Download 2. Theoretical Issues with Case and Agreement

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

Antisymmetry wikipedia , lookup

Swedish grammar wikipedia , lookup

Lithuanian grammar wikipedia , lookup

Old Norse morphology wikipedia , lookup

Causative wikipedia , lookup

Modern Greek grammar wikipedia , lookup

Arabic grammar wikipedia , lookup

Macedonian grammar wikipedia , lookup

Modern Hebrew grammar wikipedia , lookup

Navajo grammar wikipedia , lookup

Udmurt grammar wikipedia , lookup

Scottish Gaelic grammar wikipedia , lookup

Old English grammar wikipedia , lookup

Portuguese grammar wikipedia , lookup

Inflection wikipedia , lookup

Old Irish grammar wikipedia , lookup

Russian grammar wikipedia , lookup

English clause syntax wikipedia , lookup

Chinese grammar wikipedia , lookup

Spanish pronouns wikipedia , lookup

Esperanto grammar wikipedia , lookup

Spanish grammar wikipedia , lookup

Kannada grammar wikipedia , lookup

Grammatical case wikipedia , lookup

Polish grammar wikipedia , lookup

Lexical semantics wikipedia , lookup

Yiddish grammar wikipedia , lookup

Georgian grammar wikipedia , lookup

Ancient Greek grammar wikipedia , lookup

Latin syntax wikipedia , lookup

Pipil grammar wikipedia , lookup

Serbo-Croatian grammar wikipedia , lookup

Transcript
THE
RELATIONSHIP
BETWEEN CASE
AND
AGREEMENT
ANDTHETHEORETICALISSUES
RELATEDTOTHISRELATIONSHIP
OURROADMAP
vTheoreticalassumptionsandan
overviewofX-Bartheory
vCaseassignmentandcaserealization
v Casevscase
vThebasicsofagreement
vWheretherelationshipbetween
caseandagreementgets
complicated
v AlookatChoctaw
SOURCES
2
Burzio,Luigi.2000.Anatomyofgeneralization.InEric
Reuland,ed.,Argumentsandcase:ExplainingBurzio’s
Generalization,195–240.Philadelphia:JohnBenjamins.
Johnson, Kyle.2011.Introduction totransformational
grammar.Ms.UniversityofMassachusetts,Amherst.
Kratzer,Angelika.1996.Severingtheexternalargument
fromitsverb.InRooryck,Johan,andLaurieZaring,eds.
Phrasestructureandthelexicon,109-137.Springer:New
York.
Woolford,Ellen.2006.Case-agreementmismatches.In
CedricBoeckx ed.AgreementSystems,317-339.John
Benjamins.
Woolford,Ellen.2008.Isagreementreallyindependentof
case?Ms.,UniversityofMassachusetts.
Ling 222 ~ Fall 2016 ~ C. Ussery
3
Morphology (in general) in a Theoretical Context
BigPictureIssue:
Ling 222 ~ Fall 2016 ~ C. Ussery
o Forthemostpart,we’lladopttheperspectivethatsyntacticstructures
andoperationsareresponsibleforwordformation.
o Thisisnotuncontroversial.There’saschoolofthoughtwhichprettymuch
removessyntacticoperationsfromthebusinessofwordformationand
heavilyenrichesthemorphologicalcomponentofthegrammar.(Wedelve
moredeeplyintothisissueinTopicsinMorphology.)
(Re)familiarizing ourselves with the x-bar skeleton
4
•Thepositionsofheads,phrases,andspecifers canbe
switched.Forinstance,inaverb-finallanguage,we
wanttheverbtocomeaftertheobjectintheVP.
•Someheadsarelexical.Forourpurposes,words
“startoff”inthesepositions.Thewordsmightmove
tootherpositions.
•Someheadsarefunctional.Theydo“work”inthe
syntacticstructureandmayormaynothostalexical
item.Theworkthatweareconcernedwithhas
morphologicalconsequences
•Thecomplementtoaheadiswhat’srequiredbythe
head- i.e.,atransitiveverbneedsanobject.
•Adjunctsprovideadditionalinformation- e.g.,
prepositionalphrases,adjectives,adverbials.
VP:ate asavorywafflefordinner
head complement adjunct
Ling 222 ~ Fall 2016 ~ C. Ussery
•Thespecifierpositionhasdifferentfunctions.
Sometimes,aphraseisbornthere.Sometimes
