Abs ac
This abs ac a emp s o p o ide an explana ion o he wellknown cons ain , pe asi e in he
Romance languages, which disallows he subjec in wh-ques ions o in e ene be ween a
wh-ph ase and he e b. I is sugges ed ha he di e ence s ems om a undamen al s uc u al
di e ence be ween he wo ypes o languages. Romance, unlike English, p ojec s an abs ac
Cl(i ic) ca ego y abo e TP. Cl is an abs ac ope a o whose unc ion is o «ex e nalize» an
a gumen o a e b wi h espec o he ense associa ed wi h . Cl is iden i ied by ich ag eemen
o a cli ic mo pheme, which accoun s o he p esence o such a ca ego y in Romance and i s
absence in English. I is hen a gued ha he p ojec ion o a Cl-ope a o be ween CP and IP in
wh-ques ions gi es ise o a minimali y e ec . The appendix deals wi h some cases o cli ic-dou-
bling o non-speci ic wh-ph ases in modal con ex s and sugges s ha he cli ic (associa ed o Cl)
unc ions as an E- ype p onoun in such cases.
Key wo ds: wh-ques ions, cli ics, syn ax.
Resum. La p ojecció Cl(í ic) a les in e oga i es
En aques a icle s’in en a explica una es icció o ça coneguda, mol eqüen a les llengües
omàniques, que a que el subjec e a les in e oga i es-qu no pugui in e eni en e un mo -qu i
el e b. Es sugge eix que la di e ència p o é d’una di e ència es uc u al onamen al en e dos
ipus de llengües. A les llengües omàniques, a di e ència de l’anglès, es p ojec a una ca ego ia
abs ac a Cl(í ic) pe sob e del ST. El Cl és un ope ado abs ac e la unció del qual és la d’«ex e -
nali za » un a gumen d’un e b espec e al emps associa amb . El Cl s’iden i ica pe la con-
co dança ica o un mo ema clí ic, que dóna comp e de la p esència d’aques a ca ego ia a les
llengües omàniques i la se a absència en l’anglès. Es de ensa que la p ojecció d’un ope ado -
Cl en e el SC i el SI a les in e oga i es-qu dóna lloc a un e ec e de minimi a . A l’apèndix es
ac en alguns casos de doblamen de clí ic de mo s-qu no especí ics en con ex os modals i s’apun a
que el clí ic (associa amb Cl) unciona com un p onom de ipus-E en aques s casos.
Pa aules clau: in e oga i es-qu, clí ics, sin axi.
*This a icle (wi h he excep ion o he appendix) will also be published in he P oceedings o he
Con e ence «On Romance In e sion» (1998), edi ed by Hulk, A.; Pollock, J-Y., Ox o d Uni e si y
P ess. I would like o hank Aa ke Hulk o e y use ul commen s and discussion.
Ca WPL 7, 1999 253-277
The Cl(i ic) P ojec ion in Ques ions*
Ma ia Luisa Zubiza e a
Uni e si y o Sou he n Cali o nia. Depa men o Linguis ics
Los Angeles, Cali o nia 90089-1693. USA
[email p o ec ed]
Recei ed: Decembe 13 h 1998
Accep ed: Ma ch 17 h 1999
Ca WPL 7 253-277 13/6/00 12:35 Página 253
1. In oduc ion
The e is a wellknown cons ain , pe asi e in he Romance languages, which disa-
llows he subjec o in e ene be ween a on ed wh-ph ase and he e b. We illus-
a e his cons ain wi h examples om S anda d Spanish, F ench, Eu opean
Po uguese, and I alian:
(1)a.*Qué Ped ocomp ó? (Spanish)
wha Ped obough
b. *Que Pie e a ache é? (F ench)
c. *Que o Ped o comp ou? (EPo uguese)
d. *Che Pie o ha comp a o? (I alian)
The exis ence o such a cons ain can also be app ecia ed in cons uc ions like
F ench complex in e sion (whe e V- o-C has applied); i minimally con as s wi h
English, whe e he cons ain does no apply:1
(2)a. Aqui a- -il (*Pie e) pa lé?
owhom has-nom.cl. (Pie e)spoken
b. To whom did Pe e speak?
In Zubiza e a (1998), I p oposed ha he p e e bal subjec in Spanish is always
associa ed wi h a discou se- ela ed ea u e (« opic» o « ocus» o «emphasis»).
Two u he assump ions will hen conspi e o accoun o he cons ain on p e-
e bal subjec s in in e oga i es: 1) he « opic» ea u e is always p ojec ed abo e
he « ocus» ea u e a he le -edge o he clause and 2) he wh-ph ase is associa-
ed wi h he « ocus» ea u e. I he e o e ollows ha a « opic» subjec mus p ecede
he wh-ph ase when hey cooccu a he le -edge.
254 Ca WPL 7, 1999 Ma ia Luisa Zubiza e a
1. As Pole o (in p ess) a gues, he es ic ion on he posi ion o he p e e bal subjec in in e oga i-
es is independen om he V- o-C pa ame e . This is shown by examples in No he n I alian dia-
lec s such as he ollowing, whe e no V- o-C has occu ed bu he es ic ion on he p e e bal
subjec is s ill p esen .
(i)i ho domandà kol ke (??Ma io) l’ha a
o-himha e asked wha ha (Ma io) he hasdone
Suñe (1994) makes he same poin o Spanish.
Table o Con en s
Ca WPL 7 253-277 13/6/00 12:35 Página 254
1. In oduc ion
2. The Cl P ojec ion
3. The Q-ope a o and Ph-lexicaliza ion
4. The P e e bal Subjec in Romance
In e oga i es Reconside ed
5. Conclusion
Re e ences
The e a e a ious p oblems wi h such analysis. One o hem is lack o gene a-
li y. The cons ain is also ac i e in F ench, a language which lacks a « ocus» p o-
jec ion a he le -edge o he clause I is u he mo e empi ically insu icien . While
he cons ain on p e e bal subjec s holds sys ema ically in oo clauses in mos
Romance languages, he e is a iabili y in embedded clauses. While i applies in
embedded in e oga i es in Spanish and I alian, his is no he case in F ench and
Eu opean Po uguese; e.g., c . he Spanish/F ench con as below:
(3) a. *Me p egun o a quién Juan busca.
(I) wonde who Juan is looking o
b. Je me demande qui Jean che che.
Ano he ac ha he abo e analysis does no eadily accoun o is he con-
as be ween ba e and nonba e wh-ph ases. While ba e wh-ph ases ac i a e he
cons ain on p e e bal subjec s, non-ba e wh-ph ases like which N and how many
Ndo no (c . Amba (1988), Pollock e al. (1998)):
(4) a. (?) Cuál lib o Ped o comp ó?
