Inthe倉

ECM-BasedCaseTheoryandItsInteractionwiththeA-barSystem
TsuyoshiSawada
UniversityofTokyo
[email protected]
Aみ∫什αCJ
乃ね押erα巧坪e∫聯血∫J血α∫∫〟〝卯0乃力e〟加血J油川血作∫わ打eC力0椚ぶ砂〃タタ刃〟∽JJ∝α伽〃〆
砂
α〆肌椚=肌肌痛匂ぃ血=那咋血血gねαJクJ甲e巧′げα力紺dね曙拍α乎eCぴerげ血力eαd
0∂∫en万喝血co卸相加〃作∫〟招曙J伽桝α∫eα∫∫柳川g乃fα乃dA-∂α′dぷわcα伽〃,Jco〃C血虚血J血
反感㌧祓血==鮮明鮎=ダ励れ抑叩血那加q=助力路かが〟ほグ血伽〃加ね甜一作∫〟J血gわ川〃
励昭血朋圧血相如助g(EC几〃ぐ∂巧β併用如〃,わ7加成伽〃αJねr′"∫･椚肋αdガ如〃αJα∫∫〟J甲加那,
肋ec￠〝血加e加地血J加Ⅶ如叩和Ceβゐe血r如桝血抽伽.呼･明細叫那押頑血沈相成
加∂gわw加αゐ,Ore〃触れ伽研血叫ゐw乃,W油乃0〃eedげ眈肋7g肌
勒珊OJ血ク0∫ト甲e相加乃αJco花館騨相加〃,乃曙efげAゐ巧㌍,Cα∫eα∬妙椚叫放c甲如乃αJCα∫e
肋′紘一&eヱJ妙お一朗ve〃威ざわcα加〝,β柁おr加gq/甲e相加〝ぷ
1.ImtrodtlCtiom
Inthe倉ameworkpresentedbyChomsky(2000,2001),thecorepartofthecomputationis
CarriedoutbytheoperationsAgreeandMerge.Agreevaluesanunvalued(uninterpretable)
fbature,tyPically(P,OfProbePbyaninherentlyvalued(interpretable)featureofitsGoal,
henceforthG(P).WhenPiscomplete,Agreefurthervaluesanuninterpretablefbatureof
otheroperationMergetakes
G(P),tyPicallyCase,byaninterpretablefeatureofP.The
Placeundertwooccasions:(a)whenPisaccompaniedbytheEPPfeature,inwhichcase
MergeisconcomitantwithAgree,Or(b)whendrivenbythedirectneedsoftheinterfaces,e･g･
Selection,0-markingandstylisticdislocation.Inoccasion(a),theEPPchoosesaphrasethat
includesGげ)and丘xesitscopytosomeposition,CreatingtheTarget,henceforthTU))･The
Primarypurpose
ofthis
paperisto
reestablishthis
positionthatEPP
selects
as7てP)in
OCCaSion(a)ofMerge.
Forthispurpose,COnSiderthepositionalrelationbetweenPandthemostrecentlycreated
COPyOfthephraseincludingG(P)atthestageimmediatelya氏ertheoperationdrivenbyP･
WhenPhasanEPPfeature,themostrecentlycreatedcopyofthephrasethatincludesG(ア)is
●Thspaperreapsthebene丘tsofIMANISⅢTbrazuNoriko,KASAlHironobu,KUNOMasakazu,
TÅNCREDIand
Howard
LASNIK,MrYASHITA
Han皿aSa,Christopher
Especial1y;HowardLASNIK,Christopher℃ÅNCREDI,andⅥばIANABEAkiragavemeextensive
COmmentS
Onearlierversions
ofthispaper.HoweveちIcouldnotmakethebestoftheirhelpfu1
SuggeStionsandIamSOlelyresponsibleforanyremainlngerrOrSandmisunderstaJldings･Organized
responsetothesesuggestionswillperh叩Sbenndeinmyfuturepublication.
上J〃g〟ノ∫扉尺e∫gα打力ノ斬20(〃ノトj7
･二川ノ/恒､r川l…吊∫`川肋んJ
1
WÅⅠÅBABE
Akira.
P
叩)･When
does
Let
Gげ)itself
not,itis
us
the
positionalrelation
callthe
POSt-OPerationalconfiguration,COntraStingltWiththepre-OPerationalcon負guration,Whichis
dw野Sア･Gの:
T竺?fdoesnothaveEPP
P-G(P)
P-G(P)
...〝..T竺?fhasEPP
J
Pre-OPerationalconBgurat10n
J
′J
OPeration
Agree-Merge
Agree
↓
J.
J
P-G伊1
P-TU)1
post-Operationalconfi?ration
Inestablishingthepositionfor7甲),WemuStnOtlimitourobservationtothecon丘gurations
thatareassertedintheliterよuretobeP一叩),SincechoosingaP一叩)connguration丘･Omthe
inventory
ofthe
post-OPerationalcon丘gurations
stands onthe
assumptlOn
the
regarding
POSitionallowedas7げ),Whichistheverythingthatweと鱒etryipgtoreconsider.Thus,itis
importanttOeXaminenotonlyP一岬)butpost-OPerationalconRgurationsingeneral､
EvenifwelimitourattentiontooperationsthatinvoIveCaseasslgnment,atleast丘ve
di飴rentpost-OPerationalcon丘gurationshavebeenidenti丘ed,aSin(2).(2II,Ⅲ)rhanifest
P-TU)),While(2I,ⅠVV)areP-qア).Theal)breviationsqre:CmPn='n-timeiterationof
"complementto,'"spec='innerspeci丘erof;'outerspec='outerspeci丘erof'Theuseof
X-bartheoretic
notionsis
bare
deny
to
not
phrase
convenience.
struCture;itisfor
HencefoQoutlinedlettersrepresentunpronounCedcopies.
(2)I.P-CmP2(P)II.P-Sp9C@)III.P-
IVP-
VP-SPeC(cmp3(P))
Outer.SPeCU,))spec(CmP2(ぞ))
V'
/へ＼
TP
/へ＼
T'義元
T
/へ＼
vP
/へ＼
V
v'
su句
/へ､
W
畢
Ⅴ'
/へ､
/へ
T,
v'
vP
/へ＼
/へ＼
Vp
Ⅴ
Ⅹp
/へ＼
Ⅹ'
画ゆ朽吻ノ
向直呵勤一dノ
匝加β動
2
(毎加β初
(お加働ノ
(3)a.[cpC[TPMaryT[vpMaEyliked[vpVthebook]]]]
LIト｣
b･Sera
l
LI
di丘cil【TPeles
a
aprova-r-em
PrOPOSta]
Wf〟∂e櫛c〟〟 助印肋椚(坪和VピーわーjタJ肋叩托pO∬り払J甲eα〝タor噸′e∫り
｣ⅠI
I
C.itwasdi伍cult[cp甲C[TPmOtogetthebook]]]]
LIト｣
d･[cpjagkysste[TPj喝kys離⑪[vphenneinte
J
彪且好d
力er
j喝ky$離@[wV鮎@皿@]]
仰l､ピd∫/り
′70f
l
｣Ⅲ
e.Mary[vpMaqyconsiders[&V[TPittobeveryinformative]]]]
f[cpC[TPthereare[vp都椅[vpV[ppfourchaptersinthebook]]]]]
仲ご月呼)0∫0〃タβ7ごβ￠/正月b加血管〃タタタ.●ル
Againstthisstandardanalysis,Idefbndinthispaperthepositionthatthepost-OPerational
COn丘gurations
Case
to
assignment
the
uniformly
nominalsin(3)are
P-SPeC(cmp(P)).Preci声ely;Iarguethatthepositionofthenominalsin(2I,.IV)isnotcorrect,
regarding
thattheelementsspecinedaspin(2II,V)arenotthqrightonesandthatthenominalin(2III)
doesnotovertlyshowupinthepost-OPerationalcon丘gurationrelatedtoitsCaseasslgnment.
Thepost-Operationalcon丘gurations(2I-V)ar寧tObemQdi丘edas(4I-V),reSPeCtively.(4II,
IV)areP一叩)configurations,Whereas(4Ⅰ,III,V)areP-GU)).
(4)Ⅰ.クー
ⅠⅠ.ダー
ⅠⅠⅠ.クー
ⅠVクー
SPeC(cmp(P))spec(CmP(P))spec(CmP(P))spec(cmp(P))
CP/DP
Ⅴ'
/へ＼
璽
v,
/へ＼
Vp
Ⅴ,
〈
/へ＼
TP
Ⅴクー
spec(cmp(P))
VP
/〈＼
斯
〈
Ifurtherarguethatoperationsconcerningthe[+Wh]featureinspec(CP)alsogoalongthis
line:
3
Generalizing
the
toincludelong
empiricaldomin
distanCe
Case
asslgnment,the
descriptionofpost-OPerationalcon丘gurationsmayseemtobeas(6),with(4I-V)maniftsting
thespecialcase〃=l.
(6)Post-0perationalconRgurationsarerestrictedtoP-SpeC(cmpnげ)),n=1,2,....
P'
〈
委
I
〈
…Cmp乃げ)‥.
/へ
Aswewi11seeinsection8,(6)entails,underanadditionalassumption,that,Whenever
thereisapost-OPerationalcon丘gurationP-SpeC(cmp(P)),thecon丘gurationisP一叩)unless
thenominalwaspreviouslyinttoducedintothatpositionbythedirectneedsoftheinterfaces,
SuChasselection,0-markingorstylisticdislocatipn.
Thepositionofthenominalihthepost-OPerationalcon丘gurationP-SPeC(Cinp(P))in(4ⅠⅠ,
as
direct needs oftheinterfaces;SuCh
IV)cannOt be attributed toany
Ofthe
selection,
0-markingorstylisticdislocation.Thismeansthatthecon丘gurationisP一叩).Thus,We
have
a
clearinstance
ofP-7てP)thatis also P-SPeC(cmp(P)).The simplest hypothesis
of7てP),Which should be adopted unless thereis
COuntereVidence,isthattheP-7てP)configurationthatEPPselectsisuniformforarbitraryP.
From仙s,(7)払1lows:
regarding
the
nature
of EPPand
(7)7げ)=SPeC(cmp(P)),forafbitr!ryP･
Thisisthem句OrClaimofthispaper.
