 1 ) &res’(s 2 )&participant(x,s 2 )] InthespiritofHale&Keyser(1993),theinterpretationoftheeventparticipantsareentailmentsfrom the event structure. The subject of procP, then, will be interpreted as an undergoer because it is a subject of a process eventuality. The subject of initP and resP, in addition, are interpreted as an ini- tiatororresultee,respectively,simplybecausetheyaresubjectoftheinitiationorresultsubevent. Again, note that initP and resP, in and of themselves, denote merely states: whether they are inter- pretedasinitiationalorresultativedependsontheireventstructurecontext. Therelevantsemantic entailmentsarestatedin(100),adaptedfromRamchand(2008:44). 25 (100) a. If9s,e[s!e],thenbydefinitionInitiation(s) b. If9s,e[e!s],thenbydefinitionResult(s) Withrespecttotherelationshipofthelexiconandthesyntax,Ramchandassumesthatlexicalentries are specified or tagged with respect to the eventive head(s) that it lexicalizes in the syntax (e.g. an unaccusativeverbofchangeofstatewillbespecifiedas[proc,res]).Ramchand(2011)adoptsalateinser- tionaccountbywhichthelexicalitemlexicalizesthehead(s)ofpre-constructedsyntacticstructures 25 Ramchandmodelsstatesanddynamiceventsaspredicatesovereventualities,asin(xii).Forthesakeofsimplicity,Irefer tostativeeventualitiesassanddynamiceventsase. (xii) a. State(e):eisastate. b. Process(e):eisaneventualitythatcontainsinternalchange. 50 Section2.3. ModelingAktionsartinthesyntax Chapter2 in accordanceto its category features. This mechanismis known in nanosyntaxas spanning, in that asinglemorphememayspanseveraladjacentsyntacticterminals(seealsoBrody2000;Starke2009; Svenonius2012a,2016). 26 LetusseehowRamchandderivesthedifferenteventtypes. Foractivities,theauthorarguesthat their minimal ingredient is procP (e.g. (101)): if the predicate is transitive, an additional initP will be involved(e.g.(102)). (101) Activities(intransitive) a. Thebarrelrolled. b. procP DP thebarrel proc’ proc <roll> XP (102) Activities(transitive) a. Caitlinpushedthecart. b. push:[init,proc] c. initP DP Caitlin init’ init <push> procP DP thecart proc’ proc <push> XP 26 In Ramchand’s (2008) original work, the integration of the lexical item is derivational: if a verb is tagged with more than one category label (e.g. proc and res), it will first remerge as res and then remerge as proc. Note that either way the root does not occupy a single privileged position at the bottom of the tree as a complement of a verbalizing head (as in lexicalistapproacheslikeL-syntaxbutalsoinconstructionistoneslikeDistributedMorphologyortheXS-model),butit merelylexicalizesdifferentheadsthroughouttheextendedverbalstructure.Iwillgointomoredetailaboutlexicalization patternsinSection4.4.1. 51 Section2.3. ModelingAktionsartinthesyntax Chapter2 Ramchandarguesthattelicityisderivedintwoways.ThefirstwayisviaaresPprojectioncomplement of proc, whichsemanticallydeliversacomplexeventualityconsistingofaprocesseventandaresult state (see also Moens & Steedman 1988; Parsons 1990). The lexical item lexicalizes both proc and res. NotethatthesecorrespondtotheviewofachievementverbsthatIdefendedinSection2.2.5.2,i.e. as verbsthatinherentlylexicalizetelicstructure. 27 Aswithactivities,theycanbeunaccusative(e.g.(103)) ortransitive,inwhichcasethereisaninitPprojectionaswell(e.g.(104)). 28 (103) Telicverbs(unaccusative) a. break:[proc i ,res i ] b. Theglassbroke. 27 Ramchand accepts the classic view that achievements are punctual events. She argues that these verbs are instances ofverbrootsthatlexicalizeallthreeeventiveheads. Sheproposesthatsuchpredicationsareinterpretedaspunctualby virtueofhavingthesamelexicalcontentlexicalizingthethreeheadsatonce.Iprovideanexamplein(xiii). (xiii) Achievements a. Maryarrived. b. arrive:[init i ,proc i ,res i ] c. initP DP Mary init’ init <arrive> procP DP <Mary> proc’ proc <arrive> resP DP <Mary> res’ res <arrive> XP 28 As for event participants, Ramchand assumes that they may occupy more than one specifier position in the extended verbalstructure,therebyallowingforcompositeeventroles.Ramchandfurtherassumes(somewhattentatively)thatthis is specified in the lexical entry, by means of coindexing of the category features (103a). Thus, in (103c), the glass is at the sametimeanUndergoerandaResulteeofthebreakingevent. 52 Section2.3. ModelingAktionsartinthesyntax Chapter2 c. procP DP theglass proc’ proc <break> resP DP <theglass> res’ res <break> XP d. e9e 1 ,s 1 [e = e 1 ! s 1 & break(e 1 ) &participant(the glass,e 1 ) & break(s 1 ) &participant(the glass,s 1 )] (104) Telicverbs(transitive) a. Johnbroketheglass. b. initP DP John init’ init ? procP DP theglass proc’ proc <break> resP DP <theglass> res’ res <break> XP c. e9s 1 ,e 1 ,s 2 [e=s 1 !(e 1 !s 2 )&participant(John,s 1 )&break(e 1 )&participant(theglass,e 1 ) &break(s 2 )&participant(theglass,s 2 )] 53 Section2.3. ModelingAktionsartinthesyntax Chapter2 Thesecondwaytoderivetelicityistotakeanactivityverbandaddacomplementthatdelimitsit,i.e. anaccomplishmentconfiguration.Ramchandcallsthesecomplementspathsandarguestheyarethe syntacticcomplementsofproc. ThesePathscanbeDPs(105a)orDPs(106a). Theirsyntaxisprovided in(105b)and(106b)andtheirsemanticsin(105c)and(106c),respectively. (105) Accomplishments(DPPath) a. Bruceatethecake. b. initP DP Bruce init’ init <eat> procP DP <Bruce> proc’ proc <eat> DP thecake c. e9s 1 ,e 1 [e=s 1 !e 1 &participant(Bruce,s 1 )& eat(s 1 )&participant(Bruce,e 1 )& eat(e 1 )& Path(thecake,e 1 )] (106) Accomplishments(PPPath) a. Agathawalkedintotheroom. b. initP DP Agatha init’ init <walk> procP DP <Agatha> proc’ proc <walk> PP intotheroom 54 Section2.3. ModelingAktionsartinthesyntax Chapter2 c. e9s 1 ,e 1 [e=s 1 !e 1 &participant(Agatha,s 1 )&walk(s 1 )&participant(Agatha,e 1 )&walk(e 1 ) &Path(intotheroom,e 1 )] The semantic commonality of paths is that they involve a set of ordered measures—a scale—that unifieshomomorphicallywiththeeventintroducedbyproc.Thesescalescanbecontextuallyrelevant partsofaDPobjectwithcreation/consumptionverbs(anincrementalthemeinthesenseofDowty 1991; Krifka 1992, inter alia) or spatial points along a physical path, and they need to be monotonic in Schwarzschild’s (2002) sense. 29 Ramchand thus defines the path role as ‘the relation that holds betweenanentityandanevent,ifamonotonicpropertyofthatentityismonotonicwithrespectto thepart-wholestructureoftheevent’(2008:50). 30 Finally,simplestatesarederivedbyasingle inithead. Ramchandproposesthatthecomplement ofstatesisselectedby init, functioningasrhematicmaterialthatfurtherspecifiesthestate. Lestthe notationbemisleading,Notethathereinitisnotinterpretedasaninitiationalsubeventbecausethere isnofurthereventstructuretogeneratethatentailment:inisolation,initPissimplyastate. (107) States a. Katherinefearsnightmares. b. fear:[init] 29 Schwarzschild(2002)callsmonotonicapropertythattrackspart-wholerelations. Forinstance,lengthisamonotonic propertyofacable,sincelessquantityofcableentailslesslengthofcable.Thediameterofthecable,ontheotherhand,is non-monotonic,sincelessquantityofcabledoesnotentaillessdiameterforthecable. 30 Ramchand’s formalization of the Path relation, inspired by Krifka’s (1992) mapping to objects and mapping to events, is providedbelow,whereD x isthescaleofthemonotonicpropertyoftheentityx: (xiv) Path(x,e)= def 9R9D x [8e,d,d’[R(e,d)&d’d!9e’[e’e&R(e’,d’)](mappingtoevents)&8e,e’,d[R(e,d)&e’e !9d’[d’d&R(e’,d’)](mappingtomeasures) In prose, mapping to events ensures that every sub-measure of the scale of the relevant monotonic property of x (eg. the measures of the skin of an apple in the predicate peel an apple) corresponds to a sub-part of the event (eg. the peeling). Mapping to measures,inturn,ensuresthateverysub-partoftheeventcorrespondstoasub-measureoftherelevantscale. Therefore,ifthepathargumentisbounded,asisthecasein(105)and(106),telicitywillariseasasemanticentailment,since theeventisunderstoodtoendwhenthepartsoftherelevantsetofmeasuresareusedupalongtheinternalprogression oftheevent. 55 Section2.3. ModelingAktionsartinthesyntax Chapter2 c. initP DP Katherine init’ init <fear> DP nightmares d. s[participant(Katherine,s)&fear(s)&Rheme(nightmares,s)] 2.3.4.2 Criticalassessment RamchandpresentsanarticulatedsyntaxfortheVPthat,likeHale&Keyser(1993)butunliketheXS- model,assumesatightmappingbetweensyntaxandsemanticsinthatsyntacticcomplexitymirrors event complexity. However, like the XS-model, Ramchand’s model is framed in syntax proper, un- like L-syntax, which posits those syntactic configurations to be part of the lexical representation of theroot. Furthermore,thismodelmakesnoassociationbetweengrammaticalcategoryandseman- tic type, in that a verb can be an event (proc) as well as a state (init and res), which circumvents the problemsalreadydiscussedaboutHale&Keyser(1993)andBorer(2005b)inSections2.3.2.2and2.3.3.2. The obvious advantage of this model is that it only assumes one syntactic computational module, whoseassociationofsyntacticandeventcomplexityallowstoexplaintheinteractionofsyntaxand eventstructureinclassicAktionsartdiagnosticssuchasthescopeeffectsofagainandalmostwesawin Section2.2.4.2. 31 Thismodelhasgenerallycomeintocritiquefortwomainreasons:first,someauthorsbelievethat first-phase syntax (and nano-syntax, more generally) is a return to Generative Semantics in that it attemptstosemanticizesyntax(e.g. Boeckx2014). However,thecrucialdifferenceisthatfirst-phase syntaxwillnotpositasemanticallymeaningfulstructurethatisnotsyntacticallyrelevant. 32 Inarecent book, Fábregas(2016)arguesforthispointindefenseofRamchand’smodel. Hetakesasanexample the verb kill, which in Generative Semantics was assumed to be decomposed ascausebecomenot alive (see (59)). The nanosyntactic model would reject the presence of not in the decomposition of kill on syntactic grounds: kill cannot license negative polarity items by itself (e.g. (108)), which makesthepositingofanegativeoperatorunwarranted.Thus,nanosyntaxistrulyamodelofsyntactic decomposition. 31 SeealsovonStechow(1996), whodrawsfromGermandatainvolvinginterpretivedifferenceswithwordordertoex- plicitlyarguethattheambiguityofagainissyntacticallydetermined. 32 SeeTravis(2010:104)forasimilardefenseforHaleandKeyser’smodel. 56 Section2.3. ModelingAktionsartinthesyntax Chapter2 (108) *Annakilledanybody. Another polemic aspect of first-phase syntax is that it assumes that a verb may denote a maximum of three subevents (see also Parsons 1990; Hale & Keyser 1993; Travis 2000). This view has been fre- quently contended for lexical causative verbs (Fodor 1970; Fodor & Lepore 1999; von Stechow 1995, 1996; Pylkkänen 2002; Kratzer 2005; Schäfer 2008, 2009; Marantz 2013, a.o.), arguing that only two subeventsarenecessaryorwarrantedinthesemanticdecomposition. 33 Marantz(2013)isclearinhis critique: Most intuitive semantic analyses of a transitive “opening” event would identify at least three components to the event decomposition: a causing event (suppose John pushes a buttontoopenthedoor),achangeofstate(thedoorswingsonitshingesfromaclosed state)andanendstate(thedoorisopen). However, thereisnoevidencethatlanguages consistently (or perhaps ever) employ three syntactic heads corresponding to “cause”, “change”(become),and“be”,toexpressthemeaningofatransitiveverblike“open.”Marantz (2013:161) Detractorsofthethree-eventaccounthavearguedthattheadverb againonlyhastwopossibleread- ings with lexical causatives. First, there is the restitutive reading, which scopes only over the result state,andthe repetitivereading,thatscopesovertheeventofchange. Crucially,theseauthorsargue, the repetitive reading necessarily includes the agent under the scope of the adverb in the transitive version, i.e. thereisno intermediatereadinginwhichonlytherepetitionoftheeventisasserted, i.e. excludingtheparticipationoftheagent.Iillustratemyillustrationofthephenomenain(109)and(110), adaptedfromMartin&Schäfer(2014). (109) Thedoorclosedagain. a. Restitutivereading: onlytheresultstateisrepeated(thedoorwasinaclosedstateprevi- ously). b. Repetitivereading:theclosingeventanditsassociatedresultstatearebothrepeated. (110) Marthaclosedthedooragain. a. Restitutivereading: onlytheresultstateisrepeated(thedoorwasinaclosedstateprevi- ously). b. Intermediate reading: the closing event and its associated result state are both repeated, butnottheactionoftheagent. 33 Othersarguethatcausativesareinfactmono-eventive,likeDoron(2003)andalsoBorer(2005b)inSection2.3.3. 57 Section2.3. ModelingAktionsartinthesyntax Chapter2 c. Repetitivereading: theactionoftheagent, aswellastheclosingeventanditsassociated resultstatearerepeated. However,Martin&Schäfer(2014)showthat,uponcloserinspection,theintermediatereadingin(110b) isindeedpossibleinGermanandFrench.InotethesameinSpanish:aswecanseein(111),theadverb de nuevo‘otravez’iscompatiblewithanintermediatereading,inwhichPedroclosedawindowfora secondtimewithouthavingclosedithimselfthefirsttimearound. (111) La The ventana window se refl cerró closed y and se refl volvió returned a to abrir open por for un a golpe blast de of viento. wind Como Since Pedro Pedro tenía had frío, cold la it cerró he-closed de of nuevo. new ‘Thewindowclosedandopenedagainfromablastofwind.SincePedrowascold,heclosedit again.’ Martin & Schäfer (2014), whose goal is actually to argue against a three sub-event account of lexical causatives, 34 arguethatmanneradverbsposeachallengeforsuchanaccount, especiallywhencom- pared with periphrastic causatives. They discuss contrasts like (112), where, in the lexical causative (112a), slowlycannotmodifytheactionoftheagentJohnbutonlytheeventoftheshipsinking. (112b), ontheotherhand,isambiguous:slowlycanmodifytheactionofJohnorthesinkingoftheship.The authors conclude that lexical causatives do not have an extra causing event that includes only the subject. (112) a. Johnsanktheshipslowly. b. Johnmadetheshipsinkslowly. (FromMartin&Schäfer2014:220) AsIseeit,however,thisargumentonlygoesthroughifthehighercausingeventinquestionisassumed tobeeventive(i.e. dynamic). RememberthatDowtymodeledhislexicalcausativesviarelationsbe- tweensentences,inwhichthe’causingevent’ofaccomplishmentswasaneventivesentencewiththe do predicate (i,e, for Dowty, (112a) would be paraphrased as John did something that caused the ship to becomesunk).Parsons(1990)andHale&Keyser(1993)seemtobelievethecausingeventualitytobedy- namicaswell. ButRamchand’scausingeventuality(initP)isastate,anddynamicadverbslikequickly 34 Martin&Schäfer(2014)and,particularly,Alexiadouetal. (2006)andAlexiadouetal. (2015),argueforastructureinwhich thereisaseparateprojectionintroducingtheexternalargument,VoiceP,butnotaseparatesubevent,asinKratzer(1996). The process event is introduced by a vP that denotes acause relation between an event and a state (see also Pylkkänen 2002; Kratzer 2005), the latter denoted by the root, selected for by v. The configuration is then VoiceP> vP> p . The intermediatereading,then,isderivedbyhavingagainscopebetweenVoicePandvP. 58 Section2.4. Conclusions Chapter2 cannotmodifystates(e.g. *John knows Math quickly). Manneradverbs,therefore,arenota(solid)ar- gumentagainsteventdecomposition. Anotherpieceofdatathatspeaksstronglyinfavorofathreesubeventdecompositioncomesfrom locativeadverbs. Parsons(1990)notedthatthePP behind the museumin(113)isambiguousbetweena readingwhereitlocatestheagent(113a)andthetheme(113b). Hetookthisasevidencethatcausative sentenceswerebi-eventive, underthestandardassumptionineventsemanticsthatlocativemodifi- cation is modification of an event argument (Davidson 1967). Crucially, however, the locative has a resultative reading, which describes the endpoint of the kite as a result of the flying. I include it in (113c). (113) Maryflewthekitebehindthemuseum. (FromParsons1990:118) a. Marywasbehindthemuseum. b. Thekitewasbehindthemuseum. c. Thekiteendedupbehindthemuseum. These data are clearly consistent with Ramchand’s model. In reading (113a), the locative attaches to initP;inreading(113b),itattachestoprocP;inreading(113c)itattachestoresP.Asitturnsout,amodel thatdoesnotassumeaminimaldegreeofcomplexityfortheVPbothforitssyntaxanditssemantics willinevitablebeunabletocapturemanyimportantempiricalfacts. 2.4 Conclusions This chapter has provided an overview of the current state of the art in the study of Aktionsart, and in particular in its syntactic modeling in generative work. I have begun with illustrating the classic typology of event types (starting with Vendler) and I have discussed the relevant aspectual diagnos- tics as they apply to Spanish, given that it will be the main language under study in this work. In sodoing, Ihavearguedagainsttheexistencesemelfactivesasadistincteventtypeaswellasagainst a characterization of achievements based on temporal punctuality: rather, achievements are simply change-of-stateverbs,whoseprocesspartmaybeconceivedofasshorterorlonger. Ihavethenpresentedsomeofthemostinfluentialsyntacticaccountsintheliteratureregarding Aktionsart. Thediscussionhasbynotmeanscomprehensive,sincetherearemanyimportantmono- graphs that I have not been able to discuss for reasons of space (e.g. Cuervo 2003; McDonald 2008; Travis2010,a.o.).IhaveinsteadfocusedonthosethatIbelieveseminalforthestudyofAktionsart:First, (i)L-syntax,forshowinghowlexicalverbscanbedecomposedintosyntacticstructures,whichderive andconstraintpossibleargumentstructureconfigurations.Acrucialcontributionofthismodelisto 59 Section2.4. Conclusions Chapter2 showthataspectualmeaningcanbederivedfromtheseargumentstructuresandhowtheinterpreta- tionofargumentsisdeterminedbythepositiontheyoccupyinanaspectually-meaningfulstructure, and hence-roles can be done without entirely. Second, the XS-model, for the radical separation ofthesyntaxandthelexiconinthestudyofeventstructure: theformercreatesargumentstructure with a certain formal interpretation, the latter merely modifies that structure, but is not its source. Last, first-phasesyntax, forintegratingthebestintuitionsofthetwopreviousmodelsintoasystem thatassumessomeminimalgrammaticalspecificationfortherootandasyntaxthatderivesargument structureandeventtypesinasystematicandelegantmanner. Theliteraturereviewofthischapterhasmadethecaseforadecompositionalviewofthesyntax oftheVP,inparticularoneinwhichargumentstructurecomplexitymirrorseventcomplexity. Inso doing,IhavedefendedtheviabilityofderivingAktionsartand-rolesfromargumentstructurewith theminimaltheoreticalassumptionsandasinglelevelofrepresentation—syntax. Ihavealsoargued, followingothers,thatconflatingcategoryandaspectualmeaningisawrongmove: statesandevents cutacrossallthetraditionallexicalcategories. As it turns out, however, this research program has not analyzed stative predicates in sufficient depth, often implicitly or explicitly severing them from the verbal domain—e.g. the XS-model or L-syntax—or simply relegating them to the simplest unit of analysis—first-phrase syntax. 35 In what follows, I put forth a view of states that puts them on a par with events in terms of syntactic and eventivecomplexity. Departingfromthegroundworklaidinthischapter, Ishowhowthedifferent syntactic primitives of event structure can operate in deriving different kinds of stative predicates with distinct aspectual interpretations, just as we have seen for eventive predicates. This inclusive viewofstates,inturn,shedslightonthepropertiesofcertainverbswhoseaspectualclassificationin theclassic Aktionsarttypologyhasprovenproblematic. Thefollowingchapterpresentsacasestudy inquestion,analyzingasetofSpanishverbsthatIlabelstativecausatives. 35 ButseeArche(2006);Husband(2012);Roy(2013);Gumiel-Molina et al. (2015);García-Pardo&Menon(submitted),a.o., forneoconstructionistapproachestotheindividual-levelandstage-leveldistinctionwithinadjectivalpredicates,whichI donotdiscussinthisdissertation. 60 3 Stativecausatives 3.1 Introduction Thepreviouschapterpresentedthetraditionalclassificationofeventtypes,wherefourdistincttypes werepresented:states,activities,accomplishmentsandachievements.Thereisasetofverbs,however, thatdoesnotseemtofullyfitintoanyofthesefourcategories,i.e.theAktionsarttestsdonotgivecon- clusiveresultsforaclear-cutclassification.Theverbsinquestionareexemplifiedin(114)andIreferto themasgovern-typeverbs. TheseverbsaredisconcertingfromaVendlerianstandpointbecausethey show properties usually associated to events (e.g. agentivity) and states (e.g. habitual reading in the presenttense),andalsotoatelic(e.g. acceptabilityof for-phrases)andtelicpredicates(e.g. formation ofadjectivalpassives,tobediscussedindepthinChapter4). (114) Govern,surveil,protect,control,direct,preside,reign,supervise... Althoughtheseverbshavenotreceivedadequateattentionintheliteratureuntilveryrecently(Fábre- gas & Marín 2012, 2017), their oddity has not gone unnoticed. In his classic work, Vendler discusses howactivitiesandothernon-stativeeventtypescanbecome’states’inthehabitualpresenttense(e.g. Areyousmoking? vs. Doyousmoke?),henoticesthefollowingabouttheverbtorule: Nowthecuriousthingisthatwhilecabdrivers–thatis,peopleofwhomonecanalways saythattheydriveacab–sometimesareactuallydrivingacab, rulers–thatis, peopleof whom one can always say that they rule a country–are never actually ruling a country, i.e. theyareneverengagedinthespecificactivityofrulingacountrycomparabletothe specificactivityofdrivingacab. Vendler(1957:151) Dowty (1979) also noticed the existence of these verbs, and was also unsure about how to classify them. In his discussion of how his system could account for the different event types and thematic interpretations(seeSection2.3.1),hegoesasfollowswhendiscussingactivities: 61 Section3.2. Non-dynamicevents:Fábregas&Marín(2017) Chapter3 B. Activities 1. Simpleactivities:DO( 1 ,[ n ( 1 ,..., n )]).(Johniswalking.) 2. Agentive Stative Causatives (?): [DO( 1 ,[ m ( 1 ,..., m )]) CAUSE n ( 1 ,..., n )]. (TheexistenceofthisclasswassuggestedtomebyHarmonBoertienandwould includeexampleslikeHeishousinghisantiquecarcollectioninanoldbarn,which areagentiveandpresumablycausativebutdonotentailanychangeofstate.) Dowty(1979:124) As it turns out, the classic typology of Aktionsart is not sufficient in light of these verbs and it needs to be rethought. This chapter will review critically the only piece of work (to my knowledge) that has focused on characterizing these verbs aspectually and I will provide an alternative analysis. In a nutshell, I argue that these verbs are stative causatives. They are formed syntactically via an initP anda resPprojection(Ramchand2008,seeSection2.3.4),whichintroduceastateeachandareglued togetherviaacausalrelation.Myaccountisschematizedin(115). (115) a. Berta Berta gobierna governs el the país. country ‘Bertagovernsthecountry.’ b. [ initP Berta[ init’ gobierna[ resP elpaís[ res’ gobierna]]]] c. XXXXXXXXs 1 XXXX!XXXXs 2 XXXXX My account adopts a view of cause (‘!’) as a generalized relation between eventualities that arises configurationally,includingbetweentwostates.Inthislight,IreviewcriticallytheproposalinNeele- man&vandeKoot(2012)thatthereisanadditionalrelationmaintainintheeventstructureontology thatappliestogovern-typeverbs. Thischapterisorganizedasfollows:Section3.2reviewscriticallyFábregas&Marín’s(2017)anal- ysisoftheseverbs.Section3.3articulatesmyproposal,whileSection3.4defendstheviewofageneral- izedcauserelationforeventsemantics.Section3.5discussesitscompatibilitywithpreviousaccounts forthesyntaxoftheVP.Finally,Section3.6concludesthechapter. 3.2 Non-dynamicevents:Fábregas&Marín(2017) 3.2.1 AnoverviewofFábregas&Marín(2017) Fábregas&Marín(2017)(startinginFábregas&Marín2012)focusonthe Aktionsartpropertiesofthe Spanishverbsin(116).Letuscallthemthegobernar‘govern’-typeverbs. 62 Section3.2. Non-dynamicevents:Fábregas&Marín(2017) Chapter3 (116) a. aguantar‘bear’,mantener‘maintain’,sostener‘support’,sujetar‘hold’,sustentar‘support’. b. conservar‘preserve’,cuidar‘takecare’,guardar‘keep(safe)’,preservar‘preserve’,proteger‘pro- tect’,resguardar‘shelter’. c. controlar‘control’, coordinar‘coordinate’, dirigir‘direct’, gobernar‘govern’, presidir‘head’, su- pervisar‘supervise’,vigilar‘oversee’. d. bloquear‘block’,evitar‘avoid’,impedir‘prevent’,inhibir‘inhibit’,prohibir‘forbid’. The authors show that these verbs are atelic, like activities and states. They do not allow in x time phrases (e.g. (117a)), they are incompatible with the periphrasis terminar de ‘finish’, compatible only withtelicpredicates(e.g.(117b)). (117) a. *Juan *Juan conservó kept el the recuerdo souvenir en in un one mes. month intended:‘Juankeptthesouvenirinonemonth.’ b. *Juan *Juan terminó finished de of {sujetar/ holding preservar/ preserving coordinar/ coordinating prohibir} forbidding lo. it intended:‘Juanfinished{holding/preserving/coordinating/forbidding}it.’ However,classifyingthemaseitherstatesoractivitiesprovesdifficult.Ontheonehand,theybehave like states in the following respects. First, they are not compatible with the periphrasis parar de ‘to stop’,butwithdejarde‘tostop’,justlikestativeverbs(e.g.(118a),andrememberthediscussioninSection 2.2.4.1). Also,theseverbsfulfillthestrictsubintervalproperty(Bennett&Partee1972;Rothstein2004; see also Section 2.2.1). Govern-type verbs are also incompatible with velocity adverbials (e.g. (118b)). Also,likestates,gobernar-verbsdonothaveahabitualreadinginthepresenttense.Finally,theseverbs allowpassiveparticipleformationwiththecopulaestar(e.g.(118d)),atraitoftelicverbsandstates,but notofactivities(Marín2000,2004;seealsoChapter4forathoroughanalysisofthisconstruction). (118) a. Juan Juan {??/*paró/ {??/*finished dejó} left de of {aguantar/ holding conservar/ preserving supervisar/ coordinating inhibir} forbidding lo. it intended:‘Juanstopped{holding/preserving/coordinating/forbidding}it.’ b. *Juan * gobierna rules España Spain rápidamente. rapidly intended:‘JuanrulesSpainquickly.’ c. Juan Juan dirige directs la the empresa company (ahora). (now ‘Juandirectsthecompanynow.’ 63 Section3.2. Non-dynamicevents:Fábregas&Marín(2017) Chapter3 d. España Spain está is estar gobernada ruled por by un an inconsciente. irresponsible ‘Spainisruledbyanirresponsibleperson.’ Ontheotherhand, theauthorsargue, gobernar-verbsalsopatternasactivities, andnotasstates, ac- cording to other diagnostics. They allow the progressive (e.g. (119a)) and, furthermore, they allow mannerandplacemodifiers(e.g. (119b)and(119c),respectively). Also,theycanbeantecedentsof hap- pen-anaphors(e.g.(119d))andcomplementsofperceptionverbs(e.g.(119e)). (119) a. Juan Juan está is gobernando ruling España. Spain ‘JuangovernsSpain.’ b. Juan Juan sujetó held el the cuadro picture firmemente. firmly ‘Juanheldthepicturefirmly.’ c. Juan Juan vigila watches.over a acc Luis Luis desde from su his despacho. office ‘JuansurveilsLuisfromhisoffice.’ d. Juan Juan supervisaba supervised a acc Luis. Luis. Esto This sucedía happened mientras... while... ‘JuansupervisedLuis.Thishappenedwhile...’ e. Vi I.saw a acc Juan John sujetar hold el the cuadro. painting ‘IsawJohnholdthepainting.’ Giventhatthislastsetofpropertieshavebeentypicallyassociatednotjusttoactivities,butto(non- stative)eventsmoregenerally,theseauthorsarguethatthiseventtypeischaracterizedbybeingeven- tive, but not dynamic. For them, an event is an eventuality that can appear in the progressive, take mannerandplacemodifiersandbeanantecedentofhappen,asopposedtostates.Dynamicityisonly apossible,butnotanecessarypropertyofevents,againstthestandardview. Intheirsyntacticanalysis,vencodesaneventandintroducesanexternalargument.Theinternal argument is introduced by a central coincidence preposition (P cc ), which is stative (Hale & Keyser 2002). TheeventencodedinvisinterpretedasmaintainingthelowersubeventencodedinP cc . The verbstartsoffasacomplementofP cc andisultimatelyverbalizedbyv. Thestructureisprovidedin (120). 64 Section3.2. Non-dynamicevents:Fábregas&Marín(2017) Chapter3 (120) Non-dynamicevents(gobernar-typeverbs) v DP v v P cc DP P cc P cc X Foraneventtobedynamic,theseauthorsargue,vneedstohaveageneralizedPathcomplement,un- derstoodas‘thetypeofsemanticobjectthatcandefineabstractmovementacrossadimension’.Inthe spiritofHale&Keyser(2002)andRamchand(2008),theyproposethataPathcanbeofthreetypes: i) anominalcomplementthatprovidesmereologicaldivisionsfortheevent,asin(121a)(andseeRamc- hand’sanalysisofAccomplishmentsonSection2.3.4);ii)anunderlyingscalaradjective,givingriseto adeadjectivalverbthatdenoteschangealongthescaleoftheadjective, asin(121b); iii)apreposition denotingchangeoflocation,asin(121c). (121) Dynamicevents(TypesofgeneralizedPaths) a. Johnpaintedthewall. v v D/N b. Thescreencleared. v v A c. Johnpushedthecarttothewindow. v v P Path The structure in (120) cannot be dynamic because the central coincidence P does not denote a path along which a change or movement could be defined. Again, the structure is not stative, either: a strictly stative structure with a central coincidence P (e.g. The parrot is in the cage) should have the structurein(122),withaverbalizerStdenotingastate,notanevent. 65 Section3.2. Non-dynamicevents:Fábregas&Marín(2017) Chapter3 (122) St DP St St P cc DP P cc P cc X Tosummarize,IprovideanoverviewofthetypologyofAktionsarttheseauthorsassume,andthetest theyadduceinitsfavor,inTable3.1. Table3.1:TypologyofeventtypesinFábregas&Marín(2017) Tests Atelic Telic States Events Non-dynamic Dynamic Inxtime-phrases 7 7 7 3 Complementofterminarde‘finish’ 7 7 7 3 Complementofpararde‘stop’ 7 7 3 3 Velocityadverbials 7 7 3 3 Habitualreadinginthepresenttense 7 7 3 3 Participleswithestar‘tobe’ 3 3 7 3 Theprogressive 7 3 3 3 Mannermodifiers 7 3 3 3 Placemodifiers 7 3 3 3 Antecedentsofhappen-anaphors 7 3 3 3 Complementsofperceptionverbs 7 3 3 3 Agentivity 7 3 3 3 3.2.2 ReviewofFábregas&Marín(2017) Fábregas&Marín(2017)constitutesaveryimportantpieceofworkonAktionsart.Itsmainmerit,asI readit,hasbeentobroadenandredefinetheempiricalscopeofthisareaofresearch. Whileitisun- deniablethattherehasbeensubstantialtheoreticalprogressinourunderstandingofeventstructure 66 Section3.2. Non-dynamicevents:Fábregas&Marín(2017) Chapter3 in the past decades, the empirical domain has remained, for the most part, unchanged, the major- ityofauthorsassumingtheVendleriandistinctionbetweenstates,activitiesandtelics.YetFábregas& Marín(2017)showconvincinglyandsystematicallythatthetraditionaltypologyfallsshortwhenfaced withthisclassofverbs(the gobernar-type),whichstraddlebetweenactivitiesandstatesaccordingto thestandarddiagnostics. Notwithstanding, I find their theoretical implementation somewhat problematic. The idea that dynamicityisinducedbyageneralizedPathcomplementofavheadmayappearsoundinthelight of the examples provided in (121), where the event denotes change along a scale that is provided by the Path complement. However, there are plenty of dynamic verbs which do not appear to have a Pathcomplementatall. Suchisthecaseofintransitiveactivitiesingeneral,asshownin(123a). Also, activity verbs that are typically associated with a kind of Path do not necessarily require that Path to be dynamic. Such is the case of verbs of bodily motion like correr ‘run’, which, although typically associatedwithspatialPaths(e.g. treskilómetros‘threekilometers’,alatienda‘tothestore’...),canappear in configurations where their notional Path can not only be absent, but explicitly denied, and yet dynamicity is preserved (e.g. (123b), where it is asserted that there is no spatial displacement in the running). Forthesereasons,positingacovertPathprojectioninthesedynamicconstructionsisnot onlyunjustified,butproblematic. (123) a. Pedro Pedro {canta/ {sings silba/ whistles trabaja/ works juega/ plays grita/ screams habla/ talks escupe...}. spits ‘PedroJuan{sings/whistles/works/plays/screams/talks/spits...}. b. Carlos Carlos estaba was corriendo running rápidamente quickly sin without moverse moving del from.the sitio. place ‘Carloswasrunningquicklyinplace.’ 3.2.2.1 Eventivitytests? I furthermore have my doubts that the evidence separating states from non-dynamic events is too compelling. Let us focus on each test separately. Regarding the progressive, a closer look reveals thatstativeverbsdoallowtheprogressive,albeitwithlesseasethaneventiveverbs.Thesentencesin (124)areallcompletelynaturalinSpanish. Furthermore,itturnsoutthat gobernar-typeverbsdonot alwaysallowtheprogressiveeasily:thesentencesin(125)areratheroddwithoutanadequatecontext (intuitively,acontextthatsetstheeventualityaspartofamoretransitoryorcontextuallysalientstage). Itlooksliketheprogressiveisnotassensitiveaspreviouslythoughtintermsofthe Aktionsartofthe 67 Section3.2. Non-dynamicevents:Fábregas&Marín(2017) Chapter3 verb(inlinewithLevin&Rappaport-Hovav1995), 36 anditsrestrictivenesswithstatesmayturnoutto beanepiphenomenonofapragmaticconstraintofthisparticularconstruction. (124) a. Daniel Daniel está is viviendo living en in Caracas Caracas ahora. now ‘PedrolivesinCaracasnow.’ b. Estoy I.am deseando wishing que that prohíban they.forbid el the tabaco. tobacco ‘Ihopetheybantobaccoproducts.’ c. Estamos we.are dudando doubting entre between irnos going de of vacaciones vacation en in junio June o or en in agosto. August ‘WeareunsureastowhethertogoonvacationinJuneorinAugust.’ d. Estoy I.am teniendo having problemas problems para to encontrar find un a reloj watch nuevo. new ‘Iamhavingtroublefindinganewwatch.’ e. A dat Pedro Pedro le him.dat está is doliendo hurting la the cadera hip últimamente. lately ‘Pedrohasbeenhavinghippainlately.’ f. La the película movie me me.dat está is gustando. liking ‘Iamenjoyingthemovie.’ (125) a. #Putin #Putin está is gobernando governing Rusia. Russia intended:‘PutinisgoverningRussia.’ b. #Bill #Bill Gates Gates está is dirigiendo directing Microsoft. Microsoft intended:‘BillGatesisdirectingMicrosoft.’ c. #La #the sal salt está is conservando conserving la the carne. meat intended:‘Thesaltisconservingthemeat.’ 36 InLevin&Rappaport-Hovav’swords: ‘Theabilitytobeusedinthepresentprogressiveisnotatestfornonstativeness, butratherisatestforanonmomentarypredicate.’(1995:170) 68 Section3.2. Non-dynamicevents:Fábregas&Marín(2017) Chapter3 d. #Estas #these columnas columns están are sosteniendo holding el the tejado. roof intended:‘Thesecolumnsareholdingtheroof.’ Anothertestthatdoesnotseemtogiveclearresultsuponcloserinspectionistheimpossibilityofstates ofbeingantecedentsof happen. WhileIagreewiththeseauthorsthatstativeverbsaregenerallyout asantecedentsofhappen,wecannonethelessfindcounterexamples(e.g. (126)). Moreproblematically, itturnsoutthat gobernar-verbsdonotreallyshowasignificantlygreaterabilitytobeantecedentsof happenincomparisontostativeverbs. Theauthorsthemselvesacknowledgethis,arguingthatthere arefurther(unclear)pragmaticconstraintsthatpreventtheseverbsfrombeingantecedentsofhappen (e.g.(127))butalsoeventiveverbs(e.g.(127)).Ifthisisso,thenthepossibilityofbeinganantecedentof happenisnotatestfortellingapartstatesandevents. (126) a. El the niño kid estuvo was toda all la the tarde afternoon sin without hablarnos. talking-us.dat Esto this sucedió happened porque because le him.dat quitamos remove la the consola. console ‘The kid did not speak to us all afternoon. This happened because we took the console awayfromhim.’ b. Pedro Pedro tiene has gripe. flu. Esto This le him.dat sucede happens por for no not cuidarse. take.care.refl ‘Pedrohastheflu.Thishappenedtohimbecausehedoesn’ttakecareofhimself.’ (127) a. La the columna column apoyaba supported el the techo. roof ??Esto ??this sucedía happened mientras... while.. b. Juan Juan apoyaba supported a acc María. María Esto this sucedía happened mientras.. while... (128) a. El the asteroide asteroid volaba flew por around el the universo. universe ??Esto ??this sucedía happened mientras... while.. b. El the pájaro bird volaba flew en in el the jardín. garden Esto this sucedía happened mientras.. while... Theeventivitytestregardingthepossibilityofbeingacomplementofperceptionverbsisnotwithout problemseither.First,asFábregasandMarínthemselvesacknowledge,itisnottruethatalleventive verbscanbecomplementsofperceptionverbs(see,e.g.Ramchand2005:365)(e.g.(129)). 69 Section3.2. Non-dynamicevents:Fábregas&Marín(2017) Chapter3 (129) a. *Vi *I.saw a acc Juan Juan darse give.refl cuenta count del of.the error. error intended:‘IsawJuannoticethemistake.’ b. *Pedro *Pedro vio saw a acc María María decidir decide su his castigo. punishment intended:‘PedrosawMaríadecidehispunishment.’ Moreproblematically,gobernar-verbscannotbecomplementsofperceptionverbswithasmuchease astheseauthorsclaim. Theevidencetheypresentisgivenin(130). However,Inotethatchangingthe subject(131a,d)ortheobject(131b,c)degradesthesentencestothepointofungrammaticalityandmany gobernar-verbs, like gobernar itself, hardly allow it at all. The authors remark that perception verbs requireoftheircomplementnotonlythatitbeeventive,butthattheeventitdenoteshassomesortof externalmanifestation(rememberVendler’squoteonpage61aboutthedifficultyofpointingoutwhat specific actions verbs like governing involve). But this is not essentially different from the eventive verbsin(129),whicharealsooutbecauseeventssuchasnoticinganddecidingarehardlyperceptible. One wonders, then, whether perception verbs really disallow stative verbs as complements or their acceptability—or lack thereof—is solely determined by whether the eventuality, stative or eventive, canbeperceived. 37 37 Although stative verbs cannot indeed be complements of perception verbs as a general rule, this does not extend to non-verbalpredicates.Asyoucanseein(xva),stativeadjectivalpredicatesaregrammaticalascomplementsofperception verbs provided the copula is not included. Note also that adjectival predicates that denote individual-level properties (Carlson1977),whichgowiththecopulaserinSpanish,areungrammaticalascomplementsofperceptionverbswithand without the copula (xvb). Perhaps the restrictions of perception verbs have not to do so much with the event vs. state divide, but with whether the predicateis individual-level or stage-level (asthe predicate in (xva), which takes the stage- levelcopulaestar).Sincestativeverbsdenoteindividual-levelpredicates(Kratzer1995),theirunavailabilityascomplements ofperceptionverbs,evenwhentheydenoteperceptiblestates(e.g.(xvc)),wouldbeexplained. (xv) a. Vi I.saw a acc Pedro Pedro (*estar) (*)be muy very triste. sad ‘IsawPedroverysad.’ b. *Vi *I.saw a acc Juan Juan (ser) (be esquimal. Eskimo intended:‘IsawJuan(be)anEskimo.’ c. *Vi *I.saw a acc Juan Juan vivir (live en in una a casa house amarilla. yellow intended:‘IsawJuanliveinayellowhouse.’ 70 Section3.2. Non-dynamicevents:Fábregas&Marín(2017) Chapter3 (130) a. Vi I.saw a acc Juan Juan sujetar hold el the cuadro. painting ‘IsawJuanholdthepainting.’ b. Vi I.saw a acc Juan Juan cuidar take.care a acc María. María ‘IsawJuantakecareofMaría.’ c. Vi I.saw a acc Juan Juan dirigir direct el the tráfico. traffic ‘IsawJuandirecttraffic.’ d. Vi I.saw a acc Juan Juan impedir prevent un a robo. theft ‘IsawJuanpreventatheft.’ (131) a. *Vi *I.saw el the clavo nail sujetar hold el the cuadro. painting intended:‘Isawthenailholdthepainting.’ b. ?Vi ?I.saw a acc Juan Juan cuidar take.care sus his modales. manners intended:‘IsawJuanmindhismanners.’ c. ?Vi I.saw a acc Juan Juan dirigir direct la the compañía. company intended:‘IsawJuandirectthecompany.’ d. *Vi I.saw los the escombros debris impedir prevent el the paso. pass intended:‘Isawthedebrisblocktheway.’ Thefinaltestthattheauthorsproposetodifferentiate gobernar-verbsfromstativeverbsistheavail- abilityofmannerandplacemodifiers. Indeed, stativeverbsappeartodisallowsuchmodifier, aswe canseein(132). (132) a. *Juan *Juan posee owns una a finca ranch ostentosamente. ostentatiously intended:‘Juanownsaranchostentatiously.’ b. *María *María sabe knows matemáticas Math en in Roma. Rome intended:‘MaríaknowsMathinRome.’ 71 Section3.2. Non-dynamicevents:Fábregas&Marín(2017) Chapter3 However, recentworkinEnglisharguesthatstativeverbsdoallowmannerandplacemodification, their limited availability being due to pragmatic factors. Ernst (2016), arguing against Maienborn (2007)distinctionbetweenKimianandD(avidsonian)states, 38 showsthatthereasonthatstatestypi- callydisallowlocativesisbecausethereisapragmaticconstraint,statedin(133),wherethesubscript ckstandsforcontextualknowledge. (133) Thestate-locationdefaultaxiom(SLDA):informalversion:Insentenceswherefocuson alocativeexpressionispossible, inalldefault ck worlds, ifanindividualisinastateatsome location,thenthatindividualwouldbeinthatstateatanylocation. (Ernst2016:242) AlthoughhisdataisdrawnforEnglish,IobservethattheyapplytoSpanishaswell,soIwillbeusing Spanishexamplestoillustrate. Theidea,then,isthat(134)isoutbecausethatlocationisirrelevantto theindividualbeinginthatstate,sinceifRobertoweretogoanywhereelse,hisstateofknowingMath wouldremainunchanged. However,whenthereisasalientvariabilityinthecontext—i.e. whenthe contextimpliesthatthestatemaynotholdineverylocation,ornotholdtothesamedegree—locatives becomefine,aswecanseein(135). (134) *María *María sabe knows matemáticas Math en in Roma. Rome intended:‘MaríaknowsMathinRome.’ (135) a. La the silla chair está is muy very inestable unstable en in esta this parte part del of.the suelo. floor ‘Thechairisveryunstableonthispartofthefloor.’ b. Sorprendentemente, amazingly el the niño baby estuvo was callado quiet en in el the avión. plane ‘Amazingly,thebabywasquietontheplane. 38 Maienborn(2005,2007)proposesthattherearetwokindsofstates,ontologicallydistinct: Kimianstates(e.g. be, know, own, resemble, weigh...), which she defines as ‘abstract objects for the exemplification of a property P at a holder x and a timet’,andD-states(e.g. sit,stand,lie,wait,gleam,sleep...),whichcontainaDavidsonianeventargument.Maienbornclaims thatthedefiningpropertiesofeventualitiesarethattheyareperceptible(thuscanbecomplementsofperceptionverbs), canbelocatedinspace(hencethepossibilityoflocativemodification)andcanvaryinthewaytheyarerealised(i.e. they allow manner modification). Although Fábregas and Marín followed Maienborn in their early work on gobernar-verbs, characterizingthemastransitiveD(avidsonian)states(Fábregas&Marín2012),theyenduprejectingforthreereasons: i) gobernar-verbsaretypicallyagentive,unlikeD-statessuchasdream,waitorlie;ii)itistheoreticallyundesirabletointroduce anewentityintheaspectualontology(theKimianstate);iii)theso-calledD-statesactuallymaydenotechangingsituations (e.g.gleam),whereastheirgobernar-verbs,duetotheirlackofdynamicity,cannot.Thisdissertationsideswiththeseauthors andothers(Ramchand2005;Ernst2016,a.o.)inthatalleventualities,stativeoreventive,containaneventvariable. 72 Section3.2. Non-dynamicevents:Fábregas&Marín(2017) Chapter3 c. Catalina Catalina es is más more rápida fast en in Los Los Ángeles Angeles que than en in la the cima summit del of.the Everest. Everest ‘CatalinaisfasterinLosAngelesthaninthesummitoftheEverest.’ d. Nuestro Our perro dog es is callado quiet en in la the ciudad, city pero but ruidoso noisy en in nuestra our casa house de of campo. countryside ‘Ourdogisquietinthecitybutnoisyinourcountrysidehouse.’ Withregardstomannermodification, theauthorarguesthatthereasonthatmanneradverbsareat firstsightunavailablewithstativeverbsisbecausemannermodificationismorereadilyavailablewith predicatesthatareconceptuallycomplex,ormultidimensional. Stativeverbsaremuchlessconcep- tuallycomplexthandynamicpredicates,theauthorargues,sincetheylackpropertiessuchaschange throughtime,beginningpoints,endpoints,causationandresult. Beasitmay, mannermodificationofstativeverbsisfarfromimpossible, iftherightdimension ofthepropertydenotedbytheverbisidentified. 39 Iprovideexamplesin(136). (136) a. Estos these zapatos shoes casan match perfectamente. perfectly ‘Theseshoesmatchperfectly.’ 39 The author argues that stative adjectives are easier to be adverbially modified than stative verbs, since adjectives are generallymoremultidimensionalthanstativeverbs,whichtypicallymerelydenotementalstatesorabstractrelationships. He groups the type of manner adverbs possible with stative adjectives into three main semantic classes: i) adverbs that connecttospeaker’sorsubject’sattitudes(xvia);ii)characterizationofsubpartsofthepropertythatdescribestheadjective (xvib); iii) other adverbs such as those that characterize prototypical properties of the adjective, exocomparatives and temporalones(xvic). (xvi) a. pleasantlyarchaic,obnoxiouslyidiosyncratic,eerilyreminiscent,confusinglydifferent,repellentlywicked,te- diously slow, mysteriously beautiful, disturbingly weird, paranoically suspicious, calmly busy, happily sore, smuglydelighted,exuberantlyhairy... b. evenly blue, patchily opaque, smoothly luminous, regularly curvilinear, randomly distributed, complexly dense, flexibly strong, loudly insistent, massively stable, sinuously elegant, stiffly erect, dully reflective, vio- lentlycarnivorous,gracefullyslender,fluidlygraceful,gorgeouslypregnant... c. openly contemptuous, quietly demonstrative, brazenly ambitious, ostentatiously wealthy, frankly interested, studiouslyunkempt,unobtrusivelyattentive,classicallyfeminine,characteristicallydemocratic,prototypically avian, similarly intricate, differently dangerous, identically intense (reactions), consistently attentive, endur- inglyrich(musicalheritage),lastinglypenetrating(smell)... However,thedataonmanneradverbswithEnglishstativeadjectivesthatErnstpresentscannotbeeasilycarriedoverto Spanish,forreasonsthatarenotcleartome.Ileavetheissuehereforfurtherresearch,notingthatitmayturnouttobea syntacticrestrictionagainstadverbialmodificationwithattributiveadjectivesinSpanishingeneral. 73 Section3.2. Non-dynamicevents:Fábregas&Marín(2017) Chapter3 b. Sandra Sandra pertenece belongs al to.the club club fortuitamente. fortuitously ‘Sandrabelongstotheclubfortuitously.’ c. Juan Juan ama loves a acc su his novio boyfriend abiertamente. openly ‘Juanloveshisboyfriendopenly.’ d. María María posee owns las those casas houses solapadamente, covertly, mediante via empresas companies fantasma. ghost ‘Maríaownsthosehousescovertly,viashellcorporations.’ Underthislight,wecanalsomakesenseoftherelativepermissivenessofmanneradverbsingobernar- verbswithrespecttostativeverbs: gobernar-verbsareconceptuallymorecomplexthanclassicstative verbs.Forinstance,theyhaveanagentivitycomponentthatotherstatesdonot,andassuchtheyallow manneradverbsorientedtotheagent(e.g.(137)).Ontheotherhand,sincegobernar-verbsdonotdenote dynamicchange,theydonotallowvelocityadverbials(e.g. (138)). Notethattheseconsiderationsare reminiscentofVendler’squoteatthebeginningofthischapteronpage61,regardingtheconceptual impoverishmentofverbslikegovernwithrespecttodynamicactivitieslikewriteorpaint. (137) a. Pedro Pedro gobierna governs el the país country prudentemente. prudently ‘Pedrogovernsthecountryprudently.’ b. Sandra Sandra protege protects su her bolso purse celosamente. zealously ‘Sandraprotectsherpursezealously.’ (138) a. *Catalina *Catalina cuida takes.care sus her negocios businesses rápidamente. quickly intended:‘Catalinatakescareofherbusinessesquickly.’ b. *Pedro *Pedro sujeta holds el the libro book despacio slowly intended:‘Pedroholdsthebookslowly.’ Inshort,theevidencepresentedinFábregas&Marín(2017)toclassifygobernar-verbsasnon-dynamic events,asopposedtostatesanddynamicevents,isnottoostrongundercloserscrutiny.Thetestthey adduce to classify non-dynamic eventualities in two types, eventive and stative, do not hold much 74 Section3.3. Myproposal Chapter3 water since the they can be shown to be sensitive to factors different from eventivity. Rather, gob- ernar-verbspatternasstatesaccordingtoothertests. AndyetIbelievethattheseauthorsarerightin positingthegobernar-classasaseparateaspectualtype,differentfromactivitiesandstatesproper. In thefollowingsectionIwilloutlinemyproposalfortheseverbs. 3.3 Myproposal IhaveshownintheprecedingsectionhowthetestsadducedinFábregas&Marín(2017)tosupportan eventivevs.stativedistinctionwithinnon-dynamicpredicateswerenotreallyvalid,andhencethere isnogoodreasontoassumethateventscomeintwotypesaccordingtodynamicity:rather,statesare eventualitiesdefinedbytheirlackofdynamicity,whereaseventsareeventualitiesdefinedbyhaving dynamicity,asinthereceivedview. Sincethesetestsaretheonesbywhichgobernar-verbspatterned witheventsproper, 40 Iconcludethatgobernar-verbsareinfactstative. 3.3.1 Definingtraitsofgobernar-verbs However,gobernar-verbsdonotbehavelikeclassicstatesinseveralcrucialrespects.Thefirsttestisthe abilitytoformadjectivalpassiveswithestar’tobe’:althoughtheauthorsfollowMarín(2000,2004)in thatonlystativeandtelicverbsaregoodinputsforadjectivalpassives,thesituationisactuallymore restrictive: only telics and gobernar-verbs are good inputs for adjectival passives, whereas activities and states are out. 41 I provide examples of well-formed adjectival passives with telics and gobernar- verbs in (139a) and (139b), respectively, and ill-formed adjectival passives with states and activities in (139c)and(139d),respectively. 40 SeemydiscussiononagentivityinSection5.5. 41 Thesituationisactuallyabitmorecomplexforadjectivalpassives.Withinatelicverbs,notonlygobernar-verbsaregood inputs, but also locative verbs like obstruct or cover and object-experiencer psychological verbs such as worry or amuse, whichIdiscussinChapter5(e.g.(xvii)). (xvii) a. La the finca ranch está is rodeada surrounded por by tres three colinas. hills ‘Theranchissurroundedbythreehills.’ b. Pedro Pedro está is preocupado worried por by la the crisis crisis económica. economic ‘Pedroisworriedbytheeconomiccrisis.’ However,mydiscussionofadjectivalpassivesonChapter4willnotfocusontheconstructionsin(xvii),whichItaketo beinstancesoflocativealternation.SeeSection6.3formorediscussion. 75 Section3.3. Myproposal Chapter3 (139) a. La the ciudad city está is destruida. destroyed ‘Thecityisdestroyed.’ b. El the museo museum está is protegido. protected ‘Themuseumisprotected.’ c. *Pedro *Pedro está is amado. loved intended:‘Pedroisloved.’ d. *La *the cuna cradle está is mecida. rocked intended:‘Thecradleisrocked.’ Asecondpropertythatseparatesgobernar-verbsfromclassicstativeverbsisthepossibilityofhaving different scope readings with the adverb again, a property they share with telic verbs. A sentence suchas(140)isthusambiguousbetweenawidescopeandnarrowscopereadingoftheadverb. This phenomenonhasnotbeenpreviouslyobservedforgobernar-verbs,asfarasIamaware. (140) La the presa dam impide preventes el the paso pass de of agua water de of nuevo. new ‘Thedampreventstheflowofwateragain.’ 1. Wide scope reading (again > dam: The dam prevented the flow of water in the past, stoppeddoingitatsometimeandnowitisdoingitagain. 2. Narrow scope reading (dam > again): The flow of water was prevented in the past (e.g. bysomedebris),thatsituationstoppedandnowthereisadamthatpreventstheflowof water. Thisscopeambiguity, interestingly, doesnotcarryovertoclassicstativeverbs. Sentence(141)isnot ambiguous: theadverb againnecessarilytakesscopeoverthesubject. Thatis,(141)cannotmeanthat PedrowaslovedbeforebysomeoneelseandnowDavidloveshim,butonlythatDavidlovedPedro before,somehowhestoppedlovinghimatsomepoint,andthenfellbackinlovewithhim. (141) Carmen ?Carmen ama loves a acc Pedro Pedro de of nuevo. new intended:‘CarmenlovesPedroagain.’ 76 Section3.3. Myproposal Chapter3 Another characteristic property of gobernar-verbs is their inability to participate in the transitive– unaccusative alternation: these verbs are strictly transitive (Fábregas & Marín 2012, 2017), as we can seein(142). Thisisatraitthattheyseemtosharewithclassicstates(143),butnotwithtelicverbs(144) oractivities(145). 42 (142) a. El the ejército army protege protects la the ciudad. city ‘Thearmyprotectsthecity.’ b. *La *the ciudad city (se) (refl protege protects intended:‘Thecityprotects.’ (143) a. Gabriel Gabriel respeta respects las the leyes. laws ‘Gabrielrespectsthelaws.’ b. *Las *the leyes laws (se) (refl respetan respect intended:‘Thelawsrespect.’ (144) a. Pedro Pedro rompió broke la the ventana. window ‘Pedrobrokethewindow.’ b. La the ventana window se refl rompió. broke ‘Thewindowbroke.’ (145) a. Laura Laura botó bounced el the balón. ball ‘Laurabouncedtheball.’ b. El the balón ball botó. bounced ‘Theballbounced.’ 42 InSpanishandmanyotherlanguages,theunaccusativevariantofthetransitive–unaccusativealternationissometimes markedwithreflexivemorphology(seeVivanco2015forarecentoverview).theexamplesin(142)and(143)includeoptional reflexivemorphologytoshowthattheirungrammaticalityisnotduetothepresenceorabsenceofsuchmorphology. 