Which book Ped o bough
b. (?) Quel li e Pie e a ache é
c. Que li o o Ped o comp ou?
(5) a. A cuán a gen e Ma ía in i ó a la ies a?
How many people Ma ía in i ed o he pa y
b. Combien de gens Ma ie a in i é à la ê e?
In his a icle, I will a icula e a di e en p oposal. I will sugges ha he e
a e wo basic ing edien s which in e ac o accoun o he unexcep ional appli-
ca ion o he cons ain in mos Romance languages and i s a iabili y in embed-
ded in e oga i es. I) The exis ence o he so-called cli ic-le disloca ion
cons uc ion pe asi e in he Romance languages bu inexis en in English. We
assume (in he spi i o Ia idou (1990)) ha he exis ence o he cli ic-le dislo-
ca ion (CLLD) cons uc ion in he Romance languages indica es ha he p edi-
ca e-a gumen s uc u e may be syn ac ically a icula ed in hese languages in a
di e en way han in English. While he e bal a gumen s in English a e sys e-
ma ically p ojec ed wi hin he VP, in he Romance languages unde discussion
hey may be me ged wi h an abs ac (lambda) ope a o gene a ed abo e TP. Such
ope a o in u n binds an a gumen - a iable wi hin he VP. We will e e o i as he
Cl ope a o because he a gumen - a iable i binds is ypically iden i ied by a cli ic
o «s ong» in lec ional elemen .2See sec ion 2. II) The Q ea u e in he le edge o
The Cl(i ic) P ojec ion in Ques ions Ca WPL 7, 1999 255
2. No e ha wi hin his iew he opic na u e o he «cli ic-le disloca ed» a gumen is en i ely an-
gen ial o i s g amma ical p ope ies. The discou se p ope y o « opichood» is ex insic o he
g amma ical p ope ies o he cons uc ion.
Ca WPL 7 253-277 13/6/00 12:35 Página 255
he CP ield (which exp esses he in e oga i e o ce o he sen ence) mus be
phonologically lexicalized, bu he e a e mo e op ions a ailable o lexicalizing
a Q in embedded han in ma ix in e oga i es (see sec ion 3). When he Q- ea-
u e binds a a iable, i acqui es he s a us o a syn ac ic ope a o and in e ac s
wi h he Cl ope a o gi ing ise o a minimali y e ec om which he p e e bal
subjec cons ain ollows immedia ely (see sec ion 4).
2. The Cl P ojec ion
Wha is his Cl-p ojec ion? I is nei he a scope-posi ion associa ed wi h a +speci-
ici y ea u e (as sugges ed by Spo iche (1992, 1998)) no an Ag posi ion (as sug-
ges ed by F anco (1993)). We sugges ins ead ha :
(6) Cl is an abs ac ope a o whose unc ion is o «ex e nalize» an a gumen o a
e b wi h espec o he ense associa ed wi h .
Ex e naliza ion o an a gumen ia a Cl ope a o hus gi es ise o he ollo-
wing ype o s uc u e:
(7) [ DPi[ Cli[T [VP ... [V ei ....
The p oposal is hen ha he e is mo e han one way in which he p edica e-
a gumen ela ion may be syn ac ically ins an ia ed:
(8) a. In he amilia way, ia p ojec ion wi hin VP, in which case he DP a gu-
men is me ged wi h V o wi h a p ojec ion o V: [V DP] o [ DP [ V ...]]
b. The DP a gumen is me ged wi h he p ojec ion o a Cl ope a o ha binds
an a gumen a iable wi hin he VP. The DP is coindexed wi h Cl ia
Spec/head ag eemen and he e o e wi h he a gumen a iable ha Cl binds:
[ DPi[ Cli.... [VP ...ei... ]]]
We may e e o he s uc u e ob ained ia p ojec ion o DP a gumen s wi hin
VP as he lexical o small p oposi ion and o he s uc u e ob ained ia me ging o
DP wi h he Cl ope a o as he syn ac ic o la ge p oposi ion. (8b) has as i s p e-
decesso he analysis o he CLLD cons uc ion pu o h by Ia idou (1990).
Al hough he de ails o he p oposal pu o h he e di e s in impo an espec s
om ha o Ia idou’s, hey sha e a basic insigh , namely ha he p edica e-a gu-
men ela ions in he CLLD cons uc ion is a icula ed in a undamen ally di e-
en way om o he cons uc ions.
The mos con incing a gumen in a o o (8b) a e he binding ac s i s epo -
ed in Zubiza e a (1993) o Spanish and ep oduced in o he languages wi h a
CLLD cons uc ion such as A abic (Aoun and Benmamoun (1998)) and G eek
(P. Schneide -Zioga, p.c.). See also Cecche o (1995) and Guas i (1996) on I alian.
256 Ca WPL 7, 1999 Ma ia Luisa Zubiza e a
Ca WPL 7 253-277 13/6/00 12:35 Página 256
This da a shows ha he e is a p e- e bal/pos - e bal subjec asymme y in cases
whe e he on ed accusa i e opic (which con ains he in ended bindee) and he
subjec (which con ains he in ended binde ) o igina e in he same clause.3
(9) a. A suihijo, cada mad eidebe á acompaña loiel p ime
acc his child each mo he mus accompany-acc.cl he i s
día de escuela.
day o school
b. ?*El p ime día de escuela, a suihijo debe á acompaña loi
The i s day o school acc his son will accompany-acc.cl
cada mad ei
each mo he
(10) a. A suip opio hijo, ningún pad e loiquie e cas iga .
acc his own child no a he acc.cl wan s o punish
‘No a he wan s o punish his own child.’
b. ?*A suip opio hijo no loiquie e cas iga ningún pad ei
acc his own child no acc.cl wan s o punish no a he
‘No a he wan s o punish his own child.’
No asymme y is ound when he binde is he ma ix subjec and he bindee
is con ained wi hin a on ed accusa i e objec ha o igina es in he embedded
clause. See (11).
(11) a. A suihijo, ninguna mad eidebió acep a que
acc his son, no mo he should ha e accep ed ha
se lo egañe.
inde . subj.cl acc.cl scold
b. A suihijo, no debió acep a ninguna mad eique
acc his son, no should ha e accep ed any mo he ha
se lo egañe.
inde . subj.cl acc.cl scold
‘No mo he should ha e accep ed ha someone scold his son.’