Sections2-5handleseachtypeofnominallistedin(2,3).Sections2-4areorderedbythe
StruCturalpositionofthenominaltobearguedfor:thelower,theear1ier.Namely,SeCtions2,
has been
3and4dealwith
what
regarded as
cmp(vP),Outer SPeC(vP)and spec(TP),
respectively.Section5dealswithamorecomplicatedcase:theexpletived2eT?-COnStruCtion･
Section6extendstheobservationtotheA-barsystem.Section7isabriefspeculationthat
4
mightcometominda氏erthepreviousobservations.Takingovertheseobservations,SeCtion
8derivesthe
m再orconclusion(7).In
overconclusion(7),an
section9,furthertaking
interestlngSPeCulationismadeonhowderivationproceeds･Theargumentsaresummarized
insectionlO.
2･0叫ects(一in-Situ)inSimple取ansitives
Inthis
directobjectinsimpletransitive
section,WereeXaminethebasepositionofthe
COnStruCtionsbyanalogy丘omthedireQtObjectinotherconstructions.SinceLarson(1988),
itis
widely
that
accepted
a
dative
construCtion
has
a
VP
shellstruCture.With
slight
modi丘cationregardingheaddislocationto丘tthe舟ameworkofChomsky(2000,2001),an
exampleofthisconstruCtionwillbeasin(8a).Aconstructionwithadirectobjectanda
ClausalcomplementneedstobeanalyzedlikewisebytheVP
shell,aSin(8b),atleastto
maintainbinarybranChing.-InbothconstruCtions,thedirectoqectislocatedinspec(VP).
(8)a.JolmgaveabooktoMary
vP
/へ､
∬⑬払皿
Ⅴ'
/へ＼
VP
gave
/へ
V'
abook
/へ､
V
PP
/へ
toMaⅣ
5
b･JolmtoldMary[cpPROtoreadthebook].
vP
/へ
∬⑬肋皿
Ⅴ,
/へ＼
VP
told
/へ
Maけ
Ⅴ'
/
＼
V
Cp
ノ′へ
PROtoreadthebook
Chomsky(1995),amOng.･Others,eXtendstheVPshe11analysistosimpletranSitives:He
assumesthatthedirectobjectsofsimpletranSitivesareincmp(W).However,aSSumingthe
(absolute)UniversalThetaAssignmentHypothesis(Baker(1997)),the direct o旬ectofa
SimpletranSitive,WhichbearSathemerole,willbenatural1ylocatedratherinspec(VP):
(9)Maryr$adthebook
VP
/へ＼
1庖_フ
V'
/へ
read
Vp
/へ
tbebook
V
Thisanalysiscompe1stheexistenceofacategorywithaspeci丘erandwithoutacomplement,
butthisdoesnotseemtocastanyproblem.
AsWatanabeAkirab.c.)andHowardLasnik(p.c.)pointout,thedistinctionbetweenthe
COmPlementandthespecifierwithinVPlin(9)maybetrivialunderthebarephrasestruCture
theoryifwea1lowacategorywithaspeci鮎randwithoutacomplement.Ifwedismissthe
distinctlOn,ltiswellwithourpurpose.Alternatively,understandingthat
COmPlementandspeci丘eraredeterminedonlyrelativelywithrespecttothestructureinwhich
suchnotions
itappearS,Wemaydistinguishthetwonotionsbyassumingdiagnostics(10):
(10)Anominalargumentpositionistheinnerspeci負erofaO-aSSigninghead.
Undertheformeralternative,thepostべ)perationalcon丘gurationforProbevandthedirect
6
as
objectinaiimpletrrsitive･construCtionisP-SPeC(CmP(P))in(4Ⅰ)aswellas祉isP-CmP2p)J
in(2Ⅰ).Underthelatteralternative,the con丘gurationisnolongerPicmp2(P)butis
P-SPeC(cmp(P)).
3.(So-Called)OuterSpec(▼P)
We
Inthi;section,WeeXaminewhathasbeenclaimedtobeinthe10uterSPeC(vP)二
COnClude(i)thatneitherAccusativeCaseassignmentnorobjectagreementisthedrivingforce
Ofdislocationtosuchapositionand(ii)thattheseitemsarenotactuallylocatedintheouter
SPeC(vP)butarelocatedinsomehigherposition.Under(i),thereisnoreasontoassumethat
AccusativeCaseasslgnmentand､Objectagreementwithashiftedo句ectisdoneinadiffbrent
way丘･Om
thatwith
non-Shifted
eStablish that
o切ects.Strictly speaking,tO
the
Case
assignmentconfigurationis.v-SPeC(VP)withshiftedoqects,independentevidencemustbe
shown･However,We
may
POSt-OPerationalcon丘gurationfor
that
note
atleast
Accusative
Case
the
simplest
asslgnmentin
that
assumptlOnis
these
casesis
the
also
V-SPeC(VP),aSWeCOnCludedforsimpletranSitivesinsection2.
j.ノ.0むgcfJ肋節
Itisclaimedintheliterature(Chomsky(1995,2001))thatshiftedo切ectsinScandinavian
langu喝eSarelocatedintheouterspec(VP)ノandthatldislocationtosuchapositionistriggered
eXamPles
byobjectAgreementandAccusativeCaseasslgnment.Thefo1lowlng
maybe
analyzedinthiswayunderBobatjik's(1995)assumptionthatnegationandsentenceadverbs
inScandinavianlanguagesarea句oinedtovRdisregardingtheteclmicalquestionofwhethera
SpeCCanbelocatedoutsideofan叫iunct.
(11)a.[cpjagkysste[TPj喝kys釦母【vphenneinte[vpj喝ky$離⑪旭m鳩】]
J触ed
加r
御e成殉
〃OJ
henneinte
b.[cp[vpj喝kysst払孤皿@】har[TPjag払訂[vp
彪∬ed
力αγβ 才
力er 〃0′
肛vpj喝晦s釦母肋@m嘲]]】
作わ肋∂e曙〃タタタごJ,りノ
In(11b),thevPistopicalized.
However;aSHolmberg(1999)notes,Observationofembeddedclauseswithanauxiliary,
which,unlikethatofmatrixclauses,isnotdislocatedtoC,reVealsthatnegationandsentence
adverbsintheScandinavianlanguagescannotbelowerthantheauxiliary,sbaseposition,and
thus,COntraBobatiik,thattheyarehigherthananyWhereana句uncttovPcanbe.
7
(12)a･*detarm句1igt[cpatt[TPPerT[vp
har[inte[vpP母釘kysst
henne]]]]]
触gdゐer伸γβd殉
b･detarmqjligt[cpatt[TPPerT[inte[vphar[vpP母釘kysst
henne]]]]]
ブ′ 由ク"裏地
肋αJ
力∬〝OJ
C･[cpjaghar[TPj喝払紺【inte[vp ぬ紺【vpj喝kyssthenne]]]]]
才
力の′e
彪∬edカβ7
〃0′
但0肋あe曙〃タタタごノ,￠ノ
Ifshiftedobjectsaretobeidenti鮎dbybeinglocatedhigherthananegationorasentence
advetb,itcarm?tbemaintain9dthattheyarelocatedintheouterspec(vP).
Thereis
also
evidence
the
agalnSt
that
assumptlOn
oq9Ct Shi氏is concomitant
AccusativeCaseasslgnment･Tlms,eVenNominativeobjectsinIcelandic,Whichcanbe
foundinergative,PaSSive,OrpSyCh-VefbconstruCtions,pndeIBOObjectshi氏:
with
hdn/t61van
(13)a.m6r likar
ekki
娩e･CO叩P〟ねだ肋桝乃Of世e加7成り
胡乱β伽仁九転･2掬f≠
げわ肋∂e曙〃タタタご2`抄
Observationofcross-1ingulSticcorrelationbetweenwordorderandtheavailability
of
Objectshi氏incertainconstruCtionsimpliesthatobjectshi氏isaderivationaloperationbut
CannOtbeapurelysyntactica11ydrivenoperationdoneasconcomitantwithCaseasslgnment
andagreement･Thatis,ObjectshiftcorrelateswithwhetherthereisaphonologlCallyvisible
elementwithinvPatthepolntOfderivationwhereittakesplace.First,theoqectpronounin
a
particle construCtion
verb
obligatori1y
the
shi氏sin
Scandinavianlanguages▲With
O叫ect-Particleorder,WhereasitdoesnotintheScandinavianlanguagewithparticle･OqeCt
Order,namelySwedish･Eveninthelatterlanguage,eXtraCtionoftheparticleenablesobject
shi氏.
(14)a･jegskrev(*opp
Jw和お
b･de
det/detopp)
仰フ〝(癖励
呼if
kastet
megikke[vpd@ka離離m雫gut]
/力り一班Ⅳ--′桝e
仙/
OJJ∠
det/detop)
C.jegskrev(*op
J
w和ね
呼fJ
P伽メ殉
d･jegskrevdetmaskeikke[vpj喝如酎d醜OP]
Jw和おiJ椚呼∂e〃Of
e･jagskrev(upp
J′w和お
呼
det/*detupp)
f′
呼
f*domkastademqiinte[vp
肋町政柑W
g/【cpUT
桝e
伸γβ成功ノ
d⑬mka離ad@ut
〃OJ
m画]
o〟J
kastade[TPdomka$他舶【mqjinte[vpd⑳mka離ad⑰,,叫甲d]]]
■ 1頻y ■
徽∴醐
∴■
の〟j励柁-γ
伽か混信｢ルック.･Jイノノ
8
Thatthe
correlationis
notpeculiartotheverb
particleconstruCtion
canbe
seenbythe
Variabilityoftheo句ectshiRblocker:
(15)a.*【cpjaghar
J
【TPj喝払紺[henneinte[vpj喝kysst 払⑪皿@]]
血Ⅳe
力er
b.*…[cp如t
肋d
〃OJ
仰㈹血叫
[TPjagT[henneinte[vpj喝kysste払弧皿@]]
J
tbat
力er
触ed
〝OJ
C.*[cpjagtalade [TPj咽せ劇ad@[hermeinte[vpj咽せaAad⑪med払@皿皿@]]
J乎0鹿
カer
d.*[cpjaggav
J
ブJ
gαVe
e･[cpvemgav[TPdu
W力og甘Ve
乃OJ
wブ肋
【TPj喝gaV[deninte[vpj喝g附EIsad@n]]
〃0′
g肘[deninte[vp血皿gaV
ブf
γ0〝
Ve!md@皿]]
〝OJ
げわぁ血曙〝タタタごノひ
Second,Shiftingthesu句ectofasmallclause(SC)ispossibleinGerman,WhichhasO-V
Orderand
accordingly
SC-V
Swedish,Which
ordeちWhereasitisimpossiblein
hasthe
COnVerSeWOrdordeちbecausetheverbwillblockobjectshi氏inSwedish.Thattheremnant
VP
toplCalization does
not
enable
o句ect shiftin
the
Swedish
exampleimplies
thatthe
Veri丘cationoftheexistenceofaphonologicallyvisibleelementwithinvPisdoneatthepolnt
Ofderivationwher60bjectshi氏takesplace.