77 Section3.3. Myproposal Chapter3 TheempiricalpicturethatemergesfromthediscussioninSection3.2.2.1,aswellashere,isgivenon Table3.2. Tests1-3characterizegobernar-verbsasatelic,asalreadyshowninFábregas&Marín(2017). Tests4-6characterizegobernar-verbsasnon-dynamic,asalsoshowninFábregas&Marín(2017).Tests 7and8differentiate gobernar-verbsfromclassicstatesand,crucially,patternthemwithtelics. Test9 patternsstatesandgobernar-verbstogether. Table3.2:Eventtypesandtheirtests Tests Telics Activities Gobernar-verbs States 1. Inxtime-phrases 3 7 7 7 2. Forxtime-phrases 7 3 3 3 3.Progressiveentailsperfect 7 3 3 3 4.Universalreadingintheperfect 7 7 3 3 5.Subintervalproperty 7 7 3 3 6.Habitualreadinginthepresenttense 3 3 7 7 7.Scopeambiguitywithagain 3 7 3 7 8.Participleswithestar’tobe’ 3 7 3 7 9.(Anti-)causativealternation 3 3 7 7 3.3.2 Thesyntaxofgobernar-verbs Iproposethatgobernar-verbshaveaRamchandiansyntacticstructureasin(146b),fromexample(146a), andaneventsemanticsdecompositionasin(146c). (146) a. Berta Berta gobierna governs el the país. country ‘Bertagovernsthecountry.’ 78 Section3.3. Myproposal Chapter3 b. initP DP Berta init’ init <gobierna> resP DP elpaís res’ res <gobierna> XP c. s9s 1 ,s 2 [s=(s 1 !s 2 )&gobierna(s 1 )&participant(Berta,s 1 )&gobierna(s 2 )&participant(el país,s 2 )] Gobernar-verbs,then,lexicalize initand resheads,butnot proc. Theinmediateconsequenceofthisis thatthestructurecannotbedynamic,sinceonlyprocPencodesadynamiceventintheVPdecompo- sition. Ontheotherhand,thepresenceofinitPandresP,inacompositionalapproachsuchasfirst-phase syntax,willcreateacomplexstativeeventuality,composedoftwosub-states(bothinitPandresPare stativeprojectionsinRamchand’ssystem).RamchandfollowsHale&Keyser(1993)inthatverbalpro- jectionsthatareinahead-complementrelation(e.g. (147a))giverisetoacausativerelationbetween the eventualities that such projections denote, the eventuality from the higher projection being in- terpreted as bringing about/ causing the eventuality from the lower projection (the causal relation representedwith‘!’in(147b),aswellasin(146c)). Attheend,theresultisamacroeventecomposed of smaller subevents that are causally related (see (147a), where e stands for either a stative or even- tiveeventuality). BytheEventCompositionRule,then,theconfigurationinitP>resPderivesastative causativepredicate,i.e.aneventualitycomposedoftwosub-statesthatarerelatedcausally. 43 43 Thenotionofstativecausationinthe AktionsartliteraturehasbeendevelopedbyauthorslikePylkkänen(2000), Arad (1998b) and Rothmayr (2009) to account for the stative readings of aspect-alternating verbs such as object-experiencer psychologicalverbs(e.g. worry,amuse,frighten...)andlocativeverbs(e.g. surround,cover,obstruct...).Myanalysisofgobernar- verbs differs from these in crucial respects: I posit a transitive syntax for stative causatives while these authors assume anunaccusativesyntax. Also, Idonotbarthepossibilityofanagentivesubject, unliketheseauthorsdofortheirstative causativestructures. SeeChapter5forananalysisofobject-experiencerandlocativeverbs,andparticularlySection5.4.3, whereIarguethattheseverbsarenotcausativeinSpanish. 79 Section3.3. Myproposal Chapter3 (147) a. V’ V VP b. e 1 !e 2 (Hale&Keyser1993:69) c. EventCompositionRule:e=e 1 !e 2 econsistsoftwosubevents,e 1 ,e 2 suchthate 1 causallyimplicatese 2 (Ramchand2008:44) A remark is in order about this structure: Fábregas & Marín (2017) review a similar proposal made by García-Pardo (2015), who proposes that gobernar-verbs are formed via two stative heads, which syntactically combined give rise to a stative causative configuration. I show the structure in (148). Fábregas & Marín (2017) reject this analysis on two main grounds: i) that it could not explain why gobernar-verbspasseventivitytests;ii)thatcombiningtwostativeheadstogetherwouldbeaninstance ofVacuousProjection. (148) VP V ST VP DP V’ V ST p Their first objection is refuted in Section 3.2.2.1, where I show that the eventivity tests are too inconsistenttowarrantaseparationofnon-dynamicpredicatesintoeitherstativeoreventive. With respecttothesecondpoint,IconcurthattheanalysisinGarcía-Pardo(2015)hasproblemsofitsown, which I will not get into here for reasons of space. However, I do believe that the basic idea can be translatedtoRamchand’s(2008)system,asIdohere. Inherfirst-phasesyntax,shedoesnotpropose thatinitandresareformallythesamehead:forinstance,sheexplicitlyproposesthatinitishigherthan resintheuniversalorderingofverbalheadsandthatinitPassignsaccusativecase,unlikeres(2008:56). In isolation, it is true that both initP and resP equally denote a state. However, although Ramchand doesnotdiscussstativecausativesinherwork,nothinginherarchitectureexplicitlypreventsaninitP > resP configuration: in fact, stative causatives are a natural extension of her system. As I see it, the burden is on detractors of a syntactically derived stative causative configuration (Ramchandian or not)toshowwhyitisnotviable. 3.3.3 Derivingthepropertiesofgobernar-verbs InSection3.3.1,IpresentedwhatIconsideredtobethedefiningtraitsof gobernar-verbs. Thesewere the scope ambiguity with again, the availability of adjectival passives and agentive subjects and the 80 Section3.3. Myproposal Chapter3 impossibility of undergoing the transitive-unaccusative alternation. In this section I show how my proposedstructurecanaccountforthatphenomena. AsfarasIcantell,Fábregas&Marín(2017)can accountforthatphenomenatoo,exceptforone:theanti-causativizationpattern. The scope effects of again are captured by positing that the adverb again may attach in the syn- tax either above both initP and resP (the wide scope reading in (149a)) or between initP and resP (the narrow scope reading in (149b)) (see also von Stechow (1995) for the original proposal that the scope ambiguitieswith almostinchange-of-stateverbsisderivedbythedifferentattachmentheightofthe adverb). Fábregas & Marín ’s proposal can also derive the scope ambiguity in the syntax, given that theyposittwodifferentsyntacticprojectionsthathosttheeventandtheassociatedresultstate(vand P cc ,respectively). (149) Scopeambiguityofagainwithstativecausatives a. Widescope: again>initP>resP b. Narrowscope: initP>again>resP The availability of adjectival passives can be explained by the fact that gobernar-verbs have an asso- ciated result state. If adjectival passives are resultative constructions, in the sense that they denote theresultstateoftheunderlyingeventualitywithintheparticiples(fromaprocesseventoracausing state,aswithgobernar-verbs),thenitisnomysterythatgobernar-verbscanbegoodinputsforadjecti- valpassives(seealsoKratzer(2000)andfootnote41). IencodetheresultstateinaRamchandianresP, whereasFábregas&Marín(2017)encodeitinP cc ,butthenotionisthesame. Myproposalalsoexplainsthe(anti-)causativizationfacts,i.e.thestricttransitivityoftheseverbs. SincestativecausativesinherentlyinvolveaninitPprojection—introducingthecausativestate—itfol- lowsthatanexternalargumentmustnecessarilybeintroducedinthestructure:otherwisewesimply cannothaveastativecausativepredicate.ThisissomethingnotpredictedinFábregas&Marín(2017). Rememberthattheirproposedstructuresinvolveavheadforevents,whichcanthenbedynamicor notdependingontheirPcomplement,asshownin(150). (150) StructuresforeventsinFábregas&Marín(2017) a. Non-dynamicevents(Gobernar-typeverbs) v DP v v P cc DP ... 81 Section3.3. Myproposal Chapter3 b. Dynamicevents v DP v v Path DP ... Ifthisvheadisthesameinbothstructures,thenitisamysterywhychange-of-stateverbsoractivities can participate in the alternation (remember (144) and (145)) but gobernar-verbs cannot. That is, why mustvnecessarilyintroduceasubjectwhenitscomplementisaP cc ,butnotwhenitscomplementis aPath?Ifthisvheadisthesameinbothstructures,thenitisamysterywhychange-of-stateverbsor activitiescanparticipateinthealternation(remember(144)and(145))butgobernar-verbscannot.That is, why must v necessarily introduce a subject when its complement is a P cc , but not when its com- plementisaPath(i.e. withdynamicverbsthatcanundergothetransitive-unaccusativealternation)? Theirproposalneedstostipulatethissince,asitstands,itcannotderiveitnaturally. 44 Afinalwordaboutagents.Thefactthatanimatesubjectsintheseconstructionsaretypicallyagents isstraightforwardlyexplainedbythecomplexcausativestructurethatIproposefortheseverbs:sim- plyput,ananimatecauserisnaturallyinterpretedasanagent. Fábregas&Marín(2017)characterize therelationshipbetweentheeventandtheassociatedresultasmaintain,ratherthancause,but,ter- 44 Thequestionremainsthatsimplestatesseemtobeunabletoundergothealternationeither.Thishasastraightforward explanationinRamchand’sframework. Forher,simplestatesareinstancesof initP,whichinisolationdenotesmerelya state,notacausationalsubevent.Sincethetransitive-unaccusativealternationisderivedinhermodelbythepresencevs. absenceofinitP,itisimmediatelyobviousthatatransitivestativeverbcannothaveanunaccusativecounterpart. (xviii) a. Berta Berta sabe knows geografía. Geography ‘BertaknowsGeography.’ b. initP DP Berta init’ init sabe DP geografía AnotherremainingissueiswhywecannothaveanalternationconsistingofabareresP—notjustwithstativecausatives, butwithtelicverbsaswell,inasmuchastheyalsocontain resPintheirdecomposition. Igobacktothisissueinthefinal chapterofthisdissertation(Section6.3). 82 Section3.3. Myproposal Chapter3 minologyaside,itisequallyeasytoseehowanimatesubjectsinthemaintainrelationcanbeagents. SeeSection3.4.2forspecificcriticismofthenotionofmaintain. Idowanttoemphasize,however,thatthepossibilityofhavinganagentivesubjectdoesnotconsti- tuteanargumentinfavorofthenon-dynamiceventanalysisinFábregas&Marín(2017). Itisknown at least since Dowty (1979) that some stative verbs such as sit, stand and lie can be perfectly agentive withhumansubjects. Also, recentworkonlocativeverbs(surround, cover, block...), whichcanhavea change-of-locationaswellasastativereading,showsthatthestativeversioncanequallyhaveagen- tive subjects (García-Pardo 2016b). Agentivity and eventivity are thus separate notions, and neither one implicates the other. The same conclusion is reached in Levin & Rappaport-Hovav (2005), who defendthatagentivityis‘orthogonaltoaspectualclassification,withagentiveandnonagentivepred- icatesbeingfoundineveryaspectualclass’(2005:89). 3.3.4 Comparisonwithatomicapproaches InSection2.3,Idiscussedthattherearetwomainapproachesintheliteraturewithrespecttothemod- elingofAktionsartinthesyntax:atomicanddecompositional.Atomicapproachespositdifferenttypes or‘flavors’ofvheadsthatsemanticallyencodethedifferenteventtypes,simpleorcomplex. Decom- positionalapproaches,ontheotherhand,assumethateventcomplexitymirrorssyntacticcomplex- ity, inasmuchaseventualitiesarebuiltcompositionallyinthesyntax. Myaccountof gobernar-verbs clearly belongs to the latter kind of approach: the stative causative relation is built compositionally viatwostate-denotingprojections,initPandresP.Inthissection,Iwillcomparemyaccountofstative causativestoanatomicone. An atomic approach would then posit that the stative causative meaning would be incoded in a syntactichead,callitv CAUSE .Atomicmodelsgenerallycomeintwotypeswithrespecttothesyntactic relationtheyassumebetweentheAktionsart-denotingheads:itiseitherintroducedinaseparatepro- jection(Kratzer1996,2000;Alexiadou et al. 2006;Schäfer2008, a.o.) oritisintroducedbythesame head(Arad2002;Folli&Harley2005,a.o.). Iadaptthesetwoapproachestostativecausativesin(151), where I assume that the internal argument is introduced as complement of the root, as is standard intheseapproaches. Thestructurescorrespondtothesentencein(146a), repeatedbelowforconve- nience. (146a) Berta Berta gobierna governs el the país. country ‘Bertagovernsthecountry.’ 83 Section3.3. Myproposal Chapter3 (151) a. Atomicapproach(I):Externalargumentandcauseinseparateprojections VoiceP DP Berta Voice’ Voice vP v CAUSE p P p gobierna DP elpaís b. Atomicapproach(II):Externalargumentandcauseinthesameprojection vP DP Berta v’ v CAUSE p P p gobierna DP elpaís Asitturnsout,neitheroneoftheseatomicapproachesisagoodfitforstativecausativesand,inpar- ticular, for gobernar-verbs. The first approach in (151a) cannot explain why it is that gobernar-verbs do not anticausativize, i.e. why could we not have unaccusative versions of these verbs by omitting VoiceP.Awayaroundthiswouldbetostipulatethattheseheadscomewithsomefeaturethatrequired theprojectionofVoiceP.Butifso,itwouldbeverysuspiciousthatawholeeventtyperequiredsuch featureandothereventtypes,likechange-of-stateverbs,didnot. The second option illustrated in (151b) would appear to better capture the properties regarding anticausativization: by having the projection of v CAUSE also introduce the external argument, we would capture the obligatory transitivity we have observed in gobernar-verbs. However, note that this account runs into the opposite problem with respect to the previous one: the mistery now is whychange-of-stateverbsoractivitiesmayanti-causativize,whichwouldrequiresuppresionofthe externalargument(e.g.bynotprojectingaspecifier). 84 Section3.4. Againstmaintain:causeandthetemporalrelationbetweenevents Chapter3 Furthermore, it is unclear how this second approach would capture the scope ambiguity with again. If, aswesaw, againcanscopeeitherovertheresultstatewithitsinternalargumentorhigher up,overthecausingstatewithitsexternalargument,thenwerunintocomplicationstocaptureboth readingssyntacticallyinastructurewheretheexternalargumentandtheresultstateareintroduced inthesameprojection.Atomicaccounts,inshort,areill-equippedtocapturethesyntacticproperties ofgobernar-verbs. 3.4 Againstmaintain:causeandthetemporalrelationbetweenevents Inthiswork, followingproposalsstartingwithLewis(1973), Iamtakingthelinguisticexpressionof causation—whichIrefertoascausefromnowon—tobearelationbetweenevents.Inlinguisticthe- ory,itisgenerallyassumed(butseldomexplicitlydefended)thatcauserequiresatemporalsequence ofthecausingeventandthecausedevent,mainlybecausethestudyofcausehasfocusedondynamic predicatesand,inparticular,change-of-stateverbs. However,someauthorshavearguedthatstative predicates can indeed involve acause relation between states, in which there would not be tempo- ralsequencingbutrathertemporalcoextensivity(Pylkkänen2000;Arad1998b,2002;Kratzer2000). 45 Thatis,thecaused/resultstatedoesnotholdafterthecausingstateholds,butratherbothstateshold strictlyatthesametimeinterval. Although,tothebestofmyknowledge,thisisthefirstworkwhereastativecausationapproach has been explicitly defended for gobernar-verbs, there is precedent of a piece of work that, arguing against the notion of stative causation, briefly considers gobernar-type verbs in English—let us call them govern-verbs—and ultimately rejects a causative analysis of these. Such work is Neeleman & vandeKoot(2012)andIwillreviewitcriticallyinwhatfollows. 3.4.1 causeandtemporalsequencing: Neeleman&vandeKoot(2012) Neeleman&vandeKoot(2012)discusshowcausationisrepresentedinlanguage.Theydefinecausa- tionnotjustasastrictlylinguisticnotion,butasageneralone,asin(152)—basingtheirdefinitionon Lewis(1973). (152) a. Causationisarelationbetweentwoevents:acausingeventandacausedevent. b. Causationhasatemporaldimension:thecausingeventmustprecedethecausedevent. 45 Theseauthorsmostlydrewtheirdatafromstativeobject-experiencerpsychologicalverbsand,toalesserextent,locative verbs. However,IwillargueinChapter5,Section5.4.3thatthesetwoverbtypesdonothaveacausecomponentintheir stativeversion. 85 Section3.4. Againstmaintain:causeandthetemporalrelationbetweenevents Chapter3 c. Causationiscounterfactual:ifthecausingeventhadnotoccurred,thecausedeventwould havenotoccurredeither. (Neeleman&vandeKoot2012:21) Forthem,causativeverbs—i.e. thosethatlexicallyencodecausation—arelinguisticallycomposedof twoingredients,(153a)and(153b)below.Underthisview,note,causationisacompositerelation,which ariseswhenthesetwofactorscombinetogether. (153) Linguisticcomponentsofcausativeverbs: a. Acrucialcontributingfactor(CCF),mergedexternally. b. Theculminationofaneventineitheranendstateorinaresultantactivity. Letusillustratetheirproposalwiththeexamplesin(154). (154) a. Johnbrokethewindow. b. Johnspunthebottle. c. Johnfell. ForNeeleman&vandeKoot(2012),only(154a)and(154b)arecausative. Bothinvolvetheculmination of an event: in the case of the change-of-state verb in (154a), the culmination is an end state, that of thewindowbeingbroken;inthecaseoftheactivityverbin(154b),theresultantactivityisthebottle spinning.Also,bothhaveaCCF,John,whoseparticipationiscrucialfortheresultanteventualitiesto comeabout:hedoessomethingthateithercausesthewindowtobreakorthebottletospin.Sentence (154c),however,isnotcausative: althoughthereisaculminationofaneventinaresultstate,thereis noapparentCCF. 46 TheauthorsthenmovetothesetofverbswhoseSpanishanaloguesIhavebeendiscussinginthis chapter(i.e. gobernar-verbs).Theygivetheexamplesin(155). (155) a. Thewallprotectsthecity. b. John’sunclesupportshimfinancially. c. Thebeamcarriesthewallaboveit. d. Thesheriffupholdsthelaw. (Neeleman&vandeKoot2012:39) Theyarguethat,althoughtheseverbsindeedhaveaCCF,theycannotbeanalyzedascausativebecause they lack the temporal dimension presented in (152b). Instead, they argue, the relation at play with theseverbsisoneofmaintenance,whichtheydefinein(156). 46 ButseeSection3.5.2,fortheviewthatunaccusativechange-of-stateverbsdohaveacausecomponent. 86 Section3.4. Againstmaintain:causeandthetemporalrelationbetweenevents Chapter3 (156) a. Maintenance is a relation between two eventualities: a maintaining state or event and a maintainedstate. b. Maintenancelacksatemporaldimension:themaintainingstateoreventmustbecontem- poraneouswiththemaintainedstate. c. Maintenance is counterfactual: if the maintaining state or event were absent, the main- tainedstatewouldnotexisteither.XXXXXXXXXX(Neeleman&vandeKoot2012:38–39) Likeothercausativeverbs,theseverbshavearesultstate,whichintheircaseisamaintainedstate(e.g. thestateofacitybeingkeptsafebymeansofaprotection). Thus,theauthorsclaim,theseverbscan be assigned very similar semantic representations: cause verbs would have the lexical semantics in (157a)andmaintainverbswouldhavethatin(157b). (157) a. yx[[ e x[ s ...y...]]&x=CCF] b. yx[[ s x[ s ...y...]]&x=CCF] (Neeleman&vandeKoot2012:41) The maintaining eventuality, the authors argue, can be construed as either a state or an event. For instance, in a sentence like (155a) the maintaining eventuality is clearly a state (that of the wall being inastrategiclocationfortheprotectiontohold);inasentencesuchas the cavalry protects the city,on theotherhand,thereadingofthemaintainedeventualityiseventive:thecavalryispresumablyriding aroundthecitytokeepitprotected. 3.4.2 ProblemswithNeeleman&vandeKoot(2012) As I see it, the reductive view of cause as necessarily involving temporal sequencing of the even- tualities involved leads into several empirical and theoretical problems. First of all, the requisite of temporalsequencingbetweenthecausingeventandthecausedeventwasnotincludedintheoriginal definitionofcausebyLewis(1973),despiteNeeleman&vandeKoot’s(2012)claimtothecontrary:all thatwaspresentinLewis’originalworkwerepoints(152a)—causationasarelationbetweenevents— and(152c)—causationascounterfactual. Itisperfectlylegitimate,ofcourse,toadoptadefinitionofcausethatexcludestemporalcoexten- sivity,astheauthorsdo. Doingso,however,raisesmanyproblematicquestions. First,onewonders what to do with transitive activities: they claim that sentences like (154b) are causative: something donebyJohn,theCCF,bringsabouttheresultantactivityofthebottlespinning.Theactivitydoneby John(e.g.holdingthebottlefrombothsides,twistinghishandsandlettinggo)precedesthespinning ofthebottle(i.e.thebottlespinsoncetheparticipationofJohninthewholeeventualityisconcluded). However, many other transitive activities (in fact, the majority of them) actually involve temporal coextensivitybetweenthecausingeventandtheresultantactivity,asin(158). 87 Section3.4. Againstmaintain:causeandthetemporalrelationbetweenevents Chapter3 (158) a. Johndrovethecar. b. Johnrockedthecradle. c. Johnshooktheshakers. d. Johnwheeledthewheelbarrow. Sentence(158a),forinstance,cannotmeanthatJohndidsomethingandafterhedidthatthecarmoved aroundasaresult: Johnhastobeactivelyinvolvedthroughouttheresultantactivity. Underthepro- posalinNeeleman&vandeKoot(2012),therelationinvolvedshouldbemaintain,ratherthancause. Butmaintainisdefinedasmaintainingastate, notanevent, sothisdefinitionshouldatleastbere- workedtoaccomodateresultanteventsaswell. Arelatedissueisthat,if,asNeeleman&vandeKoot(2012)contend,maintainverbscanalsohave an eventive (dynamic) maintaining eventuality (e.g. the cavalry protects the city), then it is confusing what the difference between these eventive versions and change-of-state verbs would be. In other words,theissueishowwewouldderivetheAktionsartdifferencesbetweenmaintainverbswithady- namicmaintainingeventandchange-of-statecauseverbswithadynamicmaintainingevent: under Neeleman&vandeKoot’s(2012)account,theirlexicalsemanticswouldbeasin(157a). (157a) yx[[ e x[ s ...y...]]&x=CCF] Awayaroundthiswouldbetoassumethatmaintainverbsneveractuallyhavedynamicmaintaining events,andthatthepurportedeventivereadingswithsentenceslikethecavalryprotectsthecitymerely involveworld-knowledgeaboutwhatcavalriesusuallydo(thisisindeedmytakeonthematter). Thedeeperproblemoftheirproposal,asIseeit,isthattheirdefinitionofcauseandmaintainis partly Aktionsart-based (cause requires temporal sequencing, maintain requires temporal coexten- sivity)yetinnowaydocauseandmaintaindetermineAktionsartinthisaccount—anditisnotclear whatdoesdetermineAktionsart.Inotherwords,causeandmaintainaremeredescriptorsofdifferent Aktionsart configurations between events that otherwise have the same property, i.e. ‘causal’ depen- dency,definedcounterfactually. Onewonders,then,whatistobegainedtheoreticallyfrompositing bothacauseandamaintenancerelation,orwhatexplanatorypowersuchatakehas. Asafinalobservation,notethattheirproposalalsofailstocapturetheexternalargumentasym- metriesbetweengovern-verbsandchange-of-stateverbsnotedinthischapter:alexicalsemanticssuch asthosein(157)doesnotderivewhygovern-verbsdonotallowforexternalargumentsuppressionbut (some)change-of-stateverbsdo. 88 Section3.4. Againstmaintain:causeandthetemporalrelationbetweenevents Chapter3 3.4.3 Analternativeproposal Asdiscussedintheprevioussection,thereappearstobenogoodreasontoassumebothacauseand amaintenancerelation.Letusrejecttheideaandassume,asIalreadysuggestedinSection3.3.2,that causedoesnotrequiretemporalsequencing,butalsoallowstemporalcoextensivity: whatitcannot allowfor,ofcourse,istemporalprecedenceoftheresulteventoverthecausingevent,becauseitwould benonsensical(anobservationalsomadebyNeeleman&vandeKoot2012).InRamchand’smodel,as wediscussed,eventsarebuiltcompositionallyinthesyntaxthroughevent-denotingprojectionsthat, incombinationwitheachother,derivecausalrelationsbetweenthem(see(147),repeatedbelow). (147) a. V’ V VP b. e 1 !e 2 (Hale&Keyser1993:69) c. EventCompositionRule:e=e 1 !e 2 econsistsoftwosubevents,e 1 ,e 2 suchthate 1 causallyimplicatese 2 (Ramchand2008:44) In this framework, crucially, cause (represented as ‘!’ in Ramchand’s framework) builds Aktion- sart-configurations: in fact, it is the only ingredient that glues events together to build aspectually- meaningfulmacroevents.Sincecause,inandofitself,doesnotdefinetemporalrelations,Ramchand posits that there are additional temporal coherence conditions on causally-related events. I present themin(159)and(160),slightlymodifiedfromRamchand(2008). 47 47 Ramchand’sproposalontemporalcoherenceisgivenin(xix)and(xx). (xix) Init-ProcCoherence: Givenadecompositione 1 !(e 2 !e 3 ),e 1 maytemporallyoverlape 1 . (xx) Proc-ResCoherence: Given a decomposition e 1 ! ( e 2 ! e 3 ), e 3 must not temporally overlap e 2 (although they may share a transition point). (FromRamchand2008:130) Ramchand’sdefinitionoftheProc-Rescoherencein(xx)isweakerthanthatstatedin(160).Ramchandarguesthatwhether theprocesseventnecessarilyabutstheresultstateornotdependsonwhetherthesamerootlexicalizesbothheads(e.g. breakinJohnbroketheglass)orjusttheprochead,theresheadbeingnulloridentifiedbyaseparateelement(e.g.theactivity verb run in a resultative construction such as Mary ran her shoes ragged). In the former case, the temporal dependence regardingthetransitionpointisstrict;inthelattercase,itisnot.Ramchandstatesthisrequirementin(xxi). (xxi) Temporaldependenceandlexicalidentification Temporaldependenceisrequiredbysubeventsidentifiedbythesamelexicalcontent. (FromRamchand2008:131) 89 Section3.4. Againstmaintain:causeandthetemporalrelationbetweenevents Chapter3 (159) Init-ProcCoherence: Givenadecompositions!e,smaytemporallyoverlape. (160) Proc-ResCoherence: Given a decomposition e! s, s must not temporally overlap e, but they share a transition point. How about stative causatives? Following Ramchand’s reasoning for the temporal interpretation of the sub-events in the first-phase syntax, could propose a temporal coherence condition for stative causativesasin(161). (161) Init-ResCoherence: Givenadecompositions 1 !s 2 ,s 1 ands 2 aretemporallycoextensive. Thiscoherenceconditionensuresthat,inaneventdecompositionsuchas(146c),repeatedbelow,the causingstateandtheresultstateareinterpretedastemporallycoextensive,suchthat,whenthecausing state ceases to hold (i.e. the state of Berta as aninitiator of the governing event), so will the result state(i.e.thecountrywillnolongerbeinagovernedstate). (146) a. Berta Berta gobierna governs el the país. country ‘Bertagovernsthecountry.’ b. [ initP Berta [ init’ <gobierna> [ resP elpaís [ res’ <gobierna> ]]]] c. s9s 1 ,s 2 [s=(s 1 !s 2 )&gobierna(s 1 )&participant(Berta,s 1 )&gobierna(s 2 )&participant(el país,s 2 )] Theseconditions, asfarasIcantell, aredescriptivelyadequate. Yettheissueremainsastowhether they have to be simply stipulated, or they can be derived naturally from the internal constituency of events and states. The issue of how to formally characterize events vs. states has been tackled bymanyauthors, mostnotablyfromanintervalsemanticsapproach(Bennett&Partee1972;Dowty 1979;Rothstein2004, a.o., andseethebriefdiscussioninSection2.2.1). Thechallengehereishowto I leave aside this issue here, since I will be focusing on verb phrases in which the root lexicalizes all the eventive heads. For the different temporal readings of resultatives, and the temporal effects of the different lexicalization possibilities withinthefirst-phrase, thereaderisreferredtoRamchand(2008,2014). Alsonote, finally, thatthedefinitionoftheInit- Proc coherence is in harmony with the empirical observation that, with transitive activity verbs, the participation of theexternalargumentintheeventualitymayeitherprecedetheprocessoroverlapitfullyaswellaspartially. Thiswas a problem with Neeleman & van de Koot’s (2012) analysis of causation as strictly involving temporal precedence of the causingeventoverthecausedevent,aswediscussedinSection3.4.2(seeinparticulartheexamplesin(158)). 90 Section3.5. Compatibilitywithotheraccountsofcauseandtheexternalargument Chapter3 formalizetheminaneventsemanticsframeworksuchastheonepursuedinthepresentwork,given thattheeventargumentsarenotassociatedtotemporalargumentsinthefirst-phasesyntax. Apossibleroutewouldbetodefineeventsvs. statesintermsofwhetherornottheydecompose intosmallerdifferentsubparts,recastinginevent-mereologicaltermstheideafromintervalsemantics thatpredicatesofeventsaretrueuptominimaltemporalsubintervals,whereaspredicatesofstatesare trueuptoinstants.Eventsaredivisiblesincetheirsubpartsaredistinct(theyinvolvechange),whereas statesareindivisiblesincetheyhavenodiscerniblesubparts. 48 Deliveringacoherentproposaltothis matterfallsbeyondthescopeofthisdissertation,andsoIleaveitasideforfutureresearch. 3.5 Compatibilitywithotheraccountsofcauseandtheexternalargument InthissectionIpresenttworecentapproachestothesyntaxofcauseandtheexternalargumentand discusshowtheyfitwiththeproposalputforthinthischapter. Thefirstonetobediscussedisthat ofHarley(2013),whoseparatesthesemanticintroductionoftheexternalargumentfromitssyntactic introduction. The second one is that of Pylkkänen (2002), who separates cause from the syntactic introductionoftheexternalargument. 3.5.1 VoiceP6=vP Recently, Harley (2013)—and see also Harley (2017) for further development—has proposed that the introduction of an external argument in the syntax is separate from semantic transitivity. She con- tendsthatVoicePisresponsibleforprojectingtheexternalargumentinthesyntax,in(Spec,VoiceP), butnotofintroducingitsemantically: rather,causational/agentivesemanticsareencodedin vP,the complementofVoice. InthisshedepartsfromauthorslikeChomsky(1995);Kratzer(1996);Marantz (1984,1997);Harley(1995);Folli&Harley(2005),a.m.o.,whoassumeabipartitestructurefortheVPas in (162), where the projection that introduces the external argument also introduces it semantically. Harley’stripartiteproposalissketchedin(163). 48 SeealsoHallman(2015)fortherecentproposalthatstatesdenotemomentsoftime,andarethusunbreakableintosmaller subparts. 91 Section3.5. Compatibilitywithotheraccountsofcauseandtheexternalargument Chapter3 (162) TheKratzer-Marantzapproach vP/VoiceP EA v/Voice’ v/Voice x... p P/VP (163) TheHarley(2013)approach VoiceP EA Voice’ Voice vP v x... p P Evidenceforherproposalcomesfromtheinteractionoforderandmeaningbetweenapplicativeand causative morphemes in Hiaki. Assuming the Mirror Principle (Baker 1985), whereby the linear or- derofmorphemesdirectlycorrelateswiththeirsyntacticposition,Harleynotesthattheapplicative morpheme-ria,headofanAppPwhichintroducesapplicativeargumentsinitsspecifier,followsthe causativemorphemeofthelexicalverb(e.g.(164),wherethecausativemorpheme-teisglossedasv o ). This is unexpected, given that it is independently known that the applicative argument in Hiaki is syntacticallybelowtheexternalargumentbutaboveallotherarguments. (164) Santos Santos Maria-ta Maria-acc kari-te-ria house-v o -appl ‘SantosisbuildingahouseforMaria.’ (Harley2013:43) A similar effect happens with the causative suffix -tua, which expresses a similar meaning to that of aperiphrastic/analyticalcausative. Onceagain, theapplicativemorpheme -riafollowsthecausative affix (e.g. (165)), which means that ApplP is actually sandwiched between the external argument and thecausativemorpheme. 49 Thestructurethenhastobelikein(166). (165) Nee I usi-ta child-acc avion-ta plane-acc ni’i-tua-ria-k. fly-caus-appl-prf ‘Imadethe(model)planeflyforthechild.’ (Harley2013:45) 49 Harley warns us, with respect to (165), that the applicative argument cannot be interpreted as pertaining to the flying event,buttothecausingevent. AnsentencesuchasThe plane flew for the childisnotpossibleinHiakibecausethesubject needstobeagentiveforanapplicativeargumenttobelicit. 92 Section3.5. Compatibilitywithotheraccountsofcauseandtheexternalargument Chapter3 (166) VoiceP DP Nee Voice’ ApplP DP iliusi-ta Appl’ vP VoiceP DP avion-ta Voice’ vP p ni’i v Voice v -tua Appl -ria Voice Harleyalsonotesthatpassivescausativeswith -tuaareperfectlylicit(e.g. (167)). Thecausativemor- pheme is present and yet the Causer argument is syntactically absent (Hiaki does not even have by- phrases). Similarly,passivesofverbswiththeovertcausativemorpheme -teareallowed. Thesedata arethenatoddswiththeviewinwhichthesameprojectionintroducescausativesemanticsandthe externalargument. (167) a. Aapo 3sg kaa not yo’o-taka old-being kuna-tua-wa-k. marry.f-caus-pass-prf ‘Shewasmadetomarry/marriedoffwhenshewasn’tveryold.’ b. Ume the.pl yoeme(m) man-(pl) hi’ibwa-tua-wa. eat-caus-pass ‘Themenarebeingfed’(Lit:‘Themenarebeingmadetoeat.’) (Harley2013:52) 93 Section3.5. Compatibilitywithotheraccountsofcauseandtheexternalargument Chapter3 (168) Aman There kari-te-wa house-v o -pass ‘Housesarebeingbuiltthere.’ (Harley2013:53) Further evidence for this tripartite structure comes from a type of indirect causative built with the morpheme-tevo(e.g.(169)),whichhasthepeculiaritythatitprecludestheprojectionofaCausee,and so we get a passive-like interpretation despite not having passive morphology. Note that passives of sentences with -tevo are indeed possible (e.g. (170)). The fact that the Theme Santos becomes the subjectofthepassive,stronglysuggeststhattheCausee,whichshouldintervenebetweenthepassive morphologyandtheTheme,istrulysyntacticallyabsent.Thecomplementof-tevo,then,isnotVoiceP butvP. (169) a. Nee I kari-te-tevo-k. house-make-caus.indir-prf ‘Ihadahousebuilt’(implied:bysomebody) b. Nee I uka the avion-ta plane-acc ni’i-tua-tevo-k. fly-caus-caus.indir-k ‘Ihadtheplaneflown.’(implied:bysomebody) (Harley2013:51) (170) Santos Santos hitto-tevo-wa-k. medical.care-caus.indir-pass-prf ‘(Somebody)hadSantostreated.’ (Lit:‘Santoswascausedtobetreated.’) (Harley2013:51) Harley’s(2013)proposalcouldappeartobeatoddswithmyown. Ihaveclaimedinthischapterthat theintroductionoftheexternalargumentisinherentlylinkedtothestativesubevent,whichinacon- figurationalaccountisoneoftheingredientsofcausationalmeaning. However,theyarecompatible: initP could be posited not to projecta specifier in the syntax, and hence not to beable to project an externalargumentbyitself(likeHarley’scausativevP).Itmust,however,introduceanexternalargu- mentsemantically,aswellasintroduceastativesubevent,sothatthedataIhavepresentedherethat evidencetheinherentlinkbetweentheexternalargumentandthestativesubeventcanbecaptured. TheexternalargumentwouldbeintroducedsyntacticallybyVoiceP,asinHarley’swork. Thestructurewouldlookasin(171),examplifiedfrom(146a),repeatedbelow.TheVPintheexam- ple is a stative causative, but the VoiceP > initP articulation is assumed to be the same for any other verbalpredicatesthatcontaininitPintheirdecomposition,regardlessofAktionsart(i.e.transitivesand unergatives).Thesemanticcompositionisprovidedin(172). 94 Section3.5. Compatibilitywithotheraccountsofcauseandtheexternalargument Chapter3 (146a) Berta Berta gobierna governs el the país. country ‘Bertagovernsthecountry.’ (171) VoiceP DP Berta Voice’ Voice initP init <gobierna> resP DP elpaís res’ res <gobierna> XP (172) a. JinitPK=xs9s 1 ,s 2 [s=(s 1 !s 2 )&gobierna(s 1 )&participant(x,s 1 )&gobierna(s 2 )&partic- ipant(elpaís,s 2 )] b. JVoiceK=Px[P(x)] c. JVoicePK=s9s 1 ,s 2 [s=(s 1 !s 2 )&gobierna(s 1 )&participant(Berta,s 1 )&gobierna(s 2 )&par- ticipant(elpaís,s 2 )] Harley’s proposal, then, is perfectly compatible with mine. What I commit to, to be clear, is to the impossibilityofhavingcauseandthe(semantic)externalargumentintroducedbydifferentsyntactic projections.Inotherwords,Iclaimthatastructuresuchas(173)doesnotexistinnaturallanguage. (173) x (x=externalargument) s (s=causingevent) ... In the following section, I examine another line of research that extends the cause relation to un- accusative(change-of-state)predicates,anddiscusshowitfitswiththegeneralarchitectureofevent structurearguedforthisfar. 95 Section3.5. Compatibilitywithotheraccountsofcauseandtheexternalargument Chapter3 3.5.2 Unaccusativecause Thereisaconsiderablebodyofworkthatproposesthatunaccusativeverbshaveacausecomponent, indeedpresentinallchange-of-stateverbs(Levin&Rappaport-Hovav1995;Chierchia2004;Pylkkä- nen2002;Kratzer2005;Alexiadouetal. 2006,a.m.o.).Itakeasapointofdeparturethesyntacticpro- posalinPylkkänen(2002).Pylkännenproposesthatchange-of-stateverbsarederivedwithaCauseP that denotes a relation between events (a process event and a result state). An example is given in (174).VoiceP,ontheotherhand,introducesanAgentandrelatesittothecausingevent,inthelinesof Kratzer(1996)(see(175),adaptedfromPylkkänen2002:79). (174) a. Theglassbroke. b. CauseP Cause VP V melt DP theice c. e9e’ [cause(e,e’)&Theme(theice,e’)] (175) a. Johnmeltedtheice. b. VoiceP DP John Voice’ Voice CauseP Cause VP V melt DP theice c. e9e’ [Agent(e,John)&cause(e,e’)&Theme(theice,e’)] Pylkkänen (2002) presents evidence from adversity causatives in Japanese, which cannot introduce externalargumentswith by-phrasesdespitehavingovercausativemorphology(e.g. (176)). However, theycannonethelessappearwitha by-phrasenamingacausingevent(e.g. (177)). Thissuggeststhat, 96 Section3.5. Compatibilitywithotheraccountsofcauseandtheexternalargument Chapter3 although the external argument is syntactically and semantically absent, there is nonetheless causa- tionalsemantics. (176) *Taroo-ga *Taroo-NOM Hanako-ni-yotte Hanako-BY musuko- son-ACC o die sin-ase-ta -CAUSE-PAST ‘Taro’ssonwascausedtodieonhimbyHanako.’ (Pylkkänen2002:83) (177) Taroo-ga Taroo-NOM sensoo-ni-yotte war-BY musuko-o son-ACC sin-ase-ta die-CAUSE-PAST ‘Taro’ssonwascausedtodieonhimbythewar.’ (Pylkkänen2002:82) Moreevidenceforthiskindofapproachtochange-of-stateverbsisgivenbyAlexiadouet al. (2006). They note that, although English unaccusatives do not allow by-phrases introducing agents or PP- instruments (e.g. (178a)), they nonetheless allow PPs headed by from which name the causing event (e.g.(178b)).TheynotethatthesameeffecttakesplaceinGerman(e.g.(179)). (178) a. *ThewindowbrokebyJohn/withastone. b. Thewindowcracked/brokefromtheexplosion. (Alexiadouetal. 2006:182) (179) a. Die the Vase vase zerbrach broke *von *by Peter/ Peter/ *mit *with dem the Hammer hammer b. Die the Vase vase zerbrach broke durch through ein an Erdbeben earthquakeXXXXXXXXXXx(Alexiadouetal. 2006:184–5) Theseapproaches,notewell,arefundamentallydifferentfromHarley’s(2013)account(despiteHarley’s acknowledgmentoftheirsimilarity). InHarley’saccount,theprojectionbelowVoicePdoesnotjust have causative semantics, but also an implicit external argument. However, in these unaccusative cause accounts, the projection below VoicePdoes not introducean external argument neithersyn- tacticallynorsemantically.Forinstance,theapplicative-riadiscussedbyHarley(2013)isincompatible with unaccusative predicates (e.g. (180)). This means that Appl cannot take Pylkännen’s CauseP as a complement,i.e.CausePisnotHarley’scausativevP. (180) *Uu *the tasa cup Maria-ta Maria-acc hamte-ria-k. break.intr-appl-prf ‘Thecupbrokefor/onMaria.’ (Harleyetal. 2009:44) WhattheunaccusativecauseprojectionaccountdoesistoderiveDowty’s(1979)andParsons’s(1990) CAUSEandBECOMEoperatorsfromasingleone,cause.UnderaDowty-Parsonsview,change-of- stateverbsarederivedbyaBECOMEoperatorthatdenotesatransitiontoaresultstate:whetherthe 97 Section3.5. Compatibilitywithotheraccountsofcauseandtheexternalargument Chapter3 verbistransitiveornotdependsonwhetherithasaCAUSEoperator. Inaneventsemanticsasthat ofParsons,havingCAUSE,asidefromentailingtheintroductionofanAgent,alsomeanshavingan extrasubevent. Underthisview, unaccusativechange-of-stateverbswouldhavetwosubeventsand transitiveoneswouldhavethree.Iexemplifythisin(181)). (181) TheDowty-Parsonsview: a. Johnbroketheglass. 9e,e’,e”[Agent(John,e)&CAUSE(e,e’)&BECOME(e’,e”)&Theme(theglass,e”)&break(e”)] b. Theglassbroke. 9e,e’[BECOME(e,e’)&Theme(theglass,e’)&break(e’)] Pylkkänen’sgeneralizedcauseapproach,ontheotherhand,positstwosubeventsinbothcases(e.g. (182)).causesubsumeswhatwastraditionallyunderstoodasBECOME,inthatitimplicatesaprocess event that culminates in a result state, understood in causal terms (this account is, again, restricted tochange-of-stateverbs). ItalsosubsumestheclassicCAUSEoperator,inthat,althoughitdoesnot involveanexternalargumentperse,itintroducesacausingevent,inwhichanexternalargumentcan beintegratedthematically(e.g.byKratzer’s(1996)EventIdentification). (182) Thegeneralizedcauseview: a. Johnbroketheglass. 9e,e’[Agent(John,e)&cause(e,e’)&Theme(theglass,e’)&break(e’)] b. Theglassbroke. 9e,e’[&cause(e,e’)&Theme(theglass,e’)&break(e’)] Thegeneralizedcauseapproach,asfarasIcantell,isalsocompatiblewithfirst-phasesyntax. Note that,althoughRamchand’smodelinvolvesamaximumofthreesubevents,asinParsons,shedoesnot assume two different operators, CAUSE and BECOME, but rather a singlecause/ ‘lead to’ relation, which glues subevents together. In the first-phase syntax model, there are two cause relations in a transitive change-of-state VP (the causational subevent in which the Initiator external argument is involved,whichbringsabouttheprocesssubevent,andthecausationalprocessthatleadstotheresult state).Inanunaccusativechange-of-stateconfiguration,weonlyhavetwosubevents(theprocessand theresultstate)andasinglecause,i.e.essentiallyPylkkänen’sgeneralizedviewofcause.Irepresent itschematicallyin(183). (183) a. [procP [resP ]] b. Xe 1 X!Xe 2 XXX(where‘!’representsthecauserelation) 98 Section3.6. Conclusions Chapter3 Pylkännen’s approach to cause, note well, is fundamentally atomic, in that cause is encoded in a dedicatedfunctionalprojection,CauseP.Infirst-phasesyntax,ontheotherhand,thecausesemantics arise configurationally from the combination of proc and res. The configurational approach has the welcome advantage that it does not need to posit multiple projections with multiple operators (e.g. DoPwithadooperatorforactivities,MaintainPwithamaintainoperatorforstativecausatives,and soonandsoforth). 3.6 Conclusions Thischapterhasanalyzedgobernar-verbs(e.g. (146a))asaseparateaspectualclassandhasarguedthat theyarestativecausatives. IhavearguedthatthisverbtypeisbuiltbyatransitiveVPstructurecom- posed of two syntactic projections: initP, which introduces the external argument, and resP, which introduces the external argument, as in (e.g. (146b)). This structure is interpreted semantically as a complexstativepredicatebuiltupoftwocausally-relatedsub-states: acausingstateandacausedor resultstate. Thecausingstatehastheexternalargumentasitssubject,byvirtueofwhichitisinter- preted as an Initiator, whereas the result state has the internal argument as its subject, by virtue of whichitisinterpretedasaResultee(e.g.(146c)). (146) a. Berta Berta gobierna governs el the país. country ‘Bertagovernsthecountry.’ b. initP DP Berta init’ init <gobierna> resP DP elpaís res’ res <gobierna> XP c. s9s 1 ,s 2 [e=(s 1 !s 2 )&gobierna(s 1 )&participant(Berta,s 1 )&gobierna(s 2 )&participant(el país,s 2 )] 99 Section3.6. Conclusions Chapter3 I contrasted my proposal to others that assume further subdivisions between events and states— namelystates, non-dynamiceventsanddynamicevents, asinFábregas&Marín(2017)—ordifferent flavorsofcausedependingontheAktionsartofthepredicate—essentiallyNeeleman&vandeKoot’s (2012) view. I argued that it is theoretically simpler and empirically more adequate to preserve the classic distinction between states and events in terms of dynamicity, which combined with a single generalizedcauserelation,cangeneratealltheAktionsartandargumentstructureconfigurationswith- outanyadditionalstipulations.Myproposalalsogivesfurthersupporttotheviewthatagentivitycan befoundacrossallaspectualclasses. 50 Furthermore, IcommittedtoaviewoftheVPwheretheintroductionoftheexternalargument is inherently linked to causative semantics—or more accurately, to the subevent introduced by init, whichisinterpretedascausativeifitmergeswithfurtherverbalprojections. Inthislight,Ianalyzed other work that severs cause from the introduction of the external argument, notably Pylkkänen (2002)andHarley(2013). Iconcludedthattheywereinfactcompatiblewithmine,inasmuchasthey allow to preserve the claim that both the external argument variable and the causative subevent are introducedbythesameprojectionandthuscannotbesevered. Thischapteralsoputsforthamoreinclusiveviewofresultativitythanisgenerallyfoundinthe literature.Lexicalresultativesarenotjustchange-of-stateverbslikebreakoropen—orverbsofscalar change,inthesenseofRappaport-Hovav&Levin(2010)—butalsostativecausativeverbslikethegov- ern-typeanalyzedhere,andarederivedbythesamestate-denotingprojection. Whetherthisstateis interpretedasbroughtaboutbyapreviouseventorsimplymaintaineddependsonitsconfiguration withrespecttotheothereventiveprojectionsthatcomprisetheVP. ThefollowingChapter4buildsontheseproposalsdevelopedhereinChapter3regardingthein- separability of the external argument and the causing state, as well as this comprehensive notion of resultativity. In Chapter 4, I explore how the architecture of event structure developed here deter- mines and constrains derivational morphology of verbs. I present a case study of stative participles inSpanish—withdiscussionofotherlanguagesaswell—andanalyzetheinteractionofparticiplefor- mationwiththeAktionsartandargumentstructureofresultativeverbs. 50 SeealsoSection5.5forfurtherdiscussiononagentivitywithstativeobject-experiencerpsychologicalverbs. 100 4 Stativeresultatives 4.1 Introduction ThischapterfocusesonadjectivalpassivesinSpanish.Iillustratethemin(184)forSpanish,wherethey appearinpredicativepositionlinkedtotheirsubjectbythecopularverbestar‘tobe’. (184) El the vaso glass está is estar {lleno/ {full roto/ broken cascado/ cracked coloreado/ colored escondido/ hidden vigilado/ surveilled protegido...}. protected ‘Theglassis{full/broken/cracked/colored/hidden/surveilled/protected...}.’ TheseconstructionsareoftencalledadjectivalpassivesintheEnglishliterature,giventhatthesepar- ticiples behave as adjectives according to many diagnostics (e.g. they allow un-prefixation and they areselectedforbyverbslikeseemorremain,whichselectforadjectives).Thetermsstativepassiveand resultativepassiveareoftenused,giventhatstativeparticiplesaretypicallytakentodenoteresultsfrom previousevents,especiallywhenthebaseverbistelic. Inthefirstdecadesofgenerativegrammar,itwasbelievedthatadjectivalparticipleswerederived bylexicalrulespriortotheirprojectioninthesyntax(Wasow1977;Levin&Rappaport1986),andsuch view is still upheld by some researchers (Horvath & Siloni 2008; Meltzer-Asscher 2011). The main evidenceinfavorofthisviewisthatadjectivalpassivesmayhavespecializedidiomaticmeaningfrom baseverbsthatdonothavethatmeaningintheiractiveorpassiveform(e.g. (185a), fromHorvath& Siloni2008:123). Anotherargumentisthattheadjectivalparticipledoesnotworksyntacticallylikea passiveinthattheThemesubjectseemstobetheexternalargument,andby-phrasesarenotallowed (e.g.(185b),fromMcIntyre2013:7). (185) a. Writteninwater,castinstone,mylipsaresealed,thedieiscast... b. Thedoorseemedbroken/opened/painted(*byMary). However,recentworkhaschallengedtheclaimthatadjectivalpassivesarelexicallyderived.Starting with Embick (2004), many authors working on syntactic approaches to morphology have gradually 101 Section4.1. Introduction Chapter4 arguedthatadjectivalparticiplesarenotasdifferentfromverbalpassivesintermsoftheirsyntactic complexity.Asidefromhavingaroot(Kratzer2000),Embick(2004)proposedthatadjectivalpartici- plescontainaverbalprojectionvPthatintroduceschangesemanticsandverbalizestheroot. Other authorsarguethattheseparticiplesfurtherprojectaVoicePthatintroducesanimplicitexternalar- gument(McIntyre2013;Hallman2013;Bruening2014forEnglish,Doron2013forHebrew,Alexiadou et al. 2014forGerman...). Theirmainevidencecomesfromthepossibilityofhaving by-phrasesand instrumentalsinadjectivalpassives(e.g.(186),fromBruening2014:380). (186) a. ...for300yearsthesegardenswereunseen,exceptbythefavouredfew. b. Theradioactivenucleotidesaresosmallthattheyremainunseen,evenwiththemostpow- erfulmicroscope. Someauthorsgoevenfurtherandproposethat, inadditiontothisfull-fledgedargumentstructure, adjectivalparticiplesfurtherprojectPass(ive)PandPerfP/AspP,whichbuildaverbalpassiveandin- troduceperfectiveorresultativeaspect,respectively.Onthisaccount,adjectivalparticipleswouldbe evenricherthanverbalparticiplesintermsofthestructuretheycontained.ThisisthetakeinBosque (2014)and,partially,inAlexiadouet al. (2015);Anagnostopoulou(2017). Thefullsyntacticarticulation ofanadjectivalparticiple,underthisview,wouldbeasin(187). (187) AdjP Adj PerfP PerfP PassP Pass VoiceP Voice vP v p Thischapterexaminesthegrammarofstativeparticipleswithparticularattentiontohowtheirstativ- itycomesabout,andinwhatsensetheyareresultative. Thecentralclaimisthattherearetwotypes of stative participles: The first one is derived from stative causative verbs (see Chapter 3) and con- tainafull-fledgedinitPandresPconfiguration(e.g. (188)). Thesecondoneisderivedfromatelicroot thatonlylexicalizesa resP(e.g. (188)). Thistruncatedstructure, Iargue, isstativebutnotresultative, since there is not a process event represented in the structure. Stative participles are not passive or perfectiveeither:theylacktherelevantfunctionalprojections. 102 Section4.1. Introduction Chapter4 (188) a. La the ciudad city está is estar vigilada. surveilled ‘Thecityissurveilled.’ b. CopulaP estar ‘be’ AdjP Adj <–da> initP init <vigila-> resP DP laciudad res’ res <vigila-> XP (189) a. La the ciudad city está is estar destruida. destroyed ‘Thecityisdestroyed.’ b. CopulaP estar ‘be’ AdjP Adj <-da> resP DP laciudad res’ res <destrui-> XP Ialsoshowthatpastparticiplesinattributiveposition,asin(190)—alsoknownasreducedrelativesin the literature—have very different properties from predicative past participles with estar ‘to be’ (e.g. (184)). The former are eventive verbal participles (e.g. (191)), whereas only the latter are true stative adjectivalparticiplesasin(188)and(189).Thisdistinctionisoftenblurredoroverlooked(Bosque2014; Bruening2014,a.o.),butitiscrucialforapropercharacterizationofstativeparticiples. 103 Section4.1. Introduction Chapter4 (190) El the vaso glass {lleno/ {full roto/ broken cascado/ cracked coloreado/ colored escondido/ hidden vigilado/ surveilled protegido...}. protected ‘The{full/broken/cracked/colored/hidden/surveilled/protected...}glass.’ (191) a. La the ciudad city destruida. destroyed ‘Thedestroyedcity.’ b. DP D La NP N ciudad PassP Pass <–da> initP init <destrui–> procP proc <destrui–> resP DP laciudad res’ res <destrui–> XP Thestructureofthischapterisasfollows.Section4.2presentsastateoftheartinthestudyofadjecti- valpassivesinSpanish.InSection4.3,Ipresentthedatathatshowsthesystematicdifferencesbetween predicativeandattributiveparticiples.Section4.4putsforthmytheoreticalproposal.InSection4.5,I discussby-phrasesandagent/event-orientedmodifiersinadjectivalpassivesderivedfromtelicverbs, whicharepotentialcounterexamplesformytheory. Section4.6undertakesareviewofpreviousac- countsforotherlanguagesandcomparesthemtomyown,andextendsthescopeoftheproposalwith adiscussionoftheparametricdifferencesbetweenadjectivalpassivescrosslinguistically. Section4.8 presentstheconclusionsofthischapter. 104 Section4.2. AdjectivalpassivesinSpanish:astateoftheart Chapter4 4.2 AdjectivalpassivesinSpanish:astateoftheart Forthisoverview,ItakeasapointofdeparturetheseriesofworksbyBosque(1990,1999,2014).Hepro- posestheclassificationforSpanishpastparticiplesshownin(192).Note,toavoidpotentialconfusions, thatBosqueclassifiesbothresultative(R-PPrts)andeventive(E-PPrts)participlesasverbalparticiples. Thisisunlikewhatisgenerallyassumedintheliterature, whereBosque’sR-PPrts, whileclaimedto becomplexineventiveterms,areassumedtobeadjectivalandoftenlabeled adjectival passives;what thereceivedliteraturecalls verbal passivesisforhimasubsetofverbalparticiples(VPPrts),whichhe callseventiveparticiples(E-PPrts). 51 IwillrespectBosque’s(2014)terminologyfortheoverviewofhis workinthepresentsection,butintheremainderofthischapterIwillsticktotheclassicdistinction betweenverbalandadjectivalpassives. Ialsorefertothelatterasstativeorresultativeinterchange- ably. 52 Howexactlyadjectivalpassivesarestativeandresultativewillbedevelopedinmyanalysisin Section4.4. (192) ClassificationofSpanishpastparticiplesinBosque(2014) PastParticiple (PPrts) Verbal (V-PPrts) Resultative (R-PPrts) Unbound (U-PPrts) Bound (B-PPrts) Eventive (E-PPrts) Adjectival (Adj-PPrts) The author (1990; 1999) further makes a morphological distinction between what he calls participios truncos ‘cut-short participles’ and full-fledged participles, or P-PPrts proper, illustrated in Table 4.2. Cut-shortparticiplesareperfectiveadjectivesthathaveverbalrootsandusedtobefull-fledgedver- 51 Iintuitthathischoiceofclassificationobeysthereceivedviewthatassociateseventstructurewithgrammaticalcategory: sinceresultativeparticiplesappeartohaveargumentstructureandaspectualcomplexity,itwouldfollowthattheymust beverbal.SeeChapter2andparticularlySection2.3.2.2,whereIcriticizethisviewatlength. 52 This is unlike Embick’s (2004) terminology for English, who distinguishes between stative and resultative depending ontheireventcomplexity.Seefootnote53. 105 Section4.2. AdjectivalpassivesinSpanish:astateoftheart Chapter4 bal participles in previous stages of the language, i.e. they could appear in verbal passives. These participlesarederivedfromtelicverbsandhavelosttheeventivemeaningofsuchverbandtheyonly denote its final state: Bosque labels such participles perfective adjectives in his 1990; 1999 works and Adj-PPrtsinhislater2014article. 53 Bosque argues that the shared perfectivity of perfective adjectives and participles can be shown by their ability to appear in absolute constructions (e.g. (193)). However, the eventive defectiveness of perfectiveadjectives,asopposedtothatofparticiples,becomesapparentwhenwetestthemagainst instrumental,agentiveandmannermodifiers(e.g.(194)). (193) a. {Llenado/ {filled lleno} full el the vaso glass hasta up.to el the borde, edge ... ‘Theglass(being){filled/full}uptothetop,...’ b. Una one vez time {seca/ {dry secada} dried la the toalla, towel ... ‘Oncethetowelwas{dry/dried},...’ c. Ya already {fijo/ {fixed fijado} fixed el the tornillo, nail ... ‘Thenailbeingalreadyfixed,...’ (FromBosque1999:280) (194) a. Un A cartel poster {fijado/ {sticked *fijo} *stuck con with una a brocha. brush ‘Aposter{sticked/*stuck}withabrush.’ Instrumental 53 SeeEmbick(2004)forasimilardiscussionforEnglish. IshowhisclassificationofEnglishparticiplesinTable4.1,from Embick(2004:358).HisstativeparticiplesareanalogoustoBosque’s(2014)Adj-PPrts. Table4.1:EnglishparticiplesinEmbick(2004) Root Stative Resultative Eventivepassive p bless bless-èd bless-ed bless-ed p age ag-èd ag-ed ag-ed p rot rott-en rott-ed rott-ed p sink sunk-en sunk-; sunk-; p shave (clean)-shav-en shav-ed shav-ed p open open-; open-ed open-ed p empty empty-; empti-ed empti-ed p dry dry-; dri-ed dri-ed 106 Section4.2. AdjectivalpassivesinSpanish:astateoftheart Chapter4 Table4.2:Cut-shortandfull-fledgedparticiplesinSpanish(adaptedfromBosque1990,1999) Cut-short Regular descalzo‘barefoot’ descalzadolit:‘barefooted’ desnudo‘naked’ desnudado‘undressed’ despierto‘awake’ despertado‘awaken’ disperso‘dispersed’ dispersado‘dispersed’ enfermo‘sick’ enfermado‘sickened’ exento‘exempt’ eximido‘exempted’ flojo‘loose’ aflojado‘loosened’ limpio‘clean’ limpiado‘cleaned’ liso‘flat’ alisado‘flattened’ lleno‘full’ llenado‘filled’ maduro‘ripe’ madurado‘ripened’ sano‘healthy’ sanado‘healed’ seco‘dry’ secado‘dried’ suelto‘loose’ soltado‘let.loose’ sujeto‘fixed’ sujetado‘fixed’ vacío‘empty’ vaciado‘emptied’ b. Un A cartel poster {fijado/ {sticked *fijo} *stuck por by el the bedel janitor de of la the facultad. school ‘Aposter{sticked/*stuck}bytheschooljanitor.’ Agentive c. Un A cartel poster {fijado/ {sticked *fijo} *stuck con with más more esfuerzo effort de of lo which que than parecía. seemed ‘Aposter{sticked/*stuck}withmoreeffortthanexpected.’ Manner (FromBosque1999:280) Note that there are irregular participles that look morphologically ‘cut-short’ yet syntactically and semantically function as V-PPrts too (e.g. (195)). Conversely, there are participles that do not have a ‘cut-short’formbutcannonethelessfunctionasAdj-PPrtsaswell,i.e. withouteventiveimplications (e.g. Table 4.3). What we never get are roots whose cut-short participles work as V-PPrts and their regularparticiplesasAdj-PPrts. 107 Section4.2. AdjectivalpassivesinSpanish:astateoftheart Chapter4 (195) IrregularparticiplesinSpanish Abierto ‘open’, absuelto ‘absolved’, adscrito ‘adhered’, cubierto ‘covered’, descrito ‘described’, dicho ‘said’,disuelto‘dissolved’,encubierto‘covered.up’,escrito‘written’,frito‘fried’,hecho‘done’,impreso ‘printed’,inscrito‘inscribed’,muerto‘dead/died’,provisto‘equipped’,puesto‘put’,resuelto‘resolved’, roto‘broken’,satisfecho‘satisfied’,visto‘seen’,vuelto‘returned’. Table4.3:SyncreticparticiplesinSpanish(fromBosque1999:55) Participle Meaningasaverbalparticiple Meaningasanadjectivalparticiple aislado ‘isolated’ ‘alone’ alargado ‘lengthened’ ‘long’ animado ‘encouraged’ ‘lively’ callado ‘silenced’ ‘quiet’ complicado ‘complicated’ ‘difficult’ divertido ‘amused’ ‘funny’ educado ‘educated’ ‘cultured,learned’ elevado ‘raised,upgraded’ ‘high’ equivocado ‘confused,takenwrong’ ‘wrong’ limitado ‘limited’ ‘short’ ocupado ‘occupied’ ‘busy’ reducido ‘reduced’ ‘small’ resumido ‘resumed’ ‘short’ Within V-PPrts, the first distinction is drawn between E-PPrts and R-PPrts. E-PPrts appear in ver- bal/ eventive passives, which take the auxiliary ser ‘to be’ in Spanish. They can be formed from any aspectualtypeofverb,i.e.eventiveorstative,solongastheverbistransitive(e.g.(196)). 54 (196) El the rey king fue was ser {aborrecido/ {abhorred perseguido/ persecuted ejecutado/ executed desterrado} banished (por (by sus his súbditos). subjects ‘Thekingwas{abhorred/persecuted/executed/banished}(byhissubjects).’ R-PPrts,ontheotherhand,areformedwithestar‘tobe’,andtheycanbeformedbothfromtransitive verbsandfromunaccusativeverbs(e.g.(197a)),butnotfromunergativeverbs(e.g.(197b)). 54 Notethatthismakestheterm‘eventive’passiveabitconfusing—i.e.verbalpassivizationdoesnotmakeastativepredi- cate‘eventive.’ 108 Section4.2. AdjectivalpassivesinSpanish:astateoftheart Chapter4 (197) a. La the fruta fruit está is estar {guardada/ {put.away pintada/ painted podrida/ rotten desaparecida}. disappeared ‘Thefruitis{putaway/painted/rotten/disappeared}.’ b. *Juan *Juan está is estar {trabajado/ {worked nadado/ swum bailado}. danced intended:‘Juanis{worked/swum/danced}.’ R-PPrtsarestativepredicatesderivedfromeventiveverbs,whichcometodenotetheresultstateof these verbs. Bosque (2014) argues for a notion of resultativity that is derived as an interaction of a hiddenhave,withperfectsemantics,andtheAktionsartofthebaseverb. Ifthebaseverbistelic,they giverisetoB-PPrtswhicharecharacterizedashavingaretrospectivereading,i.e.theresultstatefollows themainactiondenotedbytheverb(e.g.(198a));ifthebaseverbisatelic,theygiverisetoU-PPrtsand thereadingoftheresultstateissimultaneouswiththemainaction(e.g.(198b)). (198) a. El the niño child está is estar {castigado/ {grounded atrapado}. trapped ‘Thechildis{grounded/trapped}.’ B-PPrts b. La the ciudad city está is estar {vigilada/ {surveilled rodeada}. surrounded ‘Thecityis{surveilled/surrounded}.’ U-PPrts Althoughnotveryspecificaboutthestructure,Bosque(2014)arguesthatAdj-PPrtsaredevoidofeven- tive and argument structure (e.g. (199)). R-PPrts, on the other hand, contain a VP with an internal argument,whichmaybetelicoratelic,andthusderiveaB-PPrtoranU-PPrt,respectively. R-PPrts containanAspPthatencodesperfectaspect(Bosquecallsita‘hidden’have).Now,theunderlyingVP maybeunaccusative(e.g. (200))ortransitive,whichIrepresentasaVoicePin(201). Ifitistransitive, theR-PPrtcontainsanadditionalpassiveprojection,hostinga‘hidden’be(e.g.(201)). (199) Adj-PPrts(Noargumentstructure) AdjP Adj p (200) UnaccusativeR-PPrts PPrtP Prt AspP Asp have VP (a)telic 109 Section4.3. Comparingpredicativeandattributiveparticiples Chapter4 (201) TransitiveR-PPrts PPrtP Prt AspP Asp have PassP Pass be VoiceP Voice VP (a)telic BosquelargelyignoresE-PPrtsinhiswork. Althoughthefocusofthischapterwillbeadjectivalpas- sives(Bosque’sR-PPrts),Iwillnonethelessarguethatattributiveparticiplesareactuallyverbalpassives withoutacopula—i.e. Bosque’sE-PPrts. InSection4.3,Icomparethedifferentpropertiesofattribu- tiveandpredicativeparticiplesandinSection4.4Ipresenttheanalysisintermsofverbalvs.adjectival passives. 4.3 Comparingpredicativeandattributiveparticiples In this section I am going to analyze in depth the properties of PPrts in predicative and attributive position. Ialsorefertotheformeras estar-PPrtsandthelatterasbare-PPrts. AccordingtoBosque’s (2014)proposal, predicativePPrtswiththecopula estar areR-PPrtswithcomplexeventivestructure andtwoabstractmorphemes:apassivebeandaperfecthave.Thiseffectivelymeansthatthesentences in (202) should be equivalent in syntactico-semantic terms, at least in terms of event and argument structureandtemporalmeaning. (202) a. La the cortina curtain está is estar rasgada. torn ‘Thecurtainistorn.’ b. La the cortina curtain rasgada. torn ‘Thetorncurtain.’ c. La the cortina curtain ha has sido been rasgada. torn ‘Thecurtainhasbeentorn.’ 110 Section4.3. Comparingpredicativeandattributiveparticiples Chapter4 Thegoalofthissectionistoshowthatthisisincorrect.PPrtswithestarasin(202a)donothostabstract passiveandperfectmorphology,andnotevenchange-of-statestructures. Bare-PPrtsasin(202b),on theotherhand, docontaincomplexeventstructureandpassivemorphology. However, Ishowthat are not really perfects aspectually speaking, so we cannot claim that they are identical to sentences like(202c)either. 4.3.1 Change-of-statestructure? Asweintroducedintheprevioussection,Estar-PPrtshaveAktionsartrestrictionsthatverbalpartici- plesdonothave.Ihavearguedinthepast(García-Pardo2016a,2017)thatonlyverbsthathavearesul- tativecomponentintheireventdecompositionaregoodinputsforadjectivalpassives. Thisincludes telicverbs(e.g.(203)),butalsostativecausativeverbs(e.g.(204)),whichcontainacausative-resultative configuration (see Chapter3). Also goodinputs for estar-PPrts are object-experiencerpsychological verbs(e.g. (205a))andlocativeverbs(e.g. (205b)),discussedatlengthinChapter5. 55 Activityverbs(e.g. (206a))andtherestofstativepredicates(e.g.(206b)),ontheotherhand,cannotformestar-PPrts. (203) El the libro book está is estar {destrozado/ destroyed acabado/ finished pintado/ painted encuadernado}. bound ‘Thecityis estar {destroyed/finished/painted/bound}.’ (204) Los the trabajadores workers están are estar {vigilados/ {surveilled controlados/ controlled protegidos/ protected supervisados}. supervised ‘Thecityis estar {surveilled/controlled/protected/supervised}.’ (205) a. Pedro Pedro está is estar {preocupado/ {worried aburrido/ bored impresionado/ impressed asombrado}. astonished ‘Pedrois{worried/bored/impressed/astonished}.’ b. La The ciudad city está is estar {rodeada/ {surrounded flanqueada/ flanked cubierta/ covered cercada}. sieged ‘Thecityis{surrounded/flanked/covered/sieged}.’ (206) a. *El *the perro dog está is estar {acariciado/ {petted aplaudido/ applauded mecido/ rocked rascado}. scratched intended:‘Thedogis{petted/applauded/rocked/scratched}.’ 55 Iwillbylargeignoreestar-PPrtsderivedfromobject-experiencerpsychologicalverbsandlocativeverbsinthischapter. SeeSection6.3forsomediscussion. 111 Section4.3. Comparingpredicativeandattributiveparticiples Chapter4 b. *El *the rey king está is estar {odiado/ {hated aborrecido/ abhorred respetado/ respected mirado}. looked intended:‘Thekingis{hated/abhorred/respected/looked}.’ If R-PPrts inherit the Aktionsart of the base verb, we should expect the estar-PPrts in (203) to pass telicitytests. Thispredictionisnotborneout,however: Estar-PPrtsfailallthetelicityteststhatcan beappliedtoPPrts:theydonotacceptinxtimemodification(e.g.(207a))andtheydonotshowscope ambiguities with the adverbs again (e.g. (207b)) and almost (e.g. (207c)). Also, only the state, but not theprocessevent,islocatableintimeandspace(e.g. (207d)and(207e),respectively). 56 Thisdoesnot happen with bare-PPrts, which pass successfully all those tests and which permit the temporal and spatiallocationoftheunderlyingprocessevent(e.g.(208)). (207) a. *La *the casa house está is estar pintada painted en in cuatro four días. days intended:‘Thehouseispaintedinfourdays.’ b. El the espejo mirror está is estar cascado cracked de of nuevo. new ‘Themirroriscrackedagain.’ 3 Narrowscopereading(again>state):Themirrorisonceagaininacrackedstate. 7 Widescopereading(again>event):Themirrorhasundergoneacrackingeventonce again. c. La the casa house está is estar casi almost construida. built ‘Thehouseisalmostbuilt.’ 3 Narrow scope reading (almost > state): The house has been begun, and it is close to beingcompleted. 7 Widescopereading(almost>event):Theconstructionofthehousewasalmostbegun. d. El the espejo mirror está is estar cascado cracked {*hace {*makes dos two horas/ hours desde from hace makes dos two horas}. hours intended: ‘The mirror is cracked two hours ago/ The mirror has been cracked for two hours.’ 56 With respect to the spatio-temporal unlocatability of the process event, see von Stechow (1998); Gehrke (2012) for the sameobservationforGermanadjectivalpassives,alsoknownaszustandspassiv. Wewilldealwiththesemoreatlengthin Section4.5. 112 Section4.3. Comparingpredicativeandattributiveparticiples Chapter4 e. La the ropa clothes está are estar planchada ironed en in mi my dormitorio. bedroom ‘Theclothesareironedinmybedroom.’ 3 Theclothesarecurrentlyinmybedroom,inanironedstate. 7 Theclothesunderwentanironingeventinmybedroom. (208) a. Una a casa house pintada painted en in cuatro four días. days ‘Ahousepaintedinfourdays.’ b. Un a espejo mirror cascado cracked de of nuevo. new ‘Amirrorcrackedagain.’ 3 Narrowscopereading(again>state):Themirrorisonceagaininacrackedstate. 3 Widescopereading(again>event):Themirrorhasundergoneacrackingeventonce again. c. Una a casa house casi almost construida. built ‘Analmostbuilthouse/Ahousethatwasalmostbuilt.’ 3 Narrow scope reading (almost > state): The house has been begun, and it is close to beingcompleted. 3 Widescopereading(almost>event):Theconstructionofthehousewasalmostbegun. d. Un a espejo mirror cascado cracked {hace {makes dos two horas/ hours desde from hace makes dos two horas}. hours ‘Amirrorcrackedtwohoursago/Amirrorthathasbeencrackedfortwohours.’ e. La the ropa clothes planchada ironed en in mi my dormitorio. bedroom ‘The{ironedclothes/clothesironed}inmybedroom.’ 3 Theclothesarecurrentlyinmybedroom,inanironedstate. 3 Theclothesunderwentanironingeventinmybedroom. Thisisanintriguingparadoxof estar-PPrts: although,intuitively,asentencelike el espejo está cascado ‘themirroriscracked’meanssomethinglikethemirrorisinaresultstateofhaving(been)cracked,it sohappensthatwecannotfindtheprocesssubpartwhenwerummagewithinthePPrt. Notethatit doesnotsufficetoexplainthispuzzleawaybysayingthatthePPrtisstativized:evenifitisstativized, 113 Section4.3. Comparingpredicativeandattributiveparticiples Chapter4 itsunderlyingtelicstructureshouldbedetectedbylinguistictestsifitisreallypresentinthestructure: otherwiseitisnothingmorethanastipulation. 57 4.3.2 Externalarguments Another issue are external arguments. If estar-PPrts are passive, the underlying VP should perforce betransitive,howevermodelled.InGarcía-Pardo(2017),Iarguethatthereisanaspectualsplitwithin estar-PPrts in this sense: PPrts derived from stative causative verbs contain an implicit external ar- gument,whereasPPrtsderivedfromtelicverbsdonot(seeFernández-Ramírez1951;Gómez-Torrego 1988;Bosque1999;Fernández-Leborans1999;Conti-Jiménez2004foraprecedentofthisobsevation). EvidenceforthisisthatPPrtsderivedfromstativecausativeverbsallowby-phrasesunrestrictedly(e.g. (209))aswellasagent-orientedmodifierssuchasintentionaladverbs(e.g.(210))andinstrumentals(e.g. (211)). (209) a. El the recinto precinct está is estar custodiado surveilled por by diez ten guardias guards de of seguridad. security ‘Theprecinctissurveilledbytensecurityguards.’ b. Portugal Portugal está is estar gobernado governed por by Marcelo Marcelo Rebelo Rebelo de de Sousa. Sousa ‘PortugalisgovernedbyMarceloRebelodeSousa.’ (210) a. Los the trabajadores workers están are estar vigilados surveilled implacablemente. relentlessly ‘Theworkersarerelentlesslysurveilled.’ 57 For instance, the progressive and the perfect have been argued to be stativized predicates (see Vlach 1981; Moens 1987; Parsons 1990; Kamp & Reyle 1993 for the progressive, and Parsons 1990; Kamp & Reyle 1993; Nishiyama & Koenig 2010; Kampetal. 2015fortheperfect).Aswecanseein(xxii),whenthebaseverbsfortheseformsareeventivetheynonetheless retaintheireventiveproperties,diagnosedbytheposibilityofdynamicmanneradverbs(e.g.(xxiia))andinxtimemodifiers (e.g.(xxiib)). (xxii) a. María María está is corriendo running muy very rápido. fast Maríaisrunningveryfast.’ b. Pedro Pedro ha has reparado repaired el the ordenador computer en in tres three horas. hours ‘Pedrohasrepairedthecomputerinthreehours.’ 114 Section4.3. Comparingpredicativeandattributiveparticiples Chapter4 b. El the restaurante restaurant está is estar dirigido directed inteligentemente. intelligently ‘Therestaurantisintelligentlydirected.’ (211) a. Los the presos prisoners están are estar vigilados surveilled con with cámaras cameras de of seguridad. security ‘Theprisonersaresurveilledwithsecuritycameras.’ b. Los the aviones planes están are estar coordinados coordinated con with alta high tecnología technology ‘Theplanesarecoordinatedwithhightechnology.’ TheoppositesituationisfoundinPPrtsderivedfromtelicverbs. Theseparticiplesdonotallow by- phrases(e.g. (212)),agent-orientedadverbs(e.g. (213))andinstrumentals(e.g. (214))(butseeSection4.5 forfurtherdiscussion). (212) a. *Los *the neumáticos tires están are estar rajados slashed por by un a ladrón. thief intended:‘Thetiresareslashedbyathief.’ b. *La *the puerta door está is estar abierta opened por by Marcelo Marcelo Rebelo Rebelo de de Sousa. Sousa intended:‘ThedoorisopenedbyMarceloRebelodeSousa.’ (213) a. *La *the ciudad city está is estar implacablemente relentlessly bombardeada. bombarded intended:‘Thecityisrelentlesslybombarded.’ b. *La *the mesa table está is estar inteligentemente intelligently barnizada. varnished intended:‘Thetableisintelligentlyvarnished.’ (214) a. ??La ??the puerta door está is estar forzada forced con with una a barra bar de of hierro. iron intended:‘Thedoorisforcedwithanironbar.’ b. ??La ??the pared wall está is estar quemada burned con with una the antorcha. torch intended:‘Thewallisburnedwithatorch.’ Bare-PPrts,ontheotherhand,shownosuchAktionsartrestrictions. By-phrasesareperfectwithPPrts derivedfromtelicverbs(e.g.(215)),asareagent-orientedadverbs(e.g.(216))andinstruments(e.g.(217)). 115 Section4.3. Comparingpredicativeandattributiveparticiples Chapter4 (215) a. Los the neumáticos tires rajados slashed por by un a ladrón. thief ‘Thetiresslashedbyathief.’ b. La the puerta door abierta opened por by Marcelo Marcelo Rebelo Rebelo de de Sousa. Sousa ‘ThedooropenedbyMarceloRebelodeSousa.’ (216) a. Una a ciudad city implacablemente relentlessly bombardeada. bombarded ‘Arelentlesslybombardedcity.’ b. Una a mesa table inteligentemente intelligently barnizada. varnished ‘Anintelligentlyvarnishedtable.’ (217) a. Una a puerta door forzada forced con with una a barra bar de of hierro. iron ‘Adoorforcedwithanironbar.’ b. La the pared wall quemada burned con with una a antorcha. torch ‘Thewallburnedwithatorch.’ 4.3.3 Passives Awell-knownpropertyofverbalpassives,firstnotedinBakeretal. (1989)forEnglish,isthattheygive risetotheso-calleddisjointreferenceeffect:theinternalargumentandtheexternalargumentcannotbe coreferential.ThesameeffecttakesplaceinSpanish:verbalpassivescannothaveanexternalargument coreferentwithitsinternalargument(promotedtosubjectposition),as(218)shows. (218) El the niño kid fue was ser vestido. dressed ‘Thekidwasdressed.’ 3 Disjointreading:Someonedressedthekid. 7 Reflexivereading:Thekiddressedhimself. Kratzer (2002) noted that the disjoint reference effect does not hold with adjectival passives, which shetookasevidencethatthereisnoimplicitexternalargumentinadjectivalpassives(noVoiceP,in hertheory).IreproduceherexampleforSpanishin(219). 116 Section4.3. Comparingpredicativeandattributiveparticiples Chapter4 (219) Elescaladorestáaseguradoconunacuerda. theclimberis estar securedwitharope ‘Theclimberissecuredwitharope.’ 3 Disjointreading:Theclimberwassecuredbysomeone. 3 Reflexivereading:Theclimbersecuredhimself. ProponentsoftheVoicePaccountforadjectivalpassiveshavechallengedthisauthor’stestforother languages (McIntyre 2013; Bruening 2014 for English, Alexiadou et al. 2014 for German). In García- Pardo(inprep.),IprovideexperimentalevidencethatSpanishverbalandadjectivalpassivesdoinfact showthesamesplitobservedseparatelybyBaker et al. (1989)andKratzer(2000)forEnglish: verbal passivesshowthedisjointreferenceeffectandadjectivalpassivesdonot. MyworkinGarcía-Pardo(inprep.) forSpanish,aswellastheaforementionedliteratureonEn- glish and German, focused on PPrts derived from telic verbs. With estar-PPrts derived from stative causativeverbs,wefindaninterestingsituation.Whentheydonotappearwithaby-phrase,theread- ingofthenotionalexternalargumentisvague,i.e.itcanbeinterpretedasreflexiveordisjoint,depend- ing on the context (e.g. (220a) and (221a)). However, when the estar-PPrt has a by-phrase, it crucially cannot introduce a reflexive referent (e.g. (220b) and (221b)). In this respect, they behave like verbal passives. We cannot test whether this also holds for estar-PPrts derived from telic verbs since, as I discussed,theirdistributionofby-phrasesisveryrestricted. (220) a. Pedro Pedro está is estar coordinado coordinated con with los the otros other bailarines. dancers ‘Pedroiscoordinatedwiththeotherdancers.’ 3 Disjointreading:SomeonecoordinatedPedrowiththeotherdancers. 3 Reflexivereading:Pedrocoordinateshimselfwiththeotherdancers. b. Pedro Pedro está is estar coordinado coordinated {por {by el the director/ director *por *by sí him mismo}. self ‘Pedroiscoordinated{bythedirector/*byhimself}.’ (221) a. María María está is estar controlada. controlled ‘Maríaiscontrolled.’ 3 Disjointreading:SomeoneorsomethingcontrolsMaría. 3 Reflexivereading:Maríacontrolsherself. 117 Section4.3. Comparingpredicativeandattributiveparticiples Chapter4 b. María María está is estar controlada controlled {por {by una a colonia colony de of extraterrestres/ extraterrestrials por *by sí her misma}. self ‘Maríaiscontrolled{byacolonyofextraterrestrials/*byherself}.’ Bare-PPrts, on the other hand, differ with respect to the disjoint-reference effect depending on the base Aktionsartofthebaseverb. Withtelicsandstativecausativeverbs,thereisnodisjointreference effect without a by-phrase (i.e. just like we saw with estar-PPrts), as we can see in (222a) and (223a). With a by-phrase, however, the disjoint reference effect surfaces, just like with verbal passives (e.g. (222b)and(223b)).Withnon-causativestates,asitturnsout,thedisjointreferenceholdsbothwithand withoutaby-phrase(e.g.(224)).Thisstronglysuggeststhatthetworeadingsin(222a)areduetothefact that the participle is ambiguous between a verbal and an adjectival passive: when the participle can onlybeverbalasin(224a)—rememberthatadjectivalpassivesdonottakesimplexstates—thedisjoint referenceeffectsurfacesagain. (222) a. Un a niño kid vestido. dressed ‘Adressedkid.’ 3 Disjointreading:Someonedressedthekid. 3 Reflexivereading:Thekiddressedhimself. b. Un a niño kid vestido dressed {por {by su his madre/ mother *por *by sí him mismo}. self ‘Akiddressed{byhismother/*byhimself}.’ (223) a. Unos some actores actors coordinados directed {por {by Pedro/ Pedro *por *by sí them mismos}. selves ‘Someactorsdirected{byPedro/bythemselves}.’ b. Unos some actores actors dirigidos directed {por {by Pedro/ Pedro *por *by sí them mismos}. selves ‘Someactorsdirected{byPedro/bythemselves}.’ (224) a. Un a criminal criminal odiado. hated ‘Ahatedcriminal.’ 3 Disjointreading:Someonehatedthecriminal. 7 Reflexivereading:Thecriminalhatedhimself. 118 Section4.3. Comparingpredicativeandattributiveparticiples Chapter4 b. Un a criminal criminal odiado hated {por {by todos/ everyone *por *by sí him mismo}. self ‘Acriminalhated{byeveryone/byhimself}.’ 4.3.4 Perfectsemantics ThislastsectiontakesacloserlookattheclaimthatSpanishadjectivalpassives(R-PPrts,inBosque’s terminology)containahiddenperfect, inparticularaperfectofresult. Perfectsareusuallytakento have three interpretations (McCawley 1971, 1981; Mittwoch 1988; Michaelis 1994; Iatridou et al. 2001; Pancheva2003,a.o.),asillustratedin(225). Inthe universalreadingillustratedin(225a),whichistrig- geredbyadverbialslikesincextimeoralways,theinterpretationisoneinwhichtheeventualityholds from the time denoted by the adverbial until the reference time—which, in the case of the present perfect, coincides with the utterance time. In the experiential reading in (225b), which is the reading obtainedbydefault, theeventualityhashappenedatapasteventintimeanditdoesnotholdatthe reference time. The result reading in (225c) asserts that the result state of the event described by the verbstillholdsinthereferencetime. Theresultreadingisrestrictedtotelicverbsandtheuniversal readingtoeitherstativeverbsordynamicverbsintheprogressiveform. (225) a. Paula Paula ha has vivido lived en in Francia France desde since el the 2012. 2012 ‘PaulahaslivedinFrancesince2012.’ Universal b. Fernando Fernando ha has vivido lived en in Polonia Poland en in numerosas numerous ocasiones. occasions ‘FernandohaslivedinPolandinseveraloccasions.’ Experiential c. Felipe Felipe ha has llegado arrived a at Los Los Ángeles Angeles hace makes nada. nothing ‘FernandohasjustarrivedinLA.’ Resultative Tobesure,thepredictionisthatestar-PPrtsandbare-PPrtspatternwithverbalpassivesintheperfect ofresultintermsoftemporalinterpretation. Asitturnsout,thispredictionisnotborneout: verbal passives in the perfect do not behave like estar and bare-PPRts in temporal terms. For one, estar- PPrtsdonotaccepttemporalmodificationoftheeventinthepast(e.g.(226c)),whereasbare-PPrtsand perfectsofverbalpassivesdo(e.g.(226a–b)). (226) a. La the ciudad city ha has sido been ser bombardeada bombarded hace makes tan much solo only unos some minutos. minutes ‘Thecitywasbombardedjustafewminutesago.’ 119 Section4.3. Comparingpredicativeandattributiveparticiples Chapter4 b. Una a ciudad city bombardeada bombarded hace makes tan much solo only unos some minutos. minutes ‘Acitybombardedjustafewminutesago.’ c. *La *the ciudad city está is estar bombardeada bombarded hace makes tan much solo only unos some minutos. minutes ‘Acitybombardedjustafewminutesago.’ Second,the(Peninsular)Spanishperfectdisallowspositionaladverbialsthatlocatetheeventpriorto thesamedayasreferencetime. 58 However,bare-PPrtsallowforsuchadverbials,as(227b)shows. (227) a. *La *the ciudad city ha has sido been ser bombardeada bombarded ayer. yesterday intended:‘Thecityhasbeenbombardedyesterday.’ b. La the ciudad city bombardeada bombarded ayer. yesterday ‘Thecitybombardedyesterday.’ c. *La *the ciudad city está is estar bombardeada bombarded ayer. yesterday intended:‘Thecityisbombardedyesterday.’ Also,thesalienttemporalreadingsofverbalpassivesofperfects,bare-PPrtsand estar-passivesdiffer notably in the absence of temporal adverbs. Let us look at PPrts derived from telic verbs first: the verbalpassivein(228a)andthebare-PPrtin(228b)areambiguousbetweenanexperientialandresult reading—orvague,ifonetakestheresultreadingtobeapragmaticeffect. 59 Theonlypossiblereading oftheestar-PPrtin(228c),ontheotherhand,isoneinwhichthebuildingiscurrentlyempty. (228) a. El *the edificio building ha has sido been ser evacuado. evacuated intended:‘Thebuildinghasbeenevacuated.’ b. El the edificio evacuated evacuado. building ‘Theevacuatedbuilding.’ 58 Fordiscussiononthe24-hourruleandthehodiernalperfectinSpanish, seeSchwenter(1994);Giorgi&Pianesi(1997); Brugger(2001);Xiqués(2012). 59 ThisisindeedthetakeinMcCawley(1981);Mittwoch(1988),whoclassifyboththeExperientialandtheResultativeperfect asExistentialperfects. 120 Section4.3. Comparingpredicativeandattributiveparticiples Chapter4 c. El the edificio building está is estar evacuado. evacuated intended:‘Thebuildingisevacuated.’ With PPrts derived from stative causative verbs, this divide is more evident. With the perfect of a verbalpassive(e.g.(229a)),thedefaultreadingisoneofanteriority:thebuildinghasbeensurveilledin thepast,anditisnolongerundersurveillance. Bare-PPrtsandestar-PPrtsbehavequitedifferentlyin thisrespect:thedefaultreadingof(229b)andtheonlypossiblereadingof(229c)isoneofsimultaneity: thebuildingiscurrentlyundersurveillance. (229) a. El the edificio building ha has sido been ser vigilado. surveilled intended:‘Thebuildinghasbeensurveilled.’ b. El the edificio building vigilado. surveilled ‘Thesurveilledbuilding.’ c. El the edificio building está is estar vigilado. surveilled ‘Thebuildingissurveilled.’ Apointofconvergenceofthesethreeconstructionsisthepossibilityof since-adverbialswithPPRts derivedfromstativecausativeverbs.In(230a),thesince-adverbialdeliversaUniversalperfectreading (Iatridouetal. 2001)incombinationwiththestativePPrt:thesince-adverbiallocatesthestartingpoint oftheeventuality,whichextendstoreferencetime. Aswecansee,thebare-PPrtsand estar-PPrtsin (230b)and(230c)alsoallow since-adverbials. Butthisisnotevidencethatbare-PPrtsand estar-PPrts areperfects,butratherthattheyarestative,sincestativepredicates,verbalornot(aswellashabitual activities),canacceptsince-adverbials(e.g. (231)). Inshort,wereallyhavenosolidevidencethatbare- PPrtsandestar-PPrtscontainahiddenperfectintheirdecomposition. (230) a. El the país country ha has sido been gobernado governed por by la the extrema extreme derecha right desde since 2011. 2011 ‘Thecountryhasbeen ser governedbytheextremerightsince2011.’ b. Un a país country gobernado governed por by la the extrema extreme derecha right desde since 2011. 2011 ‘Acountrygovernedbytheextremerightsince2011.’ 121 Section4.3. Comparingpredicativeandattributiveparticiples Chapter4 c. El the país country está is estar gobernado governed por by la the extrema extreme derecha right desde since 2011. 2011 intended:‘Thecountryisgovernedbytheextremerightsince2011.’ (231) Pedro Pedro {está {is estar enfermo/ sick vive lives en in Alemania/ Germany corre runs por in las the mañanas} mornings} desde since la the semana week pasada. past ‘Pedro{issick/livesinGermany/runseverymorning}sincelastweek.’ In short, perfects of verbal passives behave very differently in temporal semantic terms from estar- PPrts and bare-PPrts. Note that we do not even have evidence that there is a subset of the present perfect readings inside these PPrts, since the tests that diagnose such readings (e.g. since-adverbials for the universal perfect, or locative adverbials for the experiential perfect) are not exclusive to the perfect. Estar-passives, however, do look a lot like perfects of result at first sight: 60 However, I will argue in Section 4.4.1 that estar-PPrts do not contain resultative aspect, but just a stative projection. Notealsothat,whereastheresultativeperfectallowsfortemporalmodificationoftheprocessevent (e.g.(232a)),estar-PPrtsdonot(e.g.(232b)). (232) a. El the edificio building ha has sido been ser evacuado evacuated hace makes unos some instantes. instants ‘Thebuildingwasevacuatedsomeinstantsago.’ b. *El the edificio building está is estar evacuado evacuated hace makes unos some instantes. instants intended:‘Thebuildingisevacuatedsomeinstantsago.’ 4.3.5 Summaryoffindings This section has provided an overview of the differences between perfects of verbal passives—E- PPrts—andPPrts,withandwithoutestar.Ihaveshownthatperfectsofverbalpassivesandbare-PPrts behave alike with respect to event and argument structure, and appear to have a rich structure that goes all the way to passivization. Estar-PPrts are impoverished in comparison, and they appear to denotemerestates.IpresentasummaryofthedatadiscussedsofarinTable4.4. 61 60 Indeed, accounts positing that estar-PPrts have—at least sometimes—a perfect of result has been proposed for other languages (Kratzer 2000; Anagnostopoulou 2003; Alexiadou et al. 2015; Anagnostopoulou 2017, a.o.). See García-Pardo (2016a,2017),aswellasSection4.6.2.4,forcriticismofthisview. 61 In Table 4.4, IA stands for internal argument, EA for external argument, StC for stative causatives and N.A. for not applicable. 122 Section4.4. Myproposal Chapter4 Table4.4:Propertiesofestar-PPrts,bare-PPrtsandperfectE-PPrts Thebaseverb ThePPrt Passive Perfect Event structure Argument structure Event structure Argument structure haber‘have’+PPrt all all all all N.A. 3 Bare-PPrt all IA all all 3 7 Estar-PPrt telic IA state noEA 7 7 StC transitive 3 4.4 Myproposal 4.4.1 Estar-PPrts Iarguethatestar-PPrtscomeintwotypesdependingontheAktionsartofthebaseverb:ifthebaseverb istelic,theestar-PPrtdenotesasingleunaccusativestate,encodedinresP(e.g.(233)).Ifthebaseverbis astativecausative,theestar-PPrthasafull-fledgedtransitiveverbalstructure,comprisedofinitPand resP(e.g. (234)). UnlikeBosque(2014),Iassumethatestar-PPrtsderivedfromtelicverbsareadjectival, derivedbyanAdjPprojection—rememberthatforBosqueonlyperfectiveadjectivesareadjectival. (233) a. La the ciudad city está is estar destruida. destroyed ‘Thecityisdestroyed.’ b. CopulaP estar ‘be’ AdjP Adj <-da> resP DP laciudad res’ res <destrui-> XP (234) a. La the ciudad city está is estar vigilada. surveilled ‘Thecityissurveilled.’ 123 Section4.4. Myproposal Chapter4 b. CopulaP estar ‘be’ AdjP Adj <–da> initP init <vigila-> resP DP laciudad res’ res <vigila-> XP Let us look at estar-PPrts derived from telic verbs first, and why they do not retain their event and argumentstructure. IarguethattheAdjheadimposestherestrictionthatthePPrtitcombineswith denotes a state. The head Adj can thus be modeled as a partial identity function whose domain is restrictedtopredicatesofstates(butseeSection4.6.3.2,whereIrefinethisidea). (235) JAdjK=P <s,t> [s.P(s)] In addition, I propose that Adj has a selectional feature [ures] that needs to be matched with a resP projection. 62 These two traits of Adj ensure that it will neither take a complement that denotes a dynamicevent—telicoratelic—norastatethatdoesnotcontainresPinitsdecomposition—i.e. pure initPverbslikeown,love,fear. Theissueishowatelicverbappearsinthesyntaxinatruncatedform,i.e.asamereresP(e.g.(233b)). Notethat,infirst-phasesyntax,verbsarenotspecifiedforaparticularAktionsartintheirlexicalentry: rather,theirlexicalentryisspecifiedforthecategoryfeaturesthatitcanlexicalizeinthesyntax,which is the domain of grammar that ultimately builds meaningful event structure configurations. A verb likedestruir‘destroy’,forinstance,hasthelexicalentryin(236). (236) destruir‘destroy’ categoryfeatures:[init,proc,res] Inthissystem,however,theverbalrootdoesnotneedtolexicalizeallthematchingcategoriesofits lexicalentryinthesyntax. Ramchand(2008), inspiredbytheSupersetPrinciple(ofwhichIgivemy 62 SeeFolli&Harley(2016)fortheproposalthatalsomakesuseofaselectional[ures]featuretoderivemanner-alternation parametersinItalianandEnglish. 124 Section4.4. Myproposal Chapter4 own wording in (237)), proposes that category features of verbs may underassociate, i.e. the may not associatewithaparticularprojectioninthetree. (237) TheSupersetPrinciple(adaptedfromStarke2009;Caha2009) A morphological exponent can lexicalize a syntactic tree if its lexically-stored features are a supersetofthefeaturesofthatsyntactictree. Going back to my estar-PPrts derived from telic verbs, my proposal is that these roots leave their proc(andinit,iftheyhaveit)featuresunderassociated. 63 Thesyntacticstructureofestar-PPrtscannot includeaprocP,givenitmustbestative.However,sinceitincludesaresP,anyverbthathasacategory featureresinitslexicalentrymayinprinciplelexicalizethesyntacticheadresleavingitsothercategory features—ifany—underassociated.Thatisthecase,Iargue,withestar-PPrtsderivedfromtelicverbs. Irepresentthestructurein(238b). (238) a. La the ciudad city está is estar destruida. destroyed ‘Thecityisdestroyed.’ b. CopulaP estar ‘be’ AdjP Adj <–da> resP DP laciudad res’ res i <destrui–> [init,proc,res i ] XP c. s[participant(laciudad,s)&destruida(s)] Crucially,then,mystructuredoesnotinvolveacomplexeventstructureforthebaseverbnorresul- tativesemanticsforthePPrt. Thus,itstandsinstarkcontrasttoaccountsthatassumeanunderlying 63 IadoptalessrestrictiveviewoftheUnderassociationprinciplethanproposedinRamchand. ForRamchand,underas- sociatedfeaturesmustbelicensedbyagreewithmatchingsyntacticfeatures,i.e. theprojectionsassociatedtocategory features of an exponent must be present even if lexicalized by a different exponent. I, on the other hand, assume that underassociatedfeaturesneednotbelicensedinthesyntax. 125 Section4.4. Myproposal Chapter4 telicconfigurationforthePPrt(Bosque2014;Arche et al. 2017forSpanish,seealsoKratzer2000for GermanandEmbick(2004)forEnglish)andaresultativesemanticsforthePPrt/adjectivizer(Bosque 1990forSpanish,aswellasKratzer2000;Maienborn2009forGerman). Thisproposalmakesthepredictionthattherearenoeventimplicationsin estar-PPrtstructures. Thispredictionisborneout,asshownintheexamplesin(239).TheseexamplesareadaptedfromGese (2011:260),whomakestheanalogousobservationforGerman. (239) a. El the tubo tube bronquial bronchial izquierdo left siempre always ha has estado been estar contraído. constricted ‘Theleftbronchialtubehadalwaysbeenconstricted.’ b. El the escultor sculptor esculpe carves formas forms a at partir from de of bloques blocks de of granito granite que that él he cree believes que that siempre always han have estado been estar escondidas hidden dentro inside de of ellos. them ‘Thesculptorcarvesshapesoutofblocksofgranitethathebelieveshavealwaysbeenhid- deninsideofthem.’ c. Los the párpados eyelids del of.the murciélago bat común common todavía still están are estar cerrados closed al at.the nacer. being.born ‘Theeye-lidslitsofthevesperbatarestillclosedatbirth.’ Letusmovetoestar-PPrtsderivedfromstativecausativeverbs. Stativecausativeverbshavealexical entry as in (240). Given that estar-PPrts may project both initP and resP, a verb with the category featuresinitandresmaylexicalizethecorrespondingsyntacticheads.Iprovideanexamplein(241).In thesemanticformulain(241c),Iassumethattheexternalargumentvariableisexistentiallyquantified intheabsenceofaby-phrase. (240) vigilar‘surveil’ categoryfeatures:[init,res] (241) a. La the ciudad city está is estar vigilada. vigilada ‘Thecityissurveilled.’ 126 Section4.4. Myproposal Chapter4 b. CopulaP estar ‘be’ AdjP Adj <–da> initP init i <vigila–> [init i ,res] resP DP laciudad res’ res i <vigila–> [init,res i ] XP c. s9s 1 ,s 2 ,x[s=(s 1 !s 2 )&vigila(s 1 )&participant(x,s 1 )&vigila(s 2 )&participant(laciudad,s 2 )] Theoutcome,then,isapassivePPrtwiththeauxiliaryestar. Thisfull-fledgedeventstructureallows for by-phrases and agent-oriented modification generally. Now, there is also a third possibility for estar-PPrts,namelythatchange-of-stateverbswithinitandrescategoryfeatures(e.g. (236))lexicalize apassivestructurelikethatin(241b),leavingits proccategoryfeatureunderassociated. 64 Iarguethat thisisapossibility,anditisindeedtheunderlyingstructureofthesentencesin(242). (242) a. Londres London está is estar contaminada contaminated {por {by el the plomo/ lead *por *by Moriarty}. Moriarty ‘Londonispolluted{bylead/byMoriarty}.’ (FromFábregas2014) b. El the camino path está is estar cortado cut por by los the manifestantes. demonstrators ‘Thepathiscutbythedemonstrators.’ (FromNavas-Ruiz1987) Theaboveexampleshavebeengenerallyexplainedbyresortingtothedifferencebetweenagentsand causers(Fábregas2014for(242a))ortheneedfortheagenttobesomehowrelevantfortheresultstate (Navas-Ruiz 1987). These examples, I argue, can be uniformly accounted for by assuming that they haveanunderlyingstativecausative estar-PPrtconfiguration. Iillustrateitfortheexample(242a)in (243). 64 Afourthpossibilityisthatastativecausativeverb(i.e. averbwithjustinitandrescategoryfeaturesinitslexicalentry) lexicalizesasingleresPstructure,i.e.anunaccusativeestar-PPrtstructurelike(238b). 127 Section4.4. Myproposal Chapter4 (243) a. CopP estar ‘be’ AdjP Adj <–da> initP initP init i <contamina–> [init i ,proc,res] resP DP Londres res’ res i <contamina–> [init,proc,res i ] XP by-P porelplomo b. s9s 1 ,s 2 [s=(s 1 !s 2 )&contaminada(s 1 )&participant(porelplomo,s 1 )&contaminada(s 2 )& participant(Londres,s 2 )] Inprose,thesemanticsin(243b)saythatthereisastativeeventualityoftheleadmaintainingLondon in a polluted state. The reason that el plomo ‘the lead’ is a felicitous agent—in the sense of Folli & Harley(2008)—forastativecausativestructureisthatithaspropertiestokeepLondoninapolluted state. Moriarty,ontheotherhand,maypourpollutingsubstancesonthewaterorgasemissionstothe air,butheisnotinitselfthemaintainingagentofthepollutedstateofthetown.Asimilarexplanation canbegivenfortheexamplein(242b).Iillustratethisin(244). (244) a. Los the manifestantes demonstrators cortaron cut el the camino path con with barricadas. barricades ‘Thedemonstratorscutthepathwithbarricades.’ b. El the camino path está is estar cortado cut por by barricadas. barricades ‘Thepathiscutbybarricades.’ 128 Section4.4. Myproposal Chapter4 In(244a),itisnotthedemonstratorsthemselveswhokeepthepathblocked,butrather,thebarricades theysetup. Thedemonstratorsarethusnotmaintainingagentsin(244a)andthereforethatscenario cannot correspond to (242b). Rather, it corresponds to (244b), since the maintaining agents are the barricades. The point, then, is that typically telic verbs can appear in stative causative configurations if the resultingformalsemanticmeaningiscompatiblewithworldknowledge. Inturn,thisexplainsphe- nomenaregardingby-phrasessuchasthosediscussedin(242)and(244). 4.4.2 Bare-PPrts Thissectionpresentsmyproposalwithbare-PPrts.WesawthatPPrtshavenoAktionsartrestrictions forthebaseverbandnoby-phraserestrictionseither,andtheymoreovershowthedisjointreference effect. I argue that bare-PPrts are in fact bare verbal passives without tense or aspect operators. I provideanexamplein(245).IassumethatverbalpassivesareformedbyaspecializedheadPass(ive)P, althoughitcouldverywellbenotatedasapassiveflavorofVoice,asHarley(2013)does.However,the externalargumentissemanticallyintroducedbelowbyinitP,andtheby-phraseadjoinstoitsaturating thevariable.Passdoesnotintroduceanexternalargument,butmerelythedisjointreferenceeffect. 65 (245) a. La the ciudad city destruida destroyed por by los the bárbaros. barbarians ‘Thecitydestroyedbythebarbarians.’ 65 ThedisjointreferenceeffectcanbemodelledasapresuppositiononthePasshead,asin(xxiii),fromSpathasetal. (2015). (xxiii) Presupposition:8f es,t .f(x)(e)!f6=theme 129 Section4.4. Myproposal Chapter4 b. DP D La NP N ciudad PassP Pass <–da> initP initP init <destrui–> procP proc <destrui–> resP DP laciudad res’ res <destrui–> XP by-P porlosbárbaros c. e9s 1 ,e 1 ,s 2 [e=(s 1 !e 1 !s 2 )&destruida(s 1 )&participant(porlosbárbaros,s 1 )&destruida(e 1 ) &destruida(s 2 )&participant(laciudad,s 2 )] Myaccountofbare-PPrtsissimilartothatputforthbyHarwood(2017)forEnglishreducedrelative clausesRRC—bare-PPrtsare,indeed,RRCs(seealsoRapp2011forGerman).Althoughwearriveatthis conclusion for our respective languages for independent reasons—his evidence is purely syntactic, whileminedrawsfromthesyntax-semanticsofgrammaticalaspectandeventstructure—Inotethat his reasoning can for the most part be applied to Spanish bare-PPrts. Harwood adopts a ‘What You See Is What You Get’ approach (Bošković 2014; Harwood 2013, 2015, a.o.), which contends that if an aspectualformisabsentfromthesentence,thenitmustalsobeabsentfromtheunderlyingsyntactic derivation. 66 Fromthisapproachitfollowsthattherecannotbeahiddenhaveinthestructure,contra Bosque(2014)andalsoagainstwhiz-deletionaccountsofRRC(Chomsky1957;Smith1961;Ross1967, 66 SeealsoFábregas(2007)forthebroaderExhaustiveLexicalizationPrinciple,whichcontendsthateverysyntacticfeature mustbelexicalized. 130 Section4.4. Myproposal Chapter4 1972),whichclaimthatRRCaresimplyformedbydeletingthewh-pronoun/complementizerandthe auxiliary. Thewhiz-account, note, cannotexplainwhywecannothaveRRCssuchasthosein(246), wheretheauxiliary haber‘have’isellided. ThefirstRRCin(246a)isanactiveperfectandthesecond oneisapassiveperfectwherethecopula—either ser fortheverbalpassiveor estar fortheadjectival passive—isnotelided(thestrikethroughtextin(246)representstheelidedmaterial). (246) a. *El *the hombre man que that ha has robado stolen los the zapatos. shoes intended:‘Themanwhohasstolentheshoes.’ b. *El *the perro dog que that ha has sido/ been ser estado been estar secuestrado. kidnapped intended:‘Thedogwhohasbeenkidnapped.’ Harwood(2017)arguesthatRRCcontainthefirstphaseofthederivation—theclause-internalphase, orCIP.Heassumes,followingpreviouswork(Harwood2013,2015;Wurmbrand2013,2014;Ramchand &Svenonius2014;Aelbrecht&Harwood2015),thatthefirstphaseboundaryintheclauseiseitherthe progressiveinflectionorvoice,asshownin(247a)and(247b),respectively. 67 68 69 (247) a. [ TP T[ PerfP Perf [ ProgP Prog[ VoiceP Voice[ VP Lex.V]]]]] b. [ TP T[ PerfP Perf [ VoiceP Voice[ VP Lex.V]]]] Note that, since grammatical aspect projections are absent from bare-PPrts, CP and TP should be absenttoo—especiallyifweadoptHarwood’saccountthatRCCsareonlycomprisedoftheCIP.There areempiricalreasonstobelievethatCPandTPlayersareindeedabsentfrombare-PPrts. Consider thefollowingdatafromnegation. (248) a. La the cocina kitchen no not ha has sido been limpiada. cleaned ‘Thekitchenhasnotbeencleaned.’ 67 The evidence in favor of the phasal view comes from there-existentials, VP ellipsis and VP fronting. I note that this evidenceisnotextrapolabletoSpanish,sincethere-existentialshaveclashingpropertiesinthetwolanguages,andtheVP ellipsisandVPfrontingexamplesinvolveseparatingtheauxiliaryhavefromtheparticiplebyeitheromissionorstranding, somethingthatcannotbedoneinSpanishforindependentreasons.Ileaveasidethisissueforfuturework. 68 SeeAkmajian&Wasow(1975);Tenny(1987);Cinque(1999),a.o.forthesamehierarchicalstructureforauxiliaries. 69 The representations in (247) are a simplification. Harwood in fact adopts a two-layered structure for aspectual pro- jections: a lower one that hosts the inflectional head with uninterpretable features, and a higher vP that introduces the auxiliaryverb,whichchecksthosefeatures.Hisfullarticulationisprovidedin(xxiv),fromHarwood(2017:28). 131 Section4.4. Myproposal Chapter4 b. La the cocina kitchen no not limpiada. cleaned ‘Theuncleanedkitchen.’ Thedatafrom(248)suggestthatnegationispossiblewithbare-PPrts,justasitiswithperfectpassives. However, I contend that the negation in (248b) is not an instance of sentential negation. Sánchez López(1999)showsthattherearearetwotypesofnegationinSpanish: morphological—alsoknown as constituent negation—and sentential negation. The former takes as its scope a single lexical unit whereas the latter has the clause as its scope. The author argues that sentential negation licenses negative polarity items, whereas constituent negation does not. As we can see in (249), perfects of verbal passives with negation can license negative polarity items such as ningún ‘any’ (e.g. (249a)), whereasbare-PPrtscannot(249b)). Also,SánchezLópez(1999)showsthatsententialnegationallows forcorrectivephrases,sinceithasscopeoverotherconstituentsintheclause.Again,itturnsoutthat bare-PPrts with negation do not allow for corrective phrases (e.g. (250b)), unlike perfects of verbal passives(e.g.(250b)). (249) a. La the cocina kitchen no not ha has sido been limpiada cleaned por by ningún any empleado. employee ‘Thekitchenhasnotbeencleanedbyanyemployee.’ b. *La *the cocina kitchen no not limpiada cleaned por by ningún any empleado. employee intended:‘Thekitchennotcleanedbyanyemployee.’ (xxiv) TP T vPperf have PerfP Perf vPprog be ProgP Prog vP be VoiceP Voice VP Lex.V 132 Section4.4. Myproposal Chapter4 (250) a. La the cocina kitchen no not ha has sido been limpiada, cleaned sino but fumigada. fumigated ‘Thekitchenhasnotbeencleaned,butfumigated.’ b. *La the cocina kitchen no not limpiada, cleaned sino but fumigada. fumigated intended:‘Thekitchennotcleaned,butfumigated.’ Following the standard assumption, starting with Laka (1990), that sentential negation in Spanish is hosted by a projection below CP and TP (e.g. (251)), it follows that at least TP should be absent too, otherwisewecouldnotexplainwhyNegcannotselectit. (251) [ CP C[ NegP Neg[ TP T...]]] It is also dubious that there is a CP projection—or a left periphery, in Rizzi’s (1997) terms. First, we cannothaveanovertcomplementizer(e.g. (252a)). Ofcourse,wecouldassumethatthecomplemen- tizer is simply elided, just as Bosque (2014) assumes for the perfecthave in these constructions. But notethatifwedidhaveaCPdomain,weshouldbeabletofindtypicalleftperipheryelementssuchas focalizedmaterial. However,thisisnotpossible,as(252b)shows(thefocalizedphraseisrepresented incapitalletters). (252) a. La the cocina kitchen (*que) (*that limpiada. cleaned intended:‘Thekitchen(that)cleaned.’ b. *La *the cocina kitchen EN in CINCO five SEGUNDOS seconds limpiada. cleaned intended:‘ThekitchenINFIVESECONDScleaned.’ In short, bare-PPrts are formed by a verbal passive projection with no other aspectual or temporal operators,andwithoutaCPlayer.Thestructureisthenasin(245),repeatedbelowforconvenience. (245) a. La the ciudad city destruida destroyed por by los the bárbaros. barbarians ‘Thecitydestroyedbythebarbarians.’ 133 Section4.4. Myproposal Chapter4 b. DP D La NP N ciudad PassP Pass <–da> initP initP init <destrui–> procP proc <destrui–> resP DP laciudad res’ res <destrui–> XP by-P porlosbárbaros c. e9s 1 ,e 1 ,s 2 [e=(s 1 !e 1 !s 2 )&destruida(s 1 )&participant(porlosbárbaros,s 1 )&destruida(e 1 ) &destruida(s 2 )&participant(laciudad,s 2 )] 4.4.3 Perfectiveadjectives InSection4.2(seeTable4.2),Ipresentedanoverviewofparticiplesthathavea‘cut-short’formanda regularform(e.g. seco‘dry’, secado‘dried’). Iproposethattheseshareanadjectivalbase: theseare, in theirverbalform,deadjectivalverbs.Theirlexicalentrythenlookslike(253). (253) secar‘dry’ categoryfeatures:[init,proc,res,A] Thecut-shortparticipialformariseswhenonlytheAPprojects,asin(254b). 70 70 InthisstructureIabstractawayfromtheextendedprojectionoftheadjective(Corver1997;Kennedy1997;Fults2006, a.o.). Furthermore, I do not discard the possibility that the subject of predication is not introduced within AP but via a separateprojectionPred(ication)P,asBaker(2003)proposesfornon-verbalpredications. 134 Section4.4. Myproposal Chapter4 (254) a. La the toalla towel está is estar seca. dry ‘Thetowelisdry.’ b. CopulaP estar AP DP latoalla A’ A <seca> [init,proc,res,A i ] XP Itisgenerallybelievedthatwhenparticipleshavetwoforms—a’cutshort’formandafull-fledged-en form—the-enformcannotappearinpredicativepositionbecauseofablockingeffect(Belletti&Rizzi 1988).However,thisisnotcompletelyaccurate.Althoughthecut-shortformisindeedmorecommon withestar-PPrts,theregularparticipialformisalsopossible.Fábregas&Marín(2015b)makethispoint withthefollowingexamplesof estar-PPrtswithparticiplesthatotherwisehavecut-shortforms(e.g. (255)). Similar data is found for Sicilian in Bentley (2017), who—although not explicitly making that point—presentstheexamplesin(256). Onmypart,Ihavefoundseveralinstancesofestar-PPrtswith -enparticiples,see(257)forsecado‘dried’alone. 71 (255) a. Y and mucho much más more cuando when esa that política, policy repito, I.repeat está is vaciada emptied de of ideología. ideology ‘Andespeciallywhenthatpolicy,Irepeat,isemptiedofideology.’ b. estos these vacíos gaps estan are llenados filled por by las the moléculas molecules c. si if usted you no not está are contentada contented con with el the servicio service que that le you han have ofrecido offered puede you.can interponer present una a reclamación complaint ‘Ifyouarenotpleasedwiththeservicetheyofferedyou,youcanpresentacomplaint’ 71 AlthoughmostoftheexamplesIretrievedonlinewerefromtextsfromthe19 th century,whichsuggeststhatestar-PPrts withtheseparticiplesweremorecommoninearlierstagesofthelanguage. Ileaveasidethisdiachronicissueforfuture research. 135 Section4.4. Myproposal Chapter4 d. es he.is un an inculto uncultivated que that no not sabe know escribir write.inf y and está is molestado annoyed ‘Heisanignorantpersonwhodoesnotknowhowtowriteandheisannoyed.’ (256) a. Stu this quatru picture.msg jè(ni) is appinnutu hang.ptcp.msg ch’ with i the chiova. nails ‘Thispictureishungwithnails.’ b. U this quatru picture.msg jè(ni) is appisu. hang.ptcp.msg ‘Thispictureishung/hanging.’ (257) a. ...Los ...the árboles trees [...] [...] están are estar secados dried por by las the nevazones. snows ‘Thetreesaredriedbythesnows.’ (EnsayosobrelageografíamilitardeChile,Volumen2,F.VidalGormaz,1886:221) El the tapioka tapioka está is estar secado dried al at.the calor. heat ‘Thetapiokaisdriedintheheat.’ (Higienealimenticiadelosenfermos,delosconvalecientesy delosvaletudinarios,J.B.Conssagrives,1865:147) No not me me trataré treat con with aquel that.one que that está is estar secado dried por by la the envidia. envy ‘Iwillnotassociatemyselfwithhewhoisdriedbyenvy.’ (Lacruz,revistareligiosade Españaydemáspaísescatólicos,LeónCarboneroySol,1879:38) Then,Isimplyproposethattheseestar-PPrtsaredeadjectivalverbsverbalizedbyreswithparticipial morphology attached—i.e. just like my proposal in Section 4.4.1 for estar-PPrts that do not have a ‘cut-short’ form. I exemplify it in (258) below. Thus, cut-short participles and -en participles do in factlexicalizedifferentstructures, andhencetheformerdonotmorphologicallyblockthelatter, as Fábregas & Marín (2015b) observe. 72 The reason why -en PPrts are not often seen with estar if they have a cut-short counterpart is because they both denote a state, and the cut-short counterpart is morphosyntacticallyasimplerstructure. 72 Fábregas & Marín’s (2015b) proposal, although within the same theoretical framework, is different from mine. They proposethatcut-shortparticiplesareinstancesofresP>APconfigurations,whereasthe-encounterpartcontainsaprocP inthestructure.IhavealreadyarguedwhythereshouldnotbeaprocP.Also,Iamnotsurehowaperfectiveadjectivecan bederivedviaaresPprojection,giventhatresisaverbalhead. 136 Section4.4. Myproposal Chapter4 (258) a. La the toalla towel está is estar secada. dried ‘Thetowelisdried.’ b. CopulaP estar AdjP Adj -da resP DP latoalla res’ res i <seca> [init,proc,res i ,A] AP A i <seca> [init,proc,res,A i ] XP Finally, what about irregular participles, i.e. those that have a ‘cut-short’ form across the board for passives and perfects? I assume that these forms have a listed entry as in (259), which includes a Adj feature. Assuch, theseexponentscanlexicalizethewholeparticiple. Thelexicalentryofthisexpo- nentsalsoincludeinitandproc,sinceitalsolexicalizesverbalpassivesandperfects. 73 (259) roto‘broken’ categoryfeatures:[Adj,init,proc,res] (260) a. El the vaso glass está is estar roto. broken ‘Theglassisbroken.’ 73 But, of course, the complication arises that in verbal passives and perfects the participle is not adjectival, and a uni- fied analysis would be desirable. See Larsson & Svenonius (2013), for a unified analysis of the -en participial suffix via a generalizedPrtheadthatisalsopresentinE-PPrtsandperfects,aswellasSection4.7,whereIdiscussthisissuemoreat length. 137 Section4.5. By-phrasesandagent/event-orientedmodifiersinAPass.BeyondSpanish Chapter4 b. CopulaP estar AdjP Adj <roto> [Adj,init,proc,res] resP DP elvaso res’ res i <roto> [Adj,init,proc,res] XP 4.5 By-phrasesandagent/event-orientedmodifiersinAPass.BeyondSpanish TheeffectsobservedinSpanishestar-PPrtsregardingby-phrasesarenotexclusivetoSpanish. Inthe literature of other languages it has been frequently pointed out by-phrases and other modifiers ori- ented to the external argument are also generally ungrammatical. I provide examples for German, HebrewandEnglishin(261),(262)and(263),respectively.Iwillrefertothecounterpartsofestar-PPrts in these other languages as adjectival passives (henceforth APass), and ser-PPrts as verbal passives (henceforthVPass). (261) a. Der the Mülleimer rubbish.bin ist is {*von {*by meiner my Nichte/ niece *langsam/ *slow(ly) *genüsslich/ *pleasurabl(e/y) *mit *with der the Heugabel} pitchfork geleert. emptied intended:‘Therubbishbinisemptiedbymyniece/slowly/pleasurably/withthepitchfork.’ (FromRapp1996:246) b. *Die *the Tür door war was von by den the Polizisten policemen geöffnet. opened ‘Thedoorwasopenedbythepolicemen.’ x (FromAnagnostopoulou2003:18,viaWinfriedLechner,p.c.) (262) a. ha-mexonit the-car rexuca washed (*al-yedey (*by maks Max / / *be-tsumetxlev *in-attention / / *be-cinor). *in-hose 138 Section4.5. By-phrasesandagent/event-orientedmodifiersinAPass.BeyondSpanish Chapter4 intended:‘ThecariswashedbyMax/carefully/withahose.’ (FromMeltzer-Asscher 2011:819) b. Even stone ha-pina the-corner tihiye be.fut munaxat laid ba-makom in+the-place (*al-yedey (by 5 5 po’alim). workmen) ‘Thecornerstonewillbelaidintheplace.’ (FromHorvath&Siloni2008:107) (263) *Thedoorseemed{broken/opened/painted}(*byMary). (FromMcIntyre2013:7) Interestingly, the same effects regarding telicity that I have discussed for Spanish can be found in theselanguages(seealsoGarcía-Pardo2017). Astheexamplesin(264)–(266)show,adjectivalpassives derivedfromstativecausativeverbsalsoallow by-phrasesintheselanguages. Theexplanationtypi- cally foundin the literatureis roughly that by-phrases areacceptable if theagent participates in the state, or is understood as maintaining the state expressed by the participle (Rapp 1996 for German, Meltzer-Asscher 2011 for Hebrew and McIntyre 2013 for English, a.o. See also Bull 1965; Hengeveld 1986; Navas-Ruiz 1987 for Spanish). As I hope is clear now from my discussion, this effect regarding theparticipationoftheagentinthestatefollowsfromtheaspectualstructureoftheparticiple:inthe stativecausativepredicate,theagent’sparticipationholdsthroughoutthestate. (264) a. Das the Land country ist is gut well regiert. governed ‘Thecountryiswell-governed.’ b. Die the Arbeiter workers sind are durch by den the Vorarbeiter foreman überwacht. supervised ‘Theworkersaresupervisedbytheforeman.’ (ThomasBorer,p.c.) (265) a. ha-ictadion the-stadium šamur guarded al-yedey by šotrim policemen xamušim. armed ‘Thestadiumisguardedbyarmedpolicemen.’ (FromMeltzer-Asscher2011:826) b. Ha-bayit the-house yihiye be.fut šamur guarded al-yedey by šloša three šomrim. guards ‘Thehousewillbeguardedbythreeguards.’ (FromHorvath&Siloni2008:107) (266) a. Theroadremained{blockedbypolice/supportedbypylons}. b. Thedictatorremained{unsupported/proppedup/underestimated}bythewarlords. (FromMcIntyre2013:7) 139 Section4.5. By-phrasesandagent/event-orientedmodifiersinAPass.BeyondSpanish Chapter4 Thereare,however,instancesofadjectivalpassivesderivedfromtelicverbsthatnonethelessaccept by-phrases,aswellasagent-orientedandevent-orientedmodifiers,somethingthatisnotpredictedby myproposalasitis. ThishasbeenobservedintheliteratureofSpanishaswellasinotherlanguages. Inwhatfollows,Iwillreviewcriticallythecurrentproposalsregardingthesefactssoastotakestock onthematter. 4.5.1 State-relevance Thestate-relevanceapproachtotheavailabilityofby-phrasesinadjectivalpassivesmaintainsthatby- phrasesarepossibleiftheagenttheyintroduceissomehowdetectableintheresultstate(Hengeveld 1986).Inthisview,theby-phrasein(267a)islicitbecausetheambassadorisdetectableinthefinalstate of the letter thanks to his signature. Similarly, (267b) is fine because the editorial that published the bookcanbedetectedinthefinalstate(e.g.bylookingatthecoverorthefrontmatterofthebook).A similarpointcanbemadeofinstrumentals(e.g.(267c))andmanneradverbials(e.g.(267d)). (267) a. El the documento document está is estar firmado signed por by el the embajador. ambassador ‘Thedocumentissignedbytheambassador.’ (FromHengeveld1986) b. El the libro book está is estar editado edited por by Susaeta. Susaeta ‘ThebookiseditedbySusaeta.’ c. Estos these trozos pieces de of madera wood están are estar unidos united con with tornillos. screws ‘Thesepiecesofwoodareheldtogetherwithscrews.’ d. La the película movie está is estar rodada shot con with mucho much gusto. taste ‘Themovieismadewithalotoftaste.’ Thisobservationhasbeenalsomadeforotherlanguages,notablyforGerman(Rapp1996,1997),He- brew(Meltzer2005;Meltzer-Asscher2011;Horvath&Siloni2008)andEnglish(McIntyre2015).Ishow therelevantexamplesin(268)forGerman,(269)forHebrewand(270)forEnglish. (268) a. Die the Zeichnung drawing ist is von by einem a Kind child angefertigt. produced ‘Thedrawingisproducedbyachild.’ (FromRapp1997:192) b. DieHaarewarenschlampiggekämmt/geschnitten. thehairsweresloppilycombedcut (FromKratzer2000:392,394) 140 Section4.5. By-phrasesandagent/event-orientedmodifiersinAPass.BeyondSpanish Chapter4 c. Der the Brief letter war was mit with einem a Bleistift pencil geschrieben. written ‘Theletterwaswrittenwithapencil.’ (AdaptedfromRapp1996:254) (269) a. ha-sefer the-book arux edited al-yedey by orex editor mecuyan. excellent ‘Thebookiseditedbyanexcellenteditor.’ b. ha-xulca the-shirt ha-zot the-this tfura sewn be-xoser in-lack mikco’iyut. professionalism ‘Thisshirtissewnunprofessionally.’ c. ha-mixtav the-letter katuv written be-et. in-pen ‘Theletteriswrittenwithapen.’ d. ha-kelev the-dog kašur tied be-recu’ in-leash ‘Thedogistiedwithaleash.’ (FromMeltzer-Asscher2011:823) (270) a. Thedogistied adj upwitharope(*byapoliceman). b. Thecarisallsprayed adj withpaint(*yesterday). c. Somewordsareunderlined adj witha{highlighter/bluepen/*shortpen}. d. Thistextiswritten adj bya{moron/non-nativespeaker/*tallperson}. ThisproposalfindsitsmostrecentformulationinMcIntyre(2015),whostatesitasfollows: (271) StateRelevanceHypothesis:Event-relatedsatellitesareunacceptablein(German,English,He- brew)adjectivalpassivesunlesstheycontributetothedescriptionofthestateexpressedbythe participle or of the theme during the interval i during which this state holds. They are most acceptableiftheyprovideinformationwhichcanbeinferredsolelybyinspectionofthetheme duringintervali. (FromMcIntyre2015:941) 4.5.2 Problemswiththestate-relevanceapproach Itisnotclearhowaspeakercoulddeterminethedegreeofdetectabilityoftheagentfromtheresult statefortheby-phrasetobelicit.Moreover,thegeneralstate-relevanceproposalisriddledforcoun- terexamples, as García-Pardo (2017) notes for Spanish. The sentences in (272) should be acceptable under this account, and yet they are not: by looking at the curtain in (272a), you could easily infer 141 Section4.5. By-phrasesandagent/event-orientedmodifiersinAPass.BeyondSpanish Chapter4 bythetypeofscratchmarksthatitwasthecatwhoscratchedit,orthatitwasscratchedwithasaw knife.Similarly,wecoulddetectthestompingactivityofthekidsin(272b)byseeingthetypeoffoot- prints on the sand. In (272c), it could be detectable by the kind of marks on the trunk that it was a bearwhoopenedit.Again,allthesentencesareverydegradedwithby-phrases,whichdiminishesthe state-relevanceproposal. (272) a. La the cortina curtain está is estar arañada scratched (??por (??by el the gato). cat intended:‘Thecurtainisscratchedbythecat.’ b. La the arena terrain está is estar pisoteada stomped (??por (??by unos some niños). kids indended:‘Theterrainisstompedonbysomekids.’ c. El the maletero trunk del of.the coche car está is estar abierto opened (??por (??by un a oso). bear intended:‘Thetrunkofthecarisopenedbyabear.’ (FromGarcía-Pardo2017:24) Thisproposal,notewell,alsofailstoexplaintheavailabilityof by-phraseswithmanyparticiplesde- rivedfromstativecausatives. Inalltheexamplesbelow,theagentintroducedbytheby-phraseisnot detectablefromthestatedenotedbytheparticiple,astheextracontextclearlyshows,andyetallthe sentencesareperfectlygrammatical—whichissomethingexpectedfrommyproposalinSection4.4.1. (273) a. El the jardín garden está is estar vigilado surveiled por by cuatro four cámaras cameras de of seguridad security que that no not se se ven. see ‘Thegardenissurveilledbyfoursecuritycamerasthatareoutofsight.’ b. El the museo museum está is estar protegido protected por by un a sistema system de of rayos rays infrarrojos infrared invisible. invisible ‘Themuseumisprotectedbyaninvisibleinfraredrayssystem.’ c. El the país country está is estar gobernado governed por by una a coalición coalition de of izquierdas lefts que that no not ha has cambiado changed nada. nothing ‘Thecountryisgovernedbyaleft-wingcoalitionthathasnotchangedanything.’ Thesameproblemsarisewithevent-relatedmodifiersandinstrumental-PPs. Thesentencesin(274) donotacceptevent-relatedmodification,eventhoughitcanbeeasilyperceivable(especiallyforage- ologist)thatarockhaserodedprogressively,orthatanonionhasbeenfriedslowly,sincesuchonions 142 Section4.5. By-phrasesandagent/event-orientedmodifiersinAPass.BeyondSpanish Chapter4 quite different from those fried quickly in high heat. The same thing happens with PP-instruments in (275): even though one can tell by the shape of mark on the wall that it was done with a lighter, (275a)isout,anddespitewecanclearlyseethatthedocumentwasdestroyedwithashredder(itiscut intocountlesslongrectangles),(275b)isdegraded. Again,nosuchproblemsarisewith estar-passives derivedfromStCverbs(e.g.(276)). (274) a. Esta this roca rock está is estar erosionada eroded ??(paulatinamente). ??(progressively intended:‘Thisrockiserodedprogressively.’ b. Esta this cebolla onion está is estar pochada pan-fried ??(muy ??(very lentamente). slowly intended:‘Thisonionispan-friedveryslowly.’ (275) a. La this pared wall está is estar quemada burnt ??(con ??(with un a mechero). lighter intended:‘Thewallisburntwithalighter.’ b. El the documento document está is estar destruido destroyed ??(con ??(with un a triturador). shredder intended:‘Thedocumentisdestroyedwithashredder.’ (276) a. El the museo museum está is estar celosamente zealously vigilado, surveilled aunque although no not lo it parezca. seems ‘Themuseumiszealouslysurveilled,althoughitdoesnotlooklikeit.’ b. La the calidad quality del of.the agua water está is estar controlada controlled con with tecnología technology muy very sofisticada. sophisticated ‘Thewaterqualityiscontrolledwithverysophisticatedtechnology.’ Withinestar-passivesderivedfromtelicverbs,wecanfinda—verylimited,duetotherestrictionson participial modification, see Section 4.5.5 for more discussion—number of them that allow manner adverbials (e.g. (277a)) and PP-instruments (e.g. (277b)) with a context that negates the perceptibility ofthemannerortheinstrumentintheresultstate. (277) a. Este the vestido dress está is estar confeccionado made manualmente, manually aunque although no not se refl aprecie. perceive ‘Thisdressismademanually,althoughyoucannottell.’ b. Estos these héroes heroes están are estar dibujados drawn a at lápiz, pencil aunque although no not se refl note. notice ‘Theseheroesarepencil-drawn,althoughyoucannottell.’ 143 Section4.5. By-phrasesandagent/event-orientedmodifiersinAPass.BeyondSpanish Chapter4 4.5.3 Event-kindsandpseudo-incorporation In a series of papers, Gehrke (2011, 2012, 2013, 2015) puts forth the proposal that adjectival passives in German denote the instantiation of the consequent state kind of an event kind (inspired by the proposalinCarlson1977forthenominaldomain). Iprovideanexamplein(278),wherethesubscript ksignalsakindentityandwhereRisCarlson’s(1977)realisationrelation—meaningthatthestatesis aninstantiationofthestatekinds k . (278) a. Die the Tür door ist is geschlossen. closed b. 9e k ,s k ,s [become(e k ,s k )^theme(e k ,door)^closed(s)^theme(s,door)^R(s,s k )] (FromGehrke2012:192) Gehrkefocusesnotonlyinby-phrases,butalsoinevent-relatedmodifiersmoregenerally—i.e.modi- fiersthatpertaintothecausingorprocesspartofthetelicevent,e.g.instrumentals,manneradverbials andsoon. TheauthornotesthatinGerman,event-relatedmodifiersinAPasssuchasinstrumentals and by-phrasestendtobediscourse-opaque. Forinstance,theindefinitesin(279a)and(279b)cannot haveaspecificreading,asshownbytheungrammaticalityofreferringtothemasconcreteentitiesin thediscourse. (279) a. Die the Karte card ist is mit with [einem ]a Bleistift] 1 pencil geschrieben. written *Er 1 *he ist is blau. blue (German,fromGehrke2015:904) b. Die the Zeichnung drawing ist is von by [einem ]a Kind] 1 child angefertigt. produced *Es 1 *it hat has rote red Haare. hairs (German,fromGehrke2015:904) In addition, these non-specific indefinites also need to be contextually well-established within the communityofspeakers.Forinstance,theexamplesin(280)areoutbecausethereisnowell-established eventkindofanieceemptyingatrashcanoramanopeningadoor,asopposedtoawritingofaletter withapenciloradrawingbyakidin(279). (280) a. *Der *the Mülleimer rubbish-bin ist is von by einer a Nichte niece geleert. emptied intended:‘Therubbishbinisemptiedbyaniece.’ (German,fromGehrke2015:923) b. *Die *the Tür door ist is von by einem a Mann man geöffnet. opened intended:‘Thedoorisopenedbyaman.’ (German,fromGehrke2015:923) 144 Section4.5. By-phrasesandagent/event-orientedmodifiersinAPass.BeyondSpanish Chapter4 Gehrke further points out that specific definites and proper names may actually be possible in by- phrases, but again, they are subject to the well-established event-kind constraint: the proper name needstorefertosomeonenoteworthyamongthecommunityofspeakers,eg. theeventofChomsky citingamanuscriptvs.Sandbergercitingitin(281). (281) Das the Manuskript manuscript ist is von by {Chomsky/ {Chomsky ?Sandberger} ?Sandberger zitiert. cited ‘Themanusriptiscitedby{Chomsky/(Sandberg)}.’ (German,fromMaienborn2009viaGehrke2015:914) The author also notes that by-phrases in APass derived from stative verbs do not show restrictions with respect to specificity or well-establishedness (e.g. (282)). She points out that by-phrases with APassderivedfromtelicandstativeverbshaveotherfundamentallydifferentproperties: theformer integrateprosodicallywiththeparticipleandthelatterdonot(seealsoSchlücker2005),andtheformer areimpossibleiftheparticipleundergoesun-prefixationwhereasthelattersurviveun-prefixation(see alsoRapp1996). (282) a. Das the Haus house ist is von by Studenten students bewohnt. in-lived b. Er he ist is von by der the Musik music beeindruckt. impressed BLANKSPACEXXXXXXXX(FromGehrke2012:190) Gehrke (2012) then propose the following structures for APass. First, the author assumes, following thetraditionalview(Wasow1977;Bresnan1982;Levin&Rappaport1986),thatAPassmaybederived by lexical adjectivization as in (283), where the adjectivizer A aff is a zero suffix (Lieber 1980; Kratzer 2000). This structure would correspond, broadly, to Bosque’s (1999) Adj-PPrts, or Embick’s (2004) stativepassives. (283) Lexicaladjectivisation AP V geschlossen A aff ? ForAPassderivedfromtelicverbswithevent-relatedmodifications,Gehrkeproposesthattheyun- dergophrasaladjectivization(e.g.(284),seealsoKratzer2000).GehrkeproposesthatabareVPdenotes an event-kind, at it is only if it is embedded under further structure (vP/VoiceP in Gehrke 2012 and tenseandaspectinGehrke2015)thataneventgetsinstantiated. InthecaseofAPass,adjectivization 145 Section4.5. By-phrasesandagent/event-orientedmodifiersinAPass.BeyondSpanish Chapter4 prevents the event from being instantiated and hence it remains in the kind domain. Event-related modifierswillonlybepossibleiftheycreateasubkindoutoftheeventkinddenotedbythelexicalverb (e.g.in(281),thesubkindofChomsky-citingeventoutoftheeventkindofciting).Inmorerecentwork, Gehrke(2015)furtherproposesthattheseevent-orientedmodifiersundergopseudo-incorporationright beforeadjectivization,fromwhichthewell-establishednessrestrictiononmodifierswouldfollow. 74 (284) Phrasaladjectivisation:event-kindmodifiers AP VP PP vonChomsky V zitiert A aff ? Finally,forstate-orientedmodifiers,Gehrke(2012)proposesalexicaladjectivizationaccount,inwhich themodifierappliestothealreadybuiltparticiple,inthelinesofRapp(1996)(e.g.(285)). (285) Statemodifiers:modifiersofAP AP PP vonderMusik A’ V t 1 A aff (un)beeindruckt 1 -? Gehrke&Sánchez-Marco(2014)extendthisproposaltoSpanishinacorpusstudy. Theyarguethat, justlikeGerman,thenominalsintroducedbyby-phrasescanonlyhavenarrowscopewhentheyare indefinites (e.g. (286a)), cannot introduce discourse referents (e.g. (286b)), do not accept intersective modifiers(e.g.(286c))orstrongdeterminers(e.g.(286d)). (286) a. Todos all los the cuadros pictures estaban were estar pintados painted por by un a niño. child 3 Narrowscopereadingoftheindefinite:‘Allthepaintingswerepaintedbyachild.’(>1 childpossible) 7 Widescopereadingoftheindefinite:‘Therewasaparticularchildthatpaintedallthe painting.’ 74 Forseminalworkonnounincorporationandpseudo-incorporation,seeMithun(1984);Baker(1988);Massam(2001).For pseudo-incorporationinSpanishandCatalan,seeEspinal(2010);Espinal&McNally(2011). 146 Section4.5. By-phrasesandagent/event-orientedmodifiersinAPass.BeyondSpanish Chapter4 b. El the cuadro picture estaba was. estar pintado painted por by [un [a niño] 1 . child *pro 1 *pro Era was pelirrojo. red-haired intended:‘Thepicturewaspaintedbyachild.Hehadredhair.’ c. *El *the cuadro picture estaba was. estar pintado painted por by un a niño child pelirrojo. intended:‘Thedrawingwaspaintedbyared-hairedchild.’ d. *El *the cuadro picture estaba was. estar pintado painted por by este this niño. child intended:‘Thedrawingwaspaintedbythischild.’ (Gehrke&Sánchez-Marco2014:194–5) The authors observe in their study that there is indeed a quantitative difference in by-phrases be- tweenadjectivalandverbalpassives—or estarand ser-passives,beingfoundinmuchlargernumbers in the latter construction. Also, event-related by-phrases introduce a much larger number of weak indefinitesandbarenominalsinestar-passivesthaninser-passives,asexpectedundertheevent-kinds approach(see(286)).Withparticiplesderivedfromstativeverbs,theyobservenoqualitativedifference betweentheby-phrasesofser-passivesandestar-passives(e.g.(287)forestar-passives). (287) a. Rufina Rufina estaba was. estar aterrada terrified por by esta this situación. situation ‘Rufinawasterrifiedbythissituation.’ b. ... ... una a casa house vecina close al to.the Campo Camp de of Marte, Mars cuyo whose quinto fifth piso floor estaba was. estar ocupado inhabited por by su his esposa wife y and sus his hijos sons ‘...ahouseclosetotheCampofMars,thefifthfloorofwhichwasinhabitedbyhiswifeand hissons’ (Gehrke&Sánchez-Marco2014:198) Gehrke&Sánchez-Marco(2014)furthernotethatby-phraseswithestar-passivestendtointroduceor- ganizationsorcollectiveentities(e.g. (288a),seealsoSchlücker2005forGerman)andmayintroduce propernamesifthereferentissomehownoteworthyor‘famousenough’withinthespeakercommu- nity (e.g. (288b), cf. (281)), which reinforces the view that estar-passives contain event kinds, and not eventtokens. (288) a. [...] [...] el the costo cost está is. estar amparado enshrined por by el the Ministerio. Ministry ‘ThecostiscoveredbytheMinistry.’ 147 Section4.5. By-phrasesandagent/event-orientedmodifiersinAPass.BeyondSpanish Chapter4 b. La the labor work de of todos all los the equipos teams está is. estar supervisada supervised por by el the doctor doctor Francisco Francisco Arquillo Arquillo Torres. Torres ‘TheworkofallteamsissupervisedbydoctorFranciscoArquilloTorres.’ x (Gehrke&Sánchez-Marco2014:198) 4.5.4 Problemswiththeevent-kindsapproach While I agree with Gehrke & Sánchez-Marco (2014) that there is a tendency for event-related mod- ifiers(i.e. with estar-passivesderivedfromtelicverbs)to benon-specificorgeneric, thereare many counterexamples to the contrary. For example, one can find specific indefinites introduced by by- phrases,unlikewhattheirhypothesispredicts.Notethatnoneoftheagentsintroducedin(289)canbe saidtobewell-establishedornoteworthyinthecommunityofspeakers—infact,theyarejustbeing introducedinthediscourse. (289) a. Todos all los the justificantes sicknotes estaban were estar firmados signed por by un a niño kid que that sabía knew falsificar forge la the firma signature de of todos all los the padres. parents ‘Allthesicknotesweresignedbyakidthatknewhowtoforgealloftheparents’signatures.’ b. Este this puente bridge está is estar diseñado designed por by un an arquitecto architect amigo friend mío. mine ‘Thisbridgeisdesignedbyanarchitectwhoisafriendofmine.’ Also,thereareinstancesofconstructionsthatdonotcontainaninstantiatedevent,andyettheyaccept referentialmodifiers. ThisisshownbyMcIntyre(2015)forEnglish(e.g. (290)), andhisexamplescan beeasilycarriedovertoSpanish(e.g.(291)). (290) a. Thischair i foldsflateasily.I’llletyoufoldit i forthefirsttime. b. Ijustmadeahookforopeningthathighwindow i .Iwillnowopenit i forthefirsttime. (FromMcIntyre2015:946) (291) a. Esta this silla i chair se se dobla folds fácilmente. easily Te you. dat dejaré allow. 1p.s que that la i it. acc dobles fold. 2p.s por for primera first vez. time ‘Thischair i foldsflateasily.I’llletyoufoldit i forthefirsttime.’ 148 Section4.5. By-phrasesandagent/event-orientedmodifiersinAPass.BeyondSpanish Chapter4 b. Acabo Finish 1p.s de of hacer make. inf un a garfio hook para for abrir open dat esa that ventana i window alta. high Ahora now la i it. acc abriré open. 1p.s.fut por for primera first vez. time ‘Ijustmadeahookforopeningthathighwindow i .Iwillnowopenit i forthefirsttime.’ Furthermore,therearemanyinstancesofby-phrasesinestar-passiveswithpropernounsthatshould inprinciplebemodifyinganeventkind—thereferentofthepropernounbeingnoteworthyamong the speaker community—and yet they are ungrammatical. For instance, (292a) is at least very de- graded,eventhoughTrumptweetingisawell-establishedeventkindintheglobalcommunity. The same goes for (292b), where the Jesus Christ healing sick people is a kind of event, and Jesus Christ noteworthyamongthecommunityofspeakers(say,theChristiansofthatera,accordingtotheNew Testament).Notethatreferentiallocationsarealsodegradedeveniftheyrefertoaneventkind,such astheexecutingofprisonersinSing-Singin(292c). (292) a. ??Este ??this mensaje message está is estar tuiteado tweeted por by Trump. Trump intended:‘ThismessageistweetedbyTrump.’ b. ??Este ??this enfermo sick.person. m está is estar sanado healed por by Jesucristo. Jesus.Christ intended:‘ThissickmanishealedbyJesusChrist.’ c. ??Este *this prisionero prisoner está is estar ejecutado executed en in Sing-Sing. Sing-Sing intended:‘ThisprisonerisexecutedinSing-SIng.’ McIntyre(2015)discussesanothersetofparticiple+modifiercombinationsthatdonotdenoteanwell- establishedevent,yettheyaregrammatical(inhisexamplesfrom(293),themodifiercombinationfor awell-estabishedeventisinbracketsattheend,whichwouldalsobegrammaticalifsubstitutedfor themodifierswithintheexamplesthatarenotwell-established).Whilehisexamplesdocarryoverto Spanish,IbelievethisisnotreallyaproblemforGehrke,butratherthatthesemodifiersapplytothe resultstate.Note,forinstance,that(294a)doesnotacceptthemodifierconguantes‘withgloves’ifitis interpretedasanevent-orientedinstrumental,i.e.somethingtheagentusedtotiethedog.Theother modifiersdonotcauseaproblembecausetheyareastate-orientedmodifier,i.e.thedogendsupwith thosematerialsaroundhim. Similarly,in(294b),itisverydegradedtosaytheletterwaswrittenwith thekeyboard,sinceitisnotamaterialthatgoesontheletter,unliketheChineseinkorthechainlube forthatmatter. 149 Section4.5. By-phrasesandagent/event-orientedmodifiersinAPass.BeyondSpanish Chapter4 (293) a. DerBriefkastenistmitArmbanduhrenvollgestopft.[mitBriefen] Theletterboxisstuffedfullwithwristwatches.[withletters] b. IhrNameistmitKettenfließfettgeschrieben.[mitroterTinte] Hernameiswrittenwithchainlubricant.[withredink] c. DerHundistmiteinerKrawattefestgebunden.[miteinerLeine] Thedogistiedupwithanecktie.[withaleash] (FromMcIntyre2015:948) (294) a. El the perro dog está is estar atado tied {con {with correa/ leash con with una a corbata/ tie *con *with guantes} gloves :‘Thedogistied{withaleash/withatie/*withgloves}.’ b. Esta this carta letter está is estar escrita written {con {with tinta ink china/ Chinese con with lubricante lubricant para for engranajes/ gears ??con ??with el the teclado keyboard de of mi my ordenador}. computer ‘Thisletteriswritten{withChineseink/withgearlubricant/??withmycomputer’skey- board}.’ Note,asafinalpoint,thatthesyntacticstructuresinGehrke(2012)assume,followingKratzer(2000), that lexical adjectivization correlates with stative predicates whereas phrasal adjectivization corre- sponds to eventive predicates. It is not clear what it explains or what that this follows from, other thantheunwarrantedassumptionthatsomehowverbalheadscanbestativebutnoteventive. Inany case,thelexicalvs. phrasaladjectivizationdividedisappearsinasyntacticapproachtoAPassforma- tionsuchastheoneweadopthere:alladjectivalparticiplesarephrasalinthesyntacticsense. 4.5.5 Wherearewe? Aswehaveseen,neitherthestate-relevancenortheevent-kindsapproachfullyexplaintheavailability of by-phrases and agent/event-oriented modification. What then, allows for this modification? Are wewronginassumingthereisnoimplicitagentinadjectivalpassives? Ianswerthelastquestionin thenegative. First, acceptingthatadjectivalpassivescontainanimplicitexternalargument, oreven causativesemanticsorientedtotheexternalargument,immediatelytakesusbacktothequestionof why by-phrasesandothermodifiersaresorestricted. Ifitistheimplicitcausativesemantics/exter- nalargumentthatisresponsibleforlicensingthesemodifiers,thenwemustassumethattheycanbe sometimesmissingandsometimesabsent,abstractingnowfromtheparticulartheoreticalimplemen- tation. 150 Section4.5. By-phrasesandagent/event-orientedmodifiersinAPass.BeyondSpanish Chapter4 Itisnotsoclear,however,thatitisimplicitcausativesemantics,orapassivevoicestructure,that licensesby-phraseswithadjectivalpassivesderivedfromtelicverbs.Conti-Jiménez(2004)notesthat by-phrases in estar-passives cannot co-occur with agent-oriented adverbs and purpose clauses (e.g. (295),fromConti-Jiménez2004:40). (295) a. El the documento document está is estar firmado signed (*/?voluntariamente) (*/?voluntarily por by Cervantes. Cervantes intended:‘ThedocumentisvoluntarilysignedbyCervantes.’ b. Este this cuadro painting está is estar pintado painted por by Velázquez Velázquez (*/?para (*/?to solicitar request la the gracia grace del of.the rey). King intended:‘ThepaintingispaintedbyVelázqueztorequesttheKing’sgrace.’ TheserestrictionsnotedbyConti-Jiménez(2004)fortheestar-passivedisappearintheser-passive,as I show in (296). Crucially, these modifiers are allowed with participles derived from StC verbs both inverbalandadjectivalpassives(e.g. (297)). Itappears,then,that by-phrasesin estar-passivesderived fromtelicverbsarenotrelatedtocausativesemanticsorpassivemorphologyinthesamewayasthose derivedfromStCverbs. (296) a. El the documento document fue was ser firmado signed voluntariamente voluntarily por by Cervantes. Cervantes ‘ThedocumentisvoluntarilysignedbyCervantes.’ b. Este this cuadro painting fue was ser pintado painted por by Velázquez Velázquez para to solicitar request la the gracia grace del of.the rey. King ‘ThepaintingwaspaintedbyVelázqueztorequesttheKing’sgrace.’ (Conti-Jiménez2004:291-92) (297) a. La the torre tower {está/ {is estar fue} was ser custodiada guarded por by diez ten soldados soldiers para to evitar avoid que that se se escape escapes el the prisionero. prisoner ‘Thetower{is/was}guardedbytensoldierssothattheprisonerdoesnotescape.’ b. El the país country {está/ {is estar fue} was ser gobernado governed por by una a dictadura dictatorship militar military para to preservar preserve los the privilegios privileges de of la the oligarquía. oligarchy ‘Thecountry{is/was}governedbyamilitarydictatorshiptopreservetheprivilegesofthe oligarchy.’ 151 Section4.5. By-phrasesandagent/event-orientedmodifiersinAPass.BeyondSpanish Chapter4 c. La the obra play {está/ {is ser fue} was ser dirigida directed voluntariamente voluntarily por by el the propio own dramaturgo. playwright ‘Theplay{is/was}voluntarilydirectedbytheplaywrighthimself.’ This situation is congruent with my proposal, which denies the existence of passive voice, or even animplicitexternalargumentinthestructureofestar-PPrts. Moreover,Ibelieveitisalsoconsistent withmyviewthatthereisnodynamiceventinthesemanticsofthe estar-PPrt. Iftherewaspassive voice/externalargumentstructure,itwouldbetheonlyinstanceofaneventiveagentivepredicatethat suppressesthedynamicportionoftheeventbutmaintainstheexternalargument.Deverbalpassive- eventive nominalizations (Grimshaw 1990; Picallo 1991, a.m.o.) are a case in point: it is not the case thattheysuppresstheprocesseventwhentheyappearwithaby-phrase(e.g.(298)),asevidencebythe possibilityofeventivemodifierssuchas inmediato‘immediate’(e.g. (298a))anddurativephrasesthat modifytheevent(e.g.(298b)). (298) a. El the inmediato immediate cierre closing de of la the unidad unit de of residuos residues tóxicos toxic por by (parte (part de) of los the agentes agents de of seguridad. security ‘Theimmediateclosingofthetoxicresidueunitbythesecurityagents.’ b. La the persecución persecution de of los the infieles infidels durante during años years por by (parte (part de) of la the Inquisición. Inquisition ‘ThepersecutionoftheinfidelsforyearsbytheInquisition.’ Stillthequestionremainsofwhatlicensesby-phrasesintheinstanceswheretheyareindeedpossible. Conti-Jiménez(2004)presentsexamplessimilartothosediscussedbyGehrke&Sánchez-Marco(2014) (althoughtheydonotquoteherdirectly)andnotesthattheby-phraseneedstobehighlyinformative: theDPneedstobeanindefiniteoradefiniteentitythatisrelevantwithinourworldknowledge(e.g. (299)), or a contrastive complement (e.g. (300a)). I further note that a definite entity modified by a relative clause is also good in this kind of examples (e.g. (300a)). This provides support to Conti- Jiménez’sview,sincetherelativeclauseishighlyinformativeinthatitsinglesouttheentityintroduced inthediscourse. (299) a. Este the cuadro painting está is estar pintado painted por by Velázquez/ Velázquez por by un a niño/ kid por by un an inexperto/ inexpert ?por by Luis. ?Luis ‘Thepaintingispainted{byVelázquez/byakid/byaninexpert/?byLuis}.’ b. El the documento document está is estar firmado signed por by Cervantes/ Cervantes por by un an desconocido/ unknown ?por ?by mi my amigo. friend 152 Section4.5. By-phrasesandagent/event-orientedmodifiersinAPass.BeyondSpanish Chapter4 ‘Thedocumentissigned{byCervantes/byanunknownperson/?bymyfriend}.’ (Conti- Jiménez2004:39) (300) a. El the documento document está is estar firmado signed por by mi my amigo, friend y and no not por by el the tuyo. yours ‘Thedocumentissignedbymyfriend,andnotbyyours.’ (Conti-Jiménez2004:40) b. El the documento document está is estar firmado signed por by el the vecino neighbor ?(que ?(who vino came a to verme see.me el the otro other día). day ‘Thedocumentissignedbytheneighborwhocametoseemetheotherday.’ (Conti- Jiménez2004:39) Still, this does not work for many other participial predicates, as (301) shows, where all types of by- phrases are ungrammatical across the board. It seems as if the participles derived from certain se- manticclassesofverbs—callthem‘creation’and‘performance’verbs,broadlyspeaking: firmar‘sign’, pintar ‘paint’, escribir ‘write’, interpretar ‘perform’ and so on—lend themselves better to certain kinds of by-phrasesthanothers, whichdonotacceptit. Perhapsthese by-phrasescouldbeanalyzedasan authorship‘by’ofsorts. (301) a. ??Este ??this teléfono phone está is estar {reventado/ {smashed reparado} repaired por by {HTC/ {HTC un a trabajador worker muy very torpe/ clumsy Felisa/ Felisa el the amigo friend del of.the que that te you.dat hablé/ talked mi my amigo, friend y and no not por by el the tuyo}. yours intended:‘Thisphoneis{assembled/repaired}by{HTC/averyclumsyworker/Felisa/the friendItoldyouabout/myfriend,notbyyours}.’ b. ??El ??the rey king está is estar {guillotinado/ {executed exiliado} banished por by {el {the pueblo/ people un an verdugo executioner cruel/ cruel Felisa/ Felisa el the amigo friend del of.the que that te you.dat hablé/ talked mi my amigo, friend y and no not por by el the tuyo}. yours intended: ‘Thekingis{executed/banished}by{thepeople/acruelexecutioner/Felisa/the friendItoldyouabout/myfriend,notbyyours}.’ Notethatevent/agent-orientedmodifiersarealsoveryrestricted—remembertheungrammaticalex- amplesin(213),whichIrepeatbelowforconvenience. Andyet,afewofthesemodifiersarepossible, asshownin(277),alsorepeatedbelow. 75 75 Kratzer(2000)alreadyobservedthatadverbialmodificationcouldbepossibleinAPass,whichshetookasevidencethat adjectivizationwasalsophrasal.Ireproduceherexamplein(xxva),notingthattheSpanishcounterpartin(xxvb)doesnot soundsogoodtomyear. 153 Section4.6. Crosslinguisticvariation Chapter4 (213) a. *La *the ciudad city está is estar implacablemente relentlessly bombardeada. bombarded intended:‘Thecityisrelentlesslybombarded.’ b. *La *the mesa table está is estar inteligentemente intelligently barnizada. varnished intended:‘Thetableisintelligentlyvarnished.’ (277) a. Este the vestido dress está is estar confeccionado made manualmente, manually aunque although no not se refl aprecie. perceive ‘Thisdressismademanually,althoughyoucannottell.’ b. Estos these héroes heroes están are estar dibujados drawn a at lápiz, pencil aunque although no not se refl note. notice ‘Theseheroesarepencil-drawn,althoughyoucannottell.’ Asaninterimsummary,Ihavearguedforaviewofestar-passivesinwhichtheeventstructureblocks thatintroduceexternalargumentsanddynamiceventsaremissingwithtelicroots.Thereremainsthe issueofhowtoaccountfortheinstancesofby-phrasesandotheragent-orientedmodifiersthatIhave discussedinthepresentsection. Whichevertheultimateanalysisofthesemodifiersturnsouttobe, theevidencepresentedsofarpointsawayfromaviewwherethereisanimplicitexternalargument ofthekindwefindinverbalpassives. 4.6 Crosslinguisticvariation 4.6.1 Permissivelanguages: anintroduction SofarwehavefocusedonSpanishestar-passives,alsomentioningotherlanguageslikeGerman,En- glishandHebrewwhereanalogousconstructionsappeartobehaveasSpanishintermsoftherestric- tionson by-phrasesandagentandevent-relatedmodification. However,thereareseverallanguages inwhichby-phrasesandagentandevent-relatedmodifiersareperfectlyfineandunrestrictednotjust withPPrtsderivedfromStCverbs—asisthecaseinSpanish,GermanandHebrewandEnglish—but (xxv) a. Die the Haare hairs waren were immer still noch sloppily schlampig combed gekämmt. ‘Thehairwasstillcombedsloppily.’ b. (?)El (?)the actor actor está is peinado combed con with poco little cuidado. care ‘Theactorissloppilycombed.’ 154 Section4.6. Crosslinguisticvariation Chapter4 alsowithPPrtsderivedfromtelicverbs. ThishasbeenreportedforGreek(Anagnostopoulou2003; Alexiadou&Anagnostopoulou2008;Alexiadouetal. 2015;Anagnostopoulou2017),Russian(Paslawska &vonStechow2003)andSwedish(Larsson2009),andseeAlexiadouetal. (2015)foranoverview. IprovideexamplesfromGreekin(302)toillustratethis. Itakeexamples(302a-c)fromAnagnos- topoulou(2003:18-19);example(302d)fromAlexiadouet al. (2015:165)and(302e)fromAlexiadouet al. (2015:184). (302) a. To the psari fish itan was tiganismeno fried apo by tin the Maria. Mary ‘ThefishwasfriedbyMary.’ (By-phrases) b. To the thisavrofilakio safe itan was prosektika cautiously anigmeno/ opened/ skopima deliberately paraviasmeno. violated ‘Thesafewascautiouslyopened/deliberatelyopened.’ (Event-orientedadverbs) c. Ta the malia hair tis the basilisas queen-GEN ine are xtenismena combed me with xrisi golden xtena. comb ‘Thehairofthequeeniscombedwithagoldencomb.’ (Instrumentals) d. To the pedhi child itan was xtenismeno combed sto in.the banio. bathroom ‘Thechildwascombedinthebathroom.’ (Locatives) e. To the spiti house ine is xtismeno built to in 1963. 1963 ‘Thehouseisbuiltin1963.’ (Temporalmodifiers) Aswecanseeintheexamplesin(302),by-phrasesandagentandevent-orientedmodificationinGreek (andalsoRussianandSwedish)isunrestrictedinawaythatevent-relatedmodificationinlanguages likeSpanish(andalsoGerman,HebrewandEnglish)isnot.LetuscalllanguageslikeGreekpermissive languagesandlanguageslikeSpanishrestrictivelanguages. Greek,likeSpanishandotherlanguages,alsohasatwo-waysplitbetweenparticiplesintermsof morphologicalandsemanticcomplexity.Anagnostopoulou(2003)distinguishestwotypesofpartici- ples:-menosparticiples,whichhaveeventimplications(e.g.(302)),and-tosparticiples,thatdonothave suchimplicationsandthusbehavemuchliketheperfectiveadjectivesinBosque(1999)(Adj-PPrtsin Bosque2014).IshowsomeexamplesofthesetwoclassesonTable4.5. Anagnostopoulou (2003) shows that -tos participles do not have event implications, whereas -menos participlesdo:assuch,acontextthatnegatesapreviouseventsuchasthatin(303)iscontradictoryin thecaseof-menosparticiples(e.g.(303a))butnotinthecaseof-tosparticiples(e.g.(303b)).Also,-menos 155 Section4.6. Crosslinguisticvariation Chapter4 Table4.5:Greek-menosand-tosparticiples(fromAnagnostopoulou2003:11) vras-menos vras-tos ‘boiled’ psi-menos psi-tos ‘grilled’ zografis-menos zografis-tos ‘painted’ skalis-menos skalis-tos ‘carved’ gram-menos grap-tos ‘written’ anig-menos anix-tos ‘opened’,‘open’ klis-menos klis-tos ‘closed’,‘close’ participlesallowevent-relatedadverbs(e.g. (304a))andby-phrases(e.g. (305a)),but-tosparticiplesdo not(e.g.(304b)and(305b))(seealso(302a)and(302b)). (303) a. #Afti #this I the varka boat ine is fusko-meni pumped alla but den not tin it exi has fuskosi pumped kanis nobody akoma. yet ‘Thisboatispumpedupbutnobodyhaspumpeditupyet.’ b. Afti this i the varka boat ine is fusko-ti pumped alla but den not tin it exi have fuskosi pumped kanis nobody akoma. yet ‘Thisboatisofthetypethatcanbepumpedupbutnobodyhaspumpeditupyet.’ (Anagnostopoulou2003:11–12) (304) a. Ta the keftedakia meatballs ine are prosektika carefully tiganis-mena. fried ‘Themeatballsarefriedcarefully.’ b. Ta the keftedakia meatballs ine are prosektika carefully tiganis-ita. fried (305) a. Ta the keftedakia meatballs ine are tiganis-mena fried apo by tin the Maria. Mary ‘ThemeatballsarefriedbyMary.’ b. Ta the keftedakia meatballs ine are tiganis-ita fried apo by tin the Maria. Mary XXXX(FromAnagnostopoulou2003:13) As a final note, Anagnostopoulou (2003) observes that the restrictions on Aktionsart regarding non- causative states and activities also apply to Greek APass. Activities are infelicitous in APass unless they appear in a ‘job-done’ context (e.g. (306)). Statives, on the other hand, either lack a participle altogether(e.g.(307))orareonlypossibleas-tosparticiples(e.g.(308)). 156 Section4.6. Crosslinguisticvariation Chapter4 (306) a. #Ta the karotsia baby-carriages ine are idhi already sprog-mena. pushed b. #I the gata cat ine is idhi already xaidhemeni. petted (307) a. O the Janis Janis kseri knows tin the apantisi. answer ‘Johnknowstheanswer.’ b. Noparticiplerelatedtotheverbksero‘know’. (308) a. O the Janis Janis gnorizi knows tin the apantisi. answer ‘Johnknowstheanswer.’ b. I the apantisi answer ine is gnos-ti. known ‘Theanswerisknown.’ (FromAnagnostopoulou2003:14) 4.6.2 Previousaccounts ThissectionpresentstheproposalinAlexiadouetal. (2015),arecentextensivework—andthefirstone, ifIamnotmistaken—thattacklesthecross-linguisticdifferencesbetweenAPass.Iwillstartdiscussing Kratzer(2000),sinceitpresentsaproposalforAPassthatwaslaterthefoundationofAlexiadouetal. ’s(2015)accountforthetypologyofAPass. 4.6.2.1 Kratzer(2000) Inherinfluentialwork,Kratzer(2000)arguesthattherearetwotypesofAPass,withdifferentseman- ticsandderivedbydistinctoperators: 1.Targetstatepassives TheyarederivedfromVPsthathaveatargetstateargumentintheirdecom- position, i.e. a reversible result state of bi-eventive VPs. 76 As a test, they accept the adverbial immer noch‘still’(e.g.(309)).Theyarederivedbytheoperatorin(310). 76 Although she does not deal with the event-modification asymmetries I have discussed in this chapter, Kratzer (2000) notesthepossibilityofformingAPassfromstativecausativeverbs,presentingexampleslike(xxvia),whichsheobserves may correspond to a telic, eventive reading or a stative reading, the latter conforming to a context such as (xxvib). For Kratzer,bothreadingshaveacausingeventualityandatargetstate,butthecausingeventualityintheeventivereadingis adynamiceventandinthestativereadingitisastativeeventuality.ThisdepartsfrommyproposalinthatIdonotbelieve thattheeventivereadingactuallyhasadynamiceventinthedecompositionoftheparticiple. 157 Section4.6. Crosslinguisticvariation Chapter4 (309) a. DieGeissleinsindimmernochversteckt. ‘Thelittlegoatsarestillhidden.’ b. DieReifensindimmernochaufgepumpt. (Kratzer2000:385) ‘Thetiresarestillpumpedup.’ (310) J;K=R <s;st> s9eR(e)(s) 2.Resultantstatepasives theyhaveperfectsemanticsandmaytakeanykindofVPexceptforsimple states,whichdonothaveaneventargument. Byresultantstateshemeansastatethatholdsforever aftertheculminationofanevent,i.e.aperfect,followingParsons(1990).Sheassumesthatthesearethe semanticsoftheperfectsofverbalpassives,theironlydifferencebeingthatinresultantstatepassives Voicedoesnotproject. Herdiagnosisforresultantstatepassivesistheunacceptabilityofimmer noch ‘still’(e.g.(311)).Theoperatoratplayisin(312). (311) a. DasTheoremist(*immernoch)bewiesen. ‘Thetheoremis(*still)proven.’ b. DerBriefkastenist(*immernoch)geleert. (Kratzer2000:386) ‘Themailboxis(*still)emptied.’ (312) J;K=Pt9e[P(e)&(e)t] As Kratzer notes, activities are bad out of the blue in APass, but they nonetheless improve if a ‘job done’or‘that’sover’contextisprovided, aswecanseein(313). Inthesecases, giventhattheylacka targetstateargument,theycombinewiththeresultantstateoperator. (313) a. DieKatzeistschongestreichelt. ‘Thecatisalreadypetted.’ b. DieserKinderwagenistschongeschoben. (Kratzer2000:388) ‘Thisbabycarriageisalreadypushed.’ CONTEXT:Itwasmyjobtopetthecatortopushthebabycarriage. (xxvi) a. Thebloodvesselwasobstructed. b. Becauseofacongenitalmalformation,tissueobstructedthebloodvessel. (FromKratzer2000:393) 158 Section4.6. Crosslinguisticvariation Chapter4 4.6.2.2 Alexiadouetal. (2015) Alexiadou et al. (2015) make use of Kratzer’s (2000) distinction between target and resultant state APasstoaccountforthecrosslinguisticdifferencesfoundinthisconstruction. Theauthorsalsofol- lowGehrke’sdistinctionbetweeneventtokensandeventkinds(seeSection4.5.3)aswellasherview (Gehrke2015)thateventkindsonlygetinstantiatedwhenverbalstructureisdirectlyembeddedunder TenseandAspect. The authors propose that Greek-type (i.e. permissive) APass are derived with an Asp operator denoting a Perfect of Result (i.e. (312)). A is merely an adjectivizer. Since Asp takes directly verbal structure (a specifierless VoiceP in their account) then the event gets instantiated and thus it can be freelymodified.APassinpermissivelanguagesarethenresultantstates. They support this idea by showing that still is ungrammatical with event-related modifiers in Greek-typelanguages(e.g.(314),fromAlexiadouetal. 2015:158).Thestructureisrepresentedin(315). (314) a. Ta the lastixa tires ine are (*akoma) (still) fuskomena inflated apo by tin the Maria. Mary ‘ThetiresarestillinflatedbyMary.’ b. Ta the lastixa tires ine are (*akoma) (still) fuskomena inflated me with tin the tromba. pump. ‘Thetiresare(*still)inflatedwiththepump.’ c. To the thisavrofilakio safe itan was (*akoma) (still) prosektika cautiously anigmeno. opened ‘Thesafewas(*still)cautiouslyopened.’ (315) Greek-typeresultantstatepassives AP A AspP Asp VoiceP Voice agent vP v RootP 159 Section4.6. Crosslinguisticvariation Chapter4 On the other hand, German-type (i.e. restrictive) APass lack Asp and hence the event does not get instantiated. As a result, event-modification is restricted (e.g. (316) for German). Their structure is givenin(317). (316) Der the Mülleimer rubbish-bin ist is (*von (*by meiner my Nichte niece / / *langsam *slowly / / *genüsslich *pleasurably /*mit / der *with Heugabel) the geleert. hayforkemptied (317) German-typeresultantstatepassives AP A VoiceP Voice agent vP v RootP Fortargetstatepassives—i.e. thosethataccept stillmodification—theyproposeauniformstructure for all languages under discussion. They observe two facts: first, target state passives freely allow by-phrases,instrumentalsandagent-orientedadverbials,whichtheytaketosignalthepresenceofa Voice projection that introduces an agent semantically, i.e. those are Voice-related modifiers. Their secondobservationisthatthosemodifiersmustsomehowrelateormodifytheresultstatedirectly,as repeatedlypointedoutintheliterature(seeSection4.5.1forreferencesanddiscussion).Theirexamples areprovidedin(318),fromAlexiadouetal. 2015:181). (318) a. To the stadio stadium ine is akomi still perikiklomeno surrounded apo by tin the astinomia. police ‘Thestadiumisstillsurroundedbythepolice.’ b. O the skilos dog ine is akomi still demenos tied me with skini. leash ‘Thedogisstilltiedwithaleash.’ c. To the stadio stadium ine is akomi still filagmeno guarded prosektika. carefully ‘Thestadiumisstillcarefullyguarded.’ d. Ta the axladia pears ine are akomi still voutigmena soaked sto in.the krasi. wine 160 Section4.6. Crosslinguisticvariation Chapter4 ‘Thepearsarestillsoakedinwine.’ Tomakesenseofthesefacts,theypositpositaVoice HOLDER phraseintroducingtheresultstate above theadjectivizer,sothatTenseandAspectcanlocatetheresultstateandallowformodificationthereof. Theirstructureisprovidedin(319): (319) Targetstatepassives(alllanguages) VoiceP Voice holder AP A vP v RootP 4.6.2.3 ProblemswithKratzer(2000) AsIseeit,Kratzer’saccounthasseveralproblems. 77 First,asithasbeenoftendiscussedintheliter- ature (Löbner 1989; Krifka 2000; Ippolito 2004; Greenberg 2006), still is an aspectual particle whose semantics include an assertion and two presuppositions, 78 as we can see in (320). The assertion and the first presupposition are taken from Krifka (2000) and the second presupposition is taken from Michaelis(1993),viaGreenberg(2006). (320) Thesemanticsofstill a. Assertion: (t). b. Presupposition1:9t’/t[(t’)]. c. Presupposition2:Itisexpected/reasonablethat:atsometimeaftert In prose, the assertion states that the proposition holds at reference time. The first presupposition saysthatatsomepriortimethatabuts(/)thereferencetime,thepropositionalsoheld. Finally,the secondpresuppositiontellsusthatitisreasonableorexpectedthatthepropositionwillnotholdat somefuturetimeafterthereferencetime. MyideahereisthattheacceptabilityeffectswithstillinAPassdoesnotreallyhavetodowithatwo- waydistinctionintheirsemantics,butratherinthesemanticsofstillitselfandthepresuppositionsit triggers.Note,forinstance,thatthesamepredicatemaygivedifferentresultsdependingonthesubject 77 SeeMaienborn(2009);Gehrke(2012)forfurthercriticismofKratzer(2000). 78 Forotherusesofstill,seeIppolito(2004). 161 Section4.6. Crosslinguisticvariation Chapter4 ittakes. Forinstance,(321a)isfinesincewecanexpectthecitytoberebuiltafter,say,abombing,but (321b)isnotsinceevidenceisexpectedtobeburned/shreddedwhenonewantstodestroyit. 79 Ifind thesameeffectsinSpanish(e.g.(322)). (321) a. Die the Stadt city ist is immer still noch noch zerstört. destroyed ‘Thecityisstilldestroyed.’ b. #Die #the Beweises evidence sind is immer still noch noch zerstört. destroyed intended:‘Theevidenceisstilldestroyed.’ (322) a. La the ciudad city todavía still está is destruida destroyed (la (the reconstrucción reconstruction todavía still no not ha has empezado). begun) ‘Thecityisstilldestroyed(reconstructionhasnotbegunyet).’ b. #Las #the pruebas proofs todavía still están are destruidas. destroyed intended:‘Theevidenceisstilldestroyed.’ When stillisnotacceptableregardlessofcontext(e.g. (311a), Iassumethattheincompatibilityisdue to the lexical semantics of the verb, but not to a difference in event decomposition—telic VPs have a result state in their decomposition regardless of whether it is transitory or permanent—or in the semanticsofAPass. 80 Forfurtherdiscussionregardingthe stilltest,seeGehrke(2012)andAlexiadou etal. (2014). Turning to Kratzer’s claim that activities become fine in APass with a ‘job done’ or ‘that’s over’ context(e.g.(313)),IobservethesameeffectinSpanish(e.g.(323)). (323) a. (?)El (?)the gato cat ya already está is acariciado. petted ‘Thecatisalreadypetted.’ (Spanish) b. (?)Este (?)this carrito carriage de of bebé baby ya already está is empujado. pushed ‘Thisbabycarriageisalreadypushed.’ CONTEXT:Itwasmyjobtopetthecatortopushthebabycarriage. 79 AlltheGermanexamplesinthissection(4.6.2.3)areduetoBeritGehrke,p.c. 80 Note that even (311b) could be acceptable in a scenario where a mailbox is supposed to be constantly filled and emp- tied:‘Themailboxisstillemptied:it’stimetofillitupagain.’ 162 Section4.6. Crosslinguisticvariation Chapter4 However,itisunclearthatthesearederivedbyaperfectaspectoperator,asKratzerhasit,thatdoes notneedtelicVPsasaninput.Ifso,itisatbestunclearwhyactivitiesinAPassneedcontexttoimprove, butactivitiesinverbalpassivesintheperfectdonot(e.g.(324-325)). (324) a. ??Die ??the Katze cat ist is gestreichelt. petted intended:‘Thecatispetted.’ (German) b. Die the Katze cat ist is gestreichelt petted worden. become ‘Thecathasbeenpetted.’ (325) a. ??El ??the gato cat está is acariciado. petted intended:‘Thecatispetted.’ (Spanish) b. El the gato cat ha has sido been acariciado. petted ‘Thecathasbeenpetted.’ FurtherevidencethatAPassdonothaveperfectsemanticsinGermancomesfromthefactthatposi- tionaladverbials(onMonday,yesterday...)arefineintheGermanpresentperfect(VPassincluded),but areoutinAPass(e.g.(326)). 81 This,ofcourse,isexpectedifGermanAPassdonothaveaninstantiated event, and hence temporal adverbials cannot locate it. The fact that these possitional adverbials are possibleintheperfectthuspointstotheimpossibilityofGermanAPasstohaveperfectsemantics. (326) a. Die the Töpfe pots sind are am on Montag Monday abgespült washed-up worden. become. ‘ThepotshavebeenwasheduponMonday.’ b. *Die *the Töpfe pots sind are am on Montag Monday abgespült. washed-up intended:‘ThepotsarewasheduponMonday.’ Similarly,APassandperfectsofverbalpassivesgivedifferentreadingswithtodayinGerman:inAPass, theunderlyingeventdoesnotneedtohavehappenedtoday(e.g. (327a)),butintheperfectsofverbal passivestheeventneedstohavehappenedtoday(e.g.(327b)). 81 IthankRoumiPanchevaforsuggestingthatItestpositionaladverbialsinAPass. 163 Section4.6. Crosslinguisticvariation Chapter4 (327) a. Heute today ist is der the Boden floor gewischt. swept ‘Thefloorisswepttoday.’(Thesweepingeventdidnotnecessarilytakeplacetoday:itcould havebeensweptbeforetoday) b. Heute today ist is der the Boden floor gewischt swept worden. become ‘Thefloorhasbeenswepttoday.’(Thesweepingeventnecessarilytookplacetoday) TheempiricalstateofaffairspresentedsofarissimilartothatdescribedforSpanishinSection4.3.4 inourdiscussionabouttheperfect. Moreover,Kratzer’sproposalfortargetstatestructurespositsa complexchange-of-statesemanticsfortheAPass,somethingIexplicitlyarguedagainstforSpanishin Section4.3.1.Again,mycriticismcanbecarriedoverforGerman. AmoreplausibleexplanationfortherelativeacceptabilityofactivitiesinAPass,asIseeit,isthat they can be coerced into telics when the context allows. ‘Job done’ contexts, for instance, favor a coercionfromactivitiesintotelicssincetheeventualityisunderstoodashavinganendpoint. Inshort,whileIthinkthatthestativizerthatKratzerproposesforhertargetstatepassivesgives ustherightsemanticsforAPass,Idonotacceptherextraperfect(i.e.resultantstate)operator,onthe basis that the nuances it intends to capture can be explained by factors not pertaining to APass per se (eg. how well the presuppositions triggered by still fare with respect to the context or the lexical semanticsofagivenverb)and,mostimportantly,becausepositingperfectsemanticsforAPassmakes thewrongpredictions,asweshowed. 