The abo e da a sugges s ha he e is econs uc ion o he on ed objec o a
posi ion wi hin he clause ha i o igina es, bu ha his posi ion is no wi hin he
The Cl(i ic) P ojec ion in Ques ions Ca WPL 7, 1999 257
3. I has been b ough o my a en ion by Xa ie Villalba ha o some speake s he binding ela ion
in (9a) and (10a) is acili a ed by he p esence o a modal. Fo such speake s, he a ailabili y o a
bound eading dec eases in he absence o a modal, as in he ollowing example:
(i) A suihijo, cada mad eilo acompañó el p ime día de escuela
acc his child, each mo he acc.cl accompanied he i s day o school
This obse a ion, i co ec , dese es u he in es iga ion.
Ca WPL 7 253-277 13/6/00 12:35 Página 257
VP (c . he ill o medness o (9b) and (10b). The on ed objec may be econs-
uc ed o a posi ion abo e he VP and below he p e e bal subjec o he clause
whe e i o igina es (c . he well o medness o (9a) and (10a). The ac ha he e
is no econs uc ion o he on ed objec wi hin he VP o he clause i o igina es
sugges s ha he on ed objec is inse ed highe up in he s uc u e, i.e. in he
Spec o Cl (c . Zubiza e a (1993, 1998)). This means ha he e is no copy o
he on ed objec in he canonical objec posi ion wi hin he VP; he binding
ac s hen ollow immedia ely.
The p oposal we a e pu ing o h he e may be summa ized as ollows:
(12) In languages wi h a nominal «cli ic» o «s ong ag » sys em, a di ec a gu-
men o he e b (nomina i e o accusa i e) is «ex e nalized» by me ging he
a gumen wi h Cl (loca ed abo e T) a he han by me ging he a gumen wi h
V (o a p ojec ion o V) and hen mo ing i ou side he scope o T.
I ollows om (12) ha he p e e bal subjec in languages wi h a CLLD cons-
uc ion o he so unde discussion is also gene a ed in Spec o Cl.4This means ha
in sen ences like (9a) and (10a), he e a e wo Cls: one ex e nalizes he nomina i e
a gumen and he o he one ex e nalizes he accusa i e a gumen . As he binding ac s
indica e, in hese examples he nom- ela ed Cl is gene a ed highe han he
acc- ela ed Cl. The disloca ed objec may he e o e « econs uc » o a posi ion wi -
hin he scope o he nomina i e subjec . Bu i is also possible o he acc- ela ed Cl
o be gene a ed highe han he nom- ela ed Cl. As p edic ed, in such cases he
on ed objec may bind in o he on ed subjec :
(13) a. A cada niño, su mad e lo acompaña á el
acc each child (obj), his mo he (subj) acc. cl accompany- u he
p ime día de escuela.
i s day o school
b. A ninguno de es os niños, su pad e lo
acc none o hese child en (obj), his a he (subj) acc. cl
quiso acompaña .
wan ed o accompany
Hence, he sen ences in (9a) and (10a) ha e he s uc u e in (14a) and he sen-
ences in (13) ha e he s uc u e in (14b).5
258 Ca WPL 7, 1999 Ma ia Luisa Zubiza e a
4. The hypo hesis ha p e e bal subjec s in some o he Romance languages does no o igina e wi -
hin he VP is no a no el one (e.g. Ba bosa (1995)). He e we a e gene alizing he assump ion o
all languages wi h a CLLD cons uc ion.
5. A ques ion ha a ises a his poin is why he s uc u es in (14) do no iola e minimali y. This
ques ion is pa icula ly ele an in iew o he ac ha , as we shall see in sec ion 3, when Cl in e-
ac s wi h Q, i does gi e ise o minimali y e ec s. I would seem ha wo iden ical ope a o s (i.e.
wo ope a o s wi h exac ly he same unc ion) do no in e e e wi h one ano he , bu wo dis inc ope-
a o s may do so. See sec ion 4.3 o u he discussion.
Ca WPL 7 253-277 13/6/00 12:35 Página 258
(14) a. [DPj[ Clj[ DPi[ Cli[T [VP [ ej[V [ ei....
b. [DPi [Cli[DPj[Clj[T [VP ej[V [ei.....
Le us u n back o he gene aliza ion in (12). Why is he e a connec ion be ween
he exis ence o Cl and o mo phological cli ics o s ong ag eemen a ixes? The
answe is ha he abs ac ca ego y Cl mus be mo phologically iden i ied in he
ollowing manne :
(15) Mo phological iden i ica ion o Cl is achie ed by some nominal mo pheme
wi h phi- ea u e con ained wi hin he local T. This may be a cli ic o a «s ong»
ag eemen a ix.
The cli ic o e bal ag eemen a ix is o cou se ela ed o he DP in Spec o
Cl. In he case o he cli ic, his ela ion is cap u ed by he ac ha a nominal cli-
ic is no hing else han a copy o he D and phi- ea u es o he a gumen - a iable
bound o he DP. In he case o he ag eemen a ix, his ela ion is cap u ed by he
Spec/head ela ion ha he a gumen - a iable bound o he DP bea s wi h he e b
a some poin in he de i a ion. See below o u he de ails.
We can now aise he ques ion o wha is he ela ion be ween he ype o mo p-
hological iden i ica ion equi ed by Cl (as summa ized in (15)) and he one impo-
sed by he Func ional P ojec ion Ac i a ion P inciple (FPAP)pu o h by Speas
(1994), Koopman (1997), Pole o (1998), and o he s:
(16) Func ional P ojec ion Ac i a ion P inciple: A ea u e in he head o a unc-
ional p ojec ion is ac i a ed i he app op ia e mo phological ma e ial occu-
pies ei he he head o he Spec o ha p ojec ion a some poin in he
de i a ion.
No e ha he equi emen s on he mo phological iden i ica ion o Cl is s ic-
e han he one imposed by he FPAP. In e ec , a DP a gumen in he Spec o Cl
is insu icien o iden i y Cl. The eason is ha Cl does no encode a (seman ic)
ea u e like o he unc ional ca ego ies. I s ole is pu ely o mal, namely ha o
es ablishing a connec ion be ween a DP in i s Spec and an a gumen - a iable o he
e b. Mo e p ecisely, Cl se es he ole o ex e nalizing a DP a gumen wi h espec
o he T associa ed wi h V. Some mo phological nominal elemen on such a T is
equi ed o iden i y Cl. Ano he way o pu ing i is ha he posi ion o syn ac ic
ac i i y o Cl and he posi ion o i s ea u e-con en a e dissocia ed; i s ea u e-
con en is no unde Cl as one migh expec . Ra he , i is con ained wi hin he T
node immedia ely below i .
The acc- ela ed Cl is ypically iden i ied by an accusa i e cli ic. The e is no
o he op ion in he Romance languages because hese languages lack mo phologi-
cal objec ag eemen on he e b. The accusa i e cli ic is a copy o he D and phi- ea-
u es o he accusa i e a gumen - a iable. I le -adjoins o V p io o he applica ion
o V- o-T (I assume wi h Kayne (1994) ha adjunc ion is always le wa d):
(17) [DPi[ Cli[[Vj acc cli[ +ag ]] T] [VP ... [Vjei....