(16)a･[cp[vp[sc$@in@｢馳適地訂rauChen]gelassen]hat[TPer血如[seineTochter
ぷ〝0お
α伽βd
力が
カe
力由
血喝加er
【nicht:vp:3C:ein号Tc;土恕ご免じつ五､5エコ毒虫ニ;弧]]]】]
｢GeJⅦα〝卿吻
仰J
b･*[cp[vpgiort[scm画
研α鹿
随Ⅷ画蝕融m和]】】]
fBrvirrad]]har[TPdenh訂【mqi[alltid肛vpか㈹
血び
≠J
椚e
αル耶
co′所′∬d
御e成成澤K抄
げわぁ才∂e曙〝タタタ.･鋸ノ
Though
oqect
asslgnment
shiftis
a
derivationaloperation,itisirrelevant
to
Accusative
Case
attributes the cruCialfactor
o句ect agreement.Rather;Holmberg
the
cross-1inguistic variation,the
obiect shi氏to
propertyトFocus].Wth
Shiaedoqecttendstobespeci且candlight,butnotabareinde丘nite,anegativeDPora
and
of
Scandinavian
Wh-Phrase･Ifitisapronoun,itmustbeaweakpronoun.Insuchcases,theshiftedoqect
mightbeatopIC･However,thisisnotnecessarilythecase,Sincealsoanexpletiveinan
idiomaticexpressioncanbeshi鮎d:
(17)hantar
he
k7kesit
detmycket
veTy
sallan[scd離1ugnt]
eag)[tadetlungt`idkeiteaq,]御edisjd
げわ肋∂e曙〝タタタ.･2.抄
Sekbm
9
Inthiscase,adiffbrentprocess,SuChascliticization,mightbeworking.Wbwi11notgo
furtherintothis
matter..Forconstructionsincluding
COnCludedthatthe
ashi氏ed
subjectis extractedoutofthe
TP
bbjectthati'stoplC,1tis
pastTbpPifwe
adoptthe
usual
assumptlOnthattopICPhraseTopPishigherthanTPbecausethesubjectpretedestheshifted
Object.OnepossibilitylStOallocatethesubjectasthespecofthefocusphraseFocPand
thesecondpositionverbastheheadFoc.ThestruCturein(11)willontheseassumptions
bemodifiedas(18):
(18)a.[F｡CPjagF｡｡kysste[TbpPhenne7卸[TPj喝intevP]]]
J
触ed
力er
〃OJ
御eゐ旬
b.[TbpP[vpj喝kyssthem@]Tbphar[F｡｡Pjag肋紺[T｡PPhenne7卸[TPj喝intevP]]]
肋ed
力のJe
J
力er
ノブ0f
ThisisconsistentwithRizzi's(1997)analysisoftheleRperipheryofaclause,inwhichTopP
caniterateaboveandbelowtheumqueoptionalFocR
10
(19)
ForceP･
/へ､
′柁ゎかβ甲Force'..
./〈＼
凡打e
TopP*
/へ＼
7卸ね
Top'
/へ＼
FocP
7卸-
/へ＼
F?C'
e甲
/へ
fわc
TopP*
/へ＼
7毎)ね
Top'
/〈＼
Top
FinP
/へ､
Fin'
/へ＼
fbr
TP
j.2.乃e月e椚〃α〝fq′鳥e〟(わg御iJ智
Thoughobjectshi氏isnotobservedinEnglishinthemosttypICalcases,therearecertain
COnStruCtionswhichhavebeentakenupasevidenceofobjectshiftinthislanguage.One
SuChconstruction,Whichmayseemtoshowthisoperationmostovertly,lSPeudogapplng･
Thesurfacee飴ctofpeudogapplngislikevPellipsISeXCePtthatitleavesbehindsomevP
internalconstituent.A土the
surface,theelidedpartofpeudogapplnglSnOtneCeSSarily
a
Slngleconstituent.
(20)a･theDAprovedJonesguiltyandtheAssistantDAwillp町⑬∇母Smith曲
b.?JolmgaveBillalotofmoney,andMarywi11由Ⅴ⑰Susana孔⑬せ⑳好m⑳m@y
C.Jolmcouldpullyououtofaplane,1ikehedidpⅧ皿hisbrother⑳血⑳『ap乱弧@
d.Gee,I'veneverseenyouoncampusbefore.---Yia!NeitherhaveIs⑪⑪皿yOu⑳m
6弧呼ⅧS恥曲釘⑪
e.Jolmwi11selectme,andBillwi11s母旭離you
fMaryhasn'tdatedBill,butshehasd如⑪dHarry
何占,eごエα∽戊〝タタ∫Jノ52-ノ∫勃c,戌′ぷ窄〝夕7￠cおdブ乃エα∫〃戊〃タタ刃ノ
11
Thus,1nOrdertomaintaintheassumpt10nthate11ipsISaPPliesonlytoaslngleconstituent,itis
necessarytoassumethattheremnantisdislocatedtooutsidethevPwhichiselided.
Apossible
candidate
ofthis
dislocationis
heavy
DP
shift.Howevef,Lasnik･(1995)
ObserveSLthattheremnantofpseudogapplngShowspropertiesthatarequitediffbrent打oma
Shi氏edheavyDRwhichcasesdoubtonthispossibility.First,theremnantofpseudogapplng
Canbe,butashi氏edheavyDPcannOtbe,anindirectobject.
(21)a.ウJohngaveBillalotofmoney,､andMarywillSusan
b.*JohngavealotofmoneythefundforthepreservationofVOSlanguages
佑郁J祓〃リリj.･/jj-/j〃ノ
Second,theremnantofpseudogapplngCanbe,butashi鮎dheavyDPcannotbe,aPrOnOun
otherthanit.
(22)a.Johnwillselectme,andBillwillyou
b.?*Billwi11selecttomorrowyou
￠∬〝戊〝タタ5ごノ5刃ノ
Having
excluded
the possibilityofheavyDP
shi氏in
pseudogapplng,Lasnik
analyzes
PeudogapplngaSO切ectshi氏followedbyellipsISOfthelowerverbphrase$hell,aSinthe
払1lowlng:
12
(23)(theDAprovedJonesguiltyand)theAssistantDAwillSmith
Agrsp
/へ
theassistantDA
Agr;
へ
A許s
TP
イヽ
willよVP
/ノT｢､＼＼
他@乱SS盈離乳此DÅ
Ⅴ,●
･〈
V
Agr｡P
/へ＼
Smith
Agr｡'
/へ､
Agr｡Ⅴ『
/へ
Ⅴ,
$血払
/〈＼
阿①Ⅶ画晦
CruCially,thisanalysisreliesontheassumpt10nthatthereisastageinthederivationwhere
thephonologlCalcontentoftheved)islocatedontheheadofthelowerved)Phraseshell･如
thepointofellipsIS,thephonologicalcontentofthevefbmustbeplacedinthisposition,
Whereastheremnantmustberaisedoutofthelowerverbphraseshell.
ThereareempiricalreasonsthattheanalysISCannOtbemaintained.First,eVenmanner
adverbs,Whicharepresumabliaqoinedtovlhannotbele氏overbyellipsis(25a,b)or
PSeudogapping(25c,d).
(24)a･JohnhasbeenヲerfouslywoundedandHarryhasbeen[vp*(
b･Jolm1iedcorrvlnClnglybutBilldid[vp姐⑮(*unconvincingly)]
C･JolmquicklycookedthericeandMary(*slowly)thebeans
d･SimonenthusiasticallycountedthegoldandJack(*nervOuSly)thediamOnds
￠Jβ相成J〝夕&㌢j動かG和甜〝夕乃ごノ叫,Cご月加血桝er〝夕7ノごノ叫,d花rα〟〝夕乃.･j￠,
α〝ciねd加血α乃≠虚ブα〃dd甜乃0〝タタりノ
ThismeansthatanythingdominatedbysomesegmentofvPcannotbearemnant･Ellipsis
Ofthelowerverbphraseshell,leavlngthehigherverbphraseshell,isimpossible･Thus,the
remnantofpseudogappingcannotbeplacedinspec(vP)oraqoinedtovRbutmustbelocated
inaprqectionthatisdistinct倉omandishigherthanvP
13
Second,aSKennedyandMerchant(1997)note,PPsandAPs,Whichclearlydonot'bear
AccusativeCase,Canbetheremnantofpseudogapplng.
(25)a.hisideamightnotseemcrazytOyOu,butitdoestome
b.Lucyhadtalkedal)OutHungarianmuSicbeforeMartindidaboutBakunin
C.WeWOuldviewitasaliability;buttheywouldasanasset
d.therewasariotonthemall,buttherewasn'tinthegarden
e.Ronalookedmoreannoyedthanshedid丘ustrated
fIwanttolivewithamanmorethanIdowithawoman
￠,′エev加〝タβ￠c如d加励〃〃eゆα〝dAゐ化力α乃f〃タタり
0血Ⅳご励〃〃eみα〃dAゐ打力α〃f〝タタ7ごノ〃ノ
Thisis
a
strong
pleCe
Ofevidence
drivingforce
againstthe､aSSumPtlOnthatthe
ofthe
dislocationoftheremnantisAccusativeCaseasslgnmentOrO句ectagreement･Rather,
【+Focus】seemstobeanineYitablepropertyoftheremnant,aSisnoticedbyBoeckx(2000):
(26)a.JolmateCHOCOLArE,buthedidn'tCOOKIES
b.Jolmatechocolate,butMarydidn't(*?chocolate)
￠ごβoec奴作0ββごノ2q〃
Takingtheseobservationsintoconsideration,.WeCOnCludethatthereisacategoryabovevR
whichisnotrelatedtoCaseasslgnmentOragre血entandwhosespechoststheremnant,and
Adoptlngtheusualass豆mptlOnthatFocPishigher
thatthiscategoryislikelytobeFocP
thanTRwemustassumethatthesu句ectinthepseudogapplngCOnStruCtionisextractedout
oftheTPpastFocPifwearetOPurSuethislipe･Incase(20a),OnePOSSibility,aSSuggeSted
byWatanabeAkira(p.c.),istoallocatethesu句占ctよstheouterspecofFocP･Underthis
analysis,thepositionoftheauxiliary,Whichissandwichedbythetwofociwillbeamystery･
LetustentativelyassumethatthepositionoftheouterspecinerandtheheadofFocPare
switched.Thestructurewi11belikethefbllowlng.