phrasesmovetospecifierspots.
5
Why X-Bar???
“…a goal of syntactic theory should be to contribute towards
structuring the universe of Gs.”
[Johnson 2011:3]
v Inessence, theX-Barmodelfacilitates(aweakerstance)ormakespossible(astrongerstance)the
acquisitionprocessbyseverelyconstrainingtherangeofpossiblegrammars.
v “IfeverylanguagelearnerisequippedwiththisX’Theory,thentheywillconvergeonmoreorlessthe
sameGL whenpresentedwiththeinformationthatbeingintheenvironmentofspeakersofL
provides.Iftherearedifferences intheGLsthatlearnersconvergeon,thesewilltracebackto
different decisionstheselearnershavemadeabouttheidentityofW,X,Y,andZorhowtheirlinear
orderisdetermined.Iftherestofamodelthatincorporatestheseconstraintsiscorrect,then,it
shouldallowanylanguagelearner topickoutaGLveryclosetotheGLgivingshapetothespeechin
thatlearner’senvironment.”[Johnson,p.4]
v Thesyntacticians’task,then,istofigureouttheminutiaethattheX-Barskeletonrepresents andto
thebestofourability,usethismodeltoaccountforavarietyofphenomena.
Ling 222 ~ Fall 2016 ~ C. Ussery
6
What are the XP categories in Xbar theory?
NounPhrase
DeterminerPhrase
VerbPhrase
AdjectivePhrase
AdverbPhrase
PrepositionalPhrase
Complementizer Phrase
TensePhrase
LightVerbPhrase,akaLittlevPhrase
Ling 222 ~ Fall 2016 ~ C. Ussery
Headsofthese
phrasesassign
case(probablynotan
exhaustivelist)
7
[Cherlon]prefers[bourbonbarrel-agedbeer].
nom acc
How does this work?
• .
Nominative
T hastwofunctions:
•(Finite)Tassigns
nominativecasetothe
subject.
• Talsotransmitstense
informationtotheverb.
•InEnglish,Vmovestov
•Insomelanguages,V
movesallthewayupto
T
•SinceThasarelationship
withboththesubjectand
theverb,theverbagrees
withthesubject.
Accusative
valsohastwofunctions:
•Itassignsaccusativecasetotheobject
•Thesyntacticjob
•Itprovidesahome– viaitsspecifier– fortheverb’s
externalargument(thesubject)
•Thesemanticjob
Important:Thisis
themodelforsimple
nominativeaccusativesystems,
independentof
whetherthecase
thatisassigned
syntacticallyis
morphologically
realized.
Case:
abstract
case:
morphological
Ling 222 ~ Fall 2016 ~ C. Ussery
§
Marantz (1984):Theentirepredicate(notjusttheverb)assignsaθ-roletothesubject.Themeaning/propertiesof
theobjectinfluencethemeaningoftheverb,which,inturn,determinesthesemanticpropertiesofthesubject.
§
Theobjectisassigneditsthetarolebytheverb.Thesubjectisassigneditsthetarolebytheentirepredicate.
The Subject-Object
Asymmetry
(1)
a.throwsupportbehindacandidate
d.throwaparty
b. throwabaseball
e.throwafit
c. throwaboxingmatch(takeadive)
(2)
a.takeabookfromtheshelf
d.takeanaspirin
b.takeabustoNewYork
e.takealetterinshorthand
•
Kratzer (1996)arguesthatweneedasyntacticposition forthesubject
thatdistancesthesubjectfromtheverb.
•
Fromasyntacticperspective,we’reusedtothinkingaboutaverb’s
argumentsastheDPs/CPs/PPsthattheverbssubcategorizesfor.
•
Fromasemanticperspective,verbsrefertoeventsandinorderto
derivethemeaningofaverb,theeventthattheverbreferstoispartof
theverb’smeaning.
•
So,verbstakeanargumentthatisanevent.[Thisisn’trepresentedin
oursyntaxtrees.WetalkalotaboutthisinSemantics.]
c.takeanap
(3)
a.killacockroach
d.killabottle(emptyit)
b.killaconversation
e.killanaudience(wowthem)
c.killaneveningwatchingTV
Ling 222 ~ Fall 2016 ~ C. Ussery
(Kratzer 1996,EX6-8)
8
IMPORTANT:Objectsare“internal”argumentsoftheverbandsemantic
subjectsare“external”arguments.
Thesemanticfunctionofv isto“introduce”theexternalargumenttothe
eventencodedintheVP.Thesyntacticfunctionofv istoassignaccusative