The Cl(i ic) P ojec ion in Ques ions Ca WPL 7, 1999 259
Ca WPL 7 253-277 13/6/00 12:35 Página 259
The iden i ica ion o he nom- ela ed Cl is no uni o m ac oss he Romance
languages. Some languages (like many o he No he n I alian dialec s) ha e a
nomina i e cli ic which is sys ema ically used o iden i y he nom-associa ed Cl.
The nomina i e cli ic is a copy o he D and phi- ea u es o he nomina i e a gumen -
a iable. Since he nom a gumen - a iable is loca ed abo e he e b in he e bal
p ojec ion, i s copy (i.e. he nom cli ic) canno le -adjoin o he e b. Ins ead i
le -adjoins o he T o which he e b has been adjoined. (As we shall see in he nex
sec ion, his di e ence be ween nomina i e and accusa i e cli ics is impo an
because i accoun s o why accusa i e cli ics a e always p ocli ic, bu nomina i-
e cli ics a e encli ics when V- o-C applies).
(18) [DPi[ Cli[nom cli[[ +ag ]jT]] [VP ei[Vj.....
In languages which lack a nomina i e cli ic (like Spanish and S anda d I alian),
he subjec ag eemen elemen is used o iden i y he nom-associa ed Cl:
(19) [DPi[ Cli[[[ ]ag i]jT] [VP ei[Vj.....
In o he languages, like S anda d F ench, i would seem ha ei he a nomina-
i e cli ic o he subjec - e b ag eemen elemen may iden i y he nom- ela ed Cl.
See (20a, 20b) below. Bu upon close sc u iny, we see ha in F ench (unlike
Spanish and S anda d I alian), he subjec ag eemen elemen is no «s ong enough»
o iden i y he Cl ope a o on i s own. I can only unc ion as an iden i ie in con-
junc ion wi h a coindexed lexical DP ha is «close enough». See (20c).
(20) a. Pie e il a endu sa oi u e.
Pie e nom.cl has sold his ca
b. Pie e a endu sa oi u e.
c. Pie e sa oi u e *(il) l’ a endu.
Pie e his ca (nom.cl) acc.cl has sold
We sugges ha he no ion o «closeness» in oked abo e be o malized in he
ollowing way:
(21) DP and Ag a e close i hey a e coindexed and hey a e con ained in he
Spec o - o in he head o - adjacen p ojec ions a he poin o spell-ou .
To summa ize, I p opose o dissocia e he Cl ca ego y om he mo phological
cli ic i sel . Cl is an abs ac ope a o whose unc ion is o «ex e nalize» an a gumen
o a e b V wi h espec o he T associa ed wi h V. This is achie ed by me ging
a DP wi h a p ojec ion o Cl ha binds an a gumen a iable o he e b. The unc-
ion o he cli ic con ained wi hin T (and, in ce ain cases, o he e bal ag eemen
a ix) is o mo phologically iden i y he abs ac Cl ca ego y. This Cl ca ego y in con-
junc ion wi h he Q-ope a o cons i u e he linchpin o ou analysis o he p e e -
bal subjec cons ain in Romance in e oga i es.
260 Ca WPL 7, 1999 Ma ia Luisa Zubiza e a
Ca WPL 7 253-277 13/6/00 12:35 Página 260
3. The Q-ope a o and Ph-lexicaliza ion
As men ioned ea lie , I assume ha an in e oga i e sen ence has a Q ea u e which
encodes he o ce o he sen ence loca ed a he igh mos edge o he CP ield. As
we shall see in he nex sec ion, when he Q ea u e is coindexed wi h a ce ain
ype o wh-ph ase (such as he ba e wh-ph ases), he Q ea u e acqui es he s a us
o a syn ac ic ope a o . I will u he mo e assume ha a ea u e ha exp esses he
o ce o he sen ence (such as he Q ea u e) mus be phonologically lexicalized.
We de ine Ph(onological)-lexicaliza ion as ollows:6
(22) Ph-lexicaliza ion: A ea u e in he head o a unc ional p ojec ion is ph-lexi-
calized i he head o he Spec o ha p ojec ion is associa ed wi h app o-
p ia e ph- ea u es a he poin o spell-ou .7
We elabo a e below on he no ion «poin o spell-ou ».
The hypo hesis ha ph-lexicaliza ion applies a he poin o spell ou can be
o malized wi hin he gene al a chi ec u al amewo k p oposed in Zubiza e a
(1998: sec ion 1.3.3). I is a gued he e ha he syn ac ic de i a ion p ope and he
PF in e p e a ion accompany each o he o a s e ch, namely, he s e ch ha s a s
wi h S-s uc u e (a poin in he de i a ion whe e he s uc u e in ol ed is a single
ph ase ma ke ) and ends wi h L-s uc u e ( he las ph ase ma ke in he de i a ion
be o e i b anches). As indica ed by he discon inuous line in he diag am below, PF
in e p e a ion begins a S-s uc u e and con inues a e he de i a ion b anches. We
will e e o he de i a ion be ween S-s uc u e and L-s uc u e as he shallow pa
o he syn ac ic de i a ion and o he s uc u es gene a ed a his poin as shallow
s uc u es.
The Cl(i ic) P ojec ion in Ques ions Ca WPL 7, 1999 261
6. I is likely ha ac i e p edica es, which p essupose he u h o i s p oposi ional complemen , con-
ain an Ass(e ion) ope a o in i s CP. This ope a o is lexicalized by he complemen ize , which
explains why i mus be obliga o ily p esen . C . John eg e s (* ha ) Ma y is bald. Complemen s
o p oposi ional a i ude e bs lack an Ass ope a o ; he e o e, hei complemen ize may be absen
in some languages. C . John hinks ( ha ) Ma y is bald.
7. We shall see la e ha no only a mo pheme bu also a ea u e may ul ill he unc ion o ph-lexi-
caliza ion.We shall see la e ha no only a mo pheme bu also a ea u e may ul ill he unc ion o
ph-lexicaliza ion.
(23) (se o ph ase-ma ke s)
Σ-s uc u e (unique ph ase-ma ke )
Λ-s uc u e
PF LF
Asse ion S uc u e
Ca WPL 7 253-277 13/6/00 12:35 Página 261
The ques ion a ises as o how he Q ea u e is ph-lexicalized in (37) and (38).