14
(27)(theDAprovedJonesguiltyand)theAssistantDAwillSmith
FocP
/へ
Foc,
theassistantDA
/へ､
f'oc'
will
/へ
TP
Smith
/へ
T'
他母乳S昏i離即池辺A
/〈＼
祀
馬貯
/へ
他℡風S$五離乳m先取Å
Ⅴ,
/〈＼
p『①∇⑮ Ⅶ
/＼
$Ⅲ血肋
Ⅴ,
/へ＼
t-/か⊥フ
ThereasonwhythevPiselidedobligatori1yinthepseudogapplngCOnStruCtionmaybe
relatedtothedrivingforceofthedislocationtospec(FocP)Foc.WbcaneaSilyconceive､that
thedrivingforceofthesedislocationsistocreateanunsaturatedpredicateinvRwhichmust
then
be
elided
under(parti?l)identitywith
the
antecedent
vP
This
situation
may
be
COmPared･tOtheobligatoryvPellipsISwithdb-insertedsentences.
(28)JolmatechocolateandBilldid(*eatchocolate)too
ⅦthoutvPellipsis,db-insertionin(28)losesitsmotivation･Sinceellipsiscanbee?SeenaS
aneconomyreq山est丘omthePFinterface,the
argumentisinlinewiththerecenttrend,
representedbyReinhart(1993)andFox(1998),tOattributethemotiveofdislo云ationsto
interfaceneeds.
4.Spec(TP)
Inthissection,WereeValuatetheheadthatAgreeswiththesubjectinspec(TP)andassigns
Casetoit.IthasbeenassumedintheliteraturethattheNominativeCaseasslgnerandthe
15
NullCaseasslgnerareT.
(29)a･NominativeCaseisassignedasareflexofAgreewith[+Fin(ite)]T.
b･NullCaseisassignedasareflexofAgreewithトFin]T.
However,aWiderrangeOfconsiderationattestsproblemswithboth(29a,b).Considerthe
VarioustypesofTPsin(30):
(30)Caseassignmenttospec(TP)+++
FinwithinT
_
distinction
+
b.inflectedinfinitive
c.conけ01i血扇ve
d.ECMinfinitive
e.raisinginfinitive
asslgnedCase
N｡mN｡mM血n｡ne
a.indicative/subjunctive
岬幽..
±Past
First,innectedin丘nitives,WhichhaveトFin]TandaNominativesubject,atteStthat(29a)is
OnlyapartialdescrlPtionofthedistributionofNominativeCase･ThistypeofTPisfoundin
languagessuchasEuropeanPortuguese.Second,aSWeSeefurtherinsection4.3,ECMand
raisingin丘mitivesposeaproblemwiththeaccountofNullCaseby(29b).
Contraryto(29),WeCOnCludethatwhatassignsCasetotheelementinspec(TP)isthehead
directlyaboveTP:CforNominativeCaseinindicativesandNullCaseandDforNominative
CaseininflectedinBnitives.
イ･ノ･戯Jむec∠げ血dkα如eCね〟∫e∫
Probably,thestrongestreasonthatT'hasbeenassumedtobetheCaseassignertospec(TP)
isthatitagreeswiththesuqectinspec(TP)inindicatives/subjunctives,and,mOreOVer,that
thesubjectCasebecomesnon-NominativewhentheTlosesagreementinin丘nitives.Thus,
thereisacorrelationbetweenthemorphologyofTandthatofspec(TP):
(31)a･[TP(Maryis/Iam/weare/youare/theyare)[vpattendingtheconfbrence]]
b･Jolmpersuaded(Maryl/me2/us3/you4/them5)[TPmO(1/2/3/4/5)tO【vpattendthe
COn危rence]】
C･Johnexpected[TP(Mary/me/us/you/them)to[vpattendtheconference]]
d･(Maryis/Iam/weare/youare/theyare)likely[TP(MaEy/Ⅰ/w@/y⑬u/他@y)to
[vpattendtheconference]]
Howeveちthis
agreement
property
ofTis
notits
privilege.First,aSis
wellknown,in
EnglishindicativeswithoutAukiliaries,V;1nSteadofTこagreeswiththesuqect:
16
(32)[TP(Mary[vplikes/Ⅰ【vplike/we[vplike/you[vplike/they[vplike]KentuckyFried
ChickenbetterthanMcDonald's]]
Second,invariousdialectsofDutch,GermanandFrisian,C,aSWellastheverb,agreeSWith
tbesu切ect:
(33)a…･[cpda-@j
[TPgie kom-t]]
血ん2臓γ0〟 γ0〟CO桝e-2馳｢陥∫J飽桝功
b…･[cpda-Q)j
[TPgunderkom-t]](Q)<t)
血払ガ句′0〟
γ0〟
C(珊ピー2PJ
[TPij kom-t]]
C…･[cpda-tj
肋αんj勒-カeカe co研ピーj馳
d…･【cpda-Q)-Se
[TP萄 kom-t]](Q)<t)
肋αトj師e
成gco桝e-j馳
e…･【cpda-n-k
[TPikkom-en]]
娩αトノ勒ナナco椚e-ノ馳
f…【cpdaTG,-me[TPWunder
kom-en]](Q)<n)
肋αトノア山γe
g…･【cpda-n-Ze
血相邦励リ
We
C()朋β_ノダJ
[TPZunderkom-en]]
ム桓′ 00昭一jPJ
砂肝≠〝タタ7ごノj動ノ
Regardingthepossibilityofagreementwiththe
suqectinspec(TP),C,Tandvareall
positive;theyallagreeatleastonsomeoccasion.Thus,1tisinvalidtoattributetheCase
assignmenttospec(TP)toTsolelyforthereasonthatitagreeswiththeelementinspec(TP).
NordoesthereseemtobeanyotherconvinclngreaSOntOaSSumeSO.Thefactisatleast
COmPatiblewiththeaccountthattheCaseassignertothespec(TP)in(31a,b)istheCoutside
theTP
Complementizeragreementseenin(33)suggeststhatitisnotunnaturaltoassume
thatthereisacompleteq)-SetinCinlanguagesevenwhentheydonotovertlyagree･
Infact,thereseemstobeacrucialcorrelationbetweentheexistenceofCandtheCaseof
SPeC(TP)･Asin(34a,b),WheneverthereisaCdirectlyabovetheTR
tbesuqectin
SPeC(TP)bearSaCasewhosesourceisobviouslywithintheCP.Wheneverthereisnot,aSin
(34c,d),thesu句ectbearSaCasethatisassignedbysomeelementinthematrixclause.,
(34)a･[cpCMaryisattendingtheconference]
b･JolmpersuadedMary[cpCmOtoattendtheconference]
C･Jolmexpected[TPMarytoattendtheconference]
d･Maryislikely[TPMaFytOattendthec9hfbrence]
ThisfactatleastsuggeststhatCistheCaseassignertospec(TP)inindicativeandcontroICPs
andthatTisnotimplicatedinCaseasslgnment.
17
イ.2.蝕むgc￡げ坤cおd瑚〝招γe∫
combinationトFin,+q)]within T,arefoundin
languagessuchas伽ropean'portuguese.AlthoughtheTムfanihflectedi血itiveisLLFih],
Inflectedin丘nitives,Which
have
the
itssubjectmay,andinfactmust,bearNominativeCase,unlikethesuqectofnoninflected
in丘nitives,WhichmaybearNu11Case･Inflectedin丘nitivesanduninnectedin丘nitivesare
distinguishedbythepresenceofanagreementmOrPhemeonT･
a
diGcil[eles
aprova-r-em
(35)a.ewIsera
PrOpOStaト
W況∂β櫛c〟〟肋邸油川聯和γβ一わーjタJ血ノダ′甲0∬J
a
b.*e甲Isera
di鮎il[contigo aprova-r-eS
prOf)OSta]
wよ〟ゐe櫛c〟ゐγ0臥0ゐJ(鞘,叩Vピーわー2詣伽ク托pO∬J
di鮎il[且⑬$aPrQYa-r-em
a
PrOPOSta]
(甲,和VピーわーjPJ伽ク′甲0ぶαJ
w払肝エc肋c.0む助ね柁櫛C〝ノブ
a
d･e甲Isera
diBcil[PRO
aprova-r
PrOPq?ta]
C･*epIse-los一色
山荘∂e顧c油
(甲,和Vピーわ血ク′て甲0∬J
鱒呼0∫0〃タβ7･●β釧
(36)inflectedinfimitivesinEuropeanPortuguese
a.cu
come-r-
エ池沼eαトわーノ掬
b.tn
come-トeS
γ0肌肋椚eαトわー2馳
c､ele
鋸伽〝
come-r-0
d.n6s
come-r-mOS
印相両地
we.地肌占αトわーJ力
e.v6s
come-r,des
γ0払Jわ桝eαトわーヱPJ
f eles
come-r-em
助郷肋椚eαトわーjPJ
Atthispoint,itisobviousthat(29a)doesnotcoverthewholedistributionofNominative
Case･(29a)asadescriptionofNbminativeCaseassignmentmustbesomehowmodiBed･
AsRaposo(1987三86)reports,innectedinnnitivesseemtoappearonlyinCasemarked
envlrOnmentS:
エ乎eCげ7ア
a
diBcil]
(37)[TP[elesaprovarem
proposta]sera
肋町(坪和Ve-わ一jアJめ叩托pO∬Jw〟エみe櫛C〟JJ
Raposoanalyzes(35a)asanextraposed-Subjectversionof(37)･
18
〟co〝甲ゐ桝e〃‖0タ
(38)ed■ehtt占i6mcasa[ppsem[os.meninosverem]]
J
g〃Je柁d肋e
力0〟ぶe
Wブ助0〟J肋e c加納〝∫eピーわーjf)J
J荘co叩Pゐ椚β〃Jわカc肋′ee椚0如ピタ作成cαね∫
(39)a.eulamento
[os､deputadosterem
trabalhadopouco]
力餅e-わーjPJwor丘gd
加ゎ
娩e`桓フ〟如∫
J聯J
b･*o
Manel-(pensa/drmo/
肋g
助成
血i研
des句ava)[os
ルね力gd
陶c如ee椚0ぞルりニー
levado
arhigost?r.em
力血タ≠e〃ゐカ肌ピーわーギブ地力
｢卸ねね桝jり仲cわmかり什0加0〃αク
dlivro】
肋e占00ゑ
(40)examplesoffactiveemotive･Predicates
lamenter,deplorer,CenSurer;aPrOVar
,
J聯オーわ(≠勿わ柁-わCe〝∬柁-わ聯和γe-わ
The
fact that
a
DP
can
factive
be the complementto
emotives,but
notto^ePISt如ics,
declaratives,and.volitionals,朋i町(41,),TClariBesthatit･isCa畠emarkingbythematri音Verb
thatisthecruCialfactorofacceptabilityin(39).