casetotheverb’sobject.Whyshouldthesetwothingsberelated?
9
Burzio’s Generalization
•Ifaverbassignsaccusativecase,thenitassignsanexternalθ-role.
•Nosemanticsubjectθ-role→noaccusativeCase
•Intechnicalterms:Whenaverbphrasecombineswithaheadthatintroducesanexternal
argument,thatheadassignsaccusativecasetotheinternalargumentoftheverb.
•Inb/dthere’sonlyaninternalargument,andtheinternalargumentsurfaces
insubjectposition.
a. Shefiredme.
b.Iwasfired.
c. Theybrokethewindow.(EX26)d.Thewindowbroke.(EX2)
(fromBurzio 2000)
•Weknowthere’snotaone-to-onemappingbetweencase,
grammaticalrelations,andthematicroles.
•Anotherinstanceofthismismatchisfoundwithunaccusatives,
whichpatternlikepassivesinsomeways.
•Noaccusativeassignedtothesemanticobject.
Ling 222 ~ Fall 2016 ~ C. Ussery
– Unergatives haveanunderlyingsemanticsubject.e.g.Maryslept.
10
– Unaccusatives haveanunderlyingsemanticobject.
– Someverbsalternatebetweenbeingtransitiveandintransitive.
– Maryfrozethepopsicles.
– Thepopsiclesfroze.Popsicles isthesemanticobject.Whydowethinkthis?
More about
Intransitives
– Intransitivesentences,resultatives modifyobjects,notsubjects.
– Johnhammeredthemetalflat.Themetalisflatasaconsequence ofbeing hammered.
– *Johnhammeredthemetalsweaty.John issweatyasaresultofhammering metal.
– Passivesallowforaresultative tomodify thesemanticobject/syntacticsubject.
– Themetalwashammeredflat.
– Resultatives arealsoallowedwiththesyntacticsubjectofsomeintransitives…
– Thepopsiclesfrozesolid.
Theunderlying
objectdoesnotget
accusativecasein
passivesorin
unaccusatives.
– …Butnotwithothers.
– *Marysleptrejuvenated.(OntheinterpretationthatMaryisrejuvenatedasaconsequenceof
sleeping.)
– The popsicles isasemanticobject– justlikethe metal inthepassive.
Ling 222 ~ Fall 2016 ~ C. Ussery
Examplestakenfrom:
http://web.mit.edu/norv
in/www/24.902/unaccus
atives.html
11
In some languages, auxiliary selection
distinguishes unaccusatives and unergatives
Unaccusative - be
Unergative - have
Italian
Italian
a.Mariaè arrivata
c.Mariaha telefonato
Mariastartsoff
Mariais arrived-fem.sg.‘
astheobject,
‘Mariahasarrived.'
butisnot
assigned
accusativecase. German
Mariais
b.DieMariaist angekommen
assigned
theMariais arrived
nominative byT.
‘Mariahasarrived.'
Examplestakenfrom:
http://web.mit.edu/norv
in/www/24.902/unaccus
atives.html
Ling 222 ~ Fall 2016 ~ C. Ussery
Mariahas telephoned
'Mariahastelephoned.’
German
b.DieMariahat telefoniert
theMariahas telephoned
'Mariahastelephoned.'
Thereisno
underlying
object.Maria
startsoffasthe
subject, andis
assigned
nominative byT.
And remember
Georgian
Student-i
12
mivida.
student-ABSwent
'Thestudent went.'
Student-ma Ceril-i
unaccusative
daCera
student-ERGletter-ABSwrote
'Thestudent wrotetheletter.'
http://www.nthuleen.com/papers/L12paper.html
unergative
Butt2006: 157
Georgian distinguishes betweenunaccusatives and
unergatives.
Ling 222 ~ Fall 2016 ~ C.Ussery
13
Case and Infinitives:
Raising vs Subject control
1.
Barnettseemed tounderstandtheformula.Raising
2.
Barnetttried[Barnett]tounderstandtheformula.SubjectControl
–
Inthesecondsentence,Barnett istheagentofthetryingandofthe
understanding.
TheCaseFilter:AllDPs
musebemarkedwitha
Weneedsomeiteminthelowerclausetosatisfythesemanticrequirementsof
Caseinordertobe
theembeddedverb.
pronounced.
–
ThetechnicaltermforthissilentDPis PRO.PROisarguedtonotgetcase.
Hence,itssilence.Non-finiteTisarguedtonotassigncase.[Wealreadyknow
thatit’smorecomplicatedthanthis.]
–
Barnetti tried[PROi tounderstandtheformula].