These sen ences ha e a salien in ona ional p ope y, namely, he e is an in ona-
ional hia us a he igh edge o he in ona ional ph ase. Mo e p ecisely, he pi ch
alls down o he base line and hen ises again, indica ing he p esence o a p o-
sodic bounda y (see Beckman and Pie ehumbe (1986)). We sugges ha he
p osodic bounda y ea u e is a ached o he Q ea u e and se es he pu pose o
ph-lexicalizing i .14
Gi en wha we ha e said abo e (c . (39)), we need o eadd ess he s a us o
he ace o he wh-ph ase in he embedded ques ion in (35). Clea ly, he ace o he
wh-ph ase que unc ions as a a iable in embedded ques ions as much as i does
in ma ix ques ions (c . (1)). We ha e p oposed ha he ace o he wh-ph ase
unc ions as a a iable i he wh-ph ase is associa ed wi h Q. When Q de ines a
p ojec ion wi hin he CP ield, his associa ion is ob ained ia Spec/head ag ee-
men (as a esul o o e mo emen ), bu when Q is adjoined o he ma ix e b, he
associa ion mus be ob ained in some o he ashion. We p opose ha he mecha-
nism ha gi es ise o ha associa ion is he same one ha unde lies he associa ion
be ween he NPI and he ma ix nega i e e b in he ollowing ype o example:
(43) Pie e dou e que Ma ie ai u pe sonne.
Pie e doub s ha Ma ie has seen anybody
In he example abo e, he NPI pe sonne is licensed by he nega i e e b doub
(= ‘no su e’). In o he wo ds, pe sonne and dou e a e associa ed in some way.
We may assume ha his is done by co e neg- ea u e mo emen . Le us assume
ha he wh-ph ase and he in e oga i e e b Q+demande a e associa ed in he
same ashion ( ia co e wh- ea u e mo emen ). The esul ing ep esen a ion would
hen be as in (44) ( a he han as in (35b)). In his ep esen a ion, he lexical in e-
oga i e e b unc ions as he ope a o ha binds he wh- a iable (i.e. he ace o
que).
(44) Je me [Q [demande]]i[quei[Whi[Jeanj[Clj[TP a [VP ej [ache é eihie ]]]]]]
The ques ion ha a ises wi h espec o such a ep esen a ion is why i does no
gi e ise o a minimali y e ec . We sugges he ollowing answe . In (44), he ope-
a o Cl does no in e ene in he ela ion be ween he ope a o demande and ej
because Cl is a unc ional ca ego y and demande is a lexical ca ego y. In o he
268 Ca WPL 7, 1999 Ma ia Luisa Zubiza e a
14. As is wellknown, on ed wh-ad e bs do no ac i a e he «p e e bal subjec cons ain » in Romance
in e oga i es:
(i) Cuando Ped o comp ó es e lib o?
When Ped o bough his book
(ii)Quand Pie e a ache é ce li e?
Gi en he ole o he Cl ope a o as an a gumen -ex e nalize , a plausible assump ion is ha Cl
does no «see» ad e bial posi ions. Ano he al e na i e (p oposed in Zubiza e a (1998)) is ha
wh-ad e bs may be di ec ly gene a ed in he CP ield.
Ca WPL 7 253-277 13/6/00 12:35 Página 268
wo ds, we sugges ha minimali y should be ela i ized no only wi h espec o
he dis inc ion be ween ph ases and heads (c . Rizzi (1990)), bu also wi h espec
o ypes o heads (lexical s. unc ional).
4.2. Ge ing a ound Minimali y: P ojec ing Cl wi hin he CP Field
Gi en he minimali y e ec c ea ed by he p ojec ion be ween CP and TP o a nom-
associa ed Cl in in e oga i e s uc u es like (34), he ques ion a ises whe he a
nom-associa ed Cl may e e coexis wi h a Q-p ojec ion. The answe is ha hey can
coexis bu i he nom-associa ed Cl is p ojec ed highe han he Q-p ojec ion.
Acco ding o Pole o (in p ess), in F ench and in many No he n I alian dialec s,
he subjec may be ealized wi hin he CP ield, as in he examples below:
(45) A qui Pie e a- -il pa lé?
o whom Pie e has-he (nom.cl) spoken
‘To whom has Pie e spoken?’ (F ench)
(46) I me ga domandà Gianni quando che el ien
nom.cl me ha e asked John when ha nom.cl comes
‘They asked me when Gianni is coming.’ (a No he n I alian a ie y)
(47) A en a che Majo ch’ a mangia pi’ an
cl need ha Majo ha nom.cl ea mo e
‘Majo mus ea mo e.’ (Piedmon ese)
The possibili y o ealizing he subjec wi hin he CP ield migh ac ually be
an op ion mo e widely a ailable wi hin he Romance amily. This op ion is made
a ailable by he ac ha he subjec is ex e nalized ia Cl in hese languages. While
Cl is gene ally p ojec ed a he le -edge o he IP ield (i.e. immedia ely abo e
TP), he g amma allows o he op ion o p ojec ing Cl wi hin he CP ield in o de
o a oid a minimali y iola ion, namely abo e he Q ope a o and below he
Wh-p ojec ion in F ench bu abo e i in he No he n I alian dialec s. Thus, he
F ench example in (45) will be associa ed wi h he s uc u e below (be o e V- o-Q
mo emen applies in shallow s uc u e):
(48) [A quij[Whj[Pie ei[Cli[ej[Qj[ilia [ei[pa lé ej]]]]]]]]
In (48), Cl is abo e Q; he e o e Cl does no in e e e in he binding ela ion
be ween Q and ej. The ques ion ha a ises is why Q does no in e e e in he bin-
ding ela ion be ween Cl and he a iable ei. I sugges ha his is due o he so-
called an i-ag eemen e ec discussed by Ouhalla (1993) and e e ences ci ed
he ein. We elabo a e below.
A a iable bound by Cl has he s a us o a p onominal a iable. This is due o
he ac ha he cli ic o «s ong» ag eemen elemen ha iden i ies Cl is [+p ono-
minal]. The e is a pa ame ic a ia ion among languages:
The Cl(i ic) P ojec ion in Ques ions Ca WPL 7, 1999 269
Ca WPL 7 253-277 13/6/00 12:35 Página 269
(49) Some languages allow a [+p on] o be bound by he in e oga i e Q-ope a-
o and o he s do no .
We shall e e o p onouns bound by he in e oga i e Q-ope a o as Qu-bound
p onouns. Spanish allows o Qu-bound p onouns unde ce ain condi ions (see
he appendix below), bu F ench and mos No he n I alian dialec s do no
(see Pole o (in p ess)).15 Thus, in he F ench s uc u e in (48), he p onominal
a iable in Spec o VP (bound by he nom-associa ed Cl) canno be bound o Q.
In o he wo ds, Q is no a possible binde o he p onominal a iable (il,e).