(41)a.n6s･1amentamOS
[DPOS pedidosdosJesuitas]
We.牒酢甘
∵･∵∴側紺叩郎呵卿れ如血仁一劇画東研戯画■
b･n6sla甲entam0竺[DPO(factode)[TPeles,terem
reCebidopoucodinl1eiro】
｢血血d∴･肋拍坤肥血村政最初融
｡.､器一志{｡enSよノY伽藍g裟慧忠霊て謡霊
和鱒血卯桓ゆ`坤γり､作0〟如〃α〃
触㌧∴∴娩油ト∴
d.*o
伽
∵血血卜㌧
Manel〈pensa/afirmo/desqiava)[｡PO(facto.de)[TPeles､receb占rem
娩甲柁Ceれ昭一わ二jアブ
∠加〃ゑ c血i椚 Wね力ed
助eカcJげ
poucodinheiro]
〟め
加0〝町
田毎0∫0〝夕β7.･夕訓
Inflectedin丘nitivalsalsqappeqrascomplementtoCRinwhichcaseitislessclear,under
OrdinaryassumptlOqS,thattheTPisasslgnedCase..Inthiscase,inversio中一isobligatoryand
theCmustbeneithermatrixC,OVertCnor【+Wh]C.
イFco〝甲ね椚β〃fわクJ呼0∫蕗b′‡αJ∝〃OHmαぬCP
(42)a･eurlamento[cpterem[osdeputados知弧tral)alhadopouc9]]
b.oManelpensJa[cpterem[TPOSamigos知弧1evadoolivro]]Rpistemiq)
19
C.OManelafirmo[cpterem[TPOSamigos触打倒孤1evadoolivro]](抜ckzrativり
d･*oManeldes句ava[cpterem[TPOSamigos鹿町◎mlevadoolivro]]在oHtionaL)
伊呼0∫0〝タβ7.●タりノ
(43)a.*【cpC
eles
a
approvarem
肋印肋桝(準p和Ve-わーjPJ血ダ叩0∬J
b.*eq)lsera
di鮎il【cpque
eles
a
aprovarem
PrOPOSta]
娩αf肋砂油川(撃p和γβ一わーjPJ肋叩J甲P∬J
W擢∂e櫛C〟〟
C･*eunao
Jゐ〝富加釧
p;oposta]
jantar]
触感〃〃ビア
助町ヱ〃Vブね一わーjPJ♪r
sei[cpquemCeles90nvidarem
W力0
ParaO
RaposoarguesthatTcanbeasslgnedCa畠eifitisdislocatedtoCandisasslgnedCasethere.
TtoCdislocationisassumedtobeimpossibleiftheTPisirrealisasip(42d),OriftheheadC
isovertasin(43b)orspec(CP)is丘11edasin(43c),thedouble-COmplementizermterbeing
thereasonfor(43b,C).Itisalsoassumedthat,血ilethematrixverbcanassignCasetoCin
(42),thereisnoelementthatassignsCasetoTorCin(43a).
BasedontheobservatioムsinI-ⅠVRaposomodi丘es(29a)asin(44)･(recastinmodern
tems):
(44)NominativeCaseisassignedasareflexofAgreewithTif
a.Tbears【+Fin],Or
b.TbearS【+q)]andTitselfisassignedsome(notnecessari1yNom)Case.
(44b)assumes
a
倉ameWOrk.Let
us
sortofCasetransmission,Whichcanhardlybeadoptedinthepresent
directly observedis
reinterpret Raposo's
account.､Whatis
that
NominativeCaseasslgnmenttOthesu句ectofaninnectedin丘nitivelSPOSSibleonlyinaCase
marked
envirdnment:The
most
faithfu1account
ofthis
observationis
thataninnected
infinitival甘Pincas由Ⅰ-IIIisinfactselectedbyacategorythatmustbeCasemarked,namely
DRandthatitsheadD
Case
enablestheNominative
asslgnment.Inconnectiontothis,
(41b)withtheomissionofbciodbis､SuggeStive.'ItclearlyshotvsthataDPcanselectan
inflectedinfihitival-TP
Therefore,OurSOlutiムntoI-IIIis(45)‥
(45)InlanguageSthatallowinflectedinfinitives,thereisD,Whichmaybeovertasin(41b)or
nullandhasthefo1lowlngPrOPerties:
a.ItcanselectトFin,+q)]TP
b･LikeotherDs,itmustagreewiththeouterenvironmentandbeasslgnedCase･
c.AnalogouslytootherDs,Whichagree扇thitscomplementNRitAgreeswithand
assignsNominativeCasetothespecifierspec(TP)ofitscomplement.
ForIV(45)cannOtbeused.Indeed,IVcanbetakenasanempirica19VidencefbrwhatI
SuggeStedinsection4.1,thatis,CalsoassignsCasetospec(TP)･Consideralso(41b)with
Pctodb･kislikelythat鹿istheCaseassigner,Which-isperhapsaC･Now(42)simply
showsthataninnectedin鮎itive,Whichhastheproperty[-Fin,+甲】,is､somehownot
20
COmPatiblewithirrealisandthattheheadCselectlnganinnectedin丘nitiveTPobligatorily
CauSeSinversion.(43)showsthatmatrixC,OVertCand[十Wh]Cdonotselect[-Fin,+q)]
Tosumup,OurmOdificat享OntO(29a)is(46)‥
(46)NominativeCaseisassignedtospec(TP)asareflexofAgreewithCorDthatselects
【+甲】TP.
4･j･助吻庭中グCb〝如才Cb〟∬∫坤r…わCぬ如朗
AsnotedbyWatanabe(1993),thein丘nitivalTaloneisinsufncienttodistinguishthecases
OfNullCase
asslgnment
VS.Accusative
Case
Case
asslgnment/no
tO
asslgnment
the
infinitivalsu句ect.
(47)a.Jolmtried[cpC[TPmOtowintherace]]
b.*Jolmbelieved[TPmOtohavewontherace]
C.*itseems(toJolm)[TPmOtobeinte11igent]
IfiobyitselfwerecapableofassigningNullCasetospec(TP),theunacceptabilityof(47b,C)
WOuldbeamystery.ThecruCialfactordistinguishingthese
casesisthepresence
ofC.
Thus,Watanabe(1993,1996)elaboratesaCasetheoryinwhichCmustbeinvoIvedinCase
asslgnmenttOSu句ectswithNullCaseorNominativeCase.Heworkswithinthe倉amework
inwhichitisbelievedthatCaseasslgnmenttakesplacebasicallyinahead-SpeCCOn丘guration.
Perhaps
solelyforthisreason,he
assumesthatthe
Case
assignerto
invoIveCinCaseasslgnment,heassumesthatTthathasasslgned
spec(TP)is
T.Tb
Casemustgointoa
fo1low-uPrelationwithC･EvenifTin(47b,C)hastheabilitytoassignNullCase,theT
WillnotbeabletobeinvoIvedinthefollow-uPprOCedureintheabsenceofC.Tlms,the
PreSenCeOfCcorrelateswithcontroIvs.ECM/raisingdistinction,thoughonlyindirectly.
Sincethereis
morphologicaldistinctionbetweenthe tosin(47),tO make
distinctionbetweenioin(47a)andioin(47b,C)canbemotivatedsolelybytheory-intemal
no
overt
a
reasons･Occam'srazorprefersthetheorythatdoesnotmakethisdistinction.Watanabe
apparentlychoosestheprefbrableoption:heassumesthattoin(47)uniformlyhastheability
toasslgnNullCase.However;thesimplicityisbrokenwhenweconsiderthein丘nitival
COmPlementizerhr:
(48)[cpfor[TPhimtosoIvetheproblem]]isnotimpossible
Thoughtoin(48)ismorphologicallyidenticaltothatin(47),itisinevitabletoassumethat
the
toin(48)is
di飴rent&om
the
onein(47)under
21
his
theory;in
thatit
has
an
Accusative-CaseasslgnlngPrOPerty.Rather,1tSeemSthatin丘nitivalC,unlikeT,COrrelates
WiththeCaseofspec(TP):
(49)a.Jolmtried[cp(C/*for)PROtowintherace]
b.itisnotimpossible[cp(*C/for)himtosoIvetheprムblem]
C.Iwa山(払りhmtoleave
Whenweexaminetheparadigmin(47-49),itbecomesclearthatthereisnoempirical
reasonto
assumethatT
playsanyroleindistinguishingthe
cases.The
factisthat,the
existenceofthenu11in丘nitivalCissufncie山toaccomplishNullCaseasslgnm6ntand､the
existenceofthein丘nitivalc?mPlementizerPrissufncienttoaccomplishAccusativeCase
asslgnment･Thisis
also shownbythe鮎tthattherole
ofTa畠a
Case'a?Slgnerin
Watahabe'stheorylStitular.Thenbe,the-simplestaccd血tisthatCit畠elfPistheCaseasslgner
tospec(TP):
(50)NullCaseisassignedasareflexofAgree､*itbthehvllCt坤亭elects[一叶Ti>･
4.イ.蝕むec∠q′且αけ
Without.anythingad￠edtothe､Ordin町y往CCOuntin(51a),BCMsubjects丘tintothelooser
to
believe､that･the
reason
account,narPely(6).･However;thereis
POSt-OPPrationalstructureisratherasin(51b),inwhichtheECMsu句ectis′disIpcatedto
VerSion
of
our
spec(VP).Thence,､thiscaseagainrealizesレtherestrictedcよsen.,=lof(6)･.