Barnettseemed[Barnett tounderstandtheformula].
–
–
–
•Caseisanecessary,not
sufficientconditionfor
pronunciation.
•Silentsubjectsin
pro-droplanguages.
ThereisnoPROintheseem sentence.Thereisonlyonesemanticsubject. •Non-finiteTdoesnot
Ling 216 ~ Winter 2016 ~ Cherlon Ussery
assigncase…maybe.
3.
Barnettpersuaded thedoctor toexamineTilman.
4.
Barnettbelieved[thedoctor/her tohaveexaminedTilman]. ECM
–
InECM,theverbinthemainclauseistransitive andtheembedded clauseisthe
objectoftheverb.
–
–
ObjectControl
14
Object Control
Theembedded subjectisaccusativeandwehaveareasonforthisexceptionality.
vs
Exceptional
Non-finiteTdoesn’t assignnominative, sotheembedded subjectgets
accusativefrom theverbinthemainclause.
Case Marking
–
ThereisnoPROinECMconstructions.Theverbinboth themainclauseandthe
verbintheembedded havesemanticsubjectsandeachsubjectispronounced.
–
ObjectControlconstructions areditransitives.Persuade hastwoobjects– theDP
‘thedoctor’ andtheclause‘PROtoexamineTilman.’
–
The doctorisboth theobjectofpersuade andthesubjectofexamine.The doctor
cannotgetcaseintheembedded clauseandisrepresented byPRO.
–
Barnettpersuadedthedoctori [PROi toexamineTilman].
–
Unlikeinsubjectcontrol,inobjectcontrol,theobjectiscoreferential withPRO.
Ling 216 ~ Winter 2016 ~ Cherlon Ussery
TheBigPictureisthatinfinitives
playabigroleinCasetheory.
“Normal” v’s and “Special” v’s
15
Normal
Icelandic
1.a.Viðlásumbókina.
we.Nom read.1plbook.the.Acc
‘We readthebook.”(Sigurðsson1996,Ex14)
Inboth theDat-Nom
Icelandicsentence
andtheErg-Abs
Gujaratisentence,the
verbagreeswiththe
nom/absobject, not
withthedat/erg
subject.
--Normalv:assignsaccusativetotheobject
b.Einum málfræðingi líkuðu þessar
hugmyndir.
one.Dat linguist.Dat liked.3plthese.Nom ideas.Nom.pl
‘Onelinguist likedtheseideas.’(SigurðssonandHolmberg 2008,EX1)
--Specialv:assignsdativetothesubject
NOTE:Thesenon-Nominativesreally aresubjects,nottopicalizedobjects.We’ll see
evidence forthat.And,I’mignoringtheoptionalityoftheverbalagreement.
Special
Gujarati(spokeninIndiabyabout50millionppl)
2.a.Sudha
awy-i.
Sudha(fem).Abscame-fem
‘Sudha came.’
--Normalv:there’snoobjectforv toassigncaseto
b.Sudha-eradiokhəridy-o.
Sudha(fem)-Ergradio(masc).Absbought-masc
‘Sudha boughtaradio.’ (Woolford 2006,EX38c/39)
--Specialv:assignsergativetothesubject
Ling 222 ~ Fall 2016 ~ C. Ussery
16
Little v-dative and ditransitives
English
Thestudentsgavetheirprofessoranecklace.
Icelandic
Ég
sendi
Hildi
fiskinn
I.Nomsent.1sgHildur.Dat.sg.fem.fish.the.Acc.sg.masc.
‘IsentHildur thefish.’
English: V→ v(dat) → v(acc)
Icelandic:V→ v(dat) → v(acc)→ T
(Icelandic is a verb-second language and negation comes
after the verb.)
Ling 222 ~ Fall 2016 ~ C. Ussery
Caseisarelationship betweenaheadandaphrase.
Interim
Summary
Aheadchecks/assigns(theparticular terminologyvaries,butthegeneral conceptisthe
same)case toaphrasethatoccupies aparticular structuralposition.
Particular headsarebornwithaparticular casetogiveaway.E.g.:
FiniteTassignsnominative,v assignsaccusative
Special v’scanassigndifferentcases
Prepositionsassignavarietyofcases(accusativeinEnglish)
There isadistinctionbetweenabstractCaseandmorphological case.
AllNPswhicharepronouncedarearguedtohaveabstractCase,whichisassigned
inthesyntax.Sometimesthere isamorphological expressionandsometimesnot.
17
Ling 222 ~ Fall 2016 ~ C. Ussery
A Look at Choctaw
Spokenbyappr.10,400peopleinLouisiana,
Mississippi, Oklahoma,andTennessee
18
Ling 222 ~ Fall 2016 ~ C. Ussery
19