Consequen ly, Q does no in e e e in he ela ion be ween Cl and eiand no
minimali y e ec s a e obse ed.
The ac ha F ench does no allow o Qu-bound p onouns in conjunc ion wi h
he analysis o non-ba e wh-ph ases pu o h in sec ion 4.1 can eadily explain he
con as in (50). In (50a), he ba e nomina i e wh-ph ase is coindexed wi h Q; he-
e o e Q is bound o he nomina i e p onominal cli ic. On he o he hand, in (50b)
he non-ba e nomina i e wh-ph ase is no coindexed wi h Q; he e o e Q is no
bound o he nomina i e p onominal cli ic.
(50) a. *Qui a- - il lu un li e?
who has- he ead a book
b. Quel é oce magicien a- - il en e mé on sec e ?
Wha e ible magicien has- he locked up you sec e
(G e isse (1980), ci ed in Spo iche (1998))
The e a e wo issues ha emain o be add essed wi h espec o he s uc u e in
(48). Why is he nomina i e cli ic obliga o y -despi e he ac ha he nomina i e
DP and he subjec ag eemen elemen a e con ained in adjacen p ojec ions a e
V- o-C applies? This migh be due o he EPP (as sugges ed in Rizzi and Robe s
(1989)). While he DP in he Spec o Cl p ojec ed wi hin he CP ield ul ills he h-
equi emen o he p edica e, i ails o ul ill he EPP equi emen :
(51) a. The EPP applies speci ically o he IP: he IP mus ha e a subjec .
b. An IP subjec is ei he he DP in Spec o TP o a cli ic/s ong ag adjoined
o T. (See Alexiadou & Anagnos opoulou (1998))
270 Ca WPL 7, 1999 Ma ia Luisa Zubiza e a
15. Al hough a e, he e a e No he n I alian dialec s in which a nomina i e encli ic coexis s wi h a
wh-subjec ; c . Bellunese chi é-lo che ien? ‘Who is i ha came?’ Such cli ics a e gene ally in a-
iable in o m in ha hey do no a y wi h espec o gende and numbe and hey also appea
wi h e bs ha lack an ex e nal a gumen ; c . quando e pa e-lo che ‘l sia i à? ‘When does i
seem o you ha he has a i ed?’ This sugges s ha such cli ics a e no ela ed o Cl a all, bu
ha e he s a us o an in e oga i e ma ke . In some dialec s, such as To inese, an in a iable encli-
ic may in ac coexis wi h an in lec ed nomina i e p ocli ic (which, I claim, is ela ed o Cl); c .
lon ch’a l’a-lo a ‘Wha has he done?’. (The abo e examples a e om Mu ano (1997)). In ac , i
migh be a o al misnome o e e o such encli ics as «nomina i e» (al hough hey only appea
in ensed con ex s). The ac ha such encli ics only su ace in ma ix in e oga i es migh be due
o his o ical easons i hey a e indeed no hing else han a his o ical elic.
Ca WPL 7 253-277 13/6/00 12:35 Página 270
The o he ques ion is why he nomina i e cli ic is an encli ic in (45). Recall
ha he nomina i e cli ic o igina es abo e V (see sec ion 2 abo e). The e o e i
canno adjoin o V; ins ead, i le -adjoins o T(a e V le -adjoins o T), gi ing
ise o he ollowing s uc u e:
(52) [Tnom cl [T[V] T]]
When V- o-Q applies in shallow s uc u e o he pu pose o ph-lexicaliza ion,
i only mo es he ca ego y wi h ph-con en , namely V (i.e. T has no ph-con en ):
(53) [ [Vi[Q]] [Tnom cl [T[Vei] T]]]
The nomina i e cli ic will hen a ach a PF o he e bal ca ego y o i s le .
Ou claim ha he nomina i e cli ic may a ach o a wo d o i s le by a phonolo-
gical, sandhi- ype ule, ecei es some suppo om he No he n I alian dialec s
whe e he nomina i e cli ic may op ionally a ach o a complemen ize o i s le
(c . Pole o (in p ess)):16
(54) a. A a che el ien. (Vene o)
Look ha nom.cl comes
b. A a ch’el ien
4.3. Simila i ies wi h Chinese wi h espec o Minimali y
I is imely o men ion he case o he inde ini e wh in Manda in Chinese, discussed
by Li (1992). Li shows ha wh-wo ds in Chinese can ha e a non-in e oga i e
inde ini e in e p e a ion as well as an in e oga i e one, depending on he seman-
ic p ope ies o i s binde . The ela ion be ween a non-in e oga i e wh and i s
licenso is a binde - a iable ela ion, like he one ha exis s be ween an in e o-
ga i e wh and a ques ion ope a o . And hese binde - a iable ela ions a e subje
o Minimali y, as illus a ed by he con igu a ions below:
(55) a. +QP .... -QP .... wh1 .... wh2
b.*+QP .... -QP .... wh1 .... wh2
c.*+QP .... -QP .... wh1 .... wh2
The Cl(i ic) P ojec ion in Ques ions Ca WPL 7, 1999 271
16. Spo iche (1998) claims ha he nomina i e encli ic has a igh e mo pho-phonological connec-
ion o he e b han a nomina i e p ocli ic and he e o e p oposes ha encli ics be lexically gene-
a ed wi h he e b. I would sugges ha his is a mo e gene al mo phological p ope y o wo ds:
mophemes o he igh o he s em a e close o he s em han he mo phemes o he le o he
s em. This is ue no only o cli ics bu also o a ixes. C . he possibili y o coo dina ing p e ixes
bu no su ixes; e.g. p e- and pos -posi ion.
Ca WPL 7 253-277 13/6/00 12:35 Página 271
The con igu a ions in (55b,c) illus a e minimali y e ec s; -QP is close o wh2
han +QP. In (55a), -QP is inac i e; he e o e i is no a po en ial an eceden and
does no igge minimali y e ec s. Li (1992:144) s a es he gene aliza ion ha
eme ges om hese and o he con igu a ions in he ollowing e ms:
(56) The linking o a wh-elemen wi h an ope a o is subjec o minimali y.
The linking o A wi h B [...A...B....] obeys Minimali y i he e is no in e -
ening C [...A...C...B] such ha C is linked o ano he elemen D, D =/= B
=/= A
Ou claim is ha he s uc u e in (34b), which we epea below, is o mally
iden ical o (55c):
(57) [ [whque]i[Whi[ ei[Qi[Jeanj[ Clj[ T [VP ej[.... ei....