Thus,Lasnikand
Saito(1991)observe,byusingBindingConditionsC
andA,thatthe
SubjectofECMin血itivesisdislocatedtoapositionthat?-C?mmandstheadjuncttothe
matrixclauseandthatthisdislocationisconnectedtoAccusativeCasecheckingbythematrix
verb,aSCanbecon丘rmedbythecontrastivebehaviorofthenominalasslgnedAccusative
Casebyanelementotherthanthematrixverb:
(52)a.?*JoanbelieveshimltObeagenius[evenmoreferventlythanBob'岳1niotherdoes]
b.?JoanWantS(C/for)himltObesuccessfu1【evenmoreferventlythanBob'slmOther
does】
c.?IbelievedthoselmentObeunreliable[becauseofeachother'sIStatementS]
d･??*Iwantedtho軍?lrP?n(C/for)tobeRred,[やecauseofeachother'sIStat?mentS]
￠ぁわ戊α〝d励まわ〝タタJづ2γ32鋸
Toyieldthecorrectc-COmmandcon丘guration,thesu句ectofthein丘nitivemustもe･locatedat
least ashighas
spec(VP),eVenifwe
acceptthe
assumptionthatthereis
athirdverbal
segment,tOWhichana句unctcana4ioin.Thus,WeareConsideringthestrudure【1Suthaslin
(53):
23
(53)
vP
/へ､
∬⑳忍皿
Ⅴ'
/〈＼
believes
Vp
/へ＼
hm
V,
/へ
V
W
〈
VP
a句unct
/へ＼
V
TP
一_′へ
払五mtobeagenius
5.A5SOCiateo一触
Chomsky(2000,2001)analysesthe･eXPletive-ihe7?COnStruCtionasinvoIvingNominative/
AccusativeCaseasslgnmentOftheassociatebyT/v
Underthisaccount,thee坪1etive
thereisnotassignedCase.
(54)a.thereislikelytobeaproblem
l
J
b.weexpecttheretobeasolution
HoweveちthereisevidencethatshowsthattheexpletiveiheTeisasslgnedCaseatleastin
EnglishandSpanish,thoughthismaynotbethecaseinIcelandicandGerman,Whichhave
expletiveconstruCtionswithunergative/tranSitiveverbs.First,mOrethanoneProbecanbe
relatedtoaslngleassociate:
(55)a.therelooks[cpasifthereisaproblemwiththisanalysis]
b.therelook[cpasifthereareproblemswiththisanalysis]
｢Gねα≠〝タタタづ勃
SincetheassociatecannotAgreewithandbeasslgnedCasebymultipleProbes,SOmething
elsemustbeAgreelngWiththeProbes.Theexpletivethe7eS,Whichcorrespondone-tO.One
withtheProbes,muStbewhatAgreeswiththeProbes.
24
Second,theassociatemaybeanAccusativecliticintheSpanishnullexpletiveconstruCtion,
whichusesthevetb`have,ratherthananunaccusativeverb.
(56)a.叩Jlo-bめia
Cエん加.鞄功飢e
b.叩Jlos-babian
Cムん加c.タんカ(Ⅳe.PJ
hombre
C.e坪Ihbiaun
力αVe α 椚α〃
d.eqlhabian
dos
hombres
カの′g･PJ加0桝e〝
仲α〃ね砂
｢G和αJ〃タタタ.･j刃ノ
TheNominativeCasemustthenbeasslgnedtosomethingelse,Whichmustbetheexpletive.
SincethesefactsshowthatexpletivesareasslgnedCase,itfo1lowsthattheassociatesare
asslgned
Case舟om
somewhere
most
else.The
strongly
accepted
alternativeisthatthe
associateisassignedPartitiveCasebytheunaccusativeverb,aSPrOPOSedbyLBelletti(1988).
ThisanalysisisalsosupportedindependentlyofthefactregardingCaseasslgnmenttOthen.
First,thede丘nitenesse飴ctisobserved;theassociate'DPmustbeweakinthe
senseof
Milsark(1974,1977).Thisfactfo1lowsiftheassociatehastobeassignedPartitiveCase.
(57)a.thereisa/*thewolfatthedoor
b.therewassomeone/*everyoneintheroom
C.*therewasthelawoftheexcludedmiddleexplained
d.*therewasFrank'sarticlementioned
e.*therewereJolmandMarycycilngalongthecreek
f*therewereal1viewpolntSCOnSidered
g･*therewaseachpackageinspected
Second,theexistentialconstructionrequlreSaCertaintypeofverb.Thisfact
seemsto
ShowthattheassociateneedstobearaCasethatcanbeprovidedonlybythislimitedrangeof
Vefbs･Theclassofvetbsincludeinherentlyunaccusativeverbs:
(58)a.weconsiderthere(1ikely)*(tobe)amanintheroom
b.Iwantthere*(tobe)someonehereat6:00
C.IwantSOmeOne(tobe)hereat6:00
d.?there*(arrived)abus
￠α∽戊〃タタ2.･jβ`抄
The
classis
parameterized
with
respect
to
thelanguage;itincludes
unaccusative)verbsinItalianbutnotinEnglish:
25
passive(derived
(59)a.る
stato
messo
血び∂ee〝ク〟≠
un/*･i卜Iibro
sultavolo
血∂00鬼0ク班￠ぬ占ゐ,仲〟頑..･
α
b.*therehasbeenputabookonthetable
直.･βビ/んメタりノ夕餉ごり.･占ご上α…丑/ノ州ユ●jリ
ThiscanbeexplainedbyassumlngthatinItalian,butnotinEnglish,･Accusative.vetbscan
asslgnPartitiveCaseandthatPartitiveCaseisnotabsotbedbypassivization･'
AsnotedbyLasnik(1992),thePartitiveCaseanalysisgivesa'uni丘?dexplaムatiohtosome
factsthat
are
explained
Chomsky(1995,知0).,.Fitst,the
unsystematicallyby
Case
requlrementOfd2eTeeXplainsthedistributionoftheT?:
Chseless
わwerthere伽elqs$
(60)a.*itseemstheretobeamanhere
b.*thereislikelytheretobeamanhere
肋H(力se
c.*ItriedCtheretob占amanh占re
Genetiveαおe′､
d.*there,菖arri)al･Ofa叩an
e.there･isamanhere肋m伽e､
fweconsidertheretobea叩anhere
AcqChe‥
.￠一Cごエム∫〃鱒〝タクマづβ2てj叫￠･_由ら椚卸什
Second,Sintetheassociateofanexpletive,aSOPPO寧edto,anECMsubject,.isassl!ned
PartitiveCas6initsbaseposition,itbecome主inactiveandlcannotmovefurtherfbrCase
reaSOnS:
(61)a.*thereislikely[someonetobe･$⑳m母⑬阻母here]
b.thereislikely[他⑪町⑪tObesomeonehere]
c.Ibelieve[someonetobe$⑬m⑪⑬m⑪here]
d.*Ibelieve[tobesomeonehere]
ThisexplahationhakesunnecessarythelaSSumptlOnOfapreference､ofMergeoverMoveandi
accordingly,theconceptofNumerationandSub⊥AmyassumedinChomsky(1995,2000)･･
However,thepictureisnotthateasy･AspointedoutbyHowardLasnik(P･C･),mOVement
ispossiblewithconstruCtionswithouttheexp16ti)e:
(62)someoneis【s⑬m⑪⑬m⑰here]
ThisshowsthatpartitiveCaseasslgnmentrequlreSaneXPletive･Sincethereisnoexpletive
in(62),partitiYeCas6assignmentcannOttakeplace･1･Hencemovementofsomeoneis
allowed.
Caseis
Hence,WeadoptthehypothesisthatPartitive
asslgnedtotheassociate･In
consideringthepost-OPerationalcon丘gurationwithrespeCttOPartitiveCaseasslgnment,the
26
POSition
ofthe
associate
the
and
Probe
must
beidenti丘ed.Regarding･theLformer,■the
extractabilityofthewh-phrasein(63a,d,e)atteststhatthecategoryfo1lowingt
expletiveconstruCtioninthesecasesisasmallclauseandnotaslngleDR-sidihgwithaclear
instanCe(63b)ofthesma11clausecomplementconstruCtionandopposin苧Cle?r?aSeS(63c,g,
h)ofsimpletransitives.Wemayassumethateveninthecaseswithoutl■ムcatives,･thereis
SOmePrq)eCtionbetweentheverb▼andtheassociateoftheexpletive.
(63)a.inwhichlakearethere[manyfishinⅦ鮎地且ak⑪]
b.inwhichlakedoyouwant[scsomefishinwhi曲孔ak母]
C･*inwhichlakedidyoudiscuss[DPm?TyB聖聖融攣摘鮎】
d･inqualelagoci
sono[moltipescl皿mquaA母且喝0]
ゐ耽椚讐γβ成 伽柁アビ
｡.芸慧力笠慧…笠椚s｡㌘岩霊草p.母S頼11諾勅
f*quantlPeSCinellagocISOnO[qu弧血p@S￠五皿戚且喝￠】
ね伽山鹿
g.*inqualelagohai[｡Pdiscusso
加w力わ力血お力卯e.γ0〝
h*.quanti,,PeSCi..･hai
月蝕椚叩メ成
moltipeSCi･in耶劇毒乱ag⑳]
d鮎c〟∬ed椚α町脚
discus亭0[DP脚弧血p昏琴6五･nel･1ago]
成dγ0〟ゐc〟∬
血伽ぬ如
i.[DPquantipescinellago]haidiscusso[,P御an仕五p@Sゐilm馴喝⑬】
何イJエα∽戊〝タタ2づ夕5-jタ￠/ふブごか仇血〝夕βりc庇df〃エα馴メ鬼〃タタ初
Regarding
the
Probe,the
fact that
Partitive
Case
asslgnmentis
a
peculiarlty
Ofthe
hnaccusative､VerbsandthatPartitiveCase･isim血unetoCaseabso痺t10ninpassivization,
whiehi､Srelatedtoguppressionofthefunctionoftheheadv,iuggeststhatthePartitiveCase
asslgne(isJtheheadVdfthe､lowervetb
segment:NoⅦ;･We
h島Vetwo
post-OPerati6nal
eon丘gurations･:in'the脇eXpletiveconstruCtions::OnerelatedtotheCaseas･Slgnmehtof･ihere
byCorvandtheotherrelatedtbLthePartitiveCaseasslgnmehtofthea畠soeiatebyV:
27
CP/vP
(64)
/〈＼
/へ
T'
/へ＼
T
l
/へ
…VP…
/へ＼
VP
Ⅴ
/へ＼
業
Ⅹp(=Smallclause)
/へ
準重責 Ⅹ
BothofthemareinstanCeSOfP-SpeC(cmp(P))con丘guration.TheagreementbetweenT/v
andtheassociateofthenremainsunsoIved.