Is agreement really independent of Case
in Choctaw? [Woolford 2008]
NO
YES
v Caseandagreementarelinked.
v Yes:Averbcanagreewitha
nominativeinmorethanoneway.
v Agreementfollowsthepattern
observedinnominative-accusative
systems.
v Theverbagreeswiththenominative.
v Thechoicebetweenusing“true”
agreementandusingaclitic depends
onanactive-stative
distinction…maybe.
v Theagreementmorphemeisused
withmoreactive/volitionalactivities.
v Thecliticisusedwithless
active/volitionalactivities.
Ling 222 ~ Fall 2016 ~ C. Ussery
v Thisdivisionisverytentative
–
Davies1986:“InChoctaw,‘theagreementsystemandthecasemarkingsystemfunction
independently…’”
–
Broadwell2006:casesystemisnominative-accusative
–
Mithun 1991:Agreementsystemisactive-stative
–
Allsubjectsarenominative.
–
Themoreagentivesubjectsare“cross-referenced”withSeriesIforms.
20
Previous Literature
Thelessagentivesubjectsare“cross-referenced”withSeriesII/IIIforms,whichalsocrossreferenceobjectsandpossessors
Woolford’sProposal:Thereisadistinctionbetween“true”
agreementandclitics.
–
•
“Trueagreement”isWoolford’stermandreferstoverbalagreement
withnominatives.
•
Anominativecanbecross-referencedwitheitheratrueagreement
morphemeorapronominalclitic.
•
•
Anaccusativecanonlybecross-referencedwithaclitic.
•
•
Woolford doesn’t committotheexactdistinction [activevs
stative/externalargumentvsinternalargument,orsomethingelse]
Verbsdon’tagreewithaccusatives
Thenominativeandaccusativeclitics lookthesame.The
morphologicalexpressiondoesnotnecessarilymatchtheabstract
Case.
Ling 222 ~ Fall 2016 ~ C. Ussery
21
Overview of C/case in Choctaw
ChoctawisSOV Subjectsarenominative Objectsareaccusative
• Accusativeisoptionally
markediftheobjectis
adjacenttotheverb.
• Accusativeisobligatorily
markediftheobject
moves.
• Condition on
morphological
realizationof case
Ling 222 ~ Fall 2016 ~ C. Ussery
•
No(morphological) dative.
Inditransitives, the
accusativemarkeris
optional fortheargument
adjacenttotheverb.
•
Free(unbound) pronouns
aremarkedfornominative
andaccusative,butthey
onlyappearwhenfocused
andcontrastive.
22
Agreement vs clitics in Choctaw
Agreement
–
Series1:cross-referencesonly,butnotall,nominatives
–
Occuratleftedge,exceptfor1sgform,whichprecedesthetensesuffix
–
(WhenwelookatDakota,we’llseethatthedistributionofpronouns
varieswithintheverbalcomplex.)
Ling 222 ~ Fall 2016 ~ C. Ussery
23
Clitics
–
-Clitics always have an abstract Case that matches the abstract Case of the noun
they cross-reference.
-Series II has both nominative and accusative clitics that are syncretic.
-But…the nominative clitic is closer to the verb than the accusative clitic.
-A clitic that cross-references a subject is closer to the verb than any other clitic
Thereisno3rd
personclitic,sothe
2nd personobject
clitic isclosestto
theverbin(23).
Ling 222 ~ Fall 2016 ~ C. Ussery
YucatecMayaisspoken by1millionish ppl inMexicoandBelize
24
Aspect,
agreement, and
clitics in
Yucatec Maya
Thenominative subjectofanintransitive
iscross-referencedbyagreementwhen
theaspectisimperfective,but byaclitic
whentheaspectisperfective.
And, theclitics aren’tdistinguished for
case.Theaccusativeform inthe
transitivesentencein(27)isthesameas
thenominativeform intheintransitive
sentencein(26).
Ling 222 ~ Fall 2016 ~ C. Ussery
25
– TheCasevscasedistinctionisveryimportantinsyntacticand
morphologicaltheory.
– Sometimes there’sanalignment betweenCaseandcase,butoften times
not.
– AgreementisarguedtobesensitivetoCaseinChoctaw
Ling 222 ~ Fall 2016 ~ C. Ussery
Summary