Li ema ks ha he e a e no minimali y e ec s when he wo ope a o s in ol ed
a e o he same ype:
(58) a. +QP .... +QP .... wh1 .... wh2
b. +QP .... +QP .... wh1 .... wh2
This obse a ion is pa icula ly in e es ing because in he case o he Cl-ope a o
we ind he same gene aliza ion, as illus a ed by he well o medness o he s uc-
u es in (14), which we epea in (59). In hese s uc u es he e a e wo Cl ope a o s
wi h iden ical unc ion and no in e en ion e ec s a e obse ed.
(59) a. [DPj[Clj[ DPi[Cli[T [VP [ ej[V ei....
b. [DPj [Clj[DPi[Cli[T [VP ei[V ej.....
5. Conclusion
In his a icle I ha e econside ed he cons ain on p e e bal subjec s in Romance
in e oga i es. I pu o h an explana ion ha is c ucially based on a pa icula
analysis o cli ics and s ong ag .
Acco ding o his iew, cli ics and s ong ag a e he mo phological mani es a ion
o an abs ac Cl pos ion, which has he s a us o an ope a o . Cl, which is p ojec-
ed abo e TP, has he unc ion o «ex e nalizing» an a gumen o he e b wi h es-
pec o he T associa ed wi h he e b. I achie es his ia a combina ion o binding
and me ge. Mo e p ecisely, Cl binds a DP a gumen o he e b, gi ing ise o a
272 Ca WPL 7, 1999 Ma ia Luisa Zubiza e a
Ca WPL 7 253-277 13/6/00 12:35 Página 272
syn ac ic p edica e; a p onominal o lexical DP me ges wi h Cl and unc ions as
he subjec o he p edica e. Cl is hus o mally compa able in s a us o a ela i e ope-
a o . Cl is mo phologically iden i ied by a p onominal cli ic o s ong ag con ained
in T. Languages wi h cli ics and s ong ag , like many o he Romance languages,
hus ha e a mechanism o he han mo emen o «ex e nalize» an a gumen , in pa -
icula he subjec . The p oposal hen is ha p e e bal (non ocused) subjec s in
Romance a e in he Spec o a Cl-ope a o and his Cl c ea es minimali y e ec s
when a ques ion is o med by mo ing a wh-ph ase abo e i . Mo e p ecisely, a
wh-ph ase mo es o he Spec o a Wh-p ojec ion h ough he Spec o a Q-p ojec ion
(loca ed a he igh mos edge o he CP ield). This c ea es a coindexing ela ion
be ween he wh-ph ase and Q, which hen binds he ace o he wh-ph ase as well,
gi ing ise o a Q-ope a o / a iable ela ion. Bu he Cl ope a o is close o he
wh- a iable han he Q-ope a o , a minimali y iola ion hus a ise.17
The absence o he p e e bal subjec cons ain in in e oga i es wi h on ed non-
ba e wh-ph ases was a ibu ed, ollowing Dob o ie-So in (1994), o he ac ha
he wh- ea u e in he speci ie o such ph ases may ail o pe cola e o he DP, in
which case he e is no coindexing o a wh ca ego y wi h Q. In such s uc u es, Q does
no unc ion as a syn ac ic ope a o and no minimali y ensues.
While Cl is gene ally mapped in he highe pa o he middle ield, namely
be ween he Q p ojec ion and TP, in ce ain cases Cl may p ojec abo e he Q p o-
jec ion, hus a oiding a minimali y e ec . This is he case o he so-called com-
plex in e sion cons uc ion in F ench and o a simila cons uc ion in he No he n
I alian dialec s.
I was sugges ed ha he di e ences among he Romance languages wi h es-
pec o he p e e bal subjec cons ain in embedded in e oga i es is o be a i-
bu ed o pa ame ic di e ences in he ph-lexicaliza ion o he Q- ea u e ha exp esses
he o ce o he sen ence.
Appendix: On Acc Cli ic Doubling in Cas ilian Spanish18
In Cas ilian Spanish an accusa i e cli ic can double a wh-ph ase only in he p e-
sence o an ac i e nega ion o in a modal con ex :
(60) a. A quién no lo aguan a ni el más pacien e
Whom (obj) no acc.cl s and-3 d-sing e en he mos pa ien
de los mo ales/ni su mad e. (Ac i e Neg)
o he mo als/e en his mo he (subj)
The Cl(i ic) P ojec ion in Ques ions Ca WPL 7, 1999 273
17. An excep ion is B azilian Po uguese. In e ec , he p e e bal subjec cons ain is absen in B azilian
in e oga i es. This is as expec ed. B azilian Po uguese lacks Cl because i has no way o mo p-
hologically iden i ying i : i has a weaken subjec ag eemen sys em and has los (o is in he p o-
cess o loosing) he 3 d pe son accusa i e cli ic. See Ma cello Modes o (1999) and e e ences ci ed
he in.
18. I would like o hank Ma ina Fe nandez Lagunilla o p o iding judgemen s on he da a epo ed
in his sec ion.
Ca WPL 7 253-277 13/6/00 12:35 Página 273
b. (?) A quién no lo aguan a nadie? (Ac i e Neg)
Whom (obj) no acc.cl s and-3 d-sing nobody (subj)
c. *A quién no lo aguan a Ma ía (subj) (Non-ac i e Neg)
(61) a. A quién (piensas que) lo debe ía cas iga
Whom (obj) ( hink-2nd-sing ha ) acc.cl should-3 d-sing punish
Ma ía/ su mad e? (modal e b / «deon ic»)
Ma ía/ his mo he (subj)
b. A quién dudas de que lo cas igue
Whom (obj) doub -2nd-sing o ha acc.cl punish-3 d-sing-subjunc
Ma ía/ su mad e? (p esen subjunc i e)
Ma ía/ his mo he (subj)
c. A quién (piensas que) lo cas iga á
Whom (obj) ( hink-2nd-sing ha ) acc.cl punish-2nd-sing- u
Ma ía/ su mad e? ( u u e indica i e)
Ma ía/ his mo he (subj)
(62) a. *A quién (piensas que) lo cas igó
Whom (obj) ( hink-2nd-sing ha ) acc.cl punish-3 d-sing-pas
Ma ía/ su mad e? (pas indica i e)
Ma ía/ his mo he (subj)
b. *A quién (piensas que) lo cas iga
Whom (obj) ( hink-2nd-sing ha ) acc.cl punish-3 d-sing-p es
Ma ía/ su mad e? (p esen indica i e)
Ma ía/ his mo he (subj)
Some ema ks abou he da a a e in o de :
1. The e is no speci ici y e ec : he wh-ph ase ha is being doubled is no in e -
p e ed as being discou se-linked.
2. The wh-ph ase (a quién) can bind he pos e bal subjec wi hou gi ing ise o
WCO e ec s. This is as expec ed since he wh-ph ase o igina es in he Spec
o Cl. See sec ion 2.