4.6.2.4 ProblemswithAlexiadouetal. (2015) My criticism of Kratzer (2000) can be carried over to Alexiadou et al. (2015) since they follow her account,buthereIwillfocusonthespecificproblemsthattheiraccountfaceswiththecrosslinguistic data. Aswesaw,theresultantvs. targetdistinctioninAPassisdiagnosedbythe(un-)acceptabilityof still.ThisviewleadsAlexiadouetal. (2015)toassumethatGermanAPassthatrefusestillareresultant statepassives,asin(328)(repeatedfrom(311a)). (328) Das the Theorem theorem ist is (*immer (*still) noch) noch bewiesen. proven ButitisunclearhowthestructurethatAlexiadouetal. (2015)proposeforGerman-typeresultantstate passivesin(317)is‘resultant’,giventhatitlackstheperfectoperatorencodedinanAsphead(e.g.(312)). Ofcourse, theauthorsdonotwanttosaythatthereisAspinGermanadjectivalparticiples, sinceit 164 Section4.6. Crosslinguisticvariation Chapter4 isAspwhattheyclaimtolicenseevent-orientedandspatio-temporalmodifiersinGreek-typeAPass (i.e.thepermissivelanguages)(e.g.(315)). Theiraccountfortargetstatesin(319)isnotwithoutproblemseither. Firstandforemost,itdoes not explain how agent-oriented modifiers (eg. by-phrases, instrumentals...) are licensed, since their Voice HOLDER introduces the subject of a result state, and not that of a causative sub-event. In other words,whatAlexiadouetal. (2015)doisdisplacetheprojectiondenotingtheresultstateabovethead- jectivizer,butthatdoesnotexplainwhyVoice-relatedmodifiers—i.e. agent-orientedmodifiers—are licensed,sincewhattheycallVoice HOLDER doesnotdenoteacausativesub-eventuality,butaresulta- tiveone. Moreover, theiraccountfortarget-state passivesmakesthewrong predictions. 82 Noticetheex- ample(329a),takenfromMcIntyre(2015),whichIreproduceforSpanishin(329b). (329) a. Thedogistiedupwitharope(*byapoliceman). b. El the perro dog está is atado tied con with una a cuerda rope (*por (*by un a policía). policeman Alexiadou et al. (2015) would take (329) to be a target state passive (cf. The dog is still tied up with a rope,andseetheirpracticallyidenticalexampleforGreekin(318b)),butyet,by-phrasesaredisallowed, contrary to their prediction that target state participles allow agent-oriented modifiers across the boardbyvirtueofhavingaVoice HOLDER projection. The data presented by Alexiadou et al. (2015), I note, is not problematic for an account that dis- penseswiththetargetvs. resultantstatedistinctionandinsteadtakesanaspect-basedapproach,asI dohere. Contrastssuchas(329)arenotsurprising,becauseatado‘tiedup’isaparticiplederivedfrom atelicVP,andthusby-phrases,whichmodifythehighersub-event,willnotbepossiblegiventhatwe aredealingwithrestrictivelanguages,butmodifierssuchasconunacuerda‘witharope’,whichmodify theresultsub-event,willbepredictedtobefine. Similarly,withtelicVPsinpermissivelanguages(e.g.(314)),still-modificationisunacceptablewith by-phrases because APass are predicates of a result state which, in the case of telic VPs, temporally followsthecausingsub-eventthatby-phrasesmodify. Therefore,theadditionofaby-phraseclashes with the presuppositions triggered by still that the proposition will stop holding at a later time: the resultstatebroughtaboutbytheagentintroducedbytheby-phrasemaybetransitoryornot,butitwill haveforeverafterbeenbroughtaboutbythatagent,andthatisundoable. Notethat stillnecessarily takesscopeovertheby-phrase. 82 IowethisobservationtoRoumiPancheva. 165 Section4.6. Crosslinguisticvariation Chapter4 The effects mentioned in theprevious paragraphs do not happen with by-phrases in APass with StC VPs, which do accept still: given that the intervention of the agent and the maintenance of the resultstatearecoextensive,wecanexpectthatthepropositionwillceaseatalatertime,i.e.whenthe agentstopsactingtomaintaintheresultstate. Thisexplainsthegrammaticalityof(318a)andalsoof (318c),bothofwhicharemodifiersofthecausingsub-event.Thegrammaticalityof(318b)and(318d),on theotherhand,isexplainedbecausethemodifiersthatappearinthosesentencesareresult-oriented— andtheresultstatecanbeunderstoodascancellable(rememberourdiscussionof(329)). RecallourdiscussiononSection4.5.2regardingtheoftentimesconfusingobservationsintheliter- aturethatby-phrases—and,byextension,agentandevent-orientedmodifiers—areonlyacceptablein APassinrestrictivelanguagesiftheysomehowmaintain,pertaintooraredetectableintheresultstate: thisiswhatAlexiadouet al. (2015)meanttocapturewiththeirtargetstatestructurein(319)andwhat otherauthorshavetriedtocapturebyothermeans. 83 Theseobservationsarederivedfrommywork asfollows: by-phraseswithStCVPsareacceptablenotbecausethey‘maintain’or‘pertain’tothere- sultsub-event,butbecausethetimeargumentofthecausingsub-eventualitytheyareassociatedwith projects(andsincethecausingandresultsub-eventualitiesaretemporallycoextensive,weindeedget theeffectthatthe by-phrase‘maintains’or‘pertains’totheresultstate). Othermodifiers,inturn,are acceptable because they indeed pertain or are detectable in the result state: they are result-oriented modifiers, and since result states are always spatio-temporally instantiated, modification thereof is fullygrammatical.Thereremain,still,restrictedinstancesofby-phraseswithestar-PPrtsderivedfrom telic verbs in restrictive languages that remain to be accounted for, but which do not of themselves makeastrongcaseforVoice,orananalogousprojection,withinthePPrt:See4.5.2fordiscussion. 4.6.3 Analternativeaccount IarguedinSection4.6.2.4thatthetargetvs.resultantstatedistinctionisnotusefulforaccountingfor APasswithinoracrosslanguages.Ihavealsoarguedthroughoutthischapterthattherearenoperfect oranterioritysemanticsin-builtinAPassinrestrictivelanguages. Thecrucialdivide,Ihaveargued, isaspectual: inrestrictivelanguages,APassderivedfromStCverbsretaintheirexternalargumentin theireventstructureandthosederivedfromtelicverbsdonot,as(330)and(331)illustrate. (330) a. APassderivedfromtelicverbs (restrictivelanguages) (331) a. APassderivedfromStCverbs (restrictivelanguages) 83 Gehrke (2012), for instance, posited that there are two kinds of by-phrases: one that is event(-kind)-oriented, which can only introduce kind-level participants, and a second one that is result-oriented, which can introduce any kind of participants.ThiswasdiscussedinSection4.5.3andillustratedontheexamples(284)and(285). 166 Section4.6. Crosslinguisticvariation Chapter4 b. AdjP Adj resP b. AdjP Adj initP init resP ForSpanish,IencodedtherestrictionsonAktionsart(theunderlyingverbmusthavearesultstateand be itself stative) in the adjectivizer that derived estar-passives. I propose that this can be extended to restrictive languages as well (German, Hebrew), inasmuch as they also show the event structure restrictionsobservedforSpanish. Permissive languages like Greek, on the other hand, are more permissive in terms of Aktionsart: theydisallow(non-StC)statesandactivities(unlesscoerced),buttheycanhavefull-fledgedtelicverbs, i.e. withoutatruncatedverbalstructureasinrestrictivelanguages. Theirmaximalstructureisasin (332a). (332) a. APass(permissivelanguages) b. AdjP Adj PassP Pass initP init procP proc resP 4.6.3.1 Option1. ParametrizationofAdj(tobediscarded) A possible analysis, which I will eventually discard, is that the locus of crosslinguistic variation is the head Adj. This head has a universal property and a parametric property. The universal one is schematizedin(333a),anselectionalcategorialfeature,[ures],thatneedstomatchasyntacticcategory ofthesametype,i.e. resP.ThisderiveswhyinAPasscrosslinguisticallythereistherestrictionagainst activities and non-causative states: they do not contain resP in their decomposition and hence the uninterpretableresfeatureinAdjcannotbedeleted. The parametric feature is illustrated in (333b), already introduced in Section 4.4.1 (see also (235)). Basically, it is a partial identity function that poses the condition in the verbal predicate it takes be stative. As such, telic predicates are also out, since they contain procP. This is the case with restric- 167 Section4.6. Crosslinguisticvariation Chapter4 tive languages: I argued that in APass derived from typically telic verbs in such languages, the Adj takes as its complement a truncated verbal structure that only contains the result state. In permis- sivelanguages,however,nosuchrestrictionapplies:theAdjheadissemanticallyvacuousandhasno restrictionagainsteventivepredicates:assuch,atelicbasecanbeadjectivized. (333) ParametricvariationofAdj(tobediscarded) a. Uninterpretablecategorialfeature:[ures] Universal b. Semanticrestrictiontostates:P <s,t> [s.P(s)] Onlyinrestrictivelanguages 4.6.3.2 Option2. StatPandAdjP ThefirstoptionsketchedonSection4.6.3.1facesdifficultieswhencontrastedwithlanguageslikeChi- chewa, a Bantu language. In Chichewa, as described by Dubinsky & Simango (1996) (and references therein),therearetwokindsofagent-demotingconstructionsanalogousintheirsyntactico-semantic properties to VPass and APass in restrictive languages: the passive is exemplified in (334a) and the stativein(334b).Thepredicateinbothconstructions,crucially,isverbal. (334) a. Nyemba beans zi-na-phik-idwa agr-past-cook-pass ‘Thebeanswerecooked.’ b. Nyemba beans zi-na-phik-ika agr-past-cook-stat ‘Thebeanswerecooked.’ (FromDubinsky&Simango1996:751) They report from Mchombo (1993) the finding that, in passives, by-phrases are optional (e.g. (335a)), but are disallowed in statives (e.g. (335b)). Also, passives allow purpose clauses controlled by their implicit agent (e.g. (336a)) and agent-oriented adverbs (e.g. (337a)), whereas the stative construction allowsneither(e.g.(336b)and(337b)). (335) a. Mbale plates zi-na-tsuk-idwa agr-past-wash-pass (ndi (by Naphiri) Naphiri ‘Theplateswerewashed(byNaphiri).’ b. *Mbale *plates zi-na-tsuk-ika agr-past-wash-stat ndi by Naphiri Naphiri ‘TheplateswerewashedbyNaphiri.’ (336) a. Chakudya food chi-na-phik-idwa agr-past-cook-pass kuti [so].that anthu people a-sa-fe agr-neg-die ndi from njala hunger 168 Section4.6. Crosslinguisticvariation Chapter4 ‘Thefoodwascookedsothatpeopleshouldnotdieofstarvation.’ b. *Chakudya *food chi-na-phik-ika agr-past-cook-stat kuti [so].that anthu people a-sa-fe agr-neg-die ndi from njala hunger ‘Thefoodwascookedsothatpeopleshouldnotdieofstarvation.’ (337) a. Chitseko door chi-na-tsek-edwa agr-past-close-pass mwadala deliberately ‘Thedoorwascloseddeliberately.’ b. *Chitseko *door chi-na-tsek-eka agr-past-closed-stat mwadala deliberately ‘Thedoorwascloseddeliberately.’ (FromMchombo1993,viaDubinsky&Simango1996:751) Ontheirpart,Dubinsky&Simango(1996)furtherobservethatinstrumentphrasesareonlyallowed inthepassive(e.g. (338a)), butnotinthestative(e.g. (338b)). Furthermore, theyshowthatthestative constructionis subjectto Aktionsartrestrictions, whereasthe passiveis not. Thus, while change-of- state verbs can have a passive and a stative form (see Table 4.6), non-change-of-state verbs can only formthepassive,butnotthestative(seeTable4.7).Again,weseehowthestativeconstructionbehaves similarlytoAPassintherestrictivelanguageswehavediscussed. 84 (338) a. Kalata letter i-na-lemb-edwa agr-past-write-pass (ndi (with pensulo) pencil ‘Theletterwaswritten(withapencil).’ b. Kalata *letter i-na-lemb-eka agr-past-write-stat (ndi (with pensulo) pencil ‘Theletterwaswrittenwithapencil.’ (FromDubinsky&Simango1996:752) Table4.6:Change-of-stateverbsinChichewawithstativeandpassive 84 For a full parallelism between the Chichewa stative construction and APass in restrictive languages, it remains to be seenwhetherstativecausativeverbsinChichewa(govern,control,protect,etc.)allowthestativeconstructionornot.These verbsarenotdiscussedinDubinsky&Simango(1996),butlackofdatapreventsmefromknowingwhethertheseinfact cannotformstativesortheydobutwereoverlookedbytheauthors—rememberthatacommonassumptionforAPassin theliteratureisthattheyareonlyformedwithchange-of-state/telicverbs. 169 Section4.6. Crosslinguisticvariation Chapter4 verb stative passive a. phika ‘cook’ phik-ika phik-idwa b. swa ‘break’ sw-eka sw-edwa c. kumba ‘dig’ kumb-ika kumb-idwa d. pinda ‘bend’ pind-ika pind-idwa e. meta ‘shave’ met-eka met-edwa Table4.7:Non-change-of-stateverbsinChichewawithstativeandpassive verb stative passive a. luma ‘bite’ *lum-ika lum-idwa b. omba ‘slap’ *omb-eka omb-edwa c. ombela ‘shoot’ *ombel-eka ombel-edwa d. kuwiza ‘jeer’ *kuwiz-ika kuwiz-idwa e. kumbatila ‘embrace’ *kumbatil-ika kumbatil-idwa ItakethestativeconstructioninChichewatobetheverbalcounterpartofadjectivalpassivesinre- strictive languages (Spanish, German, Hebrew...). If so, the inevitable analysis is one in which the stativizerandtheadjectivizerfunctionsaredividedintwodifferentprojections. Letuscallthesta- tivizerprojectionStat(ivizer)P,morphologicallyverbal,andtheadjectivizerAdjP,withtheproperties in(339).ThetypologyofAPassobservedsofar,inturn,canbederivedasin(340). (339) a. JAdjK=Semanticallyvacuous b. JStatK=P <s,t> [s.P(s)] (340) a. Permissivelanguages(Greek,Swedish...)AdjP b. Restrictivelanguages(Spanish,German...)AdjP>StatP c. ChichewaStatP TheadjectivalparticiplesinpermissivelanguageslikeGreekandSwedishcancontainadynamicevent in their decomposition, and therefore they are not formed with a StatP (e.g. (341a)). Restrictive lan- guages(German,Spanish,Hebrew...)requirethattheparticiplebestative,andhencetheyhaveaStatP inadditiontoAdjP(e.g.(341b)).Finally,Chichewa(andotherpotentiallanguageslikeit)hasastativity requirement,buttheverbalbaseisnotadjectivized:thereforeitonlyhasaStatP(e.g.(341c)). 170 Section4.7. Towardsaunifiedtheoryofparticiples Chapter4 (341) a. Permissivelanguages 85 AdjP Adj init/proc/resP init/proc/res XP b. Restrictivelanguages AdjP Adj StatP Stat init/resP init/res XP c. Chichewa 86 StatP Stat init/resP init/res XP 4.7 Towardsaunifiedtheoryofparticiples So far I have discussed adjectival passives and, to a lesser extent, verbal passives. I have remained silent, for the most part, with respect to the issue of participial syncretism, which extends not only topassivesbuttoperfects. Asisknown,thesameparticipialformmayfunctionasanadjectivalpas- sive(e.g. (342a)), a verbalpassive(e.g. (343a)) oraperfect (e.g. (344a)). Inthis work, Ihavetentatively assumedthattheparticipialmorphologyofadjectivalpassivesishostedintheadjectivizingheadAdj (e.g. (342b))whereasthatofverbalpassivesishostedinthepassivizingheadPass(e.g. (343b)). Bythe 85 I assume that the difference between -menos and -tos participles in Greek (see Table 4.5 ) lies in that the former target resPandthelatterprocPand(maybe)initP(seeLarsson2009forthesameproposal).Alexiadou&Anagnostopoulou(2008) derivethelackofeventiveimplicationsin-tosparticiplesbyproposingthattheyareunverbalizedrootsattachedtoAsp. However,Anagnostopoulou&Samioti(2013)presentdatafrom-tosparticipleswithovertverbalizersthatchallengethat claim. The issue remains as to how to derive the context-sensitivity lexicalization of Adj as -tos or -menos if we do not wanttosaythattheyaredifferentheads. 86 IpredictthatChichewacanhavethestativeconstructionwithStCverbsaswellaswithtelicverbs,butasofnowIonly haveempiricalevidenceforthelatter. 171 Section4.7. Towardsaunifiedtheoryofparticiples Chapter4 same reasoning, the participial morphology in the perfect would be associated to a Perf head (e.g. (344b)). (342) a. El the portón gate está is estar cerrado. closed ‘Thegateisclosed.’ AdjectivalPassive b. AdjP Adj -ado resP (343) a. El the portón gate fue was ser cerrado. closed ‘Thegatewasclosed.’ VerbalPassive b. PassP Pass -ado initP (344) a. Pedro Pedro ha has cerrado closed el the portón. gate ‘Pedrohasclosedthegate.’ ActivePerfect b. PerfP Perf -ado VoiceP This account assumes that the morpheme -ado lexicalizes different syntactic heads whose common groundisnotimmediatelyobvious. Inotherwords,whyisitthatadjectival,verbalandperfectpar- ticiplesallenduplexicalizedbythesamemorpheme,notjustinSpanish—whichcouldverywelljust beanidiosyncrasyofthelanguage,justlikeEnglishalsouses -edforthepasttense)—butinsomany otherlanguages?Thissystematicsyncretismofparticipialmorphologycalls,Ibelieve,foraprincipled explanation. Although the issue goes beyond the scope of this chapter, here is a suggestion for an alterna- tive analysis, drawing from Larsson & Svenonius’s (2013) discussion on the syncretism between the passive—verbalandadjectival—andtheperfectparticiplesinlanguageslikeEnglish,German,Dutch and Danish. The authors argue for a single dedicated projection that hosts participial morphology withgeneralizedsemantics.Irepresenttheprojectionin(345),whichIcallPPrtP. (345) PPrtP PPrt -en VP This projection is hosted within a phase lower than that which hosts passive and perfective syntax, andthusthepredictionisthatthephonologicalrealizationofthePPrtwillbeblindtopassiveandper- fectivesyntax. Indeed,thisisborneoutnotonlyinthelanguagesdiscussedinLarsson&Svenonius 172 Section4.7. Towardsaunifiedtheoryofparticiples Chapter4 (2013), butalsoinSpanish, aswehaveshown. Thatis, wedonotencountermismatchesinthemor- phologyofparticiplesacrosspassiveandperfectenvironments,as(346a-c)exemplifies:whenthereis amismatchinform,aswith‘cut-short’participles(e.g. (346d)),wearedealingwithwhatBosquecalls ‘perfecitve’adjectives—i.e.thosewithoutanyeventstructurewhatsoever. (346) a. Pedro Pedro ha has {secado/ dried *seca} dry la the toalla towel al at.the sol. sun ‘Pedrohas{dried/dry}thetowelinthesun.’ ActivePerfect b. La the toalla towel fue was {secada/ dried *seca} dry al at.the sol. sun ‘Thetowelwas{dried/dry}inthesun.’ Verbalpassive c. La the toalla towel está was {secada/ dried *seca} dry al at.the sol. sun ‘Thetowelwas{dried/dry}at.thesun.’ Adjectivalpassive d. La the toalla towel está is seca dry (*al (*at.the sol). sun ‘Thetowelisdry(*inthesun).’ ‘Adjective Similarly,whenwehaveanirregularPPrt(e.g. roto‘broken’,andseethelistin(195),itremainsirregular across the board, i.e. a situation in which there is an irregular form for the passive and a regular form for the perfect, or viceversa, is not attested. Nor do we find such a situation between verbal and adjectival passives either. I exemplify the unattested patterns in (347) for hypothetical regular participlesandin(348)forhypotheticalirregularparticiples,wherethenon-existingparticipialform isrepresentedinitalics. (347) UnattestedpatternsinSpanishwithnon-existingregularparticiples a. Pedro Pedro ha has rompido breaked el the vaso. glass El the vaso glass {fue/ {was ser está} is estar roto. broken ‘Pedrohasbreakedtheglass.Theglass{was/is}broken.’ RegularPerfectPPrt—IrregularPassivePPrt b. María María ha has escrito written la the carta. door La the carta door {fue/ {was ser está} is ser escribida. writed 173 Section4.7. Towardsaunifiedtheoryofparticiples Chapter4 ‘Maríahaswrittentheletter.Theletteris estar writed.’ IrregularPerfectPPrt—RegularPassivePPrt c. La the mesa table fue was ser ponida. setted La the mesa table está is estar puesta. set ‘Thetablewassetted.Thetableisset.’ RegularVerbalPPrt—IrregularAdjectivalPPrt d. El the caso case fue was ser resuelto. solved.irreg El the caso case está is resolvido. solved.reg ‘Thecasewassolved.Thecaseissolved.’ IrregularVerbalPPrt—RegularAdjectivalPPrt (348) UnattestedpatternsinSpanishwithnon-existingirregularparticiples a. Fernando Fernando ha has inflado inflated.reg el the globo. balloon El the vaso balloon {fue/ {was ser está} is estar inflo. inflated.irreg ‘Fernandohasinflatedtheballoon.Theballoon{was/is}inflated.’ RegularPerfectPPrt—IrregularPassivePPrt b. Carmen Carmen ha has cargo charged.irreg el the móvil. cellphone El the móvil cellphone {fue/ {was ser está} is ser cargado. chargedreg ‘Carmenhaschargedthecellphone.Thecellphoneischarged.’ IrregularPerfectPPrt—RegularPassivePPrt c. La the falda skirt fue was ser cosida. sewed La the falda skirt está is estar cosa. sewn ‘Theskirtwassewed.Theskirtissewn.’ RegularVerbalPPrt—IrregularAdjectivalPPrt d. La the vasija vessel fue was ser molda. molded.irreg La the vasija vessel está is estar moldeada. molded.ref ‘Thevesselwasmolded.Thevesselismolded.’ IrregularVerbalPPrt—RegularAdjectivalPPrt Thisstateofaffairsdepictedin(347)and(348)receivesanaturalexplanationifindeedparticipialmor- phology lies in a phase inaccessible to perfect and passive formation. A further advantage of this approach is that, under our framework, irregular participial exponents do not need to be multiply specifiedinthelexiconfortheconstructionsthattheyappearin(e.g. forPass,Perf...). Instead,anir- regularparticiplelike abierto‘opened’wouldonlyhavetobelistedasPart,asin(349),alongwiththe restoftheverbalcategoryfeatures. 174 Section4.8. Conclusions Chapter4 (349) abierto‘opened’ categoryfeatures:[Part,init,proc,res] Whilethisapproachismorphologicallysound,itopensanothercanofwormsaltogetherforthesyn- taxandsemanticsofPPrts. IfPPrtsconstituteaphasethatisinvisibletopassivizationandperfects, how can passive formation be restricted to transitive PPrts (i.e. those that contain initP in their de- composition)? Howdoesadjectivalpassiveformationknowthatwearedealingwithastativeverbal predicate,undertheassumptionthatittakesplaceabovePart? Orhowdoweensurethattheperfect isfedaverbalparticiple,andnotanadjectivalone(e.g.stativeversionsoftelicverbs)? Awayaroundthiswouldbetogiveuptheideathat,whilethelocusofparticipialmorphologyis asinglePPrthead,itsprojectiondoesnotconstituteaphaseboundary. Assuch,passiveandperfect operatorscanaccesstheinformationwithinthePPrt.Still,thequestionremainsastowhatsyntactico- semantic component of this PPrtP is common to passives—adjectival and verbal—and perfects so thatitsappearanceinallthreeconstructionscanbenaturallyaccountedfor. Suchanissuegoeswell beyondthescopeofthischapter,andsoIputitasideforfutureresearch. 4.8 Conclusions Thischapterhaspresentedacasestudyofadjectivalpassivesinthelightofthetheoryofverbalaspect developedinChapter3.IhaveshownthatSpanishpastparticipleshavedifferentpropertiesdepending on whether they appear in attributive or predicative position. Differentiating participles in such a mannerhashelpedusunderstandbetterthepropertiesofverbalandadjectivalparticiplesinSpanish. Forthelatter,Ihavemadeacasefortwomainandrelatedpoints: i)ThekeyroleofAktionsartforthe distributionofby-phrasesandotheragent-orientedmodifiersand,ultimately,thestructuralmakeup oftheparticiple; ii)Aviewofadjectivalpassiveformationthatdoesnotinvolveanperfectaspectual operatororchange-of-statestructure, butmerelyanunderlyingstativepredicatewhich, inthecase oftelicverbs,isnecessarilytruncated. MymaincontributioninthisChapterhasbeentoprovideanovelviewofwhatitmeansforthese adjectivalpassivestobe‘stative’and‘resultative’. Theresultativityofadjectivalpassives,inthestrict senseofresultfromapreviousevent,isjustaninferencefromaverbalrootthat,despitebeingableto lexicalize full-fledged telic structures in other contexts, only lexicalizes a stative projection (resP) in adjectivalpassives.Inotherwords,itisnotaresultativestructure,butonlystative.Ontheotherhand, adjectivalpassivesfromstativecausativeverbsarebothstativeandresultative,inthebroadersenseof theword: theseinvolvecausative-resultativeconfigurationswithoutadynamicevent,andthusthey arestative. Resultativity,justasIarguedforcausativityinChapter3,arisesconfigurationally,andis notencodedinasinglehead. 175 Section4.8. Conclusions Chapter4 For the theoretical implementation of the aspectual typology of adjectival passives, I adopted a late-insertion approach governed by the Superset Principle, namely that the category features of a givenrootmustbeasuperset,properornot,ofthecontiguoussyntacticheadstheylexicalizeinthe syntacticstructure.Thus,arootlikebreakislistedwiththecategoryfeaturesinit,procandres,butthat doesnotmeanthatallsuchfeaturesneedtobeassociatedtotheircorrespondingsyntactichead:they mayverywellassociatetojust procand res,whichdeliversunaccusative break(e.g. The glass broke)or simply res, whichIclaimdeliversadjectivalparticiples(e.g. The glass is broken). Suchamodelallows ustobuildstatesoutoftypicallychange-of-stateverbswithouttheneedofadditionalresultativeor perfectiveoperators,whichmoreovermakethewrongpredictions. A bigger question that deserves further treatment is how to model the link between the mor- phosyntaxandsemanticsofparticiplesinthelightoftheirpolyfunctionality, i.e. thatthesamepar- ticipialformmayfuncionasanadjectivalandverbalpassiveandaperfect.ThatisanissuethatIintend toexploreinfuturework. 176 5 Stativepsychologicalandlocativeverbs 5.1 Introduction Thelasttwochaptershavebeendevotedtostativepredicatesacrosscategorialdomains(verbsandpar- ticiples)andtheirmorphosyntacticrealization.IhavestudiedstativecausativepredicatesinChapter 3andthebuildingofadjectivalpassivesinChapter4. Insodoing,Ihavediscussedtheirimplications forthemodelingofthesyntacticstructureoftheupperandlowerlayersoftheVP,wherecausativity andresultativityariseconfigurationally,aswellastherelationshipbetweenmorphosyntax,semantics andthelexicon.Thisresearchhasbeenundertakenfromaneoconstructionistprism,pursuingaview inwhichthesyntacticstructuredeterminesformalaspectualmeaningandinwhichwords—suchas participles—arebuiltfromsyntacticstructure. Thepresentchaptercontinuesthisneoconstructionistinquiryonstativepredicates,shiftingthe focustothematicinterpretation.Forthesepurposes,Ifocusonstativeobject-experiencerpsycholog- icalverbs(henceforthOEPVs)andlocativeverbsinrelationtotheirthematicstructure. Iexemplify OEPVsin(350)andlocativeverbsin(351). (350) preocupar‘worry’,asombrar‘amaze’,aterrar‘terrify’,aburrir‘bore’,molestar‘annoy’... (351) rodear‘surround’,cubrir‘cover’,flanquear‘flank’,bloquear‘block’,cercar‘fence’,obstruir‘obstruct’... It has been often argued in the literature for many languages that stative OEPVs have a specialized structure that inherently assigns an Experiencer thematic role to the direct object. The reasoning behinditisthattheseverbsdisplaygrammaticaleffectsdifferentfromotherstativeverbs,andthese are mostly related to the object (e.g. accusative case patterns in Russian and object clitic effects in GreekandHebrew).ThesegrammaticalidiosyncrasiesofOEPVsareknownaspsycheffects. ThegoalofthischapteristoarguethatstativeOEPVsandlocativeverbsshareacommonsyntax, andassuchtheviewthatthereisspecializedstructurethatassignsanExperiencer-rolecannotbe maintained. I make my case by showing that psych effects also appear with stative locative verbs. 177 Section5.2. ‘Psych’verbs:stateoftheart Chapter5 I show how this parallelism between stative OEPVs and locative verbs is crosslinguistically robust, drawingempiricalevidencefromSpanish,Italian,Hebrew,GreekandRussian. My proposal is that both verb types lexicalize verbal structure with a lower prepositional layer. Crucially,thisPdoesnotassignathematicroletoitscomplement—nordoesanyotherhead,forthat matter,assignathematicroletoanyotherargument. Rather,thisprepositionalstructureintroduces anabstractrelationbetwentwoarguments:iftheverbthatlexicalizesPisconceptuallypsychological (e.g.(352a)),thensuchabstractrelationwillbeunderstoodasmental,whereasiftheverbislocative(e.g. (353a)), then the relation will be spatial.The event roles in the structures in (352b) and (353b) are then not thematic roles, but merely entailments from the event structure of the VP coupled with world knowledgeabouttheconceptualtypeofverbinvolved. (352) a. La the crisis crisis aterra frightens a dom las the familias. families ‘Thecrisisfrightensfamilies.’ b. VP V aterra PP Lacrisis Stimulus P’ P alasfamilias Experiencer (353) a. La the sábana sheet tapa hides la the entrada. entrance ‘Thesheethidestheentrance.’ b. VP Lasábana Figure V’ V tapa PP P laentrada Ground This chapter is structured as follows. Section 5.2 provides an overview of OEPVs and psych effects crosslinguistically. Section5.3introduceslocativeverbsandshowsthattheyalsoshow‘psycheffects’ across the board. In Section 5.4, I develop my proposal for the syntax and semantics of OEPVs and locative verbs. Before concluding, I devote Section 5.5 to argue against agentivity as a determining factorofpsycheffects,showingthatitisstativityinstead.Finally,IpresentmyconclusionsinSection 5.6. 5.2 ‘Psych’verbs:stateoftheart 5.2.1 Experiencersandthelinkingproblem Intheirbynowclassicwork,Belletti&Rizzi(1988)providedathree-wayclassificationofpsychological verbs for Italian, illustrated in (354) and (355). In (354), we have a pair of sentences with a transitive 178 Section5.2. ‘Psych’verbs:stateoftheart Chapter5 psychological verb that displays a standard nominative subject and accusative object configuration, whereasin(355)wehaveaverbthatdisplaysanominative-dativeconfigurationinwhichtheorderof argumentscangoeitherway. (354) a. Gianni Gianni teme fears questo. this b. Questo this preoccupa worries Gianni. Gianni (355) a. AGiannipiacequesto. toGiannipleasesthis b. QuestopiaceaGianni. thispleasestoGianni (Belletti&Rizzi1988:291-92) In the first class of psychological verbs illustrated in (354a), the nominative subject is the entity that experiencestheemotionalstate: itbearstheExperiencer-role. Belletti&Rizzi(1988)callit‘Class I’butIrefertothemfromnowonassubject-experiencerpsychologicalverbs(SEPVs). Inthesecond class,exemplifiedin(354b),theExperienceristheaccusativeobject(i.e.OEPVs,whichBelletti&Rizzi 1988 simply label ‘Class II’). In the third class (Class III), the Experiencer is a dative argument: I call themdative-experiencerpsychologicalverbs(DEPVs). Thisstateofaffairsisproblematicfor-theory(Chomsky1981;Stowell1981).Insuchtheory,within theGovernmentandBindingframework,verbswereassumedtohavelexicalentriesthatspecifythe numberofargumentsthattheytake,theirgrammaticalroleinthesentenceaswellasthe-rolesas- sociatedwiththem. LinkingstipulationssuchastheProjectionPrinciple(seealsofootnote12,Ch. 2) ensuresthatthegrammaticalinformationincludedinthelexicalentryisprojectedinsyntaxproper. Inturn,theviewthatthethematicrolesofaverb’sarguments(Agent,Theme,Goal...) correlatewith thegrammaticalfunctionofsucharguments(subject,object...) wascrystallizedinBaker’s(1988)Uni- formityofThematicAssignmentHypothesis(UTAH),statedin(356). (356) UniformityofThematicAssignmentHypothesis(Baker1988:46) Identicalthematicrelationshipsbetweenitemsarerepresentedbyidenticalstructuralrelation- shipsbetweenthoseitemsatthelevelofD-structure. Intuitively, something like the UTAH appears to be correct. For instance, we do not (tend to) find languageswhereThemesaresubjectsandAgentsareobjects, buttheotherwayaround. Thatis, we donotfindsentenceslike(357b),whichdepictimaginaryverbswhosesubjectwouldbeaThemeand theirobjectanAgent.UnderaUTAHapproach,Agentsareuniversallyprojectedin,say,(Spec,VP)and ThemesascomplementsofV,asin(358). 179 Section5.2. ‘Psych’verbs:stateoftheart Chapter5 (357) a. Johnhit/built/found/pushed/bought/cleaned/broke/describedthetable. b. *Thetableplit/puilt/vound/fushed/pought/bleaned/proke/tescribedJohn. (Baker 1997:76) (358) VP Agent V’ V Theme Psychological verbs, then, seem to contradict the UTAH, as it has been repeatedly observed. If we assumethatthetransitivesentencesin(359)correspondtothesamesyntacticstructure,itisamystery howtheExperiencercanbeasubjectin(359a)andanobjectin(359b)or,toputitdifferently,howwe couldhaveanAgentsubjectin(357a),anExperiencersubjectin(359a)andaThemesubjectin(359b)if weassumetheyprojectinthesamesyntacticposition. (359) a. JohnfearsMary. SEPVsXXExperiencer»Theme b. MaryfrightensJohn. OEPVsXXTheme»Experiencer Several solutions have been proposed with respect to this asymmetry. One is to adopt a relativized versionoftheUTAH(RUTAH)(Belletti&Rizzi1988;Grimshaw1990;Pesetsky1995),wherethe-roles thataverbassignsprojectinthesyntaxaccordingtotheirrelativeorderonathematichiearchy(e.g. (360),fromGrimshaw1990). 87 (360) Agent>Experiencer>Theme Approachesheredifferastowhethertheytakethenon-ExperiencerargumenttobeaThemeinboth casesornot.Belletti&Rizzi(1988)assumetheyareThemesinbothcases:forthem,theThemesubject inOEPVsisderived,i.e.OEPVsareunaccusative,andsotheThemedoesnotc-commandtheExperi- enceratD-structureinOEPVs. ForSEPVs,theauthorsassumethattheyaretransitiveD-structures, withtheExperiencerhigherthantheTheme). ThesecondstrategywithintheRUTAHapproachispursuedinGrimshaw(1990);Pesetsky(1995), whoclaimthatthesubjectsofOEPVsareinfactCausers,notThemes.InSEPVs,the-rolesofExpe- riencerandThemearerealizedassubjectandobject,whereasinOEPVstheobjectisanExperiencer andthesubjectisaCauser,derivedinPesetsky(1995)byanullcausaffix.Thethematichierarchythat Pesetsky assumes is provided in (361) (Pesetsky refers to the Theme in OEPVs as the Target/Subject Matter(T/SM)oftheemotion). 87 Formorearticulateversionsofthematichierarchies,seeVan-Valin(1990);Van-Valin&LaPolla(1997),withintheframe- workofRoleandReferenceGrammar. 180 Section5.2. ‘Psych’verbs:stateoftheart Chapter5 (361) Causer>Experiencer>Target/SubjectMatter AnothertakethatdiffersfromtheRUTAHisfoundinDowty(1991). Insteadofassumingarichhier- archyof-roles,Dowtyproposesthattherearejusttwounderspecifiedthematicroles: proto-Agent (realized as subject) and proto-Patient (realized as object), each associated to a cluster of different properties that determine the realization of a verb’s arguments as one or the other. Psych verbs, he claims,donotprovideaclear-cutchoicefortherealizationoftheirarguments:Experiencersaresen- tient, a typical property of Agents, whereas the Stimulus (following the terminology in Talmy 1985) bringsaboutareactionintheExperiencer,apropertythatisalsotypicalofAgents. Therefore,both argumentsareequally(weak)candidatesforbeingsubjects(proto-Agents). Afinalalternativewouldbetogiveup-theoryaltogetherinfavorofanaspectualroleapproach to argument structure, in the lines of Tenny (1987, 1994) and Hale & Keyser (1993). This approach fits well with stative-causative approaches to OEPVs (Pylkkänen 2000; Arad 1998b, 2002; Kratzer 2000): while the subjects of SEPVs can be generalized holders of a state (emotional or not), the subjects of OEPVsareCausersthatmaintainacertainstateinanobject,i.e. theobjectistheholderofthatstate (emotionalornot).Iwilldefendanaspectual-roleapproachinthischapter,butIwillcontendthatthe relevantstructureforderiving‘Experiencer’objectsinOEPVsislocative(inthelinesofLandau2010), ratherthanstativecausative. 5.2.2 ‘Psych’effects Inadditiontothelinkingproblemdiscussedabove, psychverbshavebeenobservedtohaveseveral propertiesthatsinglethemoutfromothertypeofverbalpredicatesinmanylanguages.Landau(2010) referstothesepropertiesas‘psych’effects. Thesepropertiesappear, forthemostpart, withOEPVs. I enumerate them below, drawing from Landau’s discussion in his 2010 monograph. The different proposalstryingtoaccountforthesepsycheffectsaretoonumeroustoreviewhere,soIwillmerely presenttheempiricaldatainthissectionandrefertospecificaccountswhenrelevantforthediscus- sionofmyown. Belletti&Rizzi(1988)notethatexperiencerobjectsareislandsforwh-extraction.In(362a),wehave aDP(ilcandidato)thatisrelativizedoutofabiggeraccusativeobjectDP(isostenitori...).However,when theaccusativeobjectisanExperiencer,itconstitutesanislandforextraction(362b).Theexamplesin (362)arefromBelletti&Rizzi(1988:325),followinganoriginalobservationbyBenincà(1986). (362) a. Il the candidato candidate di of cui whom questa this ragazza girl apprezza likes i the sostenitori. supporters b. *Il *the candidato candidate di of cui whom questa this prospettiva perspective impaurisce frightens i the sostenitori. supporters 181 Section5.2. ‘Psych’verbs:stateoftheart Chapter5 Furthermore,accusativeexperiencerobjectscanappearascliticsrelatedtoaleft-dislocatedDPwitha dativemarkera(e.g.(363a)).Thisisnotallowedwithothertransitive-accusativeverbs,SEPVsincluded (e.g. (363b)). ThisisnotedbyBelletti&Rizzi(1988:334),followinganoriginalobservationbyBenincà (1986). (363) a. A to Giorgio, Giorgio questi these argomenti arguments non not l’hanno him.have convinto. convinced b. *A *to Giorgio, Giorgio la the gente people non not lo him conosce. know. Belletti&Rizzi(1988)alsoobservethatperiphrasticcausativesinItalianarenotallowedwithOEPVs (buttheyarewithSEPVs). Thispatternisexemplifiedin(364). Theauthorsfurthernotethatreflex- ivizationofOEPVsisnotpossible,unlikeSEPVs.Thisisillustratedintheexamplesin(365). (364) a. *Questo *this lo him ha has fatto made preoccupare/ worry commuovere/ move attrarre attract ancora even più more a to Mario. Mario ‘ThismadeMarioworry/move/attracthimevenmore.’ b. Questo this lo him ha has fatto made apprezzare/ estimate temere/ fear ammirare admire ancora even più more a to Mario Mario ‘ThismadeMarioestimatefearadmirehimevenmore.’ (Belletti&Rizzi1988:303) (365) a. lo I mi myself conosco. know ‘Iknowmyself.’ b. *Io *I mi myself interesso. interest intended:‘Iinterestmyself.’ (Belletti&Rizzi1988:296) Another psych effect involves the genitive of negation. In Russian, when an object bears accusative object, it may shift to genitive under negation, as shown in (366a). Objects bearing other cases (e.g. instrumental), on the other hand, cannot shift to genitive under negation (e.g. (366b)). However, as pointedoutbyLegendre&Akimova(1993),theaccusativeobjectsofOEPVsdonotshowsuchasym- metryundernegation,andaccusativeremainsobligatory. (366) a. Ja I ne not našel found tzvety/ flowers.acc tzvetov. flowers.gen ‘Ididn’tfind(the)flowers.’ 182 Section5.2. ‘Psych’verbs:stateoftheart Chapter5 b. On he ne not upravljal managed fabrikoj/ factory.inst *fabriki. *factory.gen ’Hedidn’tmanagea/thefactory.’ (Pereltsvaig1997,viaLandau2010:25) c. Šum noise.nom ne not ogorčil upset ni no odnu one devočku/ girl.acc *odnoj *one devočki. girl.gen ‘Thenoisedidn’tupsetasinglegirl.’ (Legendre&Akimova1993,viaLandau2010:25) Moving to Greek, Anagnostopoulou (1999) notes clitic doubling with accusative objects is optional. However,accusativeexperiencersdonoteasilyallowforcliticdropping,as(367)illustrates. 8889 (367) a. O the Jannis John (tin) (cl.acc ghnorise met tin the Maria Mary.acc se in ena a party. party ‘Johnmet(her)Maryataparty.’ b. O the Jannis John ?*(tin) ?*(cl.acc endhiaferi interests tin the Maria Mary.acc pano more ap’ola than-everything ‘JohninterestsMarymorethananythingelse.’ (Anagnostopoulou1999:75) Anagnostopoulou(1999)furtherdiscussesanotherpeculiarpropertyofExperiencerobjectsinGreek regarding cliticization. Dative objects can be relativized without the preposition being pied-piped provided a resumptive clitic appears, a property that accusative objects do not have (e.g. (368)). Ex- periencer objects behave as regular dative objects in this respect: the resumptive clitic is obligatory 88 Anagnostopoulou further notes that clitic doubling in Greek is subject to the Prominence condition (i.e. the doubled objectmustbeananaphoricdefinite,notanovel/accomodativeone).Shenotesthatobjectexperiencersmayviolatethis condition. (xxvii) a. Prinapoligokeroeghrapsamiavivliokrisiajiaenakenourjovivliopanostocliticdoubling i . ‘Sometimeago,Ireviewedanewbookoncliticdoubling.’ b. #Arghotera #later-on ton cl.acc sinandisa met-I ton the sigrafea k author.acc se in ena a taksidhi trip mu. my ‘Lateron,Imethim-theauthorduringatripofmine.’ c. I the kritiki criticism mu my ton cl.acc enohlise bothered ton the sigrafea k author.acc toso such oste that na subj paraponethi complain ston to.the ekdhoti editor ‘Mycriticismbotheredtheauthorsomuchthathecomplainedaboutittotheeditor.’ (Anagnostopoulou 1999:76) 89 Therelationbetweenpsychpredicatesandtheobligatorinessofcliticdoublinghasalsobeenattestedforotherlanguages suchasAmharic(Kramer2014)andBulgarian(Krapova&Cinque2008). 183 Section5.2. ‘Psych’verbs:stateoftheart Chapter5 whentheExperiencerisrelativized,regardlessofwhetheritbearsmorphologicaldative(i.e. DEPVs) oraccusative(i.e.OEPVs).Thispatternisshownin(369). (368) a. Simbatho like.1sg ton the anthropo man.acc pu that o the Petros Peter.nom tu cl.dat edhose gave to the vivlio. book.acc ‘IlikethemanthatPetergavehimthebook.’ b. *Simpatho *(like.1sg ton the anthropo man.acc pu that ton cl.acc sinandise met.3sg o the Petros Peter.nom ‘IlikethemanthatPetermethim.’ (Anagnostopoulou1999:77) (369) a. O the anthropos man pu that *(tu) *(cl.dat aresi like.3sg i the Maria Mary.nom ine is ilithios stupid ‘ThemanthatMaryappealstoisstupid.’ b. O the anthropos man pu that *(ton) *(cl.acc endhiaferi interests i the Maria Mary.nom ine is ilithios. stupid ‘ThemanthatMaryinterestsisstupid.’ c. O the anthropos man pu that *(ton) *(cl.acc provlimatizun puzzle ta the nea news.nom bike came mesa in ‘Themanthatthenewspuzzlescamein.’ (Anagnostopoulou1999:77–78) TheserelativizationcontrastsinGreekalsoholdforHebrew,asLandau(2010)notes. 90 InHebrew,the absenceofaresumptivecliticwithrelativizedaccusativeobjectsispreferred,whereasitspresenceis required with relativized dative and oblique objects (e.g. (370a-b)). Experiencer objects behave like dativesandobliquesinthatrespect,as(370c)shows. (370) a. ha-iš 1 the-man še-Rina the-Rina hikira knew (?oto 1 ) (?him) higia. arrived ‘ThemanthatRinaknewhasarrived.’ b. ha-iš 1 the-man še-Rina that-Rina xašva thought al-*(av 1 pro 1 ) of-*(him) higia. arrived ‘ThemanthatRinathoughtofhasarrived.’ c. ha-muamadim the-candidates še-ha-toca’ot that-the-results hiftiu surprised *(otam) *(them) lo not amru said mila. word ‘Thecandidatesthattheresultssurpriseddidnotutteraword.’ (Landau2010:31) 90 LandauquotesArad(1998:199–200)andapersonalcommunicationwithSharonArmon-Lotemastheprecedentsofthis empiricalobservation. 184 Section5.2. ‘Psych’verbs:stateoftheart Chapter5 Anotherpervasivepsycheffectinvolvestheformationofverbalpassives. OEPVsdonotformverbal passivesinmanylanguages,afactthathasbeenrepeatedlyobservedintheliterature(Belletti&Rizzi 1988; Legendre 1989; Franco 1990; Legendre 1993; Grimshaw 1990; Roberts 1991; Landau 2010; Marín 2011; Fábregas & Marín 2015a, a.o.). 91 The impossibility of forming verbal passives is not obvious in languagesthatdonotdistinguishmorphologicallybetweenverbalandadjectivalpassives(e.g.French andItalian).However,manyauthorshaveshownuponcarefultestingthatstativeOEPVsindeedlack verbal passives. I will present below the discussion of Belletti & Rizzi (1988) for Italian, and see also Section5.3.2.2forHebrewaswellasLandau(2010)foramoregeneraldiscussion. Belletti&Rizzi(1988)providethefollowingargumentsforItalian. First,thepastparticipleforms derivedfromtheseverbscannothostcliticpronouns,unlikeverbalpassivesandverbsmoregenerally (e.g.(371–372)). (371) a. La the notizia news comunicata communicated a to Gianni. Gianni b. La the notizia news ignota unknown a to Gianni. Gianni (372) a. La the notizia news comunicatagli. communicatedtohim b. *La *the notizia news ignotagli. unknowntohimXXXXXXx(Chomsky1981,viaBelletti&Rizzi1988:309–10) Furtherevidencecomesfromthedifferenttypesofauxiliariesthatverbalandadjectivalpassivesmay appearwith.Whiletheybothtaketheauxiliaryesseretheverbvenire‘come’,whenusedasanauxiliary, isonlylicitwithverbalpassives(e.g. (373)). Aswecanseein(374),SEPVsallowfor venire,butOEPVs donot(e.g.(375)). (373) a. La the porta door è is chiusa closed alle at cinque. five b. La the porta door viene comes chiusa closed alle at cinque. five XXXXXXXXXXXXXXXXxx(Belletti&Rizzi1988:310) (374) a. Giannivienetemutodatutti. Giannicomesfearedbyeveryone 91 ButseeBouchard(1992);Legendre&Akimova(1993);Pesetsky(1995);Pylkkänen(2000);Gehrke&Sánchez-Marco(2015) foradifferentview. 185 Section5.2. ‘Psych’verbs:stateoftheart Chapter5 b. Giannivieneapprezzatodaisuoiconcittadini. Giannicomesappreciatedbyhisfellow-citizens c. Questasceltavienerispettatadallamaggioranzadeglielettori. Thischoicecomesrespectedbythemajorityofthevoters (Belletti&Rizzi1988:310) (375) a. *Giannivienepreoccupatodatutti. *Giannicomesworriedbyeverybody b. *Giannivieneaffascinatodaquestaprospettiva. *Giannicomesfascinatedbythisperspective c. *Giannivieneappassionatodallapolitica. *Giannicomesexcitedbypolitics (Belletti&Rizzi1988:310–1) Third,someOEPVsdonotallowaregularparticipialforms,butonlyanirregularadjectivalform. If these verbs could form verbal passives, it would be a mystery why regular participial forms are not allowed.Thereasonthatregularparticipialformsarenotallowedasadjectiveswiththeseverbsisdue, inBelletti&Rizzi’s(1988)view,toKiparsky’s(1973)BlockingPrinciple,whichblockstheformationof aregularparticipialadjectiveifthereisanirregularformavailable. (376) a. *Sonostufato/stancato/entusiasmatodallesueidee. *Iamtired/excitedbyhisideas b. Sonostufo/stanco/entusiastadellesueidee. Iamtired/excitedofhisideas (Belletti&Rizzi1988:311) Beforeconcludingthissection,acrucialclarificationistobemade:thesepsycheffectsareclaimedto onlyholdinnon-agentivereadingsofOEPVs:whenthepsychverbisusedagentively,itstopsshowing psycheffectsandbehavelikenon-psychverbs(Kayne1975;Franco1990;Arad1999;Anagnostopoulou 1999;Landau2010;Fábregas&Marín2015a,a.o.).Theagentive/non-agentivealternationhasalsobeen connected to the stative/eventive aspectual alternation in many works (Arad 1998b, 2002; Rothmayr 2009; Landau 2010; Marín 2011; Fábregas & Marín 2015a) and it is yet another trait that sets OEPVs apartfromOEPVsandIIIverbs, whichseemtobestrictlynon-agentive/stativeacrosslanguages. I illustratethisalternationin(377)forEnglish. 92 (377) a. Stormsscareme. (Non-agentive,stative) 92 Not all OEPVs alternate aspectually with the same ease both within and across languages. Pesetsky (1995) notes for Englishthatsomeverbslikescarefavoraneventivereading,whereasotherslikedepressarestrictlystative.Marín(2011),on theotherhand,notesthatmostOEPVsinSpanishareactuallystative(seealsoSection5.3.1forfurtherdiscussiononthis point). 186 Section5.2. ‘Psych’verbs:stateoftheart Chapter5 b. JohnscaredMary(onpurpose/byknockingaglassover). (Agentive,eventive) Departingfromtheseauthors,Icontendthatthedefiningtraitforpsycheffectsisstativity,ratherthan lackofagentivity: thesourceoftheconfusion,Ibelieve,stemsfromassumingastrictcorrelationof non-agentivityandstativity,whichdoesnotalwayshold. Iwillpresentmyargumentsforthisview laterinthechapter,inSection5.5.3. AsummaryofmydiscussionofpsychverbsisprovidedinTable5.1, whereIpresentalistofthe psycheffectsIhavetreatedsofarandthelanguagesinwhichtheyhavebeenobserved. Table5.1:Psycheffectscrosslinguistically Psycheffects Language(s) StativeOEPVs Othertransitive verbs Noverbalpassives Italian,Spanish, Hebrew 3 7 Noperiphrasticcausatives Spanish,Italian 3 7 Noreflexives Spanish,Italian 3 7 Dativeleft-dislocatedobjects linkedtoaccusativeclitics Italian 3 7 Norelativizationextraction outofaccusativeobjects Italian 3 7 Obligatoryresumptiveclitic withrelativizedaccusative object Greek,Hebrew 3 7 Cliticdoublingwithac- cusativeobjects Greek 3 7 Nogenitiveofnegation Russian 3 7 187 Section5.3. Psychologicalandlocativeverbs Chapter5 5.3 Psychologicalandlocativeverbs 5.3.1 Spanishpsychverbsandtheireffects SpanishpsychologicalverbscanbeclassifiedaccordingtothetripartiteclassificationputforthinBel- letti&Rizzi(1988). Ipresenttheclassificationin(378)below. TheexamplesforSEPVsandOEPVsare takenfromMarín&Sánchez-Marco(2012). 93 (378) a. SEPVs(Nominativeexperiencer) aborrecer‘abhor’,admirar‘admire’,adorar‘adore’,amar‘love’,anhelar‘yearn’,apreciar‘appre- ciate’,codiciar‘covet’,compadecer‘pity’,deplorar‘deplore’,despreciar‘despise’,detestar‘detest’, envidiar‘envy’,estimar‘esteem’,lamentar‘lament’,odiar‘hate’,padecer‘suffer’,preferir‘prefer’, querer ‘love’, respetar ‘respect’, sentir ‘feel’, sufrir ‘suffer’, temer ‘fear’, tolerar ‘tolerate’, venerar ‘venerate’. 93 Theexamplesin(xxviiib)representthoseverbswhoseexperiencerisstrictlydativeand,asitcanbeseen,theyarevery fewincomparisontoSEPVsandOEPVs.However,inStandardSpanish,manyobject-experiencerpsychverbsdisplayan accusative/dativealternation(i.e. Belletti&Rizzi’sClassIIandIIIverbs),asshownin(xxviiia). Fewerarestrictlydative (i.e.ClassIII),asshownin(xxviiib). (xxviii) a. A dom Pedro Pedro {lo/ {cl.acc le} cl.dat molestan bother.3p.pl los the niños. children ‘(The)childrenbotherPedro.’ b. A dom María María {*la/ {*cl.acc le} cl.dat gusta likes.3p.sg el the chocolate. chocolate ‘Maríalikeschocolate.’(lit.‘ChocolatepleasesMaría.’) (StandardSpanish) However,myvariety(CastilianSpanish)neutralizescasedistinctionsincliticsinfavorofgenderdistinctions,asin(xxix): le/lesisanobjectcliticformasculinereferentsandla/lasisanobjectcliticforfemininereferents,asopposedtoStandard Spanish,wheredativecliticsarele/lesregardlessofgenderandaccusativecliticsarelo/losformasculinereferentsandla/las forfemininereferents(seeFernándezOrdóñez1999,foradialectalstudyofthisphenomenaanditsgeographicdistribution inEuropeanSpanish).Insuchvariety,then,thedistinctionbetweenOEPVsandDEPVsisblurred. (xxix) a. A dom Pedro Pedro {*lo/ {*cl.acc le} cl.masc molestan bother.3p.pl los the niños. children ‘(The)childrenbotherPedro.’ b. A dom María María {la/ {cl.fem *le} *cl.dat gusta likes.3p.sg el the chocolate. chocolate ‘Maríalikeschocolate.’(lit.‘ChocolatepleasesMaría.’) (CastilianSpanish-laístaandleístavariety) 188 Section5.3. Psychologicalandlocativeverbs Chapter5 b. OEPVs(Accusativeexperiencer) aburrir ‘bore’, acongojar ‘wring’, agobiar ‘overwhelm’, aliviar, amedrentar ‘intimidate’, angus- tiar‘distress’, anonadar‘stun’, apaciguar‘calmdown’, apasionar‘bepassionateabout’, apenar ‘sadden’,apesadumbrar‘sadden’,asombrar‘astonish’,asustar‘scare’,atemorizar‘terrify’,cabrear ‘make really mad’, compungir ‘to be remorseful’, conmocionar ‘move’, confundir ‘confound’, consternar ‘dismay’, consolar ‘console’, contrariar ‘displease’, deprimir ‘depress’, (des) animar ‘encourage/discourage’,desesperar‘exasperate’,(des)honrar‘(dis)honor’,(des)ilusionar‘excite/ disappoint’, deslumbrar ‘dazzle’, (des)motivar ‘(dis)encourage’, disgustar ‘upset’, distraer ‘dis- tract’, enamorar‘makefallinlove’, enfadar‘irritate’, enfurecer‘enfuriate’, enojar‘irritate’, en- tretener ‘entertain’, entristecer ‘sadden’, entusiasmar ‘enthuse’, espantar ‘scareaway’, estimular ‘stimulate’, excitar ‘arouse’, fascinar ‘fascinate’, fastidiar ‘annoy’, frustrar ‘frustrate’, humillar ‘humilliate’, importunar ‘importune’, indignar ‘outrage’, interesar ‘interest’, molestar ‘bother’, mosquear‘annoy’,obnubilar‘bewilder’,obsesionar‘obsess’,ofender‘offend’,ofuscar‘obfuscate’, oprimir‘opress’,perturbar‘disturb’,preocupar‘worry’,seducir‘seduce’,sorprender‘surprise’. c. DEPVs(Dativeexperiencer) gustar‘like’,agradar‘please’,placer‘please’,doler‘hurt’,apetecer‘feellike’. 5.3.1.1 StativeOEPVs GiventhatpsycheffectsarefoundinstativeOEPVs,andtheseverbsareknowntoalternateaspectually acrosslanguages,wefirstneedtodeterminewhichoftheverbsin(378b)arenon-agentive/stative. In hisaspectualstudyofpsychverbsinSpanish,Marín(2011)reachestheconclusionthatmostOEPVsin Spanishareinfactstrictlystative(seealsoMarín&McNally2011;Marín&Sánchez-Marco2012):non- agentive OEPVs are always stative, he claims, whereas within agentive OEPVs we can find agentive andnon-agentiveones.Ipresentin(379)thenon-exhaustivelistofnon-agentive/stativeOEPVs,taken fromMarín(2011). 94 (379) Stative/Non-agentiveOEPVsinSpanish aburrir‘bore’, acongojar‘wring’, afligir‘sadden’, angustiar‘distress’, anonadar‘stun’, apasionar‘be passionateabout’, apenar‘sadden’, apesadumbrar‘sadden’, deprimir‘depress’, desesperar‘exasper- ate’,disgustar‘upset’,enfadar‘anger’,enfurecer‘enfuriate’,enojar‘upset’,enorgullecer‘makeproud’, entristecer‘sadden’,entusiasmar‘enthuse’,fascinar‘fascinate’,indignar‘outrage’,interesar‘interest’, obnubilar‘bewilder’,obsesionar‘obsess’,ofuscar‘obfuscate’,preocupar‘worry’. 94 ButseeSection5.5.3.2,wheretakeacloserlookattheverbsin(379)andarguethat,whiletheyareindeedstative,someof themresistaclear-cutclassificationasnon-agentive. 189 Section5.3. Psychologicalandlocativeverbs Chapter5 5.3.1.2 Psycheffects Spanish,justlikeallthelanguagesdiscussedinSection5.2.2,alsoexhibitspsycheffects.Verbalpassives are not possible with stative OEPVs (e.g. (380a)): adjectival passives, however, are acceptable (380b). The claim that passives of stative OEPVs are always adjectival has been made for other languages (Belletti&Rizzi1988;Grimshaw1990;Levin1993;Landau2010,a.o.)andforSpanishbyVanhoe(2004); Marín (2015); Fábregas & Marín (2015a). Note that Spanish differentiates morphologically between verbalandadjectivalpassives: theformertaketheauxiliary ser‘tobe’andthelattertake estar‘tobe’, hencetheyarenotambiguous,unlikewhathappensinlanguageslikeItalianandEnglishthatdonot distinguishmorphologicallybetweenthetwo(rememberthediscussioninSection5.2.2aboutItalian adjectivalpassives). (380) a. *Tus *your padres parents fueron were ser {fascinados/ {fascinated interesados/ interested desesperados/ despaired angustiados/ anguished preocupados/ worried obsesionados} obsessed (por ()by la the situación). situation intended: ‘Yourparentswere{fascinated/interested/despaired/anguished/worried/ob- sessed}(bythesituation).’ b. Tus your padres parents están are estar {fascinados/ {fascinated interesados/ interested desesperados/ despaired angustiados/ anguished preocupados/ worried obsesionados} obsessed (por ()by la the situación). situation ‘Your parents are {fascinated/ interested/ despaired/ anguished/ worried/ obsessed} (by thesituation).’ Just like we saw for Italian, Spanish does not allow periphrastic causatives with hacer ‘make’ with stativeOEPVs(e.g.(381)). (381) *Esto *this hizo made a dom María María {aterrorizar/ {terrorize impresionar/ impress fascinar/ fascinate desesperar/ exasperate preocupar} worry a dom Juan. Juan intended:‘ThismadeMaría{terrorize/impress/fascinate/exasperate/worry}Juan.’ AnotherpsycheffectthatwefindinSpanishistheimpossibilityofhavingreflexivestructures. This may not be evident at first sight, since it is very common to find stative OEPVs with reflexive mor- phology(seeparticularlyMarín&McNally2005,2011). 190 Section5.3. Psychologicalandlocativeverbs Chapter5 (382) María María se refl.3p.sg aburre/ bores desespera/ despairs angustia/ anguishes preocupa/ worries obsesiona. obsesses lit:‘Maríabores/despairs/anguishes/worries/obsessesherself.’ Upon closer inspection, however, these are not reflexive structures, but rather more akin to reflex- ivelymarkedanticausatives. Themoreobviouspieceofevidenceisthatwecannotexpressreflexive meaningwithreflexivemorphologywithstativeOEPVs.Forinstance,ifIwanttoexpressthatPedro doesnotfindhimselfinteresting,Icannotsay(383a):Ihavetoparaphraseit,asin(383b).Thesamegoes for(384a):itisungrammaticalwithareflexivereading.Forexpressingthereflexivereading,oneneeds toresorttoaparaphrase, e.g. withanadjectivalpassive, asin(384b). Reflexivestructures, therefore, arenotpossiblewithstativeOEPVsinSpanish. (383) a. *Pedro *Pedro no not se refl.3p.sg interesa interests a at sí him mismo. self intended:‘Pedrodoesn’tinteresthimself.’ b. Pedro Pedro no not tiene has interés interest en in sí him mismo. self ‘Pedrodoesn’thaveanyinterestinhimself.’ (384) a. *Pedro *Pedro se refl.3p.sg obsesiona. obsesses intended:‘Pedroobsesseshimself.’ b. Pedro Pedro está is obsesionado obsessed consigo with.him mismo. self ‘Pedroisobsessedwithhimself.’ WithagentiveOEPVs(i.e.whichcanhaveanagentivereadingwithananimatesubject),‘psych’effects disappear,asarguedforexplicitlyinFábregas&Marín(2015a).Theseverbsareillustratedin(385),and aretakeninpartfromMarín(2011),wholabelsthem‘agentiveobject-experiencerverbs’. (385) AgentiveOEPVs acosar ‘harass’, agobiar ‘overwhelm’, aliviar ‘alleviate’, amedrentar ‘intimidate’, (des-)animar ‘en- courage/discourage’,apaciguar‘pacify’,asombrar‘astonish’,asustar‘frighten’,atemorizar‘terrify’, confundir ‘confound’, consolar ‘console’, contrariar ‘displease’, (des-)honrar ‘(dis-)honor’, deslum- brar‘dazzle’, (des-)motivar‘discourage/encourage’, distraer‘distract’, entretener‘entertain’, espan- tar‘horrify’,estimular‘stimulate’,excitar‘excite’,fastidiar‘annoy’,frustrar‘frustrate’,humillar‘hu- milliate’,importunar‘annoy’,molestar‘disturb’,ofender‘offend’,oprimir‘opress’,perturbar‘disturb’, seducir‘seduce’,sorprender‘surprise’. 191 Section5.3. Psychologicalandlocativeverbs Chapter5 Interestingly,(asubsetof)theseverbsdonotshowpsycheffects. Marín(2011)alreadypointsoutthat the verbs in (385) can form verbal passives (e.g. (386a)), and I observe that they can also appear in reflexiveconfigurations(e.g.(386b))andcanformperiphrasticpassives(e.g.(386c)).Iwillnonetheless refinetheclassificationin(385)anditsrelationtopsycheffectsinSection5.5.3. (386) a. Pedro Pedro ha has sido been ser acosado/ harassed excitado/ excited humillado/ humiliated seducido seduced (por (by María). María ‘Pedrohasbeen{harrassed/excited/humiliated/seduced}byMaría.’ b. Juan Juan se refl.3p.sg distrajo distracted viendo watching una a película. movie ‘Juanamusedhimselfwatchingamovie.’ c. Azarías Azarías hizo made a dom sus his amigos friends humillar humilliate a dom Paco. Paco ‘AzaríasmadehisfriendshumilliatePaco.’ Forpsycheffects,then,Iwillrestrictmydiscussiontothestativeverbsin(379).Iwillnowdiscussthe classofverbsthatIcall‘location’verbsandshowthattheybehaveverymuchlikestativeOEPVswith respecttopsycheffects. 5.3.2 Locativeverbsandtheir‘psych’effects 5.3.2.1 Spanishlocativeverbs There is a class of verbs in Spanish (and in many other languages) that denote a spatial relationship betweentwoentities.Iprovideanon-exhaustivelistin(387). 95 (387) Locativeverbs rodear ‘surround’, cubrir ‘cover’, flanquear ‘flank’, envolver ‘wrap’, tapar ‘cover/hide from view’, bloquear ‘block’, cercar ‘fence’, obstruir ‘obstruct’, revestir ‘coat’, poblar ‘inhabit’, inundar ‘flood’, llenar‘fill’,cobijar‘shelter’,decorar‘decorate’,illuminate‘iluminar’... 95 Notethatthelexicalsemanticsoftheseverbssometimesmakesthespatialrelationmoreobvious(e.g. rodear‘surround’, wherethesubjectisaroundtheobject)andsometimesmoreobscure: forinstance,itisnotatfirstobviousthattheverb decorar‘decorate’inasentencesuchasTheflowerpotsdecoratethefrontporchislocative,yetnotethatforthesentencetobe truetheflowerpotsneedtobeinaspatialrelationtothefrontporch,justlikeinasentencesuchasTheflowerpotsflankthe frontporch.Whatisimportantforourpurposesisthattheverbsin(387)havethesamegrammaticalproperties. 192 Section5.3. Psychologicalandlocativeverbs Chapter5 ThisunderstudiedclassofverbshasbeenusuallydiscussedtogetherwithOEPVs(Kratzer2000;Roth- mayr 2009; Jaque 2013). 96 In particular, these authors have focused on the aspectual alternation that bothclassesshowbetweenastative,non-agentivereadingandaneventive,agentivereading.Thefor- mer merely denotes a stative spatial relation whereas the latter involves a change of state/location. See example (377) for OEPVs, repeated below, and (388) for locative verbs in English. This aspectual alternation also holds in Spanish, with the caveat that psych verbs in Spanish do not generally have eventiveversions,aswediscussed. (377) a. Stormsscareme. (Non-agentive,stative) b. Johnscaredme(onpurpose/byknockingaglassover). (Agentive,eventive) (388) a. Thewallsurroundsthehouse. (Non-agentive,stative) b. Thepolicesurroundedthehouse. (Agentive,eventive) (389) a. La the corrupción corruption humilla humiliates a dom los the políticos. politicians ‘Corruptionhumiliatespoliticians.’ b. María María humilló humiliated a dom Juan Juan (en (in apenas barely cinco five minutos). minutes ‘MaríahumiliatedJuaninbarelyfiveminutes.’ 96 Rothmayr(2009)callsthem‘instrumentalalternation’verbs,onthebasisthattheycanhavetwoversions:aneventiveone withanagentsubjectandaPP-instrument(e.g. (xxxa)and(xxxia))andastativeoneinwhichtheinstrumentintroduced bythePPisnowthesubject(e.g.(xxxb)and(xxxib)). (xxx) a. Die the Irmi Irmi füllt fills die the Vase vase mit with Wasser. water ‘Irmiisfillingthevasewithwater.’ b. Wasser water füllt fills die the Vase. vase ‘Waterisfillingthevase.’ (xxxi) a. Der the Poldi Poldi schmückt decorates die the Torte cake mit with Kerzen. candles ‘Poldiisdecoratingthecakewithcandles.’ b. Kerzen candles schmücken decorate die the Torte. cake ‘Candlesaredecoratingthecake.’ (Rothmayr2009:38) 193 Section5.3. Psychologicalandlocativeverbs Chapter5 (390) a. La the valla wall rodea surrounds toda all la the casa. house ‘Thewallsurroundsthewholehouse.’ b. La the policía police rodeó surrounded toda all la the casa house (en (in apenas barely cinco five minutos) minutes ‘Thepolicesurroundedthewholehouseinbarelyfiveminutes.’ Theseauthorshaveproposed,fromdifferenttheoreticalperspectives,ananalysisintermsof‘flavors’ ofcausation,sotospeak:astativecausationversionthattakesanon-agentivesubjectandaneventive causationversionthattakesanagentivesubject,mostnotablyintheworkofRothmayr(2009).These authors’proposalsweremostlyfoundedonanalogousproposalsforOEPVs(Arad1998b;Kratzer2000; Pylkkänen 2000, a.o.) (See Section 5.4.2 for further discussion as well as Section5.4.3 for arguments againstthisview). Whathasnotbeennoticedorpursued,atleastasfarasIamaware,iswhetherthequirkygram- maticaleffectsdisplayedbypsychverbs(i.e. the‘psych’effects)canalsobeobservedwiththestative version of locative verbs or not. They actually do. First, verbal passives are out under the stative reading(e.g.(391)). (391) a. *La *the casa house era was ser rodeada surrounded por by una a valla. fence intended:‘Thehousewassurroundedbyafence.’ b. *El *the porche porch era was ser decorado decorated por by macetas. flowerpots intended:‘Theporchwasdecoratedbyflowerpots.’ Second,reflexivestructuresarealsoout.Thisisshownin(e.g.(392a),wherewecannotuseareflexive structure to say that Pedro (statively) surrounds himself with stones, or (392b), in a scenario where Maríahaspaintedherselfwithsilverpaint. Keepinmindthatthesentencesin(392)arenotungram- maticalundertheeventivereading,i.e.iftheyareinterpretedashabitual. (392) a. *Pedro *Pedro se refl.3p.sg rodea surrounds (de (of piedras). stones intended:‘Pedrosurroundshimself(withstones).’ (Non-habitualreading) b. *María *María se refl.3p.sg decora decorates (con (with pintura paint plateada). silver) intended:‘Maríadecoratesherself(withsilverpaint).’ (Non-habitualreading) 194 Section5.3. Psychologicalandlocativeverbs Chapter5 JustlikewithOEPVs,these‘psych’effectsdissapearwhentheverbiseventive:inthatcase,bothverbal passives(393a)andreflexivestructures(393b)arereadilyavailable. (393) a. La the casa house fue was ser rodeada surrounded por by la the policía police (en (in un one minuto). minute ‘Thehousewassurroundedbythepolice(inoneminute).’ b. Pedro Pedro se refl.3p.sg cubrió covered (en (in un one segundo). second ‘Pedrocoveredhimself(inonesecond).’ Yet another parallelism is found with respect to OEPVs: locative verbs do not allow a periphrastic causativewith hacer‘make’intheirstativereading. Onlytheeventivereadingisavailable,i.e. onein whichsomethingmadethepolicemengetaroundthehouse,butnotoneinwhichnochangeofstate wouldbeinvolved(e.g.(394)). (394) Esto this hizo made a dom la the policía police rodear surround la the casa. house ‘Thismadethepolicesurroundthehouse.’ 7(Stativereading:thismadethepolicebearoundthehouse. 3(Eventivereading:thismadethepolicegettobearoundthehouse. 5.3.2.2 Psycheffectswithlocativeverbsinotherlanguages Interestingly, locative verbs also show psych effects in other languages. Take Hebrew, for instance. Landau(2010)arguesthatHebrew,likeallotherlanguages,doesnothaveverbalpassivesfromstative, non-agentive OEPVs generally. Only agentive verbs allow for the verbal passive, as in (395a), where theagentisintroducedby al-yedey‘by’. However,Landau(2010)notesthatthereisasmallsetofsta- tiveOEPVsthatmayappearinthisconstruction(e.g.(395a)).Landaulabelsthemfakepassivesbecause, although morphologically verbal, he claims they are actually unergative: the demoted external ar- gument cannot be introduced by the preposition al-yedei because it is in fact an internal argument, introducedbytheprepositionme-‘of’. (395) a. Gil Gil hu’alav was-insulted al-yedey by ha-bosit. the-boss ‘Gilwasinsultedbytheboss.’ b. Gil Gil hufta was-surprised me-/ of *al-yedey *by ha-xadašot. the-news ‘Gilwassurprisedat/*bythenews.’ (Landau2010:61) 195 Section5.3. Psychologicalandlocativeverbs Chapter5 Asitturnsout,Hebrewstativelocativeverbsbehaveinaverysimilarfashion. Canonicalverbalpas- siveswithademotedexternalargumentintroducedbyal-yedeiareout/degradedwiththeseverbs,as (396a)shows. However, thedemotedargumentcanbeintroducedbyadifferentpreposition, ba-‘in’ inthiscase. (396) a. ?ha-bayit ?the-house mukaf surround.caus.pass.pres al-yedei by ha-gader the-fence ha-zo. the-this intended:‘Thehouseissurroundedbythisfence.’ b. ha-bayit the-house mukaf surround.caus.pass.pres ba-gader in.the-fence ha-zo. the-this ‘Thehouseissurroundedbythisfence.’ (ItamarKastner,p.c.) Further commonalities between stative OEPVs and locative verbs in Hebrew are found in resump- tive pronouns in relative clauses. As we saw previously,when the accusative object of stative object experiencerpsychverbsisrelativized,thereisanobligatoryresumptivepronoun(see(370c),repeated below). (370c) ha-muamadim the-candidates še-ha-toca’ot that-the-results hiftiu surprised *(otam) *(them) lo not amru said mila. word ‘Thecandidatesthattheresultssurpriseddidnotutteraword.’ (Landau2010:31) Thesameeffectholdswithlocativeverbs. InHebrew, therelativizedobjectofalocativeverbneeds tohavearesumptivepronoun:itsabsencedegradestheconstructionconsiderably,as(397)shows. 97 (397) a. ze this ha-iš the-man še-ha-rihut that-the-furniture makif surrounding.pres ?(oto). ?(him ‘Thisisthemanthatthefurnituresurrounds.’ b. ze this ha-iš the-man {še-ha-panas {that-the-flashlight meir/ illuminates.pres.m še-ha-teura that-the-lighting meira} illuminates.pres.f ?(oto) ?(him ‘Thisisthemanthatthe{flashlight/lighting}illuminates.’ c. ze this ha-iš the-man še-ha-tsdafim that-the-seashells mekaštot decorating.pres ?(oto) ?(him 97 Notethatthesentencesin(397havequestionmarks,ratherthanstars.However,myinformantwouldalsogivequestion marks, rather than stars, to examples like (370c) with psych verbs. Indeed, the contrasts in (370) are not as sharp for my informant,whowouldgivequestionmarksacrosstheboard,insteadofstars,asLandaudoes. 196 Section5.3. Psychologicalandlocativeverbs Chapter5 ‘Thisisthemanthattheseashellsdecorate.’ (ItamarKastner,p.c.) ThissameeffectregardingresumptivepronounshappensinGreektoo,aswesaw(e.g.(369b),repeated below). Again,justlikeinHebrew,stativelocativeverbsshowasimilarpattern: therelativeclauseis perfectwithaclitic,whereasitsabsencedegradesthesentenceconsiderably,as(398)shows. (369b) O the anthropos man pu that *(ton) *(cl.acc endhiaferi interests i the Maria Mary.nom ine is ilithios. stupid ‘ThemanthatMaryinterestsisstupid.’ (398) a. O the anthropos man pu that ?(ton) ?(him.cl perikiklonun surrounds ta the epipla furniture ine is omorfos. handsome ‘Themanthatthefurnituresurroundsishandsome.’ b. O the anthropos man pu that ?(ton) ?(him.cl kalipti covers i the kuverta blanket ine is o the Janis. John ‘ThemanthattheblanketcoversisJohn.’ c. O the anthropos man pu that ?(ton) ?(him.cl fotizi illuminates o the provoleas spotlight ine is o the sigrafeas. writer ‘Themanthatthespotlightilluminatesisthewriter.’ (GiorgosSpathas,p.c.) Rememberthatwealsodiscussedthat,inGreek,cliticdoublingofaccusativeobjectsisgenerallyop- tional,butitbecomesobligatorywithaccusativeexperiencerobjects(e.g.(367b)).and(xxviia-b)below. Once again, these psych effects hold for locative verbs. Clitic doubling is generally disallowed with theseverbsinGreek,as(399)shows. 98 98 Myinformanthowevernotesthatcliticdoublingseemstobeoptionalwithilluminate,asshownin(xxxii).Then,notall locativeverbsbehavealikewithrespecttocliticdoublinginGreek,asituationthatIputasideforfutureinvestigation. (xxxii) O the provoleas spotlight (tin) (her.cl fotizi illuminates tin the Maria. Mary ‘ThespotlightilluminatesMary.’ Also,myinformantnotesthatcliticdoublingbecomesoptionalagainwhentheaccusativeobjectofstativelocativeverbs isinanimate,asin(xxxiii). Ofcourse,wecannottestwhetherthisholdswithOEPVstoosinceanexperiencercannotbe inanimate. (xxxiii) a. Ta the epipla furniture (to) (it.cl kiklonun encircle to the ajalma. statue ‘Thefurnituresurroundsthestatue.’ b. O the vraxos rock (tin) (her.cl mplokari blocks tin the isodo. entrance 197 Section5.3. Psychologicalandlocativeverbs Chapter5 (367b) O the Jannis John ?*(tin) ?*(cl.acc endhiaferi interests tin the Maria Mary.acc pano more ap’ola than-everything ‘JohninterestsMarymorethananythingelse.’ (Anagnostopoulou1999:75) (399) a. Ta the epipla furniture ?(ton) ?(cl.acc kiklonun encircle ton the Petro. Petro ‘ThefurnituresurroundsPetro.’ b. O the vraxos rock ?(ton) ?(cl.acc mplokari blocks ton the Petro. Peter ‘TherockblocksPeter.’ (GiorgosSpathas,p.c.) Italian is another language where yet more psych effects can be observed in stative locative verbs. For one, the objects of stative locative verbs can appear as accusative clitics referentially linked to a dative-markedleft-dislocatedDP(e.g.(400)).ThisisexactlywhatwesawforOEPVsin(363),repeated below. 99 (363) A to Giorgio, Giorgio questi these argomenti arguments non not l’hanno him.have convinto. convinced (400) A dat Giorgio, Giorgio non not lo cl.acc circondano surround i the palloni, balls, lo cl.acc circondano surround le the sedie chairs ‘Giorgioisnotsurroundedbyballs,butbychairs.’ (DeliaBentley,p.c.) Context: thereareseveralpeoplewithdifferentobjectsaroundthem. SpeakerAbelievesthat Giorgiohasballsaroundhim,butSpeakerBcorrectshimbyuttering(400). WealsosawthatinItalian,experiencerobjectsareislandsforextraction,asshownbytheimpossibility ofrelativizingaDPoutofanexperiencerobject.Asitturnsout,theobjectsoflocativeverbsconstitute islandsforextractionaswell. (362b) *Il *the candidato candidate di of cui whom questa this prospettiva perspective impaurisce frightens i the sostenitori. supporters ‘Therockblockstheentrance.’ 99 My informant reports that she prefers (400) with the nominative argument ocurring post-verbally, as in the example, andwithsuchargumentbearingconstrastivefocus. 198 Section5.3. Psychologicalandlocativeverbs Chapter5 (401) *Il *the candidato candidate di of cui whom questi these muri/ walls queste these siepi/ bushes queste these inferriate railings circondano surround i the sostenitori. supporters intended:‘Thecandidatewhosesupportersaresurroundedbythesewalls/thesebushes/these ralings.’ (DeliaBentley,p.c.) Finally,stativeOEPVsandlocativeverbsalsobehavealikeinRussianwithrespecttothegenitiveof negation. Justlikeaccusativeexperiencersblockthegenitiveunderclausematenegation(e.g. (366c)), sodotheaccusativeobjectsofstativelocativeverbs(e.g.(402)). 100 (366c) Šum noise.nom ne not ogorčil upset ni no odnu one devočku/ girl.acc *odnoj *one devočki. girl.gen ‘Thenoisedidn’tupsetasinglegirl.’ (Legendre&Akimova1993,viaLandau2010:25) (402) a. Zabor wall ne neg okružaet surrounds dom/ house.acc {?/??}doma. {?/??}house.gen ‘Thewalldoesnotsurroundthehouse.’ b. Odejalo blanket ne neg pokryvaet covers divan/ sofa.acc {?/??}divana. {?/??}sofa.gen ‘Theblanketdoesnotcoverthesofa.’ c. Rojal’uže grand.piano.already ne neg ukrašaet decorates komnatu/ room.acc *komnaty. *room.gen ‘Thegrandpianodoesnotdecoratetheroomanymore.’ d. Fonar’ streetlight ne neg osveš’aet illuminates dom/ house.acc/ ?doma. ?house.gen ‘Thestreetlightdoesnotilluminatethehouse.’ (EkaterinaChernova,p.c.) 100 An exception to this general pattern seems to be the verb pryatat’ ‘hide’, which allows for the genitive of negation just likeanyothertransitive-accusativeverb(e.g.(xxxiv)).Idonothaveanexplanationforwhythiscouldbe,soIleaveitaside forfutureresearch,noting,onceagain,thatitisanexceptiontotherule. (xxxiv) Šal’ handkerchief ne neg prjačet hides litso/ face.acc litsa face.gen devočki. girl.gen ‘Thehandkerchiefdoesnothidethegirl’sface.’ (EkaterinaChernova,p.c.) 199 Section5.3. Psychologicalandlocativeverbs Chapter5 5.3.2.3 Interimsummary A summary of the discussion in this section is provided in Table 5.2. As we can see, stative locative Table5.2:Psycheffectswithstativelocativeverbs Psycheffects Language(s) StativeOEPVs Stativelocative verbs Othertransitive verbs Noverbalpassives Spanish,Hebrew 3 3 7 Noperiphrastic causatives Spanish,Italian 3 3 7 Noreflexives Spanish,Italian 3 3 7 Dativeleft-dislocated objectslinkedtoac- cusativeclitics Italian 3 3 7 Norelativizationex- tractionoutofac- cusativeobjects Italian 3 3 7 Obligatoryresumptive cliticwithrelativized accusativeobject Greek,Hebrew 3 3 7 Cliticdoublingwith accusativeobjects Greek 3 3 7 Nogenitiveofnega- tion Russian 3 3 7 verbsconsistentlyshowthesamegrammaticaleffectsthatstativeOEPVshavebeenobservedtohave crosslinguistically, knownas‘psych’effects. Ithinkthattheseparallelismsbetweenpsychverbsand locative verbs are too many to be attributed to chance, and strongly warrant a unified grammatical analysisofthesetwoverbtypes.Iwillproceedtodosointhefollowingsection. 200 Section5.4. Myproposal Chapter5 5.4 Myproposal 5.4.1 APPforexperiencerobjects: Landau(2010) MyanalysistakesasapointofdeparturetheproposalinLandau(2010)forOEPVs,lateradoptedfor SpanishbyFábregas&Marín(2015a).Hetakestheassumptionsin(403)and(404)asapointofdeparture forhisanalysis. 101 (403) Universally,non-nominativeexperiencersbearinherentcase. (Landau2010:20) (404) Universally,inherentcaseisassignedbyP. (Landau2010:21) Notethat,asLandaupointsout,morphologicalcasedoesnotcorrelateuniformlywithsyntacticcase, i.e. anominalmarkedwithinherentcasemayappearwitheitheraccusativeordativecase. Hepro- poses that experiencer objects are introduced by a silent PP complement of V, whose head assigns inherentobliquecasetoitsDPcomplement. Thenon-experiencerargument(Target/Subjectmatter T/SM)isintroducedVP-internally(i.e. stativeclassIIverbsareunaccusative). Thestructureisgiven in(405),adaptedfromLandau(2010:9). (405) StructureofstativeOEPVs VP T/SM V’ V PP ? Experiencer Fromthisstructure, twoconsequencesfollow, shownin(406). Bydativebehaviorin(406a), Landau means inherent case behavior typically associated with dative case/ indirect objects, even when the ‘direct’objectshowsmorphologicalaccusativecase,acasetypicallyassociatedwithstructuralcase. (406) a. TheexperiencershoulddisplayPP/dativebehavior. b. Thecaseoftheexperiencershouldresistsyntacticsuppression. (Landau2010:22) Thisstructure,Landauargues,ispotentinsofarasitsucceedsinaccountingforthepsycheffectsheob- servescrosslinguistically(mostlydiscussedinSection5.2.2). Withrespecttothedative-likebehavior, theaccountexplainswhyobjectexperiencersinItaliancanbeleft-dislocatedwithdativemorphology, associatedtoanaccusativeclitic,whereasothertransitiveverbscannotdothat.Itfurtheraccountsfor 101 Withrespectto(403),seeBelletti&Rizzi(1988)fortheearlierclaimthataccusativecaseinOEPVsisinherent. 201 Section5.4. Myproposal Chapter5 theimpossibilityofaccusativeexperiencerstoswitchtogenitivecaseunderclausematenegationin Russian,unlikestructurallyaccusativeobjects.ItalsohelpsunderstandwhyGreekaccusativeexperi- encerspatternwithdativesinthattheycannotdropthecliticassociatedtothem(seeDimitriadis1999, forfurtherconstructionsinMacedonianGreekthatalsohaveobliqueaccusativeobjectsandequally requirecliticdoubling). WithrespecttothePPstructure, Landau’saccountexplainswhyexperiencerobjectsareislands forwh-extractioninItalian,andwhyinGreekandHebrewweencounterobligatoryresumptivepro- nounswithrelativizedobjects.ThereasonprovidedbyLandauforthelatterphenomenonisthatrel- ativizationinvolvesemptyoperatormovement,whichleavesagap,andGreekPscannotbestranded. Finally,theunaccusativestructureexplainswhytheseverbslackverbalpassives(universally,accord- ingtoLandau). Crucially, LandaulinksthisnullprepositiontotheassignmentoftheExperiencer-role. Inhis ownwords: HereIfollowthetraditionaldistinction,introducedbyChomsky(1981),betweeninherent and structural case. Unlike structural case, which is assigned/checked in the syntax in certainconfigurations(government,spec-head,Agreeetc.)irrespectiveofthematicroles, inherent case is assigned in the lexicon and is tied to a specific-role. Indeed, one should see inherentcaseasaninseparablereflexof-roleassignment. (Landau2010:20,emphasismine) 5.4.2 Unifyingpsychandlocativeverbs IadoptLandau’s(2010)structureoutlinedinSection5.4.1forstativeOEPVsinSpanish(seealsoFábre- gas & Marín 2015a for Spanish). 102 However, given that the grammatical effects that this analysis ac- countsforarealsopresentinstativelocativeverbs,itstandstoreasontoproposethattheyareboth associatedtothesamesyntacticstructure,asshownin(407a)and(408a). (407) a. La the crisis crisis aterra frighten a dom las the familias. families ‘Thecrisisfrightensfamilies.’ 102 Fábregas&Marín(2015a)adoptLandau’sanalysisonlypartly:whiletheyassumethattheexperiencerobject(andindeed, allexperiencers)isintroducedbyanullPP,theypositatransitivestructureforstativeOEPVs,notunaccusative. 202 Section5.4. Myproposal Chapter5 b. VP Lacrisis V’ V aterra PP ? alasfamilias (408) a. La the sábana sheet tapa hides la the entrada. entrance ‘Thesheethidestheentrance.’ b. VP Lasábana V’ V tapa PP ? laentrada Two logical options are in principle available. One is to assume that the lexicon comes with two differentflavorsofP : oneassignsanExperiencer-roleandtheotherassignsaGround-role, as in(409).StativeOEPVswouldbelexicallyspecifiedastakingtheformerandlocativeverbsthelatter. (409) a. StativeOEPVs:XP Experiencer b. Stativelocativeverbs:XP Ground I find it very counterintuitive, however, to propose that there exist two different heads that behave syntactically in exactly the same way, their only difference being the-role they assign. Why, then, only two flavors of P and not, say, five? Note, furthermore, that other accounts of flavors of mor- phosyntacticheads(e.g.Folli&Harley2005)aremeanttocapturedifferentsyntacticcontrastswithin the same grammatical category, but in our case, both of these Ps would happen to show a homoge- neoussyntacticbehavior. Asecondoption,whichIpursuehere,istoabandontheideathatthesilentP assignsa-role. 103 In thisview,thereisnoinherentcaseassignmentwithitsconcomitant-assignment,butsimplystruc- 103 Tothisrespect,notethatnominalsmarkedwithdativecase—typicallyanalyzedastheprototypicalinherentcase—donot haveauniformthematicinterpretation.Forinstance,workonapplicativearguments,whichbeardative(Pylkkänen2002; Cuervo2003;Schäfer2008;Fernández-Soriano&Mendikoetxea2011, a.o.), hasshownthattheirthematicinterpretation isdifferentdependingonwhereinthesyntaxtheymerge: iftheprojectionthatintroducesthem(Appl(icative)P)merges 203 Section5.4. Myproposal Chapter5 tural oblique case assignment of P to its complement. This proposal equally explains why these P -complementsdonotbehavelikecanonicalaccusativeobjects,whichareassignedcasebyahigher transitiveverbalprojection,plausablyVoiceP(Kratzer1996). If so, then the challenge is to determine how the thematic interpretation of arguments comes about, i.e. what the common ground is between the event structure of these two verbs that would allowforageneralizedaspectualeventroletotheirarguments. Rememberthat,withinourresearch program,thehypothesisisthatitistheeventstructureoftheverbalpredicatethatassignsthematic rolesbyentailment.Forexpositorypurposes,letusrefertotherolethesubjectofOEPVsasStimulus (followingTalmy1985)andthesubjectandobjectoflocativeverbsasFigureandGround,respectively (inthelinesofTalmy1975). (410) a. Lacrisis Stimulus aterra alasfamilias Experiencer ‘Thecrisisfrightensfamilies.’ b. Lasábana Figure tapalaentrada Ground ‘Thesheethidestheentrance.’ ManyaccountsregardingthethematicstructureofstativeOEPVshaveproposedthattheyareinfact stativecausativestructures,theirsubjectbearingaCauserrole. 104 Authorswhoadheretothecausative analysisofOEPVseitherarguethattheyheadtransitivestructures(Pesetsky1995;Arad2002forOEPVs andKratzer2000forbothpsychandlocativeverbs)orthattheyareinfactunaccusative(Arad1998a; Pylkkänen2000forOEPVsandRothmayr2009forbothpsychandlocativeverbs). Other accounts, however, have proposed that stative OEPVs are not causative. In this view, the subject is not a Causer, but a Stimuli (or Theme, to use a more neutral term). This is proposed, an implementedintechnicallydifferentways,byBelletti&Rizzi(1988)andLandau(2010).Theseauthors arguethattheseverbsareinfactunaccusative. Iamnotawareofunaccusativeaccountsforlocative verbs. With respect to the Causer analysis, I assume that a subject can only be a Causer inasmuch as it appearsinasemanticallycausativestructure(inthelinesofArad1998b,2002;Pylkkänen2000;Roth- lowasacomplementofV,thedativeargumentisinterpretedasaGoal;ifApplPselectsaresult-denotingprojection,itis interpretedasaBeneficiary/Maleficiaryoftheeventuality;ifitprojectsaboveVPinananti-causativecontext,thenit isinterpretedasanAccidentalcauser. 104 Someoftheseaccountsareaimedataccountingforthelinkingproblem,aswediscussedinSection5.2.1(Grimshaw1990; Pesetsky1995).Otheraccounts,ontheotherhand,havefocusedmoreinestablishingaparallelismbetweentheeventiveand stativereadingsoftheaspectualalternation,bothforOEPVs(Pylkkänen2000;Arad2002)andlocativeverbs(Rothmayr 2009)(seealsoKratzer2000). 204 Section5.4. Myproposal Chapter5 mayr2009).ThereaderwillrecallthatChapter3discussedthegobernar-verbtypeatlengthandargued thattheywerestativecausatives.Inthefollowingsection,Iwillargueagainstacausativeanalysisfor stativeOEPVsandlocativeverbs. Next,Iwillarguethatthesetwoverbtypesarenottransitive,but ratherunaccusative(inthelinesofPylkkänen2000;Landau2010forOEPVs). 5.4.3 Stativepsychandlocativeverbsarenotcausative AswediscussedinChapter3,gobernar-verbsarecharacterizedasbeingcausativestates,i.e.amacroeven- tuality constituted of two substates causally related. I showed that, while the classic Aktionsart tests classifiedthemasstates,theyshowed(atleast)twokeytraitsthatdistinguishedthemfromotherstates suchasknoworlove: (a)Thepossibilitytoappearinadjectivalpassives(whichtakethecopulaestar‘to be’);(b)theexistenceofascopeambiguitywithagain(seeTable3.2forreference).Irelatedtheformer propertytotheresultativecomponent,andthelattertotheirbieventivecomposition. ThestructureIproposed,aswellasthesemantics,areillustratedwithanexamplein(146),repeated below. Asisclear,thisisessentiallythestructureonewouldadoptforacausativeanalysisofOEPVs andlocativeverbs,especiallyifoneadherestoamodelwherethemappingfromsyntaxtosemantics istransparent(Hale&Keyser1993;Ramchand2008). (146) a. Berta Berta gobierna governs el the país. country ‘Bertagovernsthecountry.’ b. initP DP Berta init’ init <gobierna> resP DP elpaís res’ res <gobierna> XP c. s9s 1 ,s 2 [e=(s 1 !s 2 )&gobierna(s 1 )&participant(Berta,s 1 )&gobierna(s 2 )&participant(el país,s 2 )] StativeOEPVsandlocativeverbsare,justlikegobernar-verbs,stative. Also,likegobernar-verbs,they allowforadjectivalpassiveswithestar‘tobe’.Thisisshownin(411a)forOEPVs(seealso(380b))andin 205 Section5.4. Myproposal Chapter5 (411b)forlocativeverbs,aswellasin(118d)forgobernar-verbs,repeatedbelow.Iwilldiscussadjectival passivesmoreatlengthinChapter4. (411) a. Pedro Pedro está is estar preocupado worried por by su his hijo. son ‘Pedroisworriedabouthisson.’ b. La the aldea village está is estar rodeada surrounded por by un a valle. valley ‘Thevillageissurroundedbyavalley.’ (118d) España Spain está is gobernada ruled por by un an inconsciente. irresponsible ‘Spainis estar ruledbyanirresponsibleperson.’ Crucially,however,stativeOEPVsandlocativeverbspartwayswithgobernar-verbswithrespecttothe scopeambiguityofagain.Withtheformerverbs,againcanonlytakewidescope,butnotnarrowscope, i.e. theadverbcannothavethesubjectoutofitsscope(e.g. (412)). Gobernar-verbs,aswediscussedin Chaper3,canhoweverhavebothreadingswiththeadverbagain(e.g.(140)).ThissuggeststhatOEPVs andlocativeverbs,unlikegobernar-verbs,arenotbieventive. (412) a. La the crisis crisis preocupa worries a dom María María de of nuevo. new ‘ThecrisisworriesMaríaagain.’ 3 Widescopereading(again>crisis:ThecrisisworriedMaríainthepast,stoppeddoing itatsometimeandnowitisdoingitagain. 7 Narrow scope reading (crisis > again): María was worried in the past (e.g. by earth- quakes),thatsituationstoppedandnowthereisacrisisthatkeepsMaríaworriedonce again. b. Los the monumentos monuments rodean surround la the montaña mountain de of nuevo. new ‘Themonumentssurroundthemountainagain.’ 3 Widescopereading(again>monuments: Themonumentssurroundedthehillinthe past,stoppeddoingitatsomepoint(e.g. theywereremovedfromthesite)andnow theyaredoingitagain(e.g.theywereplacedbackbyhumanhands). 7 Narrowscopereading(monuments>again):Thehillwassurroundedinthepast(e.g. byastonewall),thatsituationstopped(e.g.thewallwasremoved)andnowthereare somemonumentsthatkeepthehillsurroundedagain. 206 Section5.4. Myproposal Chapter5 (140) La the presa dam impide preventes el the paso pass de of agua water de of nuevo. new ‘Thedampreventstheflowofwateragain.’ 3 Wide scope reading (again > dam: The dam prevented the flow of water in the past, stoppeddoingitatsometimeandnowitisdoingitagain. 3 Narrow scope reading (dam > again): The flow of water was prevented in the past (e.g. bysomedebris),thatsituationstoppedandnowthereisadamthatpreventstheflowof water. Also,gobernar-verbsandOEPVsandlocativeverbsfaredifferentlywithrespecttowhethertheyaccept PP-instruments.Theformerallowthem(e.g.(413))butthelatterrejectthemundertheirstativereading (e.g. (414)). Note that PP-instruments with gobernar-verbs are licit both with human (e.g. (413a)) and non-human(e.g.(413b))subjects.Iassumethatinstrumentsarelicensedbycausalmeaning,following proposals that characterize instruments as intermediaries in a causal chain (Talmy 1976; Croft 1991; Goldberg2002;Koenig&Davis2006). (413) a. Pedro Pedro vigila surveils el the museo museum con with unos some prismáticos. binoculars ‘Pedrosurveilsthemuseumwithapairofbinoculars.’ b. El the sistema system de of seguridad security protege protects la the entrada entrance con with rayos rays infrarrojos. infrared ‘Thesecuritysystemprotectstheentrancewithinfraredrays.’ (414) a. *Pedro *Pedro aterra terrifies a dom Juan Juan con with un an hacha. axe intended:‘PedroterrifiesJuanwithanaxe.’(Non-habitualreading) b. *Carmen *Carmen rodea surrounds a dom Ester Ester con with una a cuerda. rope intended:‘CarmensurroundsEsterwitharope.’(Non-habitualreading) 5.4.4 Stativepsychandlocativeverbsareunaccusative OEPVsandlocativeverbsalsodifferwithrespecttogobernar-verbsinothernon-trivialrespects.First, asweobserved,neitheroftheformerclassescanformverbalpassives(e.g.(380a)and(391)). Gobernar- verbs,however,can(e.g.(415)).NotethatstativeSEPVscanformthemtoo(e.g.(416)). 207 Section5.4. Myproposal Chapter5 (380a) *Tus *your padres parents fueron were ser fascinados/ fascinated interesados/ interested desesperados/ despaired angustiados/ anguished preocupados/ worried obsesionados obsessed (por ()by la the situación). situation intended:‘Yourparentswerefascinated/interested/despaired/anguished/worried/obsessed (bythesituation).’ (391) a. *La *the casa house era was ser rodeada surrounded por by una a valla. fence ‘Thehousewassurroundedbyafence.’ b. *El *the porche porch era was ser decorado decorated por by macetas. flowerpots ‘Theporchwasdecoratedbyflowerpots.’ (415) El the museo museum era was ser {vigilado/ surveiled protegido/ protected custodiado} guarded por by la the policía. police ‘Themuseumwas{surveiled/protected/guarded}bythepolice.’ (416) Carlos Carlos era was ser {amado/ loved odiado/ hated respetado} loved por by todos. all ‘Carloswas{loved/hated/respected}byeveryone.’ WealsosawthatstativeOEPVsandlocativeverbsdidnotallowreflexivestructures.Thiswasshown, and repeated here, in (382) and (392), respectively. Note that (382) would be grammatical under an anticausativereadingwithoutthereflexiveanaphorasímisma‘herself’(seediscussioninSection5.3.1.2 andcomparewith(382)). Departingonceagainfortheseverbs,weobservethat gobernar-verbsallow forreflexive structures(e.g. (417)). SodoSEPVs (e.g. (417)), asoriginallyobservedby Belletti&Rizzi (1988)forItalian. (382) *María *María se refl.3p.sg aburre/ bores desespera/ despairs angustia/ anguishes preocupa/ worries obsesiona obsesses a to sí her misma. self intended:‘Maríabores/despairs/anguishes/worries/obsessesherself.’ (392) a. *Pedro *Pedro se refl.3p.sg rodea surrounds (de (of piedras). stones intended:‘Pedrosurroundshimself(withstones).’ (Non-habitualreading) 208 Section5.4. Myproposal Chapter5 b. *María *María se refl.3p.sg decora decorates (con (with pintura paint plateada). silver) intended:‘Maríadecoratesherself(withsilverpaint).’ (Non-habitualreading) (417) a. Pedro Pedro se refl supervisa supervises a dom sí him mismo. self ‘Pedrosuperviseshimself.’ b. Juan Juan se refl.3p.sg controla controls para for no not gastar spend demasiado. too.much ‘Juancontrolshimselfsoastonotspendtoomuch.’ (418) a. Carmen Carmen se refl.3p.sg quiere loves mucho. much ‘Carmenlovesherselfalot.’ b. Pedro Pedro se refl.3p.sg detesta detests cuando when hace does eso. that ‘Pedrohateshimselfwhenhedoesthat.’ Periphrastic causatives are yet another construction where stative OEPVs and locative verbs part wayswith gobernar-verbs(andSEPVsaswell). Theformercannotbeinfinitivalcomplementsofthe causativeverbhacer‘tomake’whenusedstatively(381),whereasgobernar-verbsandSEPVscan. (381) *Esto *this hizo made a dom María María {aterrorizar/ {terrorize impresionar/ impress fascinar/ fascinate desesperar/ exasperate preocupar} worry a dom Juan. Juan intended:‘ThismadeMaría{terrorize/impress/fascinate/exasperate/worry}Juan.’ (394) Esto this hizo made a dom la the policía police rodear surround la the casa. house ‘Thismadethepolicesurroundthehouse.’ 7(Stativereading:thismadethepolicebearoundthehouse. 3(Eventivereading:thismadethepolicegettobearoundthehouse. (419) Esto this hizo made a dom Victoria Victoria gobernar govern el the país. country ‘ThismadeVictoriagovernthecountry.’ 209 Section5.4. Myproposal Chapter5 (420) Esto this hizo made a dom Juan Juan admirar admire a dom Victoria. Victoria ‘ThismadeJuanadmireVictoria.’ ThediscussionissummarizedinTable5.3. Table5.3:Comparingstativecausativeswithstativepsychandlocativeverbs Grammaticalproperties Gobernar-verbs OEPVs Locativeverbs Scopeambiguitywithagain 3 7 7 PP-Instruments 3 7 7 (Anti-)causativealternation 7 3 7 Reflexives 3 7 7 Verbalpassives 3 7 7 Periphrasticcausativewithhacer ‘tomake’ 3 7 7 Adjectivalpassives 3 3 3 ItakethelackofverbalpassivesandperiphrasticcausativestosignaltheabsenceofaninitPinthe structureoftheverb. 105 Stativecausatives,aswediscussed,areformedbyaninitPandaresP(see(146)). ForSEPVswecanassume,withRamchand(2008),thattheyareformedbyasingleinitP,andthusthey canformverbalpassivesandallowperiphrasticcausatives.Iprovideanillustrativestructurein(421). RememberthatinitP,inisolation,denotesasingle,non-causativestate. 105 Discussing the unavailability of reflexive OEPVs in Romance, Landau (2010) argues that it cannot be explained by the unaccusativityofthepredicate,sinceDEPVsarealsounaccusativeandyetallowreflexivesinmanylanguageslikeItalian andFrench.Spanishbehavesthesameway(xxxv). (xxxv) Pedro Pedro se refl gusta pleases mucho. much ‘Pedrolikeshimselfalot.’ Landau adopts an unergative analysis of reflexives (Reinhart 1997, 2000, 2001, 2002; Reinhart & Siloni 2004). Under this analysis,areductionoperationidentifiesthetwo-rolesofaverbsothatonlyoneofthemprojects,anditabsorbsthecase oftheverbalcomplement.Crucially,Landauargues,onlyaccusativeordativecasecanbeabsorbed,butnotobliquecase: therestrictionagainstreflexiveswouldthusbeexplained.Iremainagnosticheretothespecificanalysisofreflexivity,but Inotethat,inasmuchasoursystemdoesnottake-rolestobegrammaticalprimitives,thereductionoperationshouldbe recastindifferentterms. 210 Section5.4. Myproposal Chapter5 (421) a. Teresa Teresa respeta respects las the leyes. laws ‘Teresarespectsthelaw.’ b. initP DP Teresa init’ init respeta DP lasleyes I propose that OEPVs and locative verbs, in their stative reading, have a structure as in (422b) and (423b),whereabirelationalP introducesthetwoarguments.Fortheverbalizer,wecouldthinkthat it is verbalized by a single resP, which by itself denotes a state and it is syntactically unaccusative. However, IwillsimplymodelthisasanemptyunaccusativeverbalizerV—Iwilltalkabitaboutmy rationalebehindthischoicewhenIconcludethepaperinSection5.6.Thelexicalentryoftheseverbs isgivenin(422c)and(423c). (422) a. La the crisis crisis aterra frightens a dom las the familias. families ‘Thecrisisfrightensfamilies.’ b. VP V <aterra> PP DP Lacrisis P’ P <aterra> DP alasfamilias c. aterrar‘frighten’ categoryfeatures:[V,P ] (423) a. La the sábana sheet tapa hides la the entrada. entrance ‘Thesheethidestheentrance.’ 211 Section5.4. Myproposal Chapter5 b. VP V <tapa> PP DP Lasábana P’ P <tapa> DP laentrada c. tapar‘hide/cover’ categoryfeatures:[V,P ] 5.4.5 ThecommonthematicinterpretationofExperiencersandGrounds TheissueremainsastohowExperiencersandGroundsareinterpretedunderthesamestructure. I pointed out that an account that posited two grammatically identical P heads that assigned com- pletelydifferent-roleswasneithereconomicalnorintuitive. Inthissection,IproposethatExperi- encers and Grounds share a common generalized meaning, building on Landau’s insight that expe- riencersare,insomesense,mentallocations. Thiswillrequireproposingamorecomplexsemantics fortheprepositionalheadthatLandauproposedforpsychverbs. Straightforwardly, Ipositthatthe P headhasthelexicalentryin(424). (424) JP K=x,y9",V [eigen(",x)&project(V,")&endpoint(y,V)] ThisformalizationispartlyinspiredinthesemanticsthatSvenonius(2006,2008,2010,2012b)proposed forlocativestructures,butIextendittopsychologicalstatesaswellascollapsehiscartographyinto a single syntactic head. I assume a generalized variable" that can be understood as a physical or a mentalspacedependingontherootthatlexicalizesthestructure.Letusseehowthisworks. eigen(",x)denotesarelationbetweenanentity xandthespace"itoccupies. Thenwetakethat space and project a set of vectors V away from it, which I represent as project(V,"’). The label project standsforwhateverrootlexicalizestheP head. Finally,endpoint(y,V)givesusthefinalpoint yof thevectors. Letusexemplifythiswithalocativeverbfirst,likerodear‘surround’.Iprovideanexamplein(425a), withthesyntaxin(425b)—verbalstructureaside—andthesemanticsin(425c). (425) a. El the río river Nilo Nile rodea surrounds el the desierto desert de of Bayuda. Bayuda 212 Section5.4. Myproposal Chapter5 ‘TheriverNilesurroundstheBayudadesert.’ b. PP DP ElríoNilo P’ P <rodea> DP eldesiertodeBayuda c. JPPK=9",V [eigen(",eldesiertodeBayuda)&rodea(V,")&endpoint(elríoNilo,V)] In the semantic formula, we have a certain abstract space" of the Bayuda Desert, delivered by the eigenrelation,andasetofvectorsthatprojectfromit,whichhavetheNileasanendpoint.Fromthe conceptualmeaningoftheverbsurround,weinferthatsucharelationisofaspecificlocativesort:the relevantspace"oftheBayudadesertare—atleastpartof—itsborders,withvectorsthatprojectaway fromsuchspaceandstopattheriverNile.IrepresentthisideainFigure5.1. Figure5.1:AnnotatedmapoftheNilepassingaroundBayudaDesert(Sudan) Thelocativerelationisnotalwaysasclearwithotherverbs, asIalreadynotedinfootnote95. Look at the examples in (426), for instance. While in both sentences there is a spatial relation described, thelexicalitemdoesnotsomuchgiveinformationaboutthespatialpositionofoneargumentwith respecttotheother,butratheraboutthequalityoftheirlocativerelation.Forinstance,theverbdecorar ‘decorate’in(426a)describestherelationasadecorativeone:theChristmasornamentsarepositioned somewhere with respect to the house in such a way that the spatial relation can be described as a decorative one. The same applies to iluminar ‘illuminate’ in (426b): the projector may be inside or 213 Section5.4. Myproposal Chapter5 outsidethestadium,behinditorinfrontofit,onthegroundoroverhead,butwhatisrelevantisthat suchaspatialrelationinvolvesanillumination. (426) a. Los the adornos ornaments de of Navidad Christmas decoran decorate toda all la the casa. house ‘TheChristmasornamentsdecoratethewholehouse.’ b. El the proyector projector ilumina illuminates el the estadio stadium olímpico. olympic ‘Theprojectorilluminatestheolympicstadium.’ LetusexemplifyOEPVsnow,forinstancewithagobiar‘overwhelm’,asin(427a). WhatIargueisthat, whentheverbthatlexicalizesthisprepositionalstructureisconceptuallypsychological,asin(427b), then the abstract space" is understood as a mental state, rather than a locative one. The particular lexicalitemwillnotgiveusthespatialpositionofanentitywithrespecttoanotherentity,butrather thementalstateofanentitywithrespecttoanotherentity. Inthiscase, whatwehaveisthemental space of my friend Ester, which is in a state of being overwhelmed with respect to the Los Angeles climate.Byentailment,myfriendEsteristheunderstoodexperiencerandtheLosAngelesclimateis thestimulus. (427) a. El the clima climate de of Los Los Angeles Angeles agobia overwhelms a dom mi my amiga friend Ester. Ester ‘TheLosAngelesweatheroverwhelmsmyfriendEster.’ b. PP DP ElclimadeLosAngeles P’ P <agobia> DP amiamigaEster c. JPPK=9",V [eigen(",miamigaEster)&agobia(V,")&endpoint(elclimadeLosAngeles,V)] Sincethisabstractspace"isinterpretedasamentalstate,itfollowsthattheentityassociatedtoit—the complementofP—shouldbesentient,otherwisethesentencewillbepragmaticallyodd(e.g.(428)). (428) a. #El #the clima climate de of Los Los Angeles Angeles agobia overwhelms a dom los the edificios. buildings #‘TheLosAngelesweatheroverwhelmsbuildings.’ 214 Section5.4. Myproposal Chapter5 b. #La #the crisis crisis aterra frightens a dom los the coches. cars #‘Thecrisisfrightenscars.’ I define the two possible interpretations of this abstract space" as in (429), which is esentially a re- formulationinformaltermsofmydiscussionsofar: ifthestructureislexicalizedbyapsychological verb,thisabstractspacewillbeinterpretedasamentalstate,andifitislexicalizedbyalocativeverb, itwillbeinterpretedasaphysicalspace. (429) a. If9V,"[project(V,")]&project=psychologicalverb,thenbydefinition"isamentalstate. b. If9V,"[project(V,")]&project=locativeverb,thenbydefinition"isaphysicalspace. From (429), we can derive the thematic entailments for OEPVs and locative verbs as in (430) for the complementofP andasin(431)forthespecifierofPP. (430) ThematicentailmentsofthecomplementofP . a. If9x,V,"[eigen(",x)&psychologicalverb(V,")]thenbyentailmentxisanExperiencer b. If9x,V,"[eigen(",x)&locativeverb(V,")]thenbyentailmentxisaGround (431) ThematicentailmentsofthespecifierofPP. a. If9x,V,"[endpoint(x,V)&psychologicalverb(V,")]thenbyentailmentxisaStimulus b. If9x,V,"[endpoint(x,V)&locativeverb(V,")]thenbyentailmentxisaFigure 5.4.6 InterimConclusions To sum up, this account derives the thematic interpretation of the arguments of both OEPVs and locative verbs through a common syntax and its associated formal semantic composition. I show how whether an argument is an Experiencer or a Ground—or a Stimulus or a Figure—depends on the argument’s role in the event structure of the PP coupled with our conceptual knowledge about the meaning of the verb: it does not depend on inherent case, or-role assignment, or any kind of thematicstructureinthesenseoftheclassicG&Btheory: Experiencers,Groundsaresoonaremere descriptions of the event entailments, but not thematic roles. I have also argued that these are not stativecausativeconstructions—notinSpanishatleast—andthattheyaresyntacticallyunaccusative— atleastinSpanish. Beforeconcludingthischapter,Iwilldiscusstheissueofagentivityanditsrelationtopsycheffects. Authorsgenerallyassociatenon-agentivitywithpsych-effects—tacitlyorexplicitlyassociatingnon- agentivitywithstativity. IwillshowinthefollowingSection5.5howthisassociationbetweennon- agentivity and psych effects is incorrect—again, at least for Spanish. It will be argued that the true 215 Section5.5. Agentivity? Chapter5 correlationisbetweenstativityandpsycheffects: hencepsycheffectsareamatterofaspectandnot ofthematicinterpretation. 5.5 Agentivity? As we have seen throughout this chapter, authors identifying ‘psych effects’ in OEPVs have empha- sizedthattheyonlyholdwithnon-agentiveversionsoftheseverbs,whereastheagentiveversionsare eventive. In this section I will evaluate this assumption and conclude that source of psych effects is notlackofagentivity,butstativity. There are two main problems with the received view regarding the correlation between ‘psych’ effects and agentivity. One is, I believe, the lack of a clear understanding and consensus of what it meansforasubjecttobeagentive,linguisticallyspeaking.Thesecondisthecommonmisconception, startingatleastwithLakoff(1970),thatstativityandnon-agentivitygohandinhand. 5.5.1 Thereceivedviewonagentivity Letusstartwiththenotionofagentivity. Althoughnotclearlydefinedinmostpapers,theoreticians tend to agree that it holds, broadly, when a predicate has a subject that is intentional (and hence, volitional and animate). The relevant grammatical tests to that effect are then the following (Ryle 1949;Lakoff1970;Dowty1979,a.o.): i) allowingadverbialmodificationbyagent-orientedadverbssuchasdeliberately(e.g.(432)) ii) allowingtheimperativeform(e.g.(433)) iii) appearingasinfinitivalcomplementsofdecide(e.g.(434)) iv) formingpseudo-cleftswithdo(e.g.(435)) v) takinginfinitivalpurposeclauses(e.g.(436)) ThetestswereoriginallyproposedforEnglish,butsincetheyworksimilarlyforSpanishIpresentthe examplesinthelatterlanguage. (432) a. *Verónica *Verónica sabe knows matemáticas Math deliberadamente. deliberately intended:‘VeronicaknowsMathdeliberately.’ b. Verónica Verónica colocó placed las the pruebas proofs al at.the lado side de of la the puerta door deliberadamente. deliberately ‘Verónicaplacedtheevidencebesidethedoordeliberately.’ 216 Section5.5. Agentivity? Chapter5 (433) a. *¡Posee *¡own tres three apartamentos! apartments intended:‘Ownthreeapartments!’ b. ¡Come ¡eat más more deprisa! fast ‘Eatfaster!’ (434) a. *Andrés *Andres decidió decided tener have.inf una a casa house en in Ibiza. Ibiza intended:‘AndrésdecidedtoownahouseinIbiza.’ b. Andrés Andres decidió decided comprar buy.inf una a casa house en in Ibiza. Ibiza ‘AndrésdecidedtobuyahouseinIbiza.’ (435) a. *Lo (*it que that hizo did Andrés Andrés fue was creer believe.inf en in una a deidad. deity intended:‘WhatAndrésdidwasbelieveinadeity.’ b. Lo it que that hizo did Andrés Andrés fue was fregar wash.inf los the platos. dishes ‘WhatAndrésdidwaswashthedishes.’ (436) a. *Gabriela (*Gabriela duda doubts de of casi almost todo everything para for no not ser be.inf dogmática. dogmatic intended:‘Gabrieladoubtsprettymucheverythingtonotbedogmatic.’ b. Gabriela Gabriela escribe writes libros books de of autoayuda auto-help para to ayudar help.inf a dom otras other personas.. people ‘Gabrielawritesself-helpbookstohelpotherpeople.’ Agentivesubjectstendtoappearwitheventivepredicatesandnon-agentivesubjectstendtoappearin stativepredicates(see(432–436),wherethe(a)examplesarestativeandthe(b)exampleseventive).Even though this was noted as least far as Dowty (1979) to be just a tendency, rather than a correlation, 106 thereisamyriadofworkthatassumesthelatter.Thisisparticularlytrueintherealmofpsychverbs andlocativeverbs,where(non-)agentivityfeatureshavebeenbuiltinthemodelingofthesepredicates. Forinstance,Arad(1998a),inaDistributedMorphologyaccountandunderacausativeviewofstative 106 SeethetableofverbtypesinDowty(1979:184),whereheshowsthatagentivitycutsacrossaspectualtypes. Foragentive states,henotesEnglishverbsofbodilypositionwithananimatesubject,suchassit,standandlie. 217 Section5.5. Agentivity? Chapter5 OEPVs, proposes that the v head that verbalizes roots comes in two types: eventive, which licenses an agent (e.g. (438)), and stative, which licenses a stative causer (non-agentive, by definition). Her examplesaregivenin(437)andherstructuresareshownin(438)and(439). (437) a. MaryfrightenedJohn(suddenly/byturningoffthelights). (Agentive,eventive) b. Darknessfrightensme. (Non-agentive,stative) (438) Agentive,eventivestructure v 1 agent v 1 v 1 CAUS(e,s) p P p NP (439) Non-agentive,stativestructure v 2 stative causer v 2 v 2 CAUS(s,s’) p P p NP SwitchingtoGerman,Rothmayr(2009)alsoputsforthaproposalthattiesagentivityandeventivity togetherforbothOEPVsandlocativeverbs. Sheargues,inspiredbyDowty(1979),thattherearedif- ferentaspectualoperatorshostedinverbalheads,whichbuildeventstructurecompositionally.Com- montoallcausativepredicatesisthepresenceofacauseoperator,whichexpressesaresultstate.The simplest structure is that of stative causatives (i.e. the stative version of locative verbs), which lacks anyactivityortransitionoperatorsandisthusstrictlystative.Iprovideherexamplein(440a)andits correspondingstructurein(440b). (440) a. Die the Haare hair verstopfen obstruct den the Abfluß. sink ‘Hairobstructthesink.’ (Rothmayr2009:48) b. VP DP DieHaare V V cause V DP denAbfluß V verstopfen 218 Section5.5. Agentivity? Chapter5 Fortheeventiveversionoflocativeverbs, Rothmayrproposesthatthereisanagentivedooperator that introduces an agent and acts as a dynamicity inductor. In other words, agentivity and stativity areincompatible.Iprovideherexamplein(441a)anditsstructurein(369b). 107 108 (441) a. Die the Irmi Irmi verstopft obstructs die the Straße street mit with ihrem her Lastwagen. truck ‘Irmiisobstructingthestreetwithhertruck.’ (Rothmayr2009:49) 107 Inherstructurein42,sheproposesthatinstrumentalssuchasmitihremLastwagen‘withhertruck’wouldoccupythesame syntacticpositionasthesubjectofthestativecounterpart.Thisisherwayofderivingtheinstrument-subjectalternation (e.g. Thetruckobstructsthestreet). 