`.Spec(CP)
Intheprecedingsections,WereCOnSideredthepost-Op寧rationalcon丘gurationsin(2)and
defended
take theform
which
asin(4),allof
P-SPeC(cmpげ)),Inthissection,WeeXtendthedomainofconsiderationtotheA-barsystem･
the
position
that they町e
aCtual1y
Namely,WeCOnS車derthepost-OPerationa!conngurations､COnCerningthe[+Wh]featurein
SPeC(CP)andsuggestthattheyalsotaketheformP-SPeC(cpp(P))･
`J.加ね〝聯ガWα〃d月e血加e(わ紺地椚
Itiswidelyknownthatamatrixv占d)eitherprohibits(65a-d),OPtionallyallows(65e-h)or
requires(65i-1)thespeci丘erofitscomplementCPtobefi11edwitha[+Wh]element･The
divisioncrosscutsthetypeoftheembeddedclause;(65a,b,e,f:i,j)haveindicative,Whereas
(65c,d,g,h,k,1)have･COntrOlinBnitives:
(65)a.Ibelieve[cpthatMarytookaclassonsyntax]
b.*Ibelieve[cpwhichclassMarytook]
C.Iprefer[cpC即Ototakeaclassonsyntax]
d.*Iprefer【cpwhichclassCPROtotake]
e.Iknow[cpthatMarytookaclassonsymtaX]
【Ⅰknow[cpwhichclassMarytook]
g,Billdetemined[cpCmOtotakeaclassonsyntax]
28
h･Billdetemined[cpwhichclassCmOtotake]
i･*Iwonder[cpthatM4rytOOka.classonsymtaX]
j.Iwonder[cpwhichclassMarytook]
k･*letusconsider[cpCPROtotakeaclassonsyntax]
1･1etusconsider[cpwhichclassCPROtotake】
Thoughthisvariabilityofclausesisassumedtobeduetoacomplexselectionofclausetype,
itispossibletoeliminatethe[iWh]distinction倉omselectionbyattributingthesourceofthe
VariabilitytoAgree,withuninterpretable[+Wh]inX=Vin(66)asPandtheWhelementin
SPeC(CP)oftheembeddedclauseas叩):
Ⅹp
(66)
/へ＼
Ⅹ
Cp
/へ
Wh
C,
Thisanalysiscanbeextendedtomatrixinterrogativesandrelative
clausesbyassumlngSOme
externalpr毎ection,XPin(66),inthesecases,WhoseheadXisresponsibleforAgreeingwith
【+Wb】.
丘2･加お〃〝e威αおα抑′q′助ccβ戚veq′CJわβねわcα如〝
ItisknownthattheacceptabilityofWhextractionoutofanembeddedclause■dependson
thematrixverb･Erteschik(1973)observesthattheacceptabilityof(67a)variesaccordingto
Whatvetbfi11sintheunderlinedpart.Theacceptabilityisgivenin(67g-i).
(67)a･[cpwhatdidyou_((to)them)[cpv抽舶thathehaddone血]]
b･??【cpwhatdidthepapereditorialize[cpw払如thatMcGovernhaddoneⅦ他紙]]
C･??[cpwhatgooddeeds′didthepri?Steulogize[cp融威g⑳⑳dd@@dsthatthemanhよd
accomplisbed融成g⑳⑳虚血母⑪ds】】
d･??【cpwhatdidshepurr[cが細融thatFredhadgivenherⅦ他紙]】
e･*【cpwhatdidshesimper【cp血thathomeeconomicswasⅦ血如】]
f*【cpwhatdidAbehseraulutate[cpwh離thatmacrobioticcookingwoulddoⅦ也如for
you]】
g.acceptable:5tVitea7甲Orf,a777tOLLnCe
h･queStionable:grunちhollermumurmumbh,mutterTP'W;SC7℃am,ShouちSigh,SnOri,
∫加〝桝βぢWαゴムw加〝e,ね花e∬C血ブ研
i･bad:◆ク〟叩朗∬(β成ねrね肋e,β〟吻e,COO,ノgeぢr〟椚みゐ,ぷJ〝甲eぢ〟乎,ヴゆ,C和魂
成cねお,加叩∽掬,〟血ぬお,d〝f桝αルビrJ
働ね∫C加ゑ〝タ7jJ∂一組
29
Erteschikgivesapragmaticaccountofthisfact,but･iti･Stempt!ngtQeXPlainitsyntactically･
factisthattheintermediatelanding^OftheWh-Phrasein
The
a
speciner
dfCP
position
dependsontheverbofwhichtheWh-Phraseisthespeci丘erbfitscompleinent･:Thence,We
maysimplyassumethatonlythevefbsthatacceptlmterhediatelandingcan.optIOnallybeara
by
[+Wh]fbature,aCCOmPanied
an"[+anti-Wh-feature]."The[+Wh]feature
triggers
dislocationbftheWh-Phrasetotheintermediatespec(CP),Whereasthe･[+anti-Wh-fbatur6]
PrOhibitstheWh-Phrase倉ombeingle氏thereatsemanticinterpretation･･
Ⅴ'
(68)
/へ､
V
CP
/へ＼
闇租
C'
TheanalysiscanbeextendedtocapturetheappositivecasesoftheComplexNPConstraint,
SimplybyassumlngthatnounsneverbearSuChafeature･
､(69)1N'
/〈＼
N
CP
/へ＼
闇租
C'
7.Case.amdrhase
So軸,WedeclaredCaseasslgnerSSuChasvforAccusaative,CforNominative/Null,D
forNominative,andprobablyiPLforObliqueCase･-･Thesecategoriesarerem血scentofthe
PhaseheadsannOワnCedintheliter血re(Chomsky(2OOO,:2001)):=v*,′C一(andp6rbapsD)･
Oneideathatcbm由tomindisthatthepha占einth占sen畠eofChomskyis?OnneCtedtoCase
asslgnment:
(70)AcategoryisapbaseifF･itsheadassignsCaseto)SOmeCategOry･
Although,(70)correqlydescribestheexcep,fionalb'eh?Viorofw甲kv?asa如sehead,itis
notclearatthisp.ointwhetherthereisempiricalmaterial'tha;.supp6rts(70)･Ifthereisan
instance｡fC｡,D(｡rP)thatd｡eSnOtaSSignCase:andwhoseprdectiondoesnムtbehaveasa
phase,OrifthereisaninstanCeOfaheadotherthanv,C,D;andPthatasslgnS･a-Caseand
whosepr句占ctionbehavesasaphase,thatmightbeevidencefor(70)･Ifthisisontheright
track,thenthereasonforthecorrelationmightbethatassignmentofCaseinducesSpell-Out,
30
WhichisrelatedtothenatureofLPhases.Atpresent,though,(70)isonlyaspeculationand
Wewi11notgofurtherintothisissue.
乱Identi秒img珊
Aswesawintheprecedingsections,thepost-0perationalcon丘gurationsthatwereanalyzed
as(2)turnedoutbeasin(4).Al1con丘gurationsin(4),aSWe11asthepost-OPerational
COn丘gurations
concerningthe[+Wh]fbaturein
spec(CP),taketheformP-SPeC(cmp(P)).
GeneralizingtheempiricaldomaintoincludelongdistanCeCaseasslgnmerrt,thedescrlPtlOn
Ofpost-OPerationalcon丘gurationsseemstobeas(6),with(4)manifestingthespecialcasen=
‖=
(6;rePeated)Post-OPerationalconfigurationsarereStrictedtoP-SPeC(CmPnU))),n=1,2,.‥.
Inthis
section,Weidentifythe
theorem
POSition,a
that
fo1lows
positionthatEPP
負･Om(6)plays
selects
a
as7てP).In
determiningthe
cruCialrole.This
theoremis
"0peration-&ee"inthesensethatitdoesnotevenpresupposeMerge,Ortheoperationthat
amountstoit,tObebottom-uPaSisassumedinChomsky(1995,2000,2001).Nevertheless,
atleast(72)seemstobearequisiteforanytheory.Thenotation(71)isborrowed倉om
Sawada(2000b).
(71)Not如ion
a.oa:theoperation(Merger)thatcreatesα
b･Oi>qf:OPerationoipreCedesoperationq7inthederivation
(72)Assumption:mOnOtOnicityofheadinsertion
IftherearethreecategoriessuchthattheyareCOnneCtedbysubstitution,butnotby
農豊富霊琵霊慧霊黒霊謂1?謬警翠1ニ謬誓顎警
㌔1.
Now,aSIpreviouslynotedinsectionl,(6),withassumption(72),entails(73).
(73)Tbeorem
Apost-OPFrationalcon丘gurationP-SPeC(cmp(P))isP-TU')unlessthenominalinthis
COnfiguratlOnWaSPreViouslyintroducedintothatpositionbythedirectneeds
interfaces.
/へ
黍 cmp(P)
/へ
31
ofthe
Proof
AssumethatthereisaProbePISuChthatitspo･St-OPerational･COn丘gurat主onis
Pl-SPeC(CmP(Pl))andis
further.that
Pl-G(Pl).Assume
the
nominal･G(Pl)=
SPeC(cmp(Pl))isnotintroducedintothatpositionbythedirectneedsoftheinterface.
Then,G(Pl)isintroducedbyMergeconcomitanttoAgree･SinceG(Pl)existsヲPdorto
totheoperaion,invlovlngPl,theremustbesomeProbeP2SuChthatinsertionofP2
Pr9Cedesinsertion
ofPland
AgreeinvoIvingP2CreatS
G(Pl)サi.e､,G(ア1)･=7げ2)･
Under(6),岬2)mustbeoftheformSPeC(cmp?げ2)):{
(i)
/へ
P二去
…ⅩP‥.
＼
/
Ⅹ"
Gげ1)
=Sp諾‰))⊥
Ⅹ
PlmuStbeinsertedinto(i)sothatthecon負gurationPl-SpeC(CmP(Pl))=Pl-G(Pi)larises‥
(ii)㌃I
大
/へ
ア1
/
ⅩP
＼
Ⅹ'
G(Pl)
=7てP2)
=SpeC(cmp乃げ2))
=SpeC(Cmp.げ1))
/へ＼
Ⅹ
Thereis
no
way
to
construCt(ii)&om(i)without
assumpt10nSleadtocontradiction･
violating(72)∴∵Hence,}the
■
(73)tellsusthatthepost-OPerationalcon丘gurationP-SpeC(cmp(P))in(4ⅠⅠ,IV)isi一叩),
32
sincethepositionofthenominalinthesecon丘gurationscannotbeattestedtoanyoftheditect
needsoftheinterfaces,SuChasselection,0-markingorstylisticdislocation･Thus,Wehavea
con丘gurationthatisP一叩)andisP-SPeC(cmp(P)).Thesimplesthypothesisregardingthe
natureof叩)istふatthe岬)thatEPPselectsisumiformforarbitraryP.Ths‥
(7;rePeated)TV)=SpeC(cmp(P)),forarbitraryP･
Thisiswhatweconclude丘omtheargumentsintheprecedings?Ctions･
9.SpeculatiomonⅡowDerivationProceeds
SinceaTargetiscreatedasaresultofAgree-MergeinducedbyaProbe,thepropertyof
beingtheTarget7てP)ofAgree-MergeinducedbyProbePentails(74):
(74),Ir叩)=qtben㌔>0α･
Intheprevioussections,WedefendedtheclaimthatぺX)〒SpeC(CP)and7てC)=SpeC(TP)in
StruCtureSlike(75)andthatぺv)=SPeC(VP)inECMconstructions･
33
(75)
/
＼
Ⅹ
CP
/へ＼
C'
SpeC(CP)
/へ､
C
Tp
/へ､
SpeC(W)T'
/〈＼
T
vp
/へ
SpeC(ヽ中)Ⅴ'
/へ､
W
V
/〈＼
SpeC(W)V
Tbus,Weget:
(76)a.0Ⅹ>oSpe￠(Cp)
b..oC>oSpeC(TP)
c.oV>oSPeC(VP)(inECMconstruCtions).