108 Thisisnottosaythateventivepredicatesneedtobeagentive. Rothmayrnotesthispossibilityandproposesabecome operator for these cases, which denotes a gradual change of state and where the subject undergoing the change is not agentive.Herexampleisprovidedin(xxxvia)and(xxxvib). (xxxvi) a. Das the Gewebe tissue hat has nach bit und by nach bit das the Blutgefäß blood-vessel verstopft. obstructed ‘Thetissueobstructedthebloodvesselbitbybit.’ (Rothmayr2009:51) b. VP V become VP DP DasGewebe V V cause V DP dasBlutgefäß V verstopft In an earlier account of OEPVs, Arad (1998a) also notes the possibility that their eventive versions may be non-agentive (andhaveacauserargument).Shepositsthatthesubjectsofeventiveversionsoftheseverbsaretrueexternalarguments, introducedbyaneventivevthatcanintroduceanagentoracauser.Ontheotherhand,thesubjectsofthestativeversion cannotbeagents,butmerestimulus(althoughArad1998a,likeRothmayr2009,assumesthattheunaccusativeVPcanhave causativemeaning:thestimuluswouldthenbeastativecauser).IintroducethestructuresfromArad(1998a:15)in(xxxvii) and(xxxviii). 219 Section5.5. Agentivity? Chapter5 b. vP DP DieIrmi v v do VP PP mitihremLastwagen V V cause V DP dieStraße V verstopft 5.5.2 Agentivestativelocativeverbs However, García-Pardo (2016b) presents data from Spanish that challenges the claim that agentivity cannotbefoundinstativelocativeverbs.Inparticular,Ishowedthatagentivitytestsandstativitytests cancoexistwithinthesameclausethathasalocativeverbasitsmainpredicate. In(442a), wefinda non-habitualreadingofthepresenttense,astativitytest,co-occurwithapurposeclause,anagentivity test. In (442b), we have a universal reading of the present perfect, a stativity test, co-occurring with theagent-orientedadverbapropósito(onpurpose). 109 (xxxvii) Eventivestructure vP agent/causer v v VP NP V ... (xxxviii) Non-agentive,stativestructure VP stimulus V V VP NP V ... 109 Thesentencesin(442)arealsoacceptableintheprogressive,andmorenaturaltosome’snativeears(e.g.(xxxix)). 220 Section5.5. Agentivity? Chapter5 (xxxix) a. En in estos these momentos, moments los the bandidos bandits están are flanqueando flanking el the camino path para to poder can asaltar assault a ACC los the viajeros. travellers ‘Thebanditsarecurrentlyflankingthepathtobeabletoassaultthetravellers.’ b. Los the manifestantes demonstrators han have estado been obstruyendo obstructing el the acceso access al to.the banco bank a on propósito purpose desde since las the seis six de of la the tarde. evening ‘Thedemonstratorshavebeenobstructingtheaccesstothebankonpurposesince6pm.’ Note that the progressive does not block a stative reading: as a matter of fact, sentences with locative verbs in the pro- gressive can be perfectly ambiguous between a stative and an eventive reading without any apparent coercion, and the subject, whenanimate, canbeunderstoodasanagentequallyinbothreadings(e.g. (xl), fromGarcía-Pardo(2016b:293)). Alsonotethattheprogressiveisequallypossiblewithnon-animatesubjectsincontextsthatclearlyfavorastativereading (e.g.(xli)).Finally,itshouldbepointedoutthat,asMarín&McNally(2011)observe,OEPVscanalsoaccepttheprogressive andretainingtheirstativemeaning(e.g.(xlii)). (xl) La the policía police está is rodeando surrounding el the edificio building {intencionadamente/ intentionally para to atrapar catch al ACC.the asesino}. killer ‘Thepolicesurrounds/aresurroundingthebuilding{intentionally/tocatchthekiller}.’ i) Eventivereading:Thepoliceareintheprocessofgatheringaroundthebuilding. ii) Stativereading:Thepolicearecurrentlyaroundthebuilding. (xli) a. La the roca rock está is obstruyendo obstructing la the salida exit y and no not podemos we.can escapar. escape ‘Therockisobstructingthewayoutandwecannotescape.’ b. La the columna column nos us está is tapando covering la the vista view del of.the eclipse eclipse solar. solar ‘Thecolumisobstructingourviewofthesolareclipse.’ (xlii) Juan Juan se SE está is aburriendo. bored ‘Juanis(already)bored.’ (Marín&McNally2011:475) Thus, the data clearly suggests that the progressive is not at all incompatible with stative predicates and hence it can hardlybeconsideredastativitytest(Leech1970;Comrie1976;Verkuyl1989;Levin&Rappaport-Hovav1995;contraVendler 1957; Lakoff 1966; Bennett & Partee 1972; Taylor 1977; Dowty 1979, a.o.) (remember also my discussion in Section 3.2.2.1). Rather, something else seems to be at play. Levin & Rappaport-Hovav (1995:170) suggest that the progressive probes for non-momentary predicates, rather than stative predicates. Fernández de Castro (1999), on his part, proposes that the 221 Section5.5. Agentivity? Chapter5 (442) a. En in estos these momentos, moments los the bandidos bandits flanquean flank el the camino path para to poder can asaltar assault a ACC los the viajeros. travellers ‘Thebanditscurrentlyflankthepathtobeabletoassaultthetravellers.’ b. Los the manifestantes demonstrators han have obstruido obstructed el the acceso access al to.the banco bank a on propósito purpose desde since las the seis six de of la the tarde. evening ‘Thedemonstratorshaveobstructedtheaccesstothebankonpurposesince6pm.’(García- Pardo2016b:293) Crucially,theagentivereadingofthesubjectwithstativelocativeverbsdoesnotcancelpsycheffects. As(443), verbalpassivesarenotallowedinstativecontextswithaclearlyagentivesubject. Thesen- tencesin(443)wouldbefineifahabitualreadingwasimposedbyaddingafrequencyadverbsuchas todoslosdías‘everyday’,butthenofcoursewewouldhaveaneventivereadingofthepredicate,not astativeone.Notethatthepassivesinthe(443)examplesareunequivocallyverbal:Spanishmorpho- logicallymarksverbalpassiveswithadistinctauxiliary,ser‘be’. (443) a. *El *the edificio building es is ser rodeado surrounded por by la the policía police para for que that no not escape escapes el the sospechoso. suspect intended:‘Thebuildingissurroundedbythepolicesothatthesuspectdoesnotescape.’ b. *La *the entrada entrance del of.the McDonalds McDonalds es is ser tapada covered a on propósito purpose por by los the activistas activists animalistas. animalists intended:‘TheMcDonaldsentranceiscoveredonpurposebytheanimal-rightsactivists.’ Theimpossibilityofhavingareflexivestructureinastativecontext,anotherpsycheffect,alsoholds when the subject is agentive. The examples in (444) are only grammatical under a habitual reading of the present tense, i.e. when they are eventive; they crucially are ungrammatical under the non- habitual(i.e.stative)readingreadingofthepresenttense. progressivemakesreferencetoactualevents,ratherthanmoreabstract,habitualorgenericevents. Thisisinopposition totheanalysisoftheprogressiveasastativizer, whichwouldexplain, contrarytofact, whystatesrejecttheprogressive (Vlach 1981; Moens 1987; Parsons 1990; Kamp & Reyle 1993, a.o.). I reject such an analysis here, but I will not side with a specificalternativeanalysisinthiswork. 222 Section5.5. Agentivity? Chapter5 (444) a. Pedro Pedro se refl tapa covers concienzudamente thoroughly porque because tiene he.has frío. cold intended:‘Pedrocovershimselfthoroughlybecauseheiscold.’ 7(Non-habitualreading:Pedroiscurrentlycovered(stativereading). 3(Habitualreading:Pedrousuallycovershimself(eventivereading). b. María María se refl rodea surrounds de of libros books para for que that su her madre mother piense thinks que that estudia. studies ‘Maríasurroundsherselfwithbookssothathermotherthinkssheisstudying.’ 7(Non-habitualreading:Maríaiscurrentlysurroundedbybooks(stativereading). 3(Habitualreading:Maríausuallysurroundsherselfwithbooks(eventivereading). 5.5.3 Agentivityandpsychverbs 5.5.3.1 TheclassificationinMarín(2011) Now,whereasitisclearthatsubjectsofbothstativeandeventivelocativeverbscanbeeventiveinthe classicsense,thepictureissomewhatblurrierwithOEPVs. AswesawinSection5.3.1.2,Marín(2011) argues that OEPVs come in two types: non-agentive and agentive (i.e. verbs that can never have an agentivereadingwithananimatesubjectandverbswhichcan).Theformerarealwaysstativeandthe latteraremixed:somearestative(mostofthem,Marínclaims)andsomeareeventive.Iillustrateboth typesin(379)and(385),fromMarín(2011)(seealsoMartin2006forFrench). (379) Stative/Non-agentiveOEPVsinSpanish aburrir‘bore’, acongojar‘wring’, afligir‘sadden’, angustiar‘distress’, anonadar‘stun’, apasionar‘be passionateabout’, apenar‘sadden’, apesadumbrar‘sadden’, deprimir‘depress’, desesperar‘exasper- ate’,disgustar‘upset’,enfadar‘anger’,enfurecer‘enfuriate’,enojar‘upset’,enorgullecer‘makeproud’, entristecer‘sadden’,entusiasmar‘enthuse’,fascinar‘fascinate’,indignar‘outrage’,interesar‘interest’, obnubilar‘bewilder’,obsesionar‘obsess’,ofuscar‘obfuscate’,preocupar‘worry’. (385) AgentiveOEPVs acosar ‘harass’, agobiar ‘overwhelm’, aliviar ‘alleviate’, amedrentar ‘intimidate’, (des-)animar ‘en- courage/discourage’,apaciguar‘pacify’,asombrar‘astonish’,asustar‘frighten’,atemorizar‘terrify’, confundir ‘confound’, consolar ‘console’, contrariar ‘displease’, (des-)honrar ‘(dis-)honor’, deslum- brar‘dazzle’, (des-)motivar‘discourage/encourage’, distraer‘distract’, entretener‘entertain’, espan- tar‘horrify’,estimular‘stimulate’,excitar‘excite’,fastidiar‘annoy’,frustrar‘frustrate’,humillar‘hu- milliate’,importunar‘annoy’,molestar‘disturb’,ofender‘offend’,oprimir‘opress’,perturbar‘disturb’, seducir‘seduce’,sorprender‘surprise’. 223 Section5.5. Agentivity? Chapter5 Beforediscussingtheevidenceforthisagentivityclassification,letusreviewMarín’s(2011)evidence forhisaspectualclassification. Asheshows,thenon-agentiveverbsin(379)areindeedstative. First, theydonotpasstelicitytests,sotheymustbeeitheractivitiesorstates. Aswecanseein(445),these verbsdonotpasstheinxtimetest.Furthermore,theycannotappearasparticiplesinabsoluteclauses introducedbyunavez‘once’,unliketelicpredicates(e.g.(446),andseealsoMarín&McNally(2011)on thispoint). (445) a. *Esta *this situación situation ha has angustiado/ anguished obsesionado/ obsessed preocupado worried a dom tus your padres parents en in cinco five minutos. minutes intended:‘Thissituationhasanguished/obsessed/worriedyourparentsinfiveminutes.’ b. Esta this situación situation ha has angustiado/ anguished obsesionado/ obsessed preocupado worried a dom tus your padres parents durante for meses. months ‘Thissituationhasanguished/obsessed/worriedyourparentsformonths.’ (446) *Una *one vez time aburridos/ bored obsesionados/ obsessed preocupados worried tus your padres, parents nos we iremos will.go de on vacaciones. vacation intended:‘Onceyourparentsarebored/obsessed/worried,wewillgoonvacation.’ Furthertestsshowthattheseverbsare,infact,stative. Asentencelike(447)inthepresenttensedoes notyieldahabitualreading.Also,dynamicmodifierssuchaslentamente‘slowly’arerejected(e.g.(448)). Furthermore,theseverbsdonotaccept parar‘stop’,whichonlytakesdynamicpredicatesinSpanish (e.g.(449)). (447) Esta this situación situation angustia/ anguishes desespera/ despairs ilusiona excites a dom tus your padres. parents ‘Thissituationworries/despairs/excitesyourparents.’ (448) *Esta *this situación situation angustia/ anguishes desespera/ despairs ilusiona excites a dom tus your padres parents lentamente. slowly intended:‘Thissituationworries/despairs/excitesyourparentsslowly.’ (449) a. *Esta *this situación situation ha has parado stopped de of afligir/ sadden fascinar/ fascinate interesar interest a dom tus your padres. parents ‘Thissituationhasstoppedsaddening/fascinating/worryingyourparents.’ 224 Section5.5. Agentivity? Chapter5 b. Esta this situación situation ha has dejado stopped de of afligir/ sadden fascinar/ fascinate interesar interest a dom tus your padres. parents ‘Thissituationhasstoppedsaddening/fascinating/worryingyourparents.’ With respect to the agentive verbs in (385), Marín (2011) notes that there is more aspectual variation. Mostofthemarestativebothwithanon-animateandanimatesubject,exceptforafewthatseemto be telic with an animate subject (essentially two, according to Marín). The examples he gives are in (450). (450) a. AgentiveOEPVs(stative) agobiar‘overwhelm’,animar‘cheerup’,molestar‘disturb’... b. AgentiveOEPVs(telic) humillar‘humilliate’,seducir‘seduce’... Asevidenceforthisdivision,Marín(2011)notesthattheverbsin(450a)donotpasstelicitytests:they rejectenxtiempo‘inxtime’phrases(e.g.(451a))butallowdurantextiempo‘forxtime’phrases(e.g.(452a)), are unacceptable as complements of finish ‘terminar’ (e.g. (453a)) and are unacceptable in absolute clauses (e.g. (454a)). On the other hand, the verbs in (450b) pattern as telic: they allow en x tiempo ‘inxtime’phrases(e.g. (451b))andaredegradedwith durante x tiempo‘forxtime’phrases(e.g. (452b)). However,theyaredegradedascomplementsofterminar‘finish’(e.g.(453b))andinabsoluteclauses(e.g. (454b)),butstilltheyfarebetterintheseconstructionsthantheverbsin(450a).Furthermore,whenthe verbsin(450a)areintheprogressivetheyentailtheperfect(e.g. 455a),whereasthosein(450b)donot (e.g.455b).Finally,onlytheverbsin(450b)showanambiguitywithcasi‘almost’(e.g.456),whichfurther supportstheircharacterizationastelic. (451) a. *Juan *Juan ha has agobiado/ overwhelmed animado/ cheered-up molestado annoyed a dom María María en in cinco five minutos. minutes intended:‘Juanhas{overwhelmed/cheeredup/bothered}Maríainfiveminutes.’ b. Juan Juan ha has humillado/ humiliated seducido seduced a dom María María en in cinco five minutos. minutes ‘Juanhas{humiliated/seduced}Maríainfiveminutes.’ (452) a. Juan Juan ha has agobiado/ overwhelmed consolado/ consoled molestado disturbed a dom María María durante for cinco five minutos. minutes ‘Juanhas{overwhelmed/consoled/disturbed}Maríaforfiveminutes.’ b. ??Juan ??Juan ha has humillado/ humiliated seducido seduced a dom María María durante for cinco five minutos. minutes 225 Section5.5. Agentivity? Chapter5 ‘Juanhas{humiliated/seduced}Maríaforfiveminutes.’ (453) a. *Juan *Juan ha has terminado finished de of agobiar/ overwhelming consolar/ consoling molestar annoying a dom María. María intended:‘Juanhasfinished{overwhelming/consoling/disturbing}María.’ b. Juan Juan ha has terminado finished de of humillar/ humiliating seducir seducing a dom María. María ‘Juanhasfinished{humiliating/seducing}María.’ (454) a. *Una *one vez time agobiada/ overwhelmed molestada annoyed María, María ... ... intended:‘OnceMaríawas{overwhelmed/annoyed},...’ b. ?Una ?one vez time seducida/ seduced humillada humiliated María, María ... ... ‘OnceMaríawas{seduced/humiliated},...’ (455) a. Pierre Pierre está is acosando/ harassing agobiando/ overwhelming animando/ cheering-up consolando/ consoling fastidiando/ hassling importunando/ disturbing molestando/ annoying oprimiendo oppressing a dom María. María ! ! Pierre Pierre ha has acosado/ harassed agobiado... overwhelmed a dom María. María ‘Pierreis{harassing/overwhelming/cheeringup/consoling/hassling/disturbing/annoy- ing/oppressing}María.!Pierrehas{harassed/overwhelmed...}María.’ b. Pierre Pierre está is humillando/ humiliating seduciendo seducing a dom María. María ! ! Pierre Pierre no not ha has humillado/ humiliated seducido seduced a dom María. María ‘Pierreis{humiliating/seducing}María.!Pierrehasnot{humiliated/seduced}María.’ (456) a. Mario Mario casi almost agobia/ overwhelms importuna/ disturbs molesta annoys a dom María. María ‘Marioalmost{overwhelms/disturbs/annoys}María. b. Mario Mario casi almost humilla/ humiliates seduce seduces a dom María. María ‘Marioalmost{humiliates/seduces}María.’ 226 Section5.5. Agentivity? Chapter5 Evidence than the verbs in (450a) are not just atelic, but stative, comes from their incompatibility with velocity adverbs like lentamente ‘slowly’ (e.g. (457a)) and their inability to be complements of parar ‘to stop’ (e.g. (458a)). The verbs in (450b) (i.e. humillar and seducir), on the other hand, accept velocityadverbs(e.g.(457b))andarenotasdegradedascomplementsofparar‘stop’(e.g.(458b)),which reinforcestheviewthattheyaredynamic. (457) a. *Juan *Juan agobió/ overwhelmed molestó annoyed a dom María María lentamente. slowly intended:‘Juan{overwhelmed/annoyed}Maríaslowly.’ b. Juan Juan humilló/ humiliated sedujo seduced a dom María María lentamente. slowly ‘Juan{humiliated/seducedMaríaslowly.’ (458) a. *Juan *Juan paró stopped de of agobiar/ overwhelm molestar annoy a dom María. María ‘Juanstopped{overwhelming/annoying}María.’ b. ?Juan ?Juan paró stopped de of acosar/ harass humillar/ humiliate seducir seduce a dom sus his alumnos. students ‘Juanstopped{harassing/humiliating/seducing}hisstudents.’ Goingbacktotheissueofagentivity,Marínappliestheusualbatteryofteststosupporthisdivision. Crucially, however, he restricts his data of agentive verbs to the telic ones, humillar ‘humiliate’ and seducir ‘seduce’. Heshowsthat, indeed, the humillar classcanbe ananswertoa what did X do? ques- tion(e.g. (459a),andseeMartin2006forthistestinFrench),canappearinpseudo-clefts(e.g. (460a)), cantakeagent-orientedadvebs(e.g. (461a))andappearintheimperativeform(e.g. (462a)). Thenon- agentiveverbsfromthe(b)examplesin(459–462)passnoneofthosetests,asexpected. (459) ¿Qué ¿what (es (is lo it que) that ha has.3p.sg hecho? done intended:‘Whatdidhedo?’ a. *Ha *have.3p.sg obsesionado/ obsessed preocupado worried a dom sus his padres. parents intended:‘He{obsessed/worried}hisparents.’ b. Ha have.3p.sg humillado/ humiliated seducido seduced a dom Mafalda. Mafalda ‘Juan{humiliated/seducedMaríaslowly.’ 227 Section5.5. Agentivity? Chapter5 (460) a. *Lo *it que that ha has.3p.sg hecho done es is obsesionar/ obsess preocupar worry a dom sus his padres. parents intended:‘What(s)hedidwas{obsess/worry}his/herparents.’ b. Lo it que that ha has.3p.sg hecho done es is humillar/ humiliate seducir seduce a dom Mafalda. Mafalda ‘What(s)hedidwas{humiliate/seduce}Mafalda.’ (461) a. *Obsesionó/ *obsessed.3p.sg preocupó worried.3p.sg a dom sus poss.3p.pl padres parents intencionadamente. intentionally intended:‘(S)he{obsessed/worried}his/herparentsintentionally.’ b. Humilló/ obsessed.3p.sg sedujo seduced.3p.sg a dom su poss.3p.sg vecina neighbor intencionadamente. intentionally ‘(S)heseducedhis/herneighbor.’ (462) a. *Obsesiónalos/ *obsess.imp.2p.sg.them preocúpalos. worry.imp.2p.sg.them intended:‘{Obsess/worry}them!’ b. ¡Humíllala/ ¡humiliate.imp.2p.sg.her sedúcela! seduce.imp.2p.sg.her ‘{Seduce/humiliate}her!’ IprovideanexpositorysummaryofthediscussioninTable5.4,fromMarín(2011). 5.5.3.2 Agentivityand‘psych’effectsinOEPVs Marín’s(2011)mainassumptions, inlinewiththereceivedviewonpsychverbs, arethefollowing: i) non-agentive verbs are always stative; ii) (a few) agentive verbs are only telic or eventive when they haveananimatesubject. Withrespecttothelatterpoint,rememberthatwediscussedthesetelicverbsinSection5.3.1.2and showedthattheyindeedshownopsycheffects(e.g.(386),repeatedbelow). (386) a. Pedro Pedro ha has sido been ser acosado/ harassed excitado/ excited humillado/ humiliated seducido seduced (por (by María). María ‘Pedrohasbeen{harrassed/excited/humiliated/seduced}byMaría.’ b. Juan Juan se refl.3p.sg distrajo distracted viendo watching una a película. movie ‘Juanamusedhimselfwatchingamovie.’ 228 Section5.5. Agentivity? Chapter5 Table5.4:Marín’s(2011)classificationofOEPVs Non-agentive Agentive Inanimate Animate molestar ‘annoy’ humillar ‘humilliate’ inxtime – – – + forxtime + + + + terminar/acabar‘tofinish’ – – – ? absoluteclauses – – – ? habitualreadinginthepresent – – – + lentamente‘slowly’ – – – – parar‘tostop’ – – – – V-ingimpliesV-ed – – – + ambiguitywithcasi‘almost’ – – – + c. Azarías Azarías hizo made a dom sus his amigos friends humillar humilliate a dom Paco. Paco ‘AzaríasmadehisfriendshumilliatePaco.’ Sofar,thereisnothingsurprising:agentiveOEPVshavebeenrepeatedlyobservedtolosetheir‘psych’ properties,grammaticallyspeaking,anditisalsothereceivedviewthattheagentivesubjectversions areeventive.Ourchallengeistofigureoutwhetherwecanhaveagentivestativeversions,andwhether theyshowpsycheffectsornot.Letusreviewthelistofstrictlystative/non-agentiveOEPVsin(379). (379) StativeOEPVsinSpanish aburrir‘bore’, acongojar‘wring’, afligir‘sadden’, angustiar‘distress’, anonadar‘stun’, apasionar‘be passionateabout’, apenar‘sadden’, apesadumbrar‘sadden’, deprimir‘depress’, desesperar‘exasper- ate’, disgustar ‘upset’, entristecer ‘sadden’, entusiasmar ‘enthuse’, fascinar ‘fascinate’, indignar ‘out- rage’,interesar‘interest’,obnubilar‘bewilder’,obsesionar‘obsess’,ofuscar‘obfuscate’,preocupar‘worry’. Asitturnsout,afewofthesestativeverbsdopatternasagentive,albeitpartially(i.e. theyonlypass sometests). Interesar ‘interest’, forinstance, allowsfortheimperativeform(e.g. (463a))butdisallows intentionalityadverbs(e.g. (463b)). Aburrir‘bore’,ontheotherhand,cannotappearintheimperative form(e.g.(464a))butitcanbeaninfinitivalcomplementofdecide(e.g.(463b)). 229 Section5.5. Agentivity? Chapter5 (463) a. ¡Interesa ¡interest.imp.2p.sg a dom tus your clientes clients en in el the producto! product ‘Makeyourclientsinterestedintheproduct!’ b. ??Pedro ??Pedro me me está is interesando interesting intencionadamente. intentionally intended:‘Pedroisinterestingmeintentionally.’ (464) a. ??¡Abúrreme! ??¡Bore.imp.2p.sg.me intended:‘Boreme!’ b. El the cura priest hoy today ha has decidido decided aburrirnos bore.inf.us con with un a sermón sermon soporífero. sleep-inducing ‘Thepriesthasdecidedtoboreustodaywithadullsermon.’ Thus,whilethereisindeedatendencyforstativeverbstobenon-agentive,theagentivitytestsnonethe- lessgiveusinconsistentresultsacrosstheboardforthestativeverbsin(379). Crucially,thesestative OEPVsthatcanbepartiallyclassifiedasagentiveshowpsycheffectsratherstrongly:theydonotaccept verbalpassives(e.g. (465a)),theycannotappearinreflexive(e.g. (465b))andcannotformperiphrastic causatives(e.g.(465c)). (465) a. Fui was.1p.sg {??aburrido/ {??bored *interesado} *interested por by el the cura. priest intended:‘Iwas{bored/interested}bythepriest.’ b. *Yo I me me {aburrí/ {bored interesé} interested a dom mí me mismo. self intended:‘I{bored/interested}myself.’ c. *Esto *this le him hizo made {aburrir/ {bore interesar} interest a dom María. María intended:‘Thismadehim{bore/interest}María.’ It thus appears that the more precise generalization is that psych effects have stativity, rather than agentivity,astheirsource.LetusfocusagainonMarin’sagentiveverbsfrom(385),repeatedbelow. (385) AgentiveOEPVs acosar ‘harass’, agobiar ‘overwhelm’, aliviar ‘alleviate’, amedrentar ‘intimidate’, (des-)animar ‘en- courage/discourage’,apaciguar‘pacify’,asombrar‘astonish’,asustar‘frighten’,atemorizar‘terrify’, 230 Section5.5. Agentivity? Chapter5 confundir ‘confound’, consolar ‘console’, contrariar ‘displease’, (des-)honrar ‘(dis-)honor’, deslum- brar‘dazzle’, (des-)motivar‘discourage/encourage’, distraer‘distract’, entretener‘entertain’, espan- tar‘horrify’,estimular‘stimulate’,excitar‘excite’,fastidiar‘annoy’,frustrar‘frustrate’,humillar‘hu- milliate’,importunar‘annoy’,molestar‘disturb’,ofender‘offend’,oprimir‘opress’,perturbar‘disturb’, seducir‘seduce’,sorprender‘surprise’. Further evidence that agentivity is not at the heart of psych effects comes from telic agentive verbs like humillar ‘humilliate’ and seducir ‘seduce’: they show psych effects in their non-agentive reading too (i.e. with inanimate ‘causers’). I provide examples from verbal passives in (466). Examples from reflexivesandperiphrasticcausativescannotbeprovidedsincetheseconstructionsrequireagents. (466) a. Durante for muchos many años, years fui he.was. ser seducido seduced por by aquella that imagen image que that se refl concretaba specified en in las the mujeres women a dom las the que that amé. he.loved ‘Formanyyears,IwasseducedbythatimagethatmaterializeditselfinthewomenIloved.’ x (Lasombradelamor,AnteroJiménezAntonio,2006:18) Magín Magín Díaz, Díaz un an ingeniero engineer que that fue was. ser seducido seduced por by la the economía. economy ‘MagínDíaz,anengineerthatwasseducedbytheeconomy.’ x (ElDía,August28 th 2017,http://eldia.com.do) Moreover,thereappeartobemanyotherverbsin(385)thatareactuallyneitherstativenortelic,but rather behave like activities. Such are, for example, acosar ‘harass’, amedrentar ‘intimidate’, consolar ‘console’,distraer‘distract’andentretener‘entertain’. I present below the tests that show that these verbs are in fact not stative. First, they receive a habitualreadinginthepresenttense(e.g. (467a)). Second, theydonotreceiveauniversalreadingin theperfectinthepresenceofadesde‘since’-adverbial(e.g.(467b)).Third,theycanhavebothatemporal and degree reading with the modifier un poco ‘a little’ (e.g. (467c)). Fourth, the synthetic future only makes available a temporal reading for these verbs, not an epistemic one (e.g. (467d)). Finally, with aconditionalsentence,wecanonlygetaprospectivereadingofthedependentclauseinthepresent tensewhenthemainclauseisinthefuturetense(e.g.(467e)). (467) a. Fernando Fernando {consuela/ {consoles distrae/ distracts entretiene entertains a dom Herminia} Herminia (*ahora/ (*now con with frecuencia). frequency ‘Fernando{consoles/distracts/entertains}Herminia(*now/frequently).’ 231 Section5.5. Agentivity? Chapter5 b. Rafael Rafael ha has {consolado/ {consoled distraido/ distracted entretenido} entertained a dom Carmen Carmen desde since 2010. 2010 intended:‘Fernandohas{consoled/distracted/entertained}Herminiasince2010.’ c. Pedro Pedro me me {consoló/ {consoled distrajo} distracted un a poco. bit ‘Pedro{consoled/distracted}meabit.’ 3 Degreereading:Pedroconsoled/distractedmetosomedegree. 3 Temporalreading:Pedrospentsometimeconsoling/distractingme. d. Gabriela Gabriela {consolará/ {console.fut distraerá/ distract.fut entretenendrá} entertain.fut a dom Mercedes. Mercedes ‘Gabrielawill{console/distract/entertain}Mercedes.’ 3 Deonticreading:Gabrielawillconsole/distract/entertainMercedesinthefuture. 7 Epistemicreading:IfinditpossiblethatGabrielaconsoles/distracts/entertainsMer- cedes. e. Si if David David {consuela/ {consoles distrae/ distracts entretiene} entertains a dom Fernando, Fernando Daniela Daniela se refl enfadará. get.angry.fut ‘IfDavid{consoles/distracts/entertains}Fernando,Danielawillgetangry.’ 3 Prospectivereadingofthedependentclause: Davidwillconsole/distract/entertain Fernandoatsomepointinthefuture,andthenDanielawillgetangry. 7 Presentreadingofthedependentclause: Davidcurrentlyconsoles/distracts/enter- tainsFernandoatsomepointinthefuture,andthenDanielawillgetangry. Ishowbelowthattheseverbscanindeedbeagentive:itissomethingalreadyclaimedinMarín(2011), butheonlypresentedtestsforthetelicsubsetof(385). Forcompleteness’sake,Ishowthatagentivity testsmayalsoapplysuccessfullytotheseatelicverbs.Theyacceptagent-orientedadverbs(e.g.(468a)), theimperative(e.g.(468b))andcanbecomplementsofdecide(e.g.(468c)). (468) a. Pedro Pedro me me {distrajo/ {distracted amedrentó} intimidated intencionadamente. intentionally ‘Pedro{distracted/intimidated}meintentionally.’ b. ¡{Distrae/ ¡{distract consuela/ console entretén} entertain a dom tu your hermano! brother ‘{Distract/console/entertain}yourbrothers!’ 232 Section5.5. Agentivity? Chapter5 c. Paco Paco decidió decided {distraer/ {distract consolar/ console entretener} entertain a dom mi my hermano. brother ‘Pacodecidedto{distract/intimidate/entertain}mybrother.’ Theseverbs, justlikethetelic humillar-class, donotshowpsycheffects. Theyacceptverbalpassives (e.g.(469)),theycanappearinreflexivestructures((e.g.(470a),andseealso(386b))andcanfurthermore formperiphrasticcausatives(e.g.(470b)). (469) a. En in este this momento moment fui I.was. ser distraido distracted por by una a ballena whale flotando floating por on el the Canal canal de of Sir Sir Francis Francis Drake. Drake ‘In that moment, I was distracted by a whale that was floating on the Sir Francis Drake Canal.’ (Noticiasdeajedrez,April19 th 2005,http://es.chessbase.com) b. Varias several parejas couples porteñas porteñas no not se refl dejaron leave.3p.pl amedrentar intimidate por by la the superstición superstition y and se refl casaron married hoy. today ‘SeveralcouplesfromBuenosAiresdidnotletsuperstitionintimidatethemandtheymar- riedtoday.’ (LaNación,June13 th 2017, http://www.lanacion.com.ar) (470) a. Régula Régula se refl está is {distrayendo/ {distracting entreteniendo} entertaining con with el the Scalextric Scalextric para for no not pensar think.inf en in sus her graves severe problemas. problems ‘Régula is {distracting/ entertaining} herself with the Scalextric so as to not think in her severeproblems.’ b. Mi my madre mother me me hizo made {entretener/ {entertain consolar} console a dom mi my hermano. brother ‘Mymothermademe{entertain/console}mybrother.’ Crucially,thesubjectdoesnotneedtobeagentiveforpsycheffectstodisappearwiththeseverbs.Take (469a),forinstance.Thecontextisoneinwhichtwochessplayersareplayingachessgameinaterrace bytheseaandoneofthemspotsawhalefloatinginthedistance.Thisdistractshimandheultimately losesthegamebecauseofthat,orsoheclaims.Thewhale(introducedbytheby-phrase)isclearlynot 233 Section5.5. Agentivity? Chapter5 an agent since it was not acting volitionally: for all we know, maybe the whale was not even alive. Andyet,thepassiveiswell-formed.Thereceivedhypothesisthatagentivityinducesthelossofpsych effectsthusweakensfurther,atleastforSpanish. The picture that emerges from my discussion in this section is summarized in Table 5.5. As the tableshows,thecorrectempiricalgeneralizationseemstobethatpsycheffectsarisewhenOEPVsare aspectuallystative,ratherthanwhentheyarethematicallyagentive.Istatethisgeneralizationin(471). (471) Empiricalgeneralizationaboutpsycheffects Thesourceofpsycheffectswithobject-experiencerandlocativeverbsisstativity,notlackof agentivity(atleastinSpanish) Table5.5:Agentivity,aspectandpsycheffectswithOEPVs Aktionsarttype Agentivity Psycheffects Examples Stative Non-agentive 3 compungir’toberemourseful’,enorgulle- cer’makeproud’ Somewhat agentive 3 aburrir‘bore’,deprimir‘depress’,intere- sar‘interest’ Eventive Telics Agentive 7 humillar‘humiliate’,seducir‘seduce’ Non-agentive 7 Activities Agentive 7 acosar‘harass’,consolar‘console’,distraer ‘distract’,entretener‘entertain’ Non-agentive 7 5.5.4 Agentivityasteleologicalcapability: Folli&Harley(2008) Asafinalstopinourdiscussiononagentivityandpsycheffects,itwouldbeworthwhiletoconsider anpotentiallyalternativeaccount,relyingonadefinitionofagentivitydifferentfromtheclassicone. In a recent paper, Folli & Harley (2008) propose that agentivity is best seen in terms of teleological capability, rather than animacy and intentionality. Teleological capability is defined in their work, after Higginbotham (1997), as ‘the inherent qualities and abilities of the entity to participate in the eventualitydenotedbythepredicate’(2007:191). Theauthorsprovideevidencefromdifferentconstructions,mostlyfromItalianandEnglish,which hadbeenobservedtobesensitivetoanimacyand/oragencyinthepast.Theynotethatverbsofsound emission like whistle, hum and squeak, are unergativeand therefore introduced in ‘Agent’position in 234 Section5.5. Agentivity? Chapter5 the specifier of vP or an analogous projection. 110 Yet, these verbs can have both animate (e.g. (472)) andinanimate(e.g.(473))subjects.Infact,someofthemmayonlyallowinanimatesubjects(e.g.(474– 475)). Whatseemstolicensethesubjectsin(472–475)istheircapabilitytoemitthesound,ratherthan animacyperse. (472) Gianni John ha/*é has/*is fischiato whistled (473) Il The treno train ha/*é has/*is fischiato whistled (474) Questo This tavolo/#Gianni table/#Gianni scricchiola squeaks (475) Thephone/#Johnrang. (Folli&Harley2008:191–2) TheauthorsalsoreviewlocativeandpossessiveconstructionswithhaveinEnglishandItalian,which they unify as involving abstract possession. They note that possessive have, unlike commonly as- sumed,isacceptablewithbothanimateandinanimatesubjects(e.g. (476-479)). Thedifferenceisthat animatesallowpossessionofbothinalienableandalienableparts(e.g. (476)and(478)),whereasinan- imatesonlyallowforpossessionofinalienableparts(e.g. (475)and(477)). Therestrictionthenisnot so much about animacy as it is about the kinds of things that animates and inanimates can possess: animateshavetheteleologicalcapabilityofcontrollingalienableobjects,whereasinanimatesarenot. (476) a. Gianni John ha has un a braccio arm rotto broken b. Gianni John ha has una a macchina car (477) a. La The quercia oak ha has molti many rami branches b. #La The quercia oak ha has un a uccello bird c. La The quercia oak ha has una a famiglia family di of uccelli birds sul on.the ramo branch 110 Evidencefromtheirunergativitycomesfromthefactthattheytaketheauxiliary habere‘have’inItalian,andnot essere ‘be’(e.g.(472–473)). 235 Section5.5. Agentivity? Chapter5 (478) a. Johnhasabrokenarm b. Johnhasacar (479) a. Theoaktreehasmanybranches b. #Theoaktreehasafamilyofbirds c. Theoaktree i hasafamilyofbirdsinit i (Folli&Harley2008:193) They also apply their proposal to causation chain effects in causative change-of-state verbs. They claimthattheDirectCausationcondition(Shibatani1976)canberecastinteleologicalterms: thestorm in(480a)istoogeneralofacausetoqualifyasafelicitousAgentintheteleologicalsense;ontheother hand, the branch in (480c) has the teleological properties to directly cause the event (e.g. it is sturdy enoughtobreakawindowwhenitismovingfastenough).Seealsothecontrastin(480b). (480) a. #Iltemporaleharottolafinestra Thestormbrokethewindow b. ?Ilvento/Ilcolpodiventoharottolafinestra ?Thewind/Thegustofwindbrokethewindow c. Ilramoharottolafinestra Thebranchbrokethewindow (Folli&Harley2008:195) FurthersupportfortheteleologicalviewofagencycomesfromverbsofpermissioninGreek,Russian andEnglish.Theyshowhowintheselanguagesanimatesubjectsofpermissionverbsarelicensedboth with perfective (e.g. (481a)) and imperfective (e.g. (481b)) aspect. On the other hand, when they have inanimatesubjects,perfectiveaspectisout(e.g.(481c)),andonlyimperfectiveaspectisfine(e.g.(481d)). Again,theylinkthiseffecttotheteleologicalpropertiesofthesubject.Alicensegrantspermissionby itsveryexistence. Intheperfective,theeventofpermissionisassertedtobeoverbyreferencetime, although the subject and reference time overlap. This creates a conflict in the case of a license: The permissioneventisassertedtobeover,andthelicensewithitsinceitis,sotospeak,thepermission event.Henceitisacontradictiontopresupposeitstillexistsatreferencetime. (481) a. O det idioktitis owner mas us epetrepse permit.pst.pf na na exume have skili, dog ala but den neg ixame have.pst.pl skili dog ‘Theownerpermittedustohaveadog,butwedidn’thaveadog.’ b. O det idioktitis owner mas us epetrepe permit.pst.impf na na exume have skili, dog ala but den neg ixame have.pst.pl skili dog ‘Theownerpermittedustohaveadog,butwedidn’thaveadog.’ 236 Section5.5. Agentivity? Chapter5 c. #Ekini #that i det adia license mas us epetrepse permit.pst.pf na na exume have skili, dog ala but den neg ixame have.pst.pl skili dog ‘Thelicensepermittedustohaveadog,butwedidn’thaveadog.’ d. Ekini that i det adia license mas us epetrepe permit.pst.impf na na exume have skili, dog ala but den neg ixame have.pst.pl skili dog ‘Thelicensepermittedustohaveadog,butwedidn’thaveadog.’ x (Greek,fromFolli&Harley2008:196–7) Folli&Harley’s(2008)accountneatlyexplainsthedataobservedin(466), repeatedbelow. ‘Agentive’ OEPVslikeseducir‘seduce’oramedrentar‘intimidate’maytakeinanimatesubjectsiftheyareagentsin theteleologicalsense:animagemayseducebecauseithasappealingpropertiesandsuperstitionmay intimidatebecausetheyarethoughttocarryominousconsequences.Again,aswediscussed,wehave nopsycheffectsin(466),whichcouldreinforcetheideathatagentivitydissolvespsycheffects. (466) a. Durante for muchos many años, years fui was.1p.sg seducido seduced por by aquella that imagen image que that se refl concretaba specified en in las the mujeres women a dom las the que that amé. loved.1p.sg ‘Formanyyears,IwasseducedbythatimagethatmaterializeditselfinthewomenIloved.’ x (Lasombradelamor,AnteroJiménezAntonio,2006:18) b. Varias several parejas couples porteñas porteñas no not se refl dejaron leave.3p.pl amedrentar intimidate por by la the superstición superstition y and se refl casaron married hoy. today ‘SeveralcouplesfromBuenosAiresdidnotletsuperstitionintimidatethemandtheymar- riedtoday.’ (LaNación,June13 th 2017) However,theteleologicalviewofagentivity,justliketheclassicalview,failstoshedlightontheex- istenceofpsycheffectswithstativeOEPVsandlocativeverbs.Take(482),withthestativeOEPVater- rorizar ‘terrify’. In(482a),weillustrateitwithananimateandinanimatesubject. Both,notewell,are AgentsinFolli&Harley’s(2008)sense: theybothhavepropertiesthatcanarouseastateoffrightin myself. Norman Bates is a serial killer, and the Psycho house is gloomy and sinister. And yet, these Agentsubjectsfailtolicensepsycheffects,asshownin(482b).Wefindthesamesituationwithlocative verbs(e.g.(483)). 237 Section5.6. Conclusions Chapter5 (482) a. {Norman {Norman Bates/ Bates la the casa house de from Psicosis} Psycho me me aterroriza. terrifies ‘{NormanBates/thePsychohouse}terrifiesme.’ b. *Yo *I soy am aterrorizado terrified por by {Norman {Norman Bates/ Bates la the casa house de from Psicosis} Psycho intended:‘Iamterrifiedby{NormanBates/thePsychohouse}.’ (483) a. {Los {the escombros/ debris los the niños} children están are obstruyendo obstructing la the salida. exit ‘{Thedebris/thechildren}areobstructingtheexit.’ b. *La *the salida exit es is obstruida obstructed por by {los {the escombros/ debris los the niños} children intended:‘Theexitisobstructedby{thedebris/thechildren}.’ Torecap,agentivitydefinedintermsofteleologicalcapability—i.e.theinherentpropertiesofanentity thatallowsittobringabouttheeventbyitself—alsoprovesunusefultoaccountfortheunavailability ofpsycheffects. Onceagain,theonlysolidgeneralizationthatwecanextractfromourdiscussionis thatthelocusofpsycheffectsisstativity,andnotagentivity,howeverdefined.Irepeatthegeneraliza- tionin(471)below. (471) Empiricalgeneralizationaboutpsycheffects Thesourceofpsycheffectswithobject-experiencerandlocativeverbsisstativity,notlackof agentivity(atleastinSpanish) 5.6 Conclusions ThischapterhasarguedagainsttheviewthatresortstoanExperiencer-roletoexplainpsycheffects withstativeOEPVs. Ihavedrawnevidencefromstativelocativeverbscrosslinguisticallyandshown thatthesamepsycheffectsappearwiththoseverbs,despitetheirobjectisclearlynotanExperiencer. I have argued for a common syntactic structure for these verbs, whose argument structure is artic- ulated by prepositional structure. This shared syntactic structure explains the appearance of psych effects with both types of verbs. In turn, the generalized semantics encoded in the structure derive therelevanteventrolesoftheargumentsofsuchverbs. (425) a. El the río river Nilo Nile rodea surrounds el the desierto desert de of Bayuda. Bayuda ‘TheriverNilesurroundstheBayudadesert.’ 238 Section5.6. Conclusions Chapter5 b. PP DP ElríoNilo P’ P <rodea> DP eldesiertodeBayuda c. JPPK=9",V [eigen(",eldesiertodeBayuda)&rodea(V,")&endpoint(elríoNilo,V)] (427) a. El the clima climate de of Los Los Angeles Angeles agobia overwhelms a dom mi my amiga friend Ester. Ester ‘TheLosAngelesweatheroverwhelmsmyfriendEster.’ b. PP DP ElclimadeLosAngeles P’ P <agobia> DP amiamigaEster c. JPPK=9",V [eigen(",miamigaEster)&agobia(V,")&endpoint(elclimadeLosAngeles,V)] Ifontherighttrack,thisisanimportantcontributiontotheneoconstructionistviewthateventroles areentailmentsfromtheeventstructureoftheverb,ratherthangrammaticalprimitives,andthatthis eventstructure, inturn, issyntacticallydetermined. Thishasbeensomethingexploredindepthfor causativeverbs(Hale&Keyser1993;Ramchand2008,a.o.,asdiscussedinSections2.3.2and2.3.4),and by now it is generally accepted that roles such as agent/causer and undergoer/theme can be so derived. Inthislight, mychaptertakesanotherstepforwardandshowsthatrolessuchasstimulus andexperiencerarealsoentailmentsfromtheeventstructure,somethingthathadnotbeenseriously exploreduntilknow,asfarasIamaware. Still,therearequestionsthatremain.Forinstance,howshouldwederiveSEPVs,i.e.psychological verbswherethesubject,andnottheobject,istheExperiencer(e.g.(484a))?Anoptionistoassumethat thesubjectofSEPVsisinfactthesamePPthatderivesOEPVs,asFábregas&Marín(2015a)propose (e.g. (484b)), where their StatP is a stative verbal projection). Note that in their approach P assigns anExperiencer-role. Inprinciple,myanalysisofOEPVscouldbeaccomodatedtothisframework, 239 Section5.6. Conclusions Chapter5 assumingStatPismeaninglessandthetruepredicateoftheVPisinfactP . Still,itisnotclearhow lexicalizationwouldcomeabout,giventhatthelexicalitemwouldnotformaspan: VandParenot syntacticallycontiguousheads. (484) a. Juan Juan teme fears a acc María. María ‘JuanfearsMaría.’ (Fábregas&Marín2015a:258) b. StatP PP P DP Juan Stat’ Stat teme DP aMaría ExperiencerEmotionTargetofemotion Anotherissueiswhyweneverfindpsycheffectswitheventivepredicates. Thatis,asfarasweknow, psycheffectsarenotattestedwithactivitiesandchange-of-stateverbscrosslinguistically.Irepresent thenonexistentstructuresin(485)((485a)forchange-of-stateand(485b)foractivityconfigurations). (485) a. *[ procP proc[ resP res[ PP DP Stimulus/Figure [ P’ P [DP Experiencer/Ground ]]]]] b. *[ procP proc[ PP DP Stimulus/Figure [ P’ P [DP Experiencer/Ground ]]]] ThisisamatterthatdeservesmorediscussionthanwhatIamabletoofferhere,butIwillnonetheless point towards a possible solution. With OEPVs and locative verbs, I have argued, it is a preposition what articulates argument structure and its associated event structure, rather than a verb. If so, we couldthinkofP asprojectingitsownfirstphasesyntaxdomain,inwhichverbalheadssuchasinit, proc and res do not come into play. The predicate and its arguments are articulated by PP, and the verbalizerVisinsertedformorphologicalreasons—i.e.supportingtenseandaspectmorphemes. 240 6 Conclusions 6.1 Summaryoffindings ThisdissertationhasstudiedthemorphosyntaxandsemanticsofcertainstativepredicatesinSpan- ish: stativecausativeverbs,OEPVs,locativeverbsandstativeparticiples. Mymainconcernhasbeen toprovideacross-categorialaspectualtypologyofstativepredicatesandtoderivetheirthematicdi- versity from the event structures they lexicalize. I have conducted my research from the neocon- structionistviewthatthesyntaxoftheVPdeterminesitsformaleventstructureand,indirectly,the thematicinterpretationoftheirarguments,conceivedasentailmentsfromeventstructure. Iarticu- latedmyresearcharoundtheresearchquestionsin(7). (7) Researchquestionsofthisdissertation: (7a) Whatsyntactico-aspectualprimitivesderivethesubclassesofstativeverbs? =)Chapter3 (7b) Howdostativeeventstructuresassignthematicinterpretationtoverbalarguments? =)Chapter5 (7c) Whatisthestructuralrelationshipbetweenstativeparticiplesandtheirverbs? =)Chapter4 InChapter2,Ipresentedastateoftheartregardingtheaspectualtypologyofverbsandtheirsyntactic representation.Iarguedforaviewofgrammarinwhichsyntacticcomplexitymirrorsanddetermines event complexity, which is both theoretically more elegant and empirically more adequate. I also arguedthataneoconstructionistapproachtotheVPmustassumearicheventdecompositioninorder to adequately derive event types and thematic interpretation. All in all, the purpose of this chapter was to introduce the classic event typology that my research on Chapter 3 would later expand, as well as to introduce the first-phase syntax model, the theoretical framework I have adopted for my investigations. 241 Section6.2. Maincontributions Chapter6 Chapter3analyzedgobernar-typeverbsinSpanish.Iarguedthattheyarestativecausatives,com- posedoftwostate-denotingprojectionswhicharerelatedcausallybyvirtueoftheirsyntacticconti- guity: initPand resP.Iopposedthisviewtoothersthatdenythatacausalrelationcanholdbetween states,aswellastheviewthatcharacterizesgobernar-verbsasnon-dynamicevents.Ialsodiscussedthe implicationsofmyproposalregardingtheconnectionbetweentheexternalargumentandtheariseof causeasaconfigurationalphenomenonandreconciliateditotheraccountsthatproposeasplitinthe semanticandsyntacticintroductionoftheexternalargument,ontheonehand,andacauserelation inunaccusativeconfigurations,ontheother. In Chapter 4, I analyzed PPrts in Spanish. I showed that attributive PPrts are eventive—verbal passiveswithoutanauxiliary—whereaspredicativePPrtswiththeauxiliaryestararestative.Ifurther arguedthatthereisasplitwithinstativePPrtsdependingontheAktionsartofthebaseverb:PPrtsde- rivedfromstativecausativeverbsmaintaintheirfullargumentandeventstructure—twoprojections, initPandresP,andtheirassociatedstates—whereasPPrtsderivedfromtelicverbsaretruncatedstruc- tures,composedonlyofastativeresP.This,Iargued,followsfromtherequirementthattheestar-PPrt construction be stative. I also provided a crosslinguistic typology of adjectival passives in terms of languagesthatbehavelikeSpanish—notablyGermanandHebrew—andlanguagesthatdonot, such asGreekandChichewa. Chapter5challengedtheideathatthereisanExperiencer-roleinvolvedinthegrammarofstative OEPVs. I showed that the idiosyncratic grammatical properties of these verbs—the so-called psych effects—alsoappearinlocativeverbscrosslinguistically,whichclearlydonotinvolveanExperiencer. Iproposedauniformsyntacticstructureforthetwoverbtypeswithanassociatedabstractrelational semantics.Iarguedthatthisrelationcanbeunderstoodaslocativeorpsychologicaldependingonthe conceptualmeaningofverbthatlexicalizesthateventstructure.Inturn,Iproposedthattheeventrole oftheargumentsinvolvedwasanentailmentoftheeventstructurecoupledwiththelexicalmeaning oftheverb.Finally,IarguedthatthesepsycheffectsthatOEPVsandlocativeverbssharedonothave non-agentivityastheirlocus,butstativity. 6.2 Maincontributions MydissertationhasaugmentedtheclassicVendleriantaxonomyofeventtypeswiththeadditionof causativestates.Insodoing,Iarguedthat,despitethereceivedview,statescanbeaspectuallycomplex: thenotionsofcauseandresultnotonlymayapplytoevents,butalsotostates,andareindispensablefor acomprehensiveaspectualtypologythatincludes gobernar-typeverbs. Ialsoshowedthatthisricher typologycanbederivedassumingaminimumsetofsyntacticprimitives—theRamchandianprojec- tionsinitP,procPandresP,whicharehereexplotedtotheirfullpotential—andsemanticprimitives— 242 Section6.3. Futuredirections Chapter6 the classic sortal distinction between states and events and the generalized cause relation. In this view,statesarenotimpoverishedordefectiveprimitives,butfully-functioningeventualitiesthatcan formcomplexeventtypesfollowingwell-definedprinciplesofsyntacticandsemanticcomposition. Myresearch,inturn,hasmadenon-trivialpredictionsforthemorphosyntaxandsemanticsofthe VP.FortheupperpartoftheVP—theVoiceP/initParea—Ihaveexplicitlyarguedthatthesemanticin- troductionoftheexternalargumentandtheinitiationstateareinseparablesincetheyareperformed by the same head: in other words, I make the prediction that there are no languages where cause canappearwithoutanimplicitexternalargument. ForthelowerpartoftheVP—theresParea—Ial- low for telic roots to lexicalize a single resP in the syntax, following independently well-established morphosyntacticprinciples. Thismoveallowedustoexplainmanypropertiesofadjectivalpassives crosslinguistically,aswellasclarifytherecurrentconfusionbetweenstativity,resultativityandper- fectivityinthisdomain. Furthermore,mydissertationhasprovidedfurthersupportfortheviewthatthematicrolesdonot existasprimitives,butaremerelyentailmentsfromeventstructure.Thishadbeensomethinghitherto pursued for the domain of causative events, where notions such as initiator and undergoer/ resultee could be easily derived from the aspectual configuration of the subevents. However, I have shown that even roles such as Experiencer or Stimulus can be derived as entailments from event structure onceweunderstandthejointworkoftheabstractaspectualconfigurationandthelexicalmeaningof theverb. Thesefindingsthenprovideanotherblowtoclassic-theory,whichcontendsthat-roles areprimitivesofthegrammaticalsystem. Seen as a whole, the research conducted here has been a contribution to the neoconstructionist researchprogramingeneral,andtoitsresearchonstativepredicatesinparticular. Ihavereinforced theviewthateventtypesandthematicrolesarederivedsyntactically,thusaddingtocurrentneocon- structionist research on the syntax-semantics interface of the VP (Travis 2000; Borer 2005b; Ramc- hand2008;Lohndal2014,a.o.).Ithasalsocontributedtofillavoidinthisprogram,whichhasbylarge neglectedstativeverbalpredicates(butseeArad1998b;Pylkkänen2000;Alexiadou2011;Jaque2013for notableexceptions). 6.3 Futuredirections Asiscommonlythecase,thepaththatIhavebeguninthisdissertationhasopenedbifurcationsthat gobeyondthedistanceIcouldtravel. 243 Section6.3. Futuredirections Chapter6 6.3.1 Locativeinversion Anopenquestioniswhatshouldbetheanalysisoftheestar-PPrtconstructionwithPPrtsderivedfrom locative verbs and OEPVs, as was briefly mentioned in the example (xvii) from footnote 41, which I repeatbelow. (xvii) a. La the finca ranch está is estar rodeada surrounded por by tres three colinas. hills ‘Theranchissurroundedbythreehills.’ b. Pedro Pedro está is estar preocupado worried por by la the crisis crisis económica. economic ‘Pedroisworriedbytheeconomiccrisis.’ Theestar-PPrtconstructionwithPPrtsderivedfromtelicandstativecausativeverbswasthoroughly studied in Chapter 4, as were OEPVs and locative verbs in Chapter 5. However, I did not explore howOEPVsandlocativeverbsformedestar-PPrtconstructions.GiventhedifferentstructuresIposit for OEPVs and locative verbs, on the one hand, and stative causatives on the other (Chapter 3), we shouldexpectthemtoformdifferentkindsofPPrtsintheestar-construction.Itseemstobethecase: for one, the introduction of the Figure/Stimulus argument may or must be introduced by different prepositions depending on factors such as the lexical verb and the specificity of the argument (e.g. (486a-b)), whereas with stative causative verbs the Initiator argument is always introduced within a by-phrase(e.g.(486c)). (486) a. La the ciudad city está is rodeada surrounded {de/ {of ?por} ?by montañas. mountains intended:‘Thecityissurrounded{of/?by}mountains.’ b. María María está is estar alucinada astonished {con/ {with ?por} ?by el the nuevo new edificio. building intended:‘Maríaisastonished{with/?by}thenewbuilding.’ c. El The país country está is estar gobernado governed {?de/ {?of ?con/ ?with por} by militares. military intended:‘Thecountryisgoverned{?of/?with/by}themilitary.’ Myintuition,alongwithConti-Jiménez(2004),isthatestar-PPrtswithlocativeverbs—and,byexten- sion,OEPVs—areacross-categorialinstanceofthelocativealternation,i.e.atypeofargumentstructure alternationthatinvolvesalocativerelation(cf. thewell-known spray/loadalternation). Beasitmay, 244 Section6.3. Futuredirections Chapter6 itremainsto bedeterminedwhattherelation isbetweentheVP-structureI defendedforthesetwo verbtypesinChapter5andtheirassociatedestar-PPrt. 6.3.2 Nominalizations Additionally,itwouldbeinformativetostudythenominalizationsofthestativeverbsstudiedinmy dissertation.Thetheoryputforthhere—thatdeverbalmorphologycanconsistof‘defective’structures that featurally-richer verbs can lexicalize—could shed light on the typology of nominalizations put forth in Grimshaw (1990), which distinguishes between simple and complex nominals. I illustrate theirpropertiesonTable6.1(adaptedfromAlexiadou2001;Borer2013,examplesfromAlexiadou2001), noting that simple and complex nominals appear to be suspiciously parallel to estar-PPrts derived fromtelicandstativecausativeverbs,respectively. Table6.1:Grimshaw’s(1990)typologyofnominalizations Simple/ResultNominals ComplexEvent/A-SNominals INoevent-relatedroles IEvent-roles(i.e. T-roles) INoeventreading(487a) IEventreading(488a) IComplementsareoptional(487b) IComplementsareobligatory(488b) INoagent-orientedmodifiers(487c) IAgent-orientedmodifiers(488c) ISubjectsarepossessives(487d) ISubjectsarearguments(488d) IBy-phrasesarenon-arguments IBy-phrasesarearguments I(Spanishde,Hebrewšel)(487e) I(Spanishpor,HebrewPalyedey)(488e) INoimplicitargumentcontrol(487f) IImplicitargumentcontrol(488f) INoaspectualmodifiers(487g) IAspectualmodifiers(488g) IModifierslikefrequentonlywithplural IModifierslikefrequentonlywithsingular Inouns(487h) Inouns(488h) (487) Simple/ResultNominals a. *theexamatnoon b. *theexamofthepapers c. *theintentionalexamisdesirable d. (*)theinstructor’sexamination e. apicturebyapainter 245 Section6.3. Futuredirections Chapter6 f. *theexaminordertopassallthestudents g. *theexamforthreehours h. *thefrequent{*exam/exams} (488) ComplexEventNominals a. theexaminationofthestudentsatnoon. b. theexamination*(ofthepapers) c. theinstructor’sintentionalexaminationofthestudent d. theinstructor’sexaminationofthepapers e. thedestructionofthecitybytheenemy f. theassignmentofeasyproblemsinordertopassallthestudents g. theexaminationofthepapersinthreehours h. thefrequent{examination/*examinations}ofthepapers (489) a. el the aislamiento insulation hermético hermetic de of la the unidad unit de of residuos residues tóxicos toxic (*por (*by parte part de of los the expertos). experts intended:‘Thehermeticinsulationofthetoxicresiduesunitbytheexperts.’ b. el the rápido quick aislamiento insulation de of la the unidad unit de of residuos residues tóxicos toxic por by parte part de of los the expertos. experts ‘Thequickinsulationofthetoxicresiduesunitbytheexperts.’ Itremainstobeseenwhethersuchaparallelismcouldbesensiblyestablished.Anobviousdifferenceis thatestar-PPrtsderivedfromtelicverbsneverhavethefulleventreading—butseemycrosslinguistic discussion in Section 4.6—whereas with nominals they have both readings. More strikingly, with telicverbs,theeventreadingdoesnotincludethenotionalresultstateintheeventdecomposition:in (489a),withthedisambiguatingresult-orientedadverbhermético‘hermetic’,theeventreadingin(489b) becomesimpossible.Ultimately,researchonnominalizationsshouldprovetoinform—andhopefully strengthen—theproposalputforthinthisworkforthestativepredicatesunderstudy. 6.3.3 Stativityandgrammaticalcategory Finally,thereisafar-reachingquestionthatIhavenottackledfrontallyinthisdissertation,andcer- tainly deserves future explorations. Such question concerns the relationship between stativity and grammaticalcategory.Now,inSection2.3.2.2,weshowedthatstativitycanmanifestitselfinallgram- matical categories and, moreover, that there is no one category that exclusively hosts stative predi- cates.Thiswasshownin(79)–(83),repeatedbelowforconvenience. 246 Section6.3. Futuredirections Chapter6 (79) Stativeverbs a. Thebookcostsfivedollars. b. Peterfearshisparents. (80) Eventiveverbs a. Maryjumpedacrosstheriver. b. Johnchurnedthebutter. (81) Stativeadjectives a. Hansishandsome. b. Aweeklynewspaper. (82) Eventiveadjectives a. Marywasbeingunkind. b. Carmenisusuallyquitegenerous. (83) Stativenouns a. Thelovetowardsyourneighbor. b. Yourfearofgunsiswarranted. (84) Eventivenouns a. Theviolentfightbetweenthegangs. b. Mycousin’slongweddingparty. However,derivationalmorphologydoesseemtoprefernon-verbalcategorieswhenitcomestostative predicates. The examples in (490) show that we can derive telic verbs from adjectives with relative ease:in(490a),wehaveanadjectivalpredicatewhoseverbalsupportisalightverbdenotingachange ofstate.As(490b)shows,wecanderiveadeadjectivalverbwhichalsodenotesachangeofstate.This morphosyntacticflexibilityfortelicpredicatesdoesnotcarryovertostativepredicates.Theexamples in(491)areevidencethatwecannotderiveastativeverbalpredicatefromastativeadjectivalpredicate. Thatis,sentence(491b)isungrammaticalunderastativereadinganalogoustothatof(491a). (490) a. El the cielo sky se refl puso became oscuro. dark ‘Theskybecamedark.’ b. El the cielo sky se refl oscureció. darkened ‘Theskydarkened.’ (491) a. El the cielo sky está is oscuro. dark ‘Theskyisdark.’ b. *El *the cielo sky oscurece. darkens intended:‘Theskydarkens.’(disregardthehabitualreading) A similar situation is found with denominal verbs. As (492b) shows, we can easily derive telic verbs fromnominals,whichcanmoreoverbeparaphrasedwithanotherlexicalorlightverbformingatelic 247 Section6.3. Futuredirections Chapter6 predicate (e.g. (492a)). Again, this morphosyntactic flexibility is not found in the stative domain, as shownin(493),where(493b)cannothaveameaninganalogoustothatof(493a). (492) a. Vertí poured.1p.sg el the vino wine en in la the botella. bottle ‘Ipouredthewineintothebottle.’ b. Embotellé bottled.1p.sg el the vino. wine ‘Ibottledthewine.’ (493) a. El the vino wine está is en in la the botella. bottle ‘Thewineisinthebottle.’ b. *El *the vino wine embotella. bottles V intended:‘Thewinebottles V .’ Wefindthesameresistancetoformverbalstativepredicateswithintheverbaldomain.Asithasbeen extensivelydiscussedintheliterature—andinthisdissertation,tangentially—telicandactivityverbs often alternate between a transitive and an intransitive counterpart, i.e. the transitive-unaccusative or (anti-)causative alternation. This is exemplified in (494a) and (494b) with the verb reventar ‘burst’, where(494a) isthetransitiveversion withanInitiator/Agentsubject andaTheme objectand(494b) istheintransitiveversionwithaThemesubjectandnoAgent. Whatismuchlessoftenpointedout, however,isthatverbalmorphologydoesnotallowtoformastativeversionof(494b)—or(494a),for that matter—as shown in (494c). To express the result state of the bursting, we need to resort to an adjectivalpassive(seeChapter4),asshownin(494d). (494) a. María María reventó bursted el the globo. balloon ‘Maríaburstedtheballoon.’ b. El the globo balloon reventó. bursted ‘Theballoonbursted.’ c. *El *the globo balloon revienta. bursts intended:‘Theballoonbursts.’(disregardeventivereading) 248 d. El the globo balloon está is reventado. bursted ‘Theballoonisbursted.’ It seems rather clear that the grammar resists to derive verbal states. Current approaches to verbal morphosyntax,asfarasIamaware,cannotderivethisstateofaffairsotherthanbybrutestipulation. First-phasesyntaxpredicts,inprinciple,thepossibilityofhavingarootspecifiedas[(init),proc,res]to lexicalizeabare resPprojection,butthisdoesnothappeninthegrammar(e.g. (494c)). TheL-syntax model (see 2.3.2) would be forced to assume that all change-of-state verbs start off as adjectives de- notingtheresultstate,butthisisatoddswiththefactwefindverbalmorphologywithinparticiples (e.g. caramelizado‘caramelized.’).DistributedMorphology,whichpositsthatrootsareacategorialand become verbs when attached to a verbalizer, would also face the challenge to restrict certain roots to attach to stative verbalizers. It looks like we have a long way to go until we find a deeper expla- nationfortheimpossibilityofderivingtheseverbalstateswhichultimatelytranscendsthechoiceof framework. Needlesstosay,pursuingallthequestionspresentedinthissectionwouldtakethisthesistoofar afield.Ithusleavethediscussionhere,onthehopesthatfutureworkwillshedlightontheremaining issues and challenges that the study of stativity and morphosyntax still face. My quest for answers willcontinue,butthepresentdissertationendshere. 249 Bibliography Acedo-Matellán,Víctor.2016. TheMorphosyntaxofTransitions. Oxford:OxfordUniversityPress. Aelbrecht,Lobke,&Harwood,William.2015. Tobeornottobeelided: VPellipsisrevisited. Lingua, 153,66–97. Akmajian,Adrian,&Wasow,Thomas.1975.TheconstituentstructureofVPandAUXandtheposition oftheverbbe. LinguisticAnalysis,1,205–245. Alexiadou,Artemis.2001. FunctionalStructureinNominals. Amsterdam:JohnBenjamins. Alexiadou,Artemis.2011. Stativesandnominalization. RechercheslinguistiquesdeVincennes,40,25–52. Alexiadou,Artemis,&Anagnostopoulou,Elena.2008. StructuringParticiples. Pages 33–41 of: Chang, C. B., & Haynie, Hannah J. (eds), Proceedings of the 26th West Coast Conference on Formal Linguistics. Somerville,MA:CascadillaProceedingsProject. Alexiadou, Artemis, Anagnostopoulou, Elena, &Schäfer, Florian.2006. ThePropertiesofAnticaus- atives Crosslinguistically. In: Frascarelli, M. (ed), Phases of Interpretation. Berlin: Mouton de Gruyter,187-211. Alexiadou, Artemis, Gehrke, Berit, & Schäfer, Florian. 2014. The argument structure of adjectival participlesrevisited. Lingua,149,118–138. Alexiadou, Artemis, Anagnostopoulou, Elena, &Schäfer, Florian.2015. External Arguments in Transi- tivityAlternations: ALayeringApproach. Oxford:OxfordUniversityPress. Anagnostopoulou, Elena. 1999. On Experiencers. In: Alexiadou, Artemis, Horrocks, Geoffrey C., & Stavrou,Melita(eds),StudiesinGreekSyntax. Dordrecht:Kluwer,67–93. Anagnostopoulou,Elena.2003. ParticiplesandVoice. In: Alexiadou,A.,Rathert,M.,&vonStechow, A.(eds),PerfectExplorations. Berlin:MoutondeGruyter,1-36. Anagnostopoulou, Elena. 2017. Voice, manners, and results in adjectival passives. In: D’Alessandro, Roberta,Franco,Irene,&ÁngelGallego(eds),TheVerbalDomain. Oxford:OxfordUniversityPress, 105–128. 250 Anagnostopoulou,Elena,&Samioti,Yota.2013. Allosemy,Idioms,andtheirDomains:Evidencefrom AdjectivalParticiples. In:Folli,Raffaella,Sevdali,Christina,&Truswell,Robert(eds),Syntaxandits Limits. Oxford:OxfordUniversityPress,218–250. Arad,Maya.1998a. Psychnotes. UCLWorkingPapersinLinguistics,10,203–223. Arad,Maya.1998b. VPStructureandtheSyntax-LexiconInterface. Ph.D.thesis,UCL. Arad,Maya.1999. Whatcountsasaclass? Thecaseofpsychverbs. MIT Working Papers in Linguistics, 1–23. Arad, Maya. 2002. Universal Features and Language-Particular Morphemes. In: Alexiadou, A. (ed), TheoreticalApproachestoUniversals. Amsterdam:JohnBenjamins,15-39. Arche,MaríaJesús.2006. IndividualsinTime. Amsterdam:JohnBenjamins. Arche,MaríaJesús,Fábregas,Antonio,&Marín,Rafael.2017. TowardsaunifiedtreatmentofSpanish copulas. In: Perpiñán, Silvia, Heap, David, Moreno-Villamar, Itziri, & Soto-Corominas, Adriana (eds),RomanceLanguagesandLinguisticTheory11. Amsterdam:JohnBenjamins,33–52. Bach,Emmon.1986. Thealgebraofevents. LinguisticsandPhilosophy,9(1),5–16. Baker,Mark,Johnson,Kyle,&Roberts,Ian.1989. Passiveargumentsraised. Linguistic Inquiry,20: 2, 219–51. Baker,MarkC.1985. Themirrorprincipleandmorphosyntacticexplanation. LinguisticInquiry,23(4), 373–415. Baker,MarkC.1988. Incorporation: A Theory of Grammatical Function Changing. Chicago: University ofChicagoPress. Baker,MarkC.1997. Thematicrolesandsyntacticstructure. Pages73–137of: Haegeman,Lilliane(ed), Elementsofgrammar. Springer:Netherlands. Baker,MarkC.2003.LexicalCategories: Verbs,NounsandAdjectives.Cambridge:CambridgeUniversity Press. Belletti,Adriana,&Rizzi,Luigi.1988.Psych-verbsandT-theory.NaturalLanguageandNaturalLanguage andLinguisticTheory,6,291–352. Benincà,Paula.1986. IllatasinistrodellafraseItaliana. AssociationofTeachersofItalianJournal,47(57– 85). 251 Bennett,Michael,&Partee,Barbara.1972. TowardtheLogicofTenseandAspectinEnglish.SantaMonica, California:SystemDevelopmentCorporation;reprintedwithanAfterwordbyIndianaUniversity LinguisticsClub,Bloomington,1978. Bentley,Delia.2017(May22-24). Result state adjectives: valence and voice. TalkgivenattheCambridge WorkshoponVoice. Boeckx,Cedric.2014.WhatPrinciplesandParametersgotwrong. In:Picallo,M.Carme(ed),Linguistic VariationintheMinimalistFramework. Oxford:OxfordUniversityPress,155-178. Borer,Hagit.1994. Theprojectionofarguments. In: Benedicto,Elena,&Runner,Jeff(eds),University ofMassachusettsOccasionalPapersinLinguistics. Amherst:GLSA,UniversityofMassachusetts,19-47. Borer,Hagit.2005a. StructuringSense,Vol.I:InNameOnly. Oxford:OxfordUniversityPress. Borer,Hagit.2005b. StructuringSense,Vol.II:TheNormalCourseofEvents. Oxford:OxfordUniversity Press. Borer,Hagit.2013. StructuringSense,Vol.III:TakingForm. Oxford:OxfordUniversityPress. Bosque,Ignacio.1990. Sobreelaspectoenlosadjetivosyenlosparticipios. In: Bosque,Ignacio(ed), Tiempoyaspectoenespañol. Madrid:Cátedra,177-214. Bosque,Ignacio.1999. Elsintagmaadjetival.Modificadoresycomplementosdeladjetivo.Adjetivoy participio. Chap.4of:Bosque,I.,&Demonte,V.(eds),Gramáticadescriptivadelalenguaespañola,vol. 1. Madrid:EspasaCalpe,217-310. Bosque,Ignacio.2014. OnresultativepastparticiplesinSpanish. CatalanJournalofLinguistics,41–77. Bouchard, Denis. 1992. Psych constructions and linking to conceptual structures. In: Hirschbühler, Paul,&Konrad,Koerner(eds), Romance Languages and Modern Linguistic Theory. Amsterdam: John Benjamins,25–44. Bošković, Željko. 2014. Now I’m a phase, now I’m not a phase: on the variability of phases with ex- tractionandellipsis. LinguisticInquiry,45,27–89. Bresnan, JoanW.1982. Thepassiveinlexicaltheory. In: Bresnan, JoanW.(ed), The Mental Represen- tationofGrammaticalRelations. MITPressSeriesonCognitiveTheoryandMentalRepresentation. Cambridge,MA:MITPress,3–86. 252 Brody, Michael.2000. Mirror theory: Syntacticrepresentationin perfectsyntax. Linguistic Inquiry, 29–56. Bruening, Benjamin.2014. Wordformationissyntactic: adjectivalpassivesinEnglish. Natural Lan- guageandLinguisticTheory,32,363–422. Brugger,Gerhard.2001. Temporalmodification,the24-hourruleandthelocationofreferencetime. In:Guti’errez-Rexach,Javier,&Silva-Villar,Luis(eds),CurrentIssuesinSpanishSyntaxandSemantics. Berlin/NewYork:MoutondeGryuter. Bull,William.1965. SpanishforTeachers. NewYork:RonaldPress. Burzio, Luigi. 1986. Italian Syntax: A Government-Binding Approach. Springer Science and Business Media. Caha,Pavel.2009. TheNanosyntaxofCase. Ph.D.thesis,UniversityofTromsø. Carlson, Gregory N. 1977. Reference to kinds in English. Ph.D. thesis, University of Massachusetts, Amherst. [Pub.NewYork:Garland,1980.(OutstandingDissertationsinLinguisticsseries.)]. Carrier,Jill,&Randall,JanetH.1992. Theargumentstructureandsyntacticstructureofresultatives. LinguisticInquiry,23(2),173–234. Chierchia, Gennaro. 2004. A semantics for unaccusatives and its syntactic consequences. In: Alex- iadou, A., Anagnostopoulou, E., & Everaert, M. (eds), The Unaccusativity Puzzle: Explorations of the Syntax-LexiconInterface. OxfordUniversityPress,22-59. Chomsky,Noam.1957. SyntacticStructures. TheHague/Paris:Mouton. Chomsky,Noam.1981. LecturesonGovernmentandBinding. Dordrecht:Foris. Chomsky,Noam.1986a. Barriers. Cambridge,MA:MITPress. Chomsky,Noam.1986b. KnowledgeofLanguage: ItsNature,OriginandUse. NewYork:Praeger. Chomsky,Noam.1995. TheMinimalistProgram. Cambridge,MA:MITPress. Chomsky,Noam.2001.Derivationbyphase.In:Kenstowicz,Michael(ed),KenHale: ALifeinLanguage. Cambridge,MA:MITPress,1-52. 253 Chomsky,Noam.2008. Onphases. In: Freidin,Robert,Otero,CarlosP.,&Zubizarreta,MaríaLuisa (eds), Foundational Issues in Linguistic Theory. Essays in Honor of Jean-Roger Vergnaud. Cambridge, MA:MITPress,291-321. Cinque, Guglielmo. 1999. Adverbs and functional heads: a cross-linguistic perspective. Oxford: Oxford UniversityPress. Clark,Eve,&Clark,Herbert.1979. Whennounssurfaceasverbs. Language,767–811. Comrie,Bernard.1976. Aspect. Cambridge:CambridgeUniversityPress. Conti-Jiménez,Carmen.2004. Construccionespasivasconestar. ELUA,18,21–44. Corver,Norbert.1997. TheinternalsyntaxoftheDutchextendedadjectivalprojection. NaturalLan- guageandLinguisticTheory,289–368. Croft,William.1991. SyntacticCategoriesandGrammaticalRelations: TheCognitiveOrganizationofInfor- mation. Chicago,IL:UniversityofChicagoPress. Cuervo,MaríaCristina.2003. DativesatLarge. Ph.D.thesis,MIT. Davidson, Donald. 1967. The logical form of action sentences. In: The Logic of Decision and Action. Pittsburgh:UniversityofPittsburghPress,81-95. [ReprintedinDavidson,Donald.2006.TheEssen- tialDavidson.Oxford:OxfordUniversityPress,37-71]. de Miguel, Elena. 1999. El aspecto léxico. In: Bosque, Ignacio, & Demonte, Violeta (eds), Gramática descriptivadelalenguaespañola,vol.2,ch.46. Madrid:EspasaCalpe,2977–3060. Dimitriadis, Alexis. 1999. On clitics, prepositions and case licensing in Standard and Macedonian Greek. In: Alexiadou, Artemis, Horrocks, Geoffrey C., & Stavrou, Melita (eds), Studies in Greek Syntax. Dordrecht:Kluwer,95–112. Doron, Edit. 2003. Agency and voice: The semantics of the Semitic templates. Natural Language Semantics,1–67. Doron,Edit.2013.Theinteractionofadjectivalpassiveandvoice.In:H.Borer,A.Alexiadou,&Schäfer, F.(eds),TheRootsofSyntaxandtheSyntaxofRoots. Oxford:OxfordUniversityPress,164–191. Dowty,DavidR.1979. WordMeaningandMontagueGrammar. Dordrecht:Reidel. Dowty,DavidR.1991. ThematicProto-RolesandArgumentSelection. Language,67(3),546–619. 254 Dubinsky, Stanley, & Simango, Silvester Ron. 1996. Passive and stative in Chichewa: Evidence for modulardistinctionsingrammar. Language,72(4),749–781. Embick,David.2004. OntheStructureofResultativeParticiplesinEnglish. Linguistic Inquiry,35: 3, 355–92. Ernst,Thomas.2016. Modificationofstativepredicates. Language,92(2),237–274. Espinal, María Teresa. 2010. Bare nominals in Catalan and Spanish. Their structure and meaning. Lingua,120,984–1009. Espinal, María Teresa, & McNally, Louise. 2011. Bare singular nominals and incorporating verbs in SpanishandCatalan. JournalofLinguistics,47,87–128. Fábregas,Antonio.2007. TheExhaustiveLexicalizationPrinciple. Nordlyd, Tromsø Working Papers in Language&Lingusitics,34(2),165–199. Fábregas,Antonio.2014. Unanálisissintácticodedostiposdecausante. Estudios de Lingü ística de la UniversidaddeAlicante,28,191–214. Fábregas,Antonio.2016. Lasnominalizaciones. Madrid:Visor. Fábregas, Antonio, & Marín, Rafael. 2012. Syntactic and semantic properties of Davidsonian state verbs. Talk presented at the 42 Linguistic Symposium on Romance Languages (Southern Utah University). Fábregas, Antonio, & Marín, Rafael. 2015a. Deriving individual-level and stage-level psych verbs in Spanish. TheLinguisticReview,32(2),227–275. Fábregas,Antonio,&Marín,Rafael.2015b(May21–23). Howinsertioninnonterminalnodesexplainstwo participlepuzzles. TalkgivenatCGG25. Fábregas,Antonio,&Marín,Rafael.2017.Onnon-dynamiceventiveverbsinSpanish. Linguistics,55(3), 451–488. Fernández de Castro, Félix. 1999. Las perífrasis verbales en el español actual. Biblioteca románica his- pánica2.Estudiosyensayos. Madrid:Gredos. Fernández-Leborans,MaríaJesús.1999.Lapredicación:lasoracionescopulativas. Chap.37of:Bosque, Ignacio,&Demonte,Violeta(eds),Gramáticadescriptivadelalenguaespañola,vol.2. Madrid:Espasa Calpe,2357-2460. 255 Fernández Ordóñez, Inés. 1999. Leísmo, laísmo y loísmo. Chap. 21 of: Bosque, Ignacio, & Demonte, Violeta(eds),Gramáticadescriptivadelalenguaespañola,vol.1. Madrid:EspasaCalpe,1317–1398. Fernández-Ramírez, Salvador. 1951. Gramática española. El verbo y la oración. Vol. 4. Madrid: Arco Libros. Fernández-Soriano,Olga,&Mendikoetxea,Amaya.2011.Non-selecteddativesubjectsinanticausative constructions. ArchivioGlottologicoItaliano,XCVI,87–128. Filip,Hana.1999.Aspect,EventualityTypesanNominalReference.NewYork:Routledge,Taylor&Francis Group(Garland). Fodor,Jerry.1970. Threereasonsfornotderiving‘kill’from‘causetodie’. LinguisticInquiry,429–38. Fodor,Jerry,&Lepore,Ernest.1999. Impossiblewords? LinguisticInquiry,445–453. Folli, Raffaella, & Harley, Heidi. 2005. Flavors of v: Consuming results in Italian and English. In: Slabakova,R.,&Kempchinsky,P.(eds),AspectualInquiries. Dordrecht:Kluwer,95-120. Folli, Raffaella, & Harley, Heidi. 2008. Teleology and animacy in external arguments. Lingua,118(2), 190–202. Folli, Raffaella, & Harley, Heidi. 2016. Against deficiency-based typologies: Manner-alternation pa- rametersinItalianandEnglish. In: Romance Languages and Linguistic Theory 10: Selected papers from GoingRomance28,Lisbon. Amsterdam:JohnBenjamis,103–120. Franco, Jon. 1990. Towards a typology of experiencer verbs: Evidence from Spanish. MIT Working PapersinLinguistics,12,46–62. Fukui, Naoki, & Speas, Margaret. 1986. Specifiers and projection. MIT Working Papers in Linguistics, 8(DepartmentofLinguisticsandPhilosophy,MIT),128–172. Fults, Scott. 2006. The Structure of Comparison: An Investigation of Gradable Adjectives. Ph.D. thesis, UniversityofMaryland. García-Pardo,Alfredo.2015. TheAspectualStructureoftheVP:ACaseStudyinAdjectivalPassives. Qual- ifyingPaper,USC. García-Pardo,Alfredo.2016a. Theaspectualcompositionofadjectivalpassives. In: RossBurkholder, CarlosCisneros,EmilyCoppessJulianGroveEmilyHaninkCherryMeyerHilaryHeadMcMahan Natalia Pavlou Özge Sarıgul Adam Singerman, & Zhang, Anqi (eds), Proceedings of the 50 th Annual MeetingoftheChicagoLinguisticsSociety(CLS50). Chicago,IL:UniversityofChicago. 256 García-Pardo, Alfredo. 2016b. Deriving aspect-alternating verbs without-roles. Pages 287–300 of: Hammerly, Christopher, & Prickett, Brandon (eds), NELS 46: Proceedings of the Forty-Sixth Annual Meeting of the North East Linguistic Society,vol.1. Amherst,MA:GraduateLinguisticsStudentsAs- sociation. García-Pardo,Alfredo.2017. Aspectandargumentstructureinadjectivalpassives. Borealis: An Inter- nationalJournalofHispanicLinguistics,6(1),21–52. García-Pardo,Alfredo.inprep.. Laestructuraargumentaldelaspasivasconseryestar. Revista de la SociedadEspañoladeLingüística. García-Pardo,Alfredo,&Menon,Mythili.submitted. Theaspectualstructureoftheadjective.Span- ish ser and estar. In: Selected Proceedings of the 20 th Hispanic Linguistics Sympsium. Somerville, MA: CascadillaProceedingsProject. Gehrke,Berit.2011. Stativepassivesandeventkinds. In:Reich,I.,Horch,E.,&Pauly,D.(eds),Proceed- ingsofSinnundBedeutung15. Saarbrücken,Universaar:SaarlandUniversityPress,241-257. Gehrke, Berit.2012. Passivestates. In: Demonte, V., &McNally, L.(eds), Telicity, Change and State. A Cross-CategorialViewofEventStructure. Oxford:OxfordUniversityPress,185-211. Gehrke,Berit.2013. Stillpuzzledbyadjectivalpassives? In: Folli,R.,Sevdali,C.,&Truswell,R.(eds), SyntaxanditsLimits. Oxford:OxfordUniversityPress,175-191. Gehrke,Berit.2015.Adjectivalparticiples,eventkindmodificationandpseudo-incorporation.Natural LanguageandLinguisticTheory,33,897–938. Gehrke,Berit,&Sánchez-Marco,Cristina.2014. Differentby-phraseswithadjectivalandverbalpas- sives:EvidencefromSpanishcorpusdata. Lingua,149. Gehrke, Berit, &Sánchez-Marco, Cristina.2015. Laspasivaspsicológicas. In: Marín, Rafael(ed), Los predicadospsicológicos. Madrid:Visor,145–184. Gese,Helga.2011. EventsinAdjectivalPassives. In:Reich,I.,Horch,E.,&Pauly,D.(eds),Proceedingsof SinnundBedeutung15. Saarbrücken,Germany:Universaar-SaarlandUniversityPress,259-73. GiliGaya,Samuel.1961. Cursosuperiordesintaxisespañola. Barcelona:Bibligraf. Giorgi, Alessandra, & Pianesi, Fabio. 1997. Tense and Aspect. From Semantics to Morphosyntax. New York/Oxford:OxfordUniversityPress. 257 Goldberg,AdeleE.2002. Surfacegeneralizations:Analternativetoalternations. CognitiveLinguistics, 13,327–56. Gómez-Torrego,Leonardo.1988. Perífrasisverbales. Madrid:ArcoLibros. GómezVázquez,Diana,&GarcíaFernández,Luis.2013.Estadoscontrolados.In:ValÁlvaro,JoséFran- cisco(ed), De la unidad del lenguaje a la diversidad de las lenguas: actas del X Congreso Internacional de LingüísticaGeneral. Zaragoza:UniversidaddeZaragoza,335–346. Greenberg,Yael.2006. StructuringAspectualandTemporalRelationswithTwoHebrewAdverbials and the Semantics/ Pragmatics of still. In: Gyuris, B., Kálmán, L., Piñón, C., & Varasdi, K. (eds), Proceedingsofthe9 th SymposiumonLogicandLanguage. Besenyőtelek(Hungary),62-9. Grimshaw,Jane.1990. ArgumentStructure. Cambridge,MA:MITPress. Gumiel-Molina,Silvia,Moreno-Quibén,Norberto,&Pérez-Jiménez,Isabel.2015.Comparisonclasses andtherelative/absolutedistinction: adegree-basedcompositionalaccountoftheser/estaralter- nation in Spanish. Pages 955–1001 of: Gehrke, Berit, & Castroviejo, Elena (eds), Degree and Manner Modification Across Categories, vol. 33. Natural Language and Linguistic Theory Special Topic/ Com- mentIssue. Hale,KennethL.,&Keyser,SamuelJ.1993. Onargumentstructureandlexicalexpressionofsyntactic relations. In: Hale,K.,&Keyser,S.J.(eds), The View from Building 20. Cambridge,MA:MITPress, 53-109. Hale, Kenneth L., & Keyser, Samuel J. 1997. Adjectives, other stative predicates and the roots of stativity. Ms.,MassachussettsInstituteofTechnology. Hale,KennethL.,&Keyser,SamuelJ.2002. ProlegomenontoaTheoryofArgumentStructure.Cambridge, MA:MITPress. Hale,KennethL.,&Keyser,SamuelJ.2005.Aspectandthesyntaxofargumentstructure.In:Erteschik- Shir,Nomi,&Rapoport,Tova(eds),TheSyntaxofAspect. Oxford:OxfordUniversityPress,11-41. Halle, Morris, & Marantz, Alec. 1993. Distributed Morphology and the Pieces of Inflection. In: The ViewfromBuilding20. Cambridge,MA:MITPress,111-176. Hallman,Peter.2013. Predicationandmovementinpassive. Lingua,125,76–94. Hallman,Peter.2015. TheTemporalNatureofState/EventDistinction. In: GLOWWorkshoponStates andEvents. Paris(France). April18. 258 Harley,Heidi.1995. Subjects,EventsandLicensing. Ph.D.thesis,MIT. Harley,Heidi.2013. ExternalargumentsandtheMirrorPrinciple:OnthedistinctnessofVoiceandv. Lingua,125,34–57. Harley, Heidi. 2017. The "bundling" hypothesis and the disparate functions of little v. In: Ro- bertaD’Alessandro,IreneFranco,&ÁngelGallego(eds), The Verbal Domain. Oxford: OxfordUni- versityPress,3–28. Harley,Heidi,Tubino-Blanco,Mercedes,&Haugen,Jason.2009. Applicativeconstructionsandsup- pletive verbs in Hiaki (Yaqui). In: Linda Lanz, Amy Franklin, Jennifer Hoecker Elizabeth Gentry BrunnerMichelleMorrison,&Pace,Cassandra(eds),RiceWorkingPapersinLinguistics,vol.1. Hous- ton,TX:RiceLinguisticsSociety,RiceUniversity,42–51. Harwood,William.2013. Being progressive is just a phase: dividing the functional hierarchy. Ph.D.thesis, GhentUniversity. Harwood,William.2015. Beingprogressiveisjustaphase: celebratingtheuniquenessofprogressive aspectunderaphase-basedanalysis. NaturalLanguageandLinguisticTheory,33,523–573. Harwood,William.2017. Reducedrelativesandextendedphases:Aphase-basedanalysisoftheinflec- tionalrestrictionsonEnglishreducedrelativeclauses. StudiaLinguistica,1–44. Hay,Jennifer,Kennedy,Christopher,&Levin,Beth.1999. Scalestructureunderliestelicityin"degree achievements". In:Matthews,T.,&Strolovitch,D.(eds),SemanticsandLinguistictheory9.Ithaca,NY: CLCPublications,127-44. Hengeveld,Kees.1986. CopularverbsinafunctionalgrammarofSpanish. Linguistics,24,393–420. Higginbotham,JamesT.1997. Locationandcausation. UniversityofOxford. Hoekstra, Teun. 1992. Aspect and Theta Theory. In: Roca, I. M. (ed), Thematic Structure: Its Role in Grammar. Berlin:Foris,145-174. Horvath,Julia,&Siloni,Tal.2008. Activelexicon: Adjectivalandverbalpassives. In: Armon-Lotem, Sharon,Rothsten,Susan,&Danon,Gabi(eds),GenerativeApproachestoHebrew Linguistics. Amster- dam:JohnBenjamins,105-136. Husband,Matthew.2012. OntheCompositionalNatureofStates. Amsterdam:JohnBenjamins. 259 Iatridou,Sabine,Anagnostopoulou,Elena,&Izvorsky,Roumyana.2001. Observationsabouttheform andmeaningoftheperfect. In: KenHale: ALifeinLanguage. Cambridge,MA:MITPress,189-238. Ippolito,Michaela.2004.Ananalysisofstill. In:Young,R.B.(ed),ProceedingsofSemanticsandLinguistic Theory14. CornellUniversity,Ithaca,NY:CLCPublications,127-44. Jaque,Matías.2013. La expresión de la estatividad en español: niveles de representación y grados de dinami- cidad. Ph.D.thesis,UniversidadAutónomadeMadrid. Kamp, Hans, & Reyle, Uwe. 1993. From Discourse to Logic: Introduction to Modeltheoretic Semantics of NaturalLanguage,FormalLogicandDiscourseRepresentationTheory. Dordrecht:Kluwer. Kamp,Hans,Reyle,Uwe,&Rossdeutscher,Antje.2015.PerfectsasFeatureShiftingOperators.Universität Stuttgart. Kayne,Richard.1975. Frenchsyntax: TheTransformationalcycle. Cambridge,MA:MITPress. Kearns,Kate.2000. Semantics. NewYork:St.Martin’s. Kennedy,Christopher.1997.Projectingtheadjective:thesyntaxandsemanticsofgradabilityandcomparison. Ph.D.thesis,UniversityofCalifornia,SantaCruz. Kenny,Anthony.1963. Action,EmotionandWill. London:Routledge. Kiparsky, Paul. 1973. Elsewhere in Phonology. In: Anderson, Stephen R., & Kiparsky, Paul (eds), A FestschriftforMorrisHalle. NewYork:Holt,RinehartandWinston,93–106. Koenig,Jean-Pierre,&Davis,AnthonyR.2006. TheKEYtolexicalsemanticrepresentations. Journal ofLinguistics,42,71–108. Koopman,Hilda,&Sportiche,Dominique.1985. ThetaTheoryandExtraction. Koopman,Hilda,&Sportiche,Dominique.1988. Subjects. Ms.,UniversityofCaliforniaatLosAngeles. Koopman,Hilda,&Sportiche,Dominique.1991. Thepositionofsubjects. Lingua,211–258. Kramer,Ruth.2014. Cliticdoublingorobjectagreement: theviewfromAmharic. Natural Language andLinguisticTheory,32,593–634. Krapova, Iliyana, & Cinque, Guglielmo. 2008. Clitic reduplication constructions in Bulgarian. In: Kalluli, Dalina, & Tasmowski, Liliane (eds), Clitic Doubling in the Balkan Languages. Amsterdam: JohnBenjamins,257–286. 260 Kratzer, Angelika.1995. Stage-levelandIndividual-levelpredicates. In: Carlson, GregoryN., &Pel- letier,FrancisJeffry(eds),TheGenericBook. Chicago:ChicagoUniversityPress. Kratzer,Angelika.1996. Severingtheexternalargumentfromtheverb. In:Rooryck,Johan,&Zaring, Laurie(eds),PhrasestructureandtheLexicon. Dordrecht:Kluwer,109-37. Kratzer,Angelika.2000. Buildingstatives. In: Conathan,LisaJ.,Good,Jeff,Kavitskaya,Darya,Wulf, AlyssaB.,&Yu,AlanC.L.(eds),Proceedingsofthe26 th MeetingoftheBerkeleyLinguisticsSociety.Berke- ley,CA:BerkeleyLinguisticsSociety,385-99. Kratzer,Angelika.2002. TheEventArgument. Ms.,UniversityofMassachusetts. Kratzer,Angelika.2005. Buildingresultatives. In: Event Arguments in Syntax, Semantics and Discourse. Tübingen:Niemeyer,178-212. Krifka,Manfred.1989. Nominalreference,temporalconstitutionandquantificationineventseman- tics.In:Bartsch,Renate,vanBenthem,Johan,&vanEmdeBoas,Peter(eds),SemanticsandContextual Expression. Stanford,CA:CSLIPublications,75–115. Krifka,Manfred.1992. ThematicRelationsasLinksbetweenNominalReferenceandTemporalCon- stitution. In: Sag, Ivan A., & Szablocsi, Anna (eds), Lexical Matters. Stanford, CA: Center for the StudyofLanguageandInformation,29-53. Krifka, Manfred. 2000. Alternatives for aspectual particles: Semantics of still and already. Paper presentedattheBerkeleyLinguisticsSocietyMeeting. Kuroda, Sige-Yuki. 1988. Whether we agree or not: A comparative syntax of English and Japanese. LingvisticaeInvestigationes,1–47. Laka,Itziar.1990. NegationinSyntax: OntheNatureofFunctionalCategoriesandProjections.Ph.D.thesis, MIT. Lakoff,George.1966.StativeadjectivesandverbsinEnglish.In:Oettinger,A.G.(ed),MathematicalLin- guistics and Automatic Translation. TheComputationLaboratory, HarvardUniversity, Cambridge, MA,I-1-I-16. Lakoff,George.1970. IrregularityinSyntax. NewYork:Holt,RinehartandWilson. Landau,Idan.2010. TheLocativeSyntaxofExperiencers. LinguisticInquiryMonograph53. Cambridge, MA:MITPress. 261 Larsson,Ida.2009. Participles in time: the development of the Perfect tense in Swedish. Ph.D.thesis,Göte- borgUniversity. Larsson,Ida,&Svenonius,Peter.2013. Phases and categories in passive and perfect participles. Talkpre- sentedattheComparativeGermanicSyntaxWorkshop28attheUniversityofLeipzig,October4. Leech,GeoffreyN.1970. Towards a Semantic Description of English. Bloomington: IndianaUniversity Press. Legendre,Géraldine.1989. InversionwithcertainFrenchexperiencerverbs. Language,65(4),752–782. Legendre,Géraldine.1993. AntipassivewithFrenchpsych-verbs. In: Dunca,Erin,Farkas,Donka,& Spaelti,Philip(eds),ProceedingsofWCCFL12. Stanford,CA:CSLIPublications,373–388. Legendre,Géraldine,&Akimova,Tanya.1993. InversionandantipassiveinRussian. Pages 286–318 of: Avrutin, Sergey, Franks, Steven, & Progovac, Ljiljana (eds), The Second Annual Workshop on Formal ApproachestoSlavicLinguistics. Levin, Beth.1993. English Verb Classes and Alternations: A Preliminary Investigation. Chicago, IL:Uni- versityofChicago,IL:UniversityofChicagoPress. Levin,Beth,&Rappaport,Malka.1986. TheFormationofAdjectivalPassives. LinguisticInquiry,17:4, 623–61. Levin,Beth,&Rappaport-Hovav,Malka.1995. Unaccusativity: AttheSyntax-LexicalSemanticsInterface. LinguisticInquiryMonograph,26,Cambridge,MA:MITPress. Levin,Beth,&Rappaport-Hovav,Malka.2005. ArgumentRealization. Cambridge:CambridgeUniver- sityPress. Lewis,David.1973. Causation. TheJournalofPhilosophy,70(17),556–567. Lieber,Rochelle.1980. OntheOrganizationoftheLexicon. Ph.D.thesis,MIT. Lin,Jimmy.2004. EventStructureandtheEncodingofArguments: TheSyntaxoftheMandarinandEnglish VerbPhrase. Ph.D.thesis,MassachussettsInstituteofTechnology. Löbner,Sebastian.1989. Germanschön - erst - noch: Anintegratedanalysis. Linguistics and Philosophy, 12,167–212. Lohndal,Terje.2014. PhraseStructureandArgumentStructure.ACaseStudyoftheSyntax-SemanticInter- face. Oxford:OxfordUniversityPress. 262 Maienborn,Claudia.2005. OnthelimitsoftheDavidsonianapproach: Thecaseofcopulasentences. TheoreticalLinguistics,31,275–316. Maienborn,Claudia.2007. OnDavidsonianandKimianstates. In: Comorovski,I.,&vonHeusinger, K.(eds),Existence: SemanticsandSyntax. Dordrecht:Springer,107-30. Maienborn, Claudia. 2009. Building ad hoc properties: On the interpretation of adjectival passives. In:Riester,A.,&Solstad,T.(eds),ProceedingsofSinnundBedeutung13. UniversityofStuttgart,31-46. Marantz,Alec.1984. OntheNatureofGrammaticalRelations. Cambridge,MA:MITPress. Marantz, Alec. 1997. No Escape From Syntax: Don’t Try Morphological Analysis in the Privacy of YourOwnLexicon. In: UniversityofPennsylvaniaWorkingPapersinLinguistics,200-225,vol.4. Marantz,Alec.1999.Creatingwordsaboveandbelowlittlev.Ms.,MassachussettsInstituteofTechnology. Marantz,Alec.2013. Verbalargumentstructure:Eventsandparticipants. Lingua,152–168. Marín,Rafael.2000. El componente aspectual de la predicación. Ph.D.thesis,UniversidadAutónomade Barcelona. Marín,Rafael.2004. Entreseryestar. Madrid:ArcoLibros. Marín,Rafael.2011.Casitodoslospredicadospsicológicossonestativos. In:ÁngelesCarrascoGutiér- rez(ed),Sobreestadosyestatividad. Munich:LincomStudiesinTheoreticalLinguistics,26–44. Marín, Rafael. 2015. Los predicados psicológicos: debate sobre el estado de la cuestión. In: Marín, Rafael(ed),Lospredicadospsicológicos. Madrid:Visor,11-50. Marín,Rafael,&McNally,Louise.2005.TheAktionsartofSpanishreflexivepsychologicalverbs.Pages 212–225of: PRoceedingsofSinnundBedeutung9. Marín,Rafael,&McNally,Louise.2011. Inchoativity,changeofstateandtelicity:evidencefromSpan- ishreflexivepsychologicalverbs. NaturalLanguageandLinguisticTheory,29:2,467–502. Marín, Rafael,&Sánchez-Marco, Cristina.2012. Verbosynombrespsicológicos: juntosyrevueltos. Borealis: AnInternationalJournalofHispanicLinguistics,2,91–108. Martin,Fabienne.2006. Prédicatsstatifs,causatifsetrésultatifsendiscours. Ph.D.thesis,UniversitéLibre deBruxelles. 263 Martin,Fabienne,&Schäfer,Florian.2014. Causationatthesyntax-semanticsinterface. In: Copley, Bridget,&Martin,Fabienne(eds),Causation in Grammatical Structures. Oxford: OxfordUniversity Press,209-44. Massam,Diane.2001. PseudonounincorporationinNiuean. NaturalLanguageandLinguisticTheory, 19,153–197. Mateu, Jaume. 2002. Argument Structure. Relational Construal at the Syntax-Semantics Interface. Ph.D. thesis,UniversitatAutònomadeBarcelona. McCawley,JamesD.1971. TenseandtimereferenceinEnglish. In:Fillmore,C.J.,&Langendoen,D.T. (eds),StudiesinLinguisticSemantics. Holt,RinehartandWinston,97–113. McCawley,JamesD.1973.Pre-lexicalsyntax.In:GrammarandMeaning:PapersonSyntacticandSemantic Topics. Tokyo:TaishukanPublishingCompany,343-356. McCawley,JamesD.1981. NotesontheEnglishperfect. AustralianJournalofLinguistics,1,81–90. McDonald,Jonathan.2008. TheSyntacticNatureofInnerAspect. Amsterdam:JohnBenjamins. Mchombo,Sam.1993. AformalanalysisofthestativeconstructioninBantu. Journal of African Lan- guagesandLinguistics,14,5–28. McIntyre, Andrew. 2013. Adjectival passives and adjectival participles in English. In: Alexiadou, Artemis,&Schäfer,Florian(eds),Non-CanonicalPassives. Amsterdam:JohnBenjamins,21-42. McIntyre,Andrew.2015. Eventmodifiersin(German)adjectivalparticiples:RemarksonGehrke(this issue). NaturalNaturalLanguageandLinguisticTheory,33,939–953. Meltzer, Aya. 2005. Adjectival passives and adjectival decausatives in Hebrew. M.Phil. thesis, Tel Aviv University. Meltzer-Asscher,Aya.2011. AdjectivalpassivesinHebrew: evidenceforparallelismbetweenthead- jectivalandverbalsystems. NaturalLanguageandLinguisticTheory,29,815–855. Michaelis,LauraA.1993. "Continuity"withinThreeScalarModels: ThePolysemyofAdverbial Still. JournalofSemantics,10,193–237. Michaelis,LauraA.1994.TheambiguityoftheEnglishpresentperfect.JournalofLinguistics,30(111–157). Mithun,Marianne.1984. Theevolutionofnounincorporation. Language,847–894. 264 Mittwoch,Anita.1988. AspectsofEnglishaspect:ontheinteractionofperfect,progressiveanddura- tionalphrases. LinguisticsandPhilosophy,11,203–254. Mittwoch,Anita.1991. IndefenseofVendler’sachievements. BelgianJournalorLinguistics,6(71-84). Moens, Marc. 1987. Tense, Aspect and Temporal Reference. Ph.D. thesis, Center for Cognitive Science, UniversityofEdinburgh. Moens,Marc,&Steedman,Mark.1988. TemporalOntologyandTemporalReference. Computational Linguistics,14:2,15–28. Morgan,Jerry.1969. Onarguingaboutsemantics. PapersinLinguistics,49–70. Mourelatos,AlexanderP.D.1978. Events,statesandprocesses. LinguisticsandPhilosophy,2,415–434. Navas-Ruiz,Ricardo.1987. Usosde‘ser’y‘estar’. Madrid:SociedadEspañoladeLibrería. Neeleman,Ad,&vandeKoot,Hans.2012. TheLinguisticExpressionofCausation. In: Everaert,M., Marelj,M.,&Siloni,T.(eds), The Theta System: Argument Structure at the Interface. Oxford: Oxford UniversityPress,20-51. Nishiyama,Atsuko,&Koenig,Jean-Pierre.2010. Whatisaperfectstate? Language,86,611–46. Olsen, Mari Broman. 1994. The semantics and pragmatics of lexical aspect features. Studies in the LinguisticsSciences,361–375. Olsen, Mari Broman. 1997. A Semantic and Pragmatic Model of Lexical and Grammatical Aspect. New York:Garland. Pancheva, Roumyana. 2003. The aspectual makeup of the Perfect participle. In: Perfect Explorations. Berlin:MoutondeGruyter,277-306. Parsons,Terence.1990. Events in the Semantics of English. A Study in Subatomic Semantics. Cambridge, MA:MITPress. Paslawska,Alla,&vonStechow,Arnim.2003. PerfectreadingsinRussian. In: Alexiadou,A.,Rathert, M.,&vonStechow,A.(eds),Perfectexplorations. Berlin:MoutondeGruyter,307–62. Pereltsvaig,Asya.1997. Thegenitiveofnegation. Pages167–190of: ProceedingsoftheFifthMeetingofthe IsraeliAssociationofTheoreticalLinguistics. 265 Perlmutter, David M. 1978. Impersonal Passives and the Unaccusative Hypothesis. Pages 157–190 of: Proceedingsofthe4thAnnualMeetingoftheBerkeleyLinguisticsSociety. Pesetsky,David.1995. ZeroSyntax: ExperiencersandCascades. Cambridge,MA:MITPress. Picallo,M.Carme.1991. NominalsandNominalizationsinCatalan. Probus,3(279-316). Portner,Paul.2003. The(temporal)semanticsand(modal)pragmaticsoftheperfect. Linguistics and Philosophy,26,459–510. Pustejovsky,James.1991. Thesyntaxofeventstructure. Cognition,41,47–81. Pylkkänen, Liina. 2000. On Stativity and Causation. In: Tenny, C., & Pustejovsky, J. (eds), Events as GrammaticalObjects: TheConvergingPerspectivesofLexicalSemanticsandSyntax. CSLIPublications, 417-445. Pylkkänen,Liina.2002. IntroducingArguments. Ph.D.thesis,MIT. Ramchand, Gillian. 2014. Causal chains and instrumental case in Hindi/Urdu. In: Copley, Bridget, & Martin, Fabienne (eds), Causation in Grammatical Structures. Oxford: Oxford University Press, 245–278. Ramchand,Gillian,&Svenonius,Peter.2014. Derivingthefunctionalhierarchy. LanguageSciences,46, 152–174. Ramchand,GillianC.2005. Post-Davidsonianism. TheoreticalLinguistics,31(3),359–73. Ramchand,GillianC.2008. VerbMeaningandtheLexicon: AFirstPhaseSyntax.Cambridge:Cambridge UniversityPress. Ramchand, Gillian C. 2011. Minimalist Semantics. In: Boeckx, Cedric (ed), Oxford Handbook of Lin- guisticMinimalism. Oxford:OxfordUniversityPress,449-471. Rapp,Irene.1996. Zustand? Passiv? Überlegungenzumsogenannten"Zustandspassiv". Zeitschrift für Sprachwissenschaft,15(2),231–265. Rapp,Irene.1997. PartizipenundsemantischeStruktur.ZupassivischenKonstruktionendes3.Status.Tübin- gen,StauffenburgVerlag. Rapp, Irene. 2011. The attributive past participle: Structure and temporal interpretation. In: Féry, Caroline,&Sternefeld,Wolfgang(eds), Audiatur Vox Sapientiae: A Festschrift for Arnim von Stechow. Berlin:Akademie-Verlag,392–409. 266 Rappaport-Hovav,Malka,&Levin,Beth.1998. Buildingverbmeanings. In: Butt,Miriam,&Geuder, Wilhelm (eds), The Projection of Arguments: Lexical and Syntactic Constraints. Stanford, CA: CSLI Publications,97–134. Rappaport-Hovav,Malka,&Levin,Beth.2010. ReflectionsonManner/ResultComplementarity. In: Doron, E., Rappaport-Hovav, M., & Sichel, I. (eds), Syntax, Lexical Semantics and Event Structure. Oxford:OxfordUniversityPress,21-38. Reinhart, Tanya. 1997. Syntactic effects of lexical operations: Reflexives and unaccusatives. In: OTS WorkingPapersinLinguistics. UniversityofUtrecht. Reinhart, Tanya. 2000. The Theta System: Syntactic realization of verbal concepts. In: UiL-OTS WorkingPapersinLinguistics. UniversityofUtrecht. Reinhart,Tanya.2001. Experiencingdeviations. In:Hastings,Rachel,Jackson,Brendan,&Zvolensky, Zsofia(eds),ProceedingsofSALT11. Ithaca,NY:CLCPublications,127-44. Reinhart,Tanya.2002. Thethetasystem–anoverview. TheoreticalLinguistics,28,229–90. Reinhart, Tanya, & Siloni, Tal. 2004. Against the unaccusative analysis of reflexives. In: Alexiadou, Artemis,Anagnostopoulou,Elena,&Everaert,Martin(eds), The Unaccusativity Puzzle: Explorations oftheSyntax-LexiconInterface. Oxford:OxfordUniversityPress,288–331. Rizzi, Luigi. 1997. The Fine Structure of the Left Periphery. In: Haegeman, Lilliane (ed), Elements of Grammar. Dordrecht:Kluwer,281–337. Roberts,Ian.1991. NP-movement,crossoverandchainformation. In:Haider,Hubert,&Klaus,Netter (eds),RepresentationandDerivationintheTheoryofGrammar. Dordrecht:KluwerAcademicPublish- ers,17–52. Ross,JohnR.1967. Constraintsonvariablesinsyntax. Ph.D.thesis,MIT. Ross,JohnR.1972. Doubling. LinguisticInquiry,3,61–86. Rothmayr,Antonia.2009. TheStructureofStativeVerbs. Amsterdam:JohnBenjamins. Rothstein,Susan.2004. StructuringEvents: AStudyintheSemanticsofLexicalAspect. Blackwell:Oxford. Rothstein, Susan. 2012. Another look at accomplishments and incrementality. Pages 60–102 of: De- monte,Violeta,&McNally,Louise(eds), Telicity, Change and State: A Cross-Categorial View of Event Structure. Oxford:OxfordUniversityPress,60-102. 267 Roy,Isabelle.2013. Non-Verbal Predication: Copular Sentences at the Syntax-Semantics Interface. Oxford: OxfordUniversityPress. Ryle,Gilbert.1949. TheConceptofMind. London:BarnesandNoble. SánchezLópez, Cristina.1999. Lanegación. Chap. 40 of: Bosque, Ignacio, &Demonte, Violeta(eds), Gramáticadescriptivadelalenguaespañola,vol.2. Madrid:EspasaCalpe,2561–2634. Schäfer,Florian.2008. TheSyntaxof(Anti-)Causatives: ExternalArgumentsinChange-of-StateContexts. Amsterdam:JohnBenjamins. Schäfer,Florian.2009. Thecausativealternation. LanguageandLinguisticsCompass,3(2),641–81. Schlücker,Barbara.2005. Event-relatedmodifiersinGermanadjectivalpassives. In: Maier,E.,Bary, C.,&Huitink,J.(eds),ProceedingsofSinnundBedeutung9. RadboudUniversityNijmegen,417-30. Schwarzschild,Roger.2002. TheGrammarofMeasurement. In: Jackson,B.(ed), Semantics and Lin- guisticTheory12. Ithaca,NY:CLCPublications,225-45. Schwenter.1994. Thegrammaticalizationofanteriorinprogress:evidencefromaPeninsularSpanish dialect. StudiesinLanguage,18(1),71–111. Shibatani, Masayoshi. 1976. The grammar of causative constructions: a conspectus. In: Shibatani, Masayoshi (ed), Syntax and Semantics 6: The Grammar of Causative Constructions. New York: Aca- demicPress,1–40. Simpson, Jane. 1983. Resultatives. In: Levin, Lori, Hovav, Malka Rappaport, & Zaenen, Annie Else (eds),PapersinLexical-FunctionalGrammar. Bloomington,IN:IndianaUniversityLinguisticsClub. Smith,Carlota.1983. Atheoryofaspectualchoice. Language,59,479–501. Smith,Carlota.1991. TheParameterofAspect. Dordrecht:KluwerAcademicPublishers. Smith,CarlotaS.1961. AclassofcomplexmodifiersinEnglish. Language,37,342–365. Sorace,Antonella.2004.Gradienceatthelexicon–syntaxinterface:Evidencefromauxiliaryselection andimplicationsforunaccusativity. In: Alexiadou,Artemis,Anagnostopoulou,Elena,&Everaert, Martin(eds),The Unaccusativity Puzzle: Explorations of the Syntax-Lexicon Interface. Oxford: Oxford UniversityPress,243-268. Spathas,Giorgos,Alexiadou,Artemis,&Schäfer,Florian.2015. MiddleVoiceandReflexiveInterpre- tations:Afto-PrefixationinGreek. NaturalLanguageandLinguisticTheory,33(4),1293–1350. 268 Starke, Michal. 2009. Nanosyntax - A short primer to a new approach to language. In: Svenonius, Peter,Ramchand,Gillian,Starke,Michal,&Taraldsen,KnutTarald(eds), Nordlyd,vol.36. CASTL, Tromsø,1-6. Stowell,Tim.1981. OriginsofPhraseStructure. Ph.D.thesis,MIT. Svenonius,Peter.2006. Theemergenceofaxialparts. Nordlyd, Tromsø Working Papers in Language & Lingusitics: SpecialIssueonAdpositions,33.1,49–77. Svenonius,Peter.2008. ProjectionsofP. In:Asbury,Anna,Dotlacil,Jakub,Gehrke,Berit,&Nouwen, Rick(eds),SyntaxandSemanticsofSpatialP. Amsterdam:JohnBenjamins,63–84. Svenonius,Peter.2010. SpatialPinEnglish. In: Cinque,Guglielmo,&Rizzi,Luigi(eds), Cartography ofSyntacticStructures. Oxford:OxfordUniversityPress,127–160. Svenonius,Peter.2012a. Spanning. Ms.,UniversityofTromsø. Svenonius,Peter.2012b. StructuraldecompositionofspatialadpositionsinNovember24. TalkgivenatThe MeaningofPworkshopatRuhrUniversitätBochum. Svenonius, Peter. 2016. Spans and words. In: Siddiqi, Daniel, & Harley, Heidi (eds), Morphological Metatheory. Amsterdam:JohnBenjamins,201–222. Talmy, Leonard. 1975. Figure and Ground in complex sentences. In: Proceedings of the First Annual MeetingoftheBerkeleyLinguisticsSociety,419–430. Talmy,Leonard.1976. SemanticCausativeTypes. In: Shibatani,Masayoshi(ed),Syntax and Semantics 6: TheGrammarofCausativeConstructions. NewYork:AcademicPress,43-116. Talmy,Leonard.1985. Lexicalizationpatterns: Semanticstructureinlexicalforms. In: Shopen,Tim- othy (ed), Language Typology and Syntactic Description, I: Clause Structure. Cambridge: Cambridge UniversityPress,57-149. Taylor,Barry.1977. TenseandContinuity. LinguisticsandPhilosophy,1,199–220. Tenny,Carol.1987. GrammaticalizingAspectandAffectedness. Ph.D.thesis,MIT. Tenny,Carol.1992.Theaspectualinterfacehypothesis. In:Sag,IvanA.,&Szablocsi,Anna(eds),Lexical Matters. CSLIPublications,1-28. Tenny,Carol.1994. AspectualRolesandtheSyntax-SemanticsInterface. Dordrecht:Kluwer. 269 Tikagawa,Yoshihisa.1986. Subjects in Japanese and English. Ph.D.thesis,UniversityofMassachusetts, Amherst. Travis,Lisa.2000. EventStructureinSyntax. In: Tenny,Carol,&Pustejovsky,James(eds), Events as GrammaticalObjects: TheConvergingPerspectivesofLexicalSemanticsandSyntax. Stanford,CA:CSLI Publications,145–185. Travis,Lisa.2010. InnerAspect.TheArticulationoftheVP. Dordrecht:Springer. vanHout,Angeliek.2004. UnaccusativityasTelicityChecking. In: Alexiadou,A.,Anagnostopoulou, E.,&Everaert,M.(eds),TheUnaccusativityPuzzle: ExplorationsoftheSyntax-LexiconInterface.Oxford UniversityPress,60-83. Van-Valin,RobertD.Jr.1990. Semanticparametersofsplittransitivity. Language,66,221–260. Van-Valin,RobertD.Jr.,&LaPolla,Randy.1997. Syntax: Structure, Meaning and Function. Cambridge: CambridgeUniversityPress. Vanhoe, Henk. 2004. Aspectos de la sintaxis de los verbos psicol´ógicos en español. Frankfurt am Main: PeterLang. Vendler,Zeno.1957. VerbsandTimes. ThePhilosophicalReview,66(2),143–60. Verkuyl,HenkJ.1972. OntheCompositionalNatureoftheAspects. Dordrecht:Reidel. Verkuyl, Henk J. 1989. Aspectual classes and aspectual composition. Linguistics and Philosophy, 12(1), 39–94. Vivanco, Margot.2015. Causatividad y cambio de estado en español: la alternancia causativo-inacusativa. Ph.D.thesis,UniversidadComplutensedeMadrid. Vlach, Frank. 1981. The semantics of the progressive. In: Tedeschi, Philip J., & Zaenen, Annie (eds), TenseandAspect. SyntaxandSemantics14. NewYork:AcademicPress,271-292. von Stechow, Arnim. 1995. Lexical decomposition in syntax. In: Egli, Urs, Pause, Peter, Schwarze, Christoph,vonStechov,Arnim,&Wienold,Götz(eds),LexicalKnowledgeintheOrganizationofLan- guage. Amsterdam:JohnBenjamins,81-177. vonStechow,Arnim.1996. TheDifferentReadingsofWieder‘Again’: AStructuralAccount. Journalof Semantics,13,87–138. 270 vonStechow,Arnim.1998. GermanparticiplesIIinDistributedMorphology. UniversityofTübingen. Wasow,Thomas.1977. Transformationsandthelexicon. In: Culicover,P.W.,T.Wasow,&Akmajian, A.(eds),Formalsyntax. NewYork:AcademicPress,327-60. Wechsler,Stephen.2005. Resultativesunderthe‘event-argumenthomo-morphism’modeloftelicity. In:Erteschik-Shir,N.,&Rappoport,T.(eds),TheSyntaxofAspect. Oxford:OxfordUniversityPress, 255-273. Williams,Edwin.1981. ArgumentStructureandMorphology. TheLinguisticReview,1,81–114. Wurmbrand,Susi.2013.QRandselection:covertevidenceforphasehood. Pages277–290of:Proceedings ofNELS42. Amherst:UniversityofMassachusetts,GLSA. Wurmbrand,Susi.2014. TenseandaspectinEnglishinfinitives. LinguisticInquiry45,403–447. Xiqués, TeresaMaria.2012. Re-examiningthehodiernalpresentperfect: towardsaunifiedaccount. Pages 291–313 of: Boone, Enrico, Kohlberger, Marin, & Schulpen, Maartje (eds), Proceedings of Con- SOLEXX. Zubizarreta,MaríaLuisa,&Oh,Eunjeong.2007. On the Syntactic Composition of Manner and Motion. Cambridge,MA:MITPress. 271
Abstract (if available)
Abstract
This dissertation studies stative predicates within a broader, cross-categorial typology of events. I adscribe to the neo-constructionist program, which contends that verbal syntax determines event structure and the thematic interpretation of arguments, and that it is also the sole computational module of derivational morphology. In the empirical side, my research advances a comprehensive taxonomy of event types with a focus on states. In the theoretical side, I provide novel support to the neo-constructionist program in the event domain, and show how my proposed taxonomy can be syntactically derived. The main language under discussion is Castilian Spanish, but I also provide cross-linguistic evidence from a variety of other languages to further support my proposals. ❧ Chapter 1 serves as an introduction to my dissertation. In Chapter 2, I present an state of the art of the study of verbal syntax that led to the neo-constructionist program, and I introduce the theoretical framework that I use for the remainder of my dissertation: first-phase syntax. In Chapter 3, I analyze a set of Spanish verbs, the gobernar ‘govern’-type and argue that they are derived by a bi-phrasal structure that is unambiguously interpreted as a stative causative eventuality, i.e. two states related causally. Chapter 4 explores adjectival passives as a case study of derived statives. I show how these constructions are truly stative—and not perfective or resultative, as is often argued. Finally, Chapter 5 argues against the existence of a dedicated Experiencer thematic role for stative object-experiencer psychological verbs (e.g. worry, amaze...), on the basis that the grammatical tests adduced in favor of a special Experiencer structure can also be found in stative locative verbs (e.g. surround, cover...) cross-linguistically. I argue that they share a uniform, abstract relational structure, which is understood as locative or psychological depending on the type of verb that lexicalizes it. Chapter 6 presents the pertinent conclusions.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Perspective in Turkish complementation
PDF
Silence in answers: a study of ellipsis in Hindi
PDF
Processing the dynamicity of events in language
PDF
Building phrase structure from items and contexts
PDF
Syntax-prosody interactions in the clausal domain: head movement and coalescence
PDF
Exploring the effects of Korean subject marking and action verbs’ repetition frequency: how they influence the discourse and the memory representations of entities and events
PDF
Decomposing Slavic aspect: the role of aspectual morphology in Polish and other Slavic languages
PDF
A reduplicative analysis of sentence modal adverbs in Spanish
PDF
Copy theory of movement and PF conditions on spell-out
PDF
Where number lies: plural marking, numerals, and the collective-distributive distinction
PDF
Comparative iIlusions at the syntax-semantics interface
PDF
Towards the unity of movement: implications from verb movement in Cantonese
PDF
Subjectivity, commitments and degrees: on Mandarin hen
PDF
The theory of empty noun in Chinese: with special reference to the right node raising construction
PDF
When things are left unsaid: existential and anaphoric implicit objects in discourse
PDF
The grammar of individuation, number and measurement
PDF
The vanishing dead body and the rising lyric persona in early modern east Slavic poetry
PDF
The regulation, roles, and mechanism of action of mitochondrial-derived-peptides (MDPs) in aging
Asset Metadata
Creator
García-Pardo, Alfredo
(author)
Core Title
The morphosyntax of states: deriving aspect and event roles from argument structure
School
College of Letters, Arts and Sciences
Degree
Doctor of Philosophy
Degree Program
Linguistics
Publication Date
08/01/2018
Defense Date
05/02/2018
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
Aktionsart,argument structure,causatives,event structure,locative verbs,morphosyntax,OAI-PMH Harvest,participles,psychological verbs,states,verbs
Format
application/pdf
(imt)
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Zubizarreta, Maria Luisa (
committee chair
), Li, Audrey (
committee member
), Pancheva, Roumyana (
committee member
), Uzquiano, Gabriel (
committee member
)
Creator Email
alfgarpar@usc.edu,garciapa@usc.edu
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c89-43391
Unique identifier
UC11670752
Identifier
etd-GarcaPardo-6585.pdf (filename),usctheses-c89-43391 (legacy record id)
Legacy Identifier
etd-GarcaPardo-6585.pdf
Dmrecord
43391
Document Type
Dissertation
Format
application/pdf (imt)
Rights
García-Pardo, Alfredo
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA
Tags
Aktionsart
argument structure
causatives
event structure
locative verbs
morphosyntax
participles
psychological verbs
verbs