Ifweadd(77)tothenatureofsyntacticderivations,andsti11maintain(72),(76)1eadsusto
aninterestingviewconcerninghowderivationproceeds.(77b)fo1lowsthespiritofPesetsky
(1989).
a･宗野n13詳解芯欝霊泡P2beheads･Ifo"1>oSpeCq
b.earliness:AProbeinducesAgreeimmediatelya氏eritisinserted.
Firstofall,COmbining(76)with(77a)yields(78):
(7S)a.oT>0甲￠C(VP)
b.0V>oSp￠C(Ⅶ)
Next,WeCanCOnneCtthe"ordering倉agments"in(76,78)undertheconditions(72,77b)to
get
the
pICture
Ofwhat
a
derivation
of
a
clause
wouldlooklike.Interestlngly,the
nrst
assumptlOnSareStrictenoughtoleaveuswithonlytwoalternatives.The
di蝕rent録om
the ordinary
bottom-uP
accountwiththe
34
OPeration
oneis
Merge,eXCept
not
so
thatit
requlreStuCkinglnOfspeci丘ersunderahead;WhichisreminiscentoftuckinglnOfmultiple
SPeCi丘ersinthesenseofRichards(1997).
(79)oV>oV>oSpeC(VP)>oT>oSpeC(VP)>oC>oSpeC(TP)>oX>oSPeC(CP)
Yp
Z'
→
Z'
→
/へ＼
/へ＼
∠二ゝ
Z
YP
Z
YP
/〈＼
∠二ゝ
ノSPeC(YP)Y'
∠二ゝ
The
Itisthata
doesnotrequlretuCkingln,butrequlreSanambitiousview
secondone
derivationproceedsinasimpletop-downfashion･
(80)oX>oSPeC(Ci)>oC>oSPeC(TP)>oT>oSPeC(VP)>oV>oSPeC(VP)>oV
∠=ゝ→
ZP
∠=ゝ
∠=ゝ
→
ZP
ZP
/へ
/へ
SpeC(ZP)Z'
Z'
spec(ZP)
〈
Z
Yp
Thdugh'thelsecbndalternativ占-seemstobeambitioilS,itssilmplicityascomparedto.the-nrst
ohemight'suggestthat<itisth6correctone.■､InfaCt,thisLtyPe6fde;i)ationhasrecentlybさen
developedbysomeinqividuals,includingme(Sawada(2000a))･l
10.Summary
In'thispaper,☆e'teconsideredthe'con丘ghrationsthatarise'a氏6rthらoperationAgree･The
CaSeSObservPdallmanifestaP-SPeC(cmp(P))con丘guration･Takinglongdistance■Case
assignmentintbcohsideration,P-SPeC(cmpnU)))s`eemstosufnceEfo;thedescription-ofthe
tvholeミdisthbutiohoftheconngurationa姦etAgre6･Fromthis,ヤederiv盲
SPeC(Cmp(P)),WhichentailsoP>oSpeC(Cmpげ))｣forcas?占dbetbtheやPirationsp
narro☆syntax.■Assumlng･homogeneltyOfphraseconstrtlCtioh,We､areledt'obP>
ospelc(cmpの)ingeneral･ThisglVeS.u亨aninterestlngViewonhowderivationproceeds･
Deri■vationproceedseither倉omthebott6mup,requlrlngSPeCinetstotuckinbelowheads,
●
35
▲
Or
entirely倉omthetopdown,withnoneedoftucki喝1n.
Re托re瓜CeS
BakeちMark(1997)``ThematicRolesandSyntacticStruCtureS,"ElementsqfGrammar,ed.by
LilianeHaegeman,Kluwer;Dordrecht.
Belletti,Adriana(1988)"TheCaseofUnaccusatives,"Li77guisticlhqui7γ19,1-34.
Boeckx,Cedric(2000)"AnAdditionalNoteonPseudogapping,"EllわSisinCofUunction,ed.
byKerstinSchwabeandNingZhang,Tubingen,117-132,Niemeyer.
Brodie,Belinda
L.(1985)"English
Placementin
Adverb
Generalized
Phrase
StruCture
Grammar,''OhioSiateth,it,eTTゆWbrkilqPEPeT･SillLilqLis(ics31:ShL(#esiltGe71e7･a〟=ed
PhTWeSb7LCiunGmmar,1-63,ed.byMichaelGeis,OhioStateUniversity;Ohio.
Chomsky;Noam(1986)助owle4geげLaTWqge:1ね肋tu7e,Origin,m2dtke,Praeger;New
Ybrk.
Chomsky,Noam(1995)乃eMnimaHstPTtWam,MITPress,Cambridge,Mass.s
Chomsky;Noam(2000)"MinimalistInquiries,"StqpbyStqp,89-155,ed.byRogerMartinet
al.,MITPress,Cambridge,Mass.
Chomsky,Noam(2001)"DerivationbyPhase,"KenHdle;AL昨inLaTWage,1-52,ed･by
MichaelKenstowicz,MITPress,Cambridge,Mass.
Erteschik,Nomi(1973)Onihe肋tuTeqflikmdCbnstTuin加,Doctoraldissertation,MIT.
Fox,Danny(1998)励ono叩γandSbmaniiclhteTPTeiation,Doctoraldissertation,MIT.
Groat,ErichM.(1999)"RaisingtheCaseofExpletives,"WbrkiTqMnimalism,ed.bySamuel
DavidEpsteinandNotbertHornstein,MITPress,Cambridge,Mass.
蝕osu,Alexander(1975)"APleaforGreater.CautioninProposiTgFunctionalExplanations
inLinguistics,''Pqers〆刑the肋asessionon凡ncfionaHsm,ed･byGrossmanet･al･,
ChicagoLinguisticSociety;Chicago.
Hankamer;Jorge(1971)CbnstrainLs
on
Deletionin
Syniar,ノDoctoraldissertation,Yale
Universlty,PublishedasDektioninCoo7成nateSb7LCtuTeS,byGarlandPublishingInc･,
NewYbrk,1979.
Holmberg,Anders(1999)"RemarksonHolmberg'寧Generalizarion,"StudiaLi7智〃istica53,
ト39.
ImanishiT.NorikoandAsanoIchirou(1990)ShououtoSbkldyo,Taishuukan,Tokyo.
Kennedy,ChristopherandJason
and
Merchant(1997)"A肘ibutive Comparatives
Bound
Ellipsis,"Ms.,UniversltyOfCaliforniaatSantaCruz･
Larson,Richard
K(1988)"On
the
Lasnik,甲OWarq(1992)"Caseand
O句ect
Double
Expletives:Notes
エお堺′ね如血曾〟力γ23,38ト405･
36
Copstruction,"LiTWisticlhquiTy19,
toward
a
Parametric
Accqunt,"
Lasnik,Howard(1999)肋imaHstAnaかざis,Blackwell,Cambridge,Mass.
Lasnik,HowardandSaitoMamoru(1991)"OntheSu句ectofIn丘nitives,"CLS27肋tl:
乃e
GeneralSkssion,ed.byLise
M.Dobrin,旬nnNichoIs,andRosaM.Rodriguez,
324-343,ChicagoLinguisticSociety,UniversltyOfChicago.
Levin,NancyS･(1986)A血in一掩rbEllわSisin勒OkenEW,Garland,NewYbtk･
Milsark,GaryL.(1974)励istentialConstructionsinEnghsh,Doctoraldissertation,MIT.
Milsark,GaryL.(1977)"TbwardanExplanationofCertainPeculiaritiesoftheExistential
ConstruCtioninEnglish,"LiTWisticAna桓is3,1-29.
Pesetsky;David(1989)"Language-ParticularProcesses
and
the Earliness
Principle,"Ms.,
れⅡT.
Raposo,Eduardo(1987)"Case
Theory
andInn-tO-Comp:TheInflectedIn丘nitivein
EuropeanPortuguese:'Lil塔IListicDtqlLiry18,85-109.
Reinhart,Thnya(1993)"Wh-in-SituintheFrameworkoftheMinimalistProgram,"肋tuTd
エ叩′聯励椚α〝ガぴ6,29-56.
Richards,Norvin(1997)What
Mbves
W7zen
押切enin
W72ich
LaTWqge?,､Doctoral
dissertation,MIT.
Rizzi,Luigi(1997)"Fine
StruCtureOftheLe食Periphery,"且わmentsqfGTWnmar,ed.by
LilianeHaegeman,Kluwer;Dordrecht.
SawadaTsuyoshi(2000a)TheCeriainyPrinc*leand7卸TDownq,e7udon,M･4･thesis,
TbkyoMetropolitanUniversity.
SawadaTsuyoshi(2000b)"OrderingofOperationsandDirectionsofDerivations,"L吻isiic
ReseaTt:h:拝brkbqPqpersinEhghshLilWistics17,125-151,TheUniversityofTokyo
EnglishLinguisticsAssociation,Tbkyo.
Terazu,Noriko(1975)"CoordinateDeletion,GappingandRightNodeRaising,"SiuLhesin
EMLi7研istics3,19-65,AsahiShuppansha,Tbkyo.
Travis,Lisade扇ena(1989)"NotesonCaseandExpletives,,,paperpresentedattheSecond
PrincetonWbtkshoponComparativeGrammar;PrincetonUniverslty;April1989.
Watanabe,Akira(1993)AG尺TB∬ed伽eTheoTyandlLs血ieTYICtionwiihiheArbar伽tem,
Doctoraldissertation,MIT.
Watanabe,Akira(1996)CbseAbsoTPtionandW7iAg7?ement,Kluwer,Dordrecht.
Zwart,C.Jan-Wbuter(1997)舶クワho中里加Qf柁rbAゐvement:AMnimalist4wIVaChtoihe
SmiarqfDuねカ,KluweちDordrecht.
37