3. The accusa i e cli ic being +de ini e, i ende s he ques ion +de ini e. In o he
wo ds, exis ence o an indi idual co esponding o he objec is p esupposed
by bo h speake and hea e . This is shown by he ac ha «nobody» is no a
possible answe o any o he ques ions in (60) and (61).
I would like o p opose ha in o de o unde s and he phenomenon desc ibed
abo e, we mus appeal o he le el o Asse ion S uc u e in he sense o Zubiza e a
(1998). The Ass S uc u e is a pos -LF le el whe e he pa i ioning o he sen ence
in e ms o ocus and p esupposi ion is a icula ed in e ms o o de ed p oposi-
274 Ca WPL 7, 1999 Ma ia Luisa Zubiza e a
Ca WPL 7 253-277 13/6/00 12:35 Página 274
ions.19 The (sha ed) p esupposi ion is an asse ion ha in oduces an exis en ially
bound a iable. The ocus o he sen ence p o ides a alue o he a iable in oduced
in he p esupposi ion. The o me he e o e p ecedes he la e in he Ass S uc u e.
To illus a e, conside he example in (63a), wi h he F(ocus)-ma king as indica-
ed in (63b). This sen ence will be associa ed wi h he Ass S uc u e gi en in (63c).
(63) a. Ma ía in i ed Juan.
b. Ma ía in i ed [FJuan]
c. Ass S uc u e:
Ass (1): The e is an x such ha Ma ía in i ed x
Ass (2): The x such ha Ma ía in i ed x = Juan
I was no ed in Zubiza e a (1998) ha he ela ion be ween he de ini e des-
c ip ion in Ass (2) and he exis en ially-bound a iable in Ass (1) is o mally com-
pa able o he ela ion be ween an E- ype p onoun and an inde ini e in a p e ious
sen ence, as in he example below.
(64) A sailo came in. He was wea ing a ed shi .
[he = he sailo ha came in]
Capi alizing on his in ui ion, we sugges ha he accusa i e cli ic in (60) and
(61) has he s a us o an E- ype p onoun. We de elop i below.
Suppose ha nega ion and modals in oduce an independen asse ion. Conside
he examples in (60a) and (61a). The p esence o nega ion and he modal gi e ise
o Ass (1) in (65) and (66) espec i ely. Ass (2) assigns a alue (i.e. a cons an ) o
he a iable in oduced in Ass (1). We use a capi al le e o indica e a cons an ,
whe e he cons an co esponds o he possible answe s o he ques ion. Ou claim
is ha he accusa i e cli ic co esponds o he de ini e desc ip ion in Ass (2).
(65) Ass (1): The e is NO x (no e en he mos pa ien o mo als) such ha x
s ands some y
Ass (2): The y such ha he e is no x...(x s ands y) = A
lo
(66) Ass (1): SHOULD (Ma ía punish some y)
Ass (2): The y such ha (should (Ma ía punish y) = A
lo
The Cl(i ic) P ojec ion in Ques ions Ca WPL 7, 1999 275
19. I ha e also a gued in he same wo k ha he no ions opic/commen should be a icula ed a he
Ass S uc u e le el in e ms o he subjec /p edica e ela ion.
Ca WPL 7 253-277 13/6/00 12:35 Página 275
We hen each he ollowing conclusion o Cas ilian Spanish:
(67) In Cas ilian Spanish, an accusa i e cli ic may be QU-bound i he cli ic has
he s a us o an E- ype p onoun (in o he wo ds, i he accusa i e cli ic
co esponds o a de ini e desc ip ion in he Ass S uc u e o he sen ence).
Re e ences
Alexiadou, A.; Anagnos opoulou, E. (1998). «Pa ame izing AGR: Wo d O de ,
V-Mo emen , and EPP-Checking». Na u al Language and Linguis ic Theo y 16:
491-539.
Aoun, J.; Benmamoun, E. (1998). «Minimali y, econs uc ion, and PF mo emen ».
Linguis ic Inqui y 20: 141-172.
Ba bosa, P. (1995). Null Subjec s. Depa men o Linguis ics and Philosophy. Camb idge:
MIT. PhD disse a ion.
Benincà, P.; Pole o, C. (1997). «A case o do suppo in Romance». Uni e si y o
Padua. Unpublished manusc ip .
Cecche o, C. (1995). «Recons uc ion in cli ic-le disloca ion». Depa men o Cogni i e
Science, Ins i u o scien i ico San Ra aele, Milan. Manusc ip .
Chomsky, N. (1993). «A minimalis p og am o linguis ic heo y». In Hale, K.; Keyse ,
S. J. (eds.). The View om Building 20. Camb idge, Mass.: MIT P ess.
—(1994). «Ba e ph ase s uc u e». MIT Occasional Pape s in Linguis ics 5. Depa men
o Linguis ics and Philosophy. Camb idge, Mass.: MIT [Also in Webelhu h, G.
(ed.). Go e nmen and Binding Theo y and he Minimalis P og am. Ox o d:
Blackwell, 1995.]
Dob o ie-So in, C. (1994). The Syn ax o Romanian. Mou on de G uy e .
Fe nández-So iano, O. (1989). Rección y ligamen o en español: Aspec os del pa á-
me o de suje o nulo. Uni e sidad Au ónoma de Mad id. Doc o al disse a ion.
F anco, J. (1993). On objec ag eemen in Spanish. Depa men o Spanish and
Po uguese. Uni e si y o Sou he n Cali o nia, Los Angeles. Doc o al disse a ion.
Guas i, M. T. (1996). «On he con o e sial s a us o Romance in e oga i es». P obus
8: 161-180.
Ia idou, S. (1990). «Cli ics and island e ec s». Camb idge, Mass.: MIT. Unpublished
manusc ip .
Kayne, R. (1991). «Romance cli ics, e b mo emen , and PRO». Linguis ics Inqui y
22: 647-686.
— (1994). The an isymme y o syn ax. Camb idge, Mass.: MIT P ess.
Koopman, H. (1997). «Uni ying p edica e cle cons uc ions». Pape p esen ed a BLS
special session.
Li, A. (1992). «Inde ini e wh in Manda in Chinese». Jou nal o Eas Asian Linguis ics
1: 125-155.
Modes o, M. (1999). «Null subjec s wi hou ich ag eemen ». To appea in Ka o, M.;
Neg ão, E. (eds.). B azilian Po uguese and he Null Subjec Pa ame e . F ank u
and Mad id: Ve ue -Ibe oame ican.
Muna o, N. (1997). P opie a s u u ali e dis ibuzionali dei sin agmi in e oga i i in
alcuni diale i i aliani se en ionali. Dipa imen o di Linguis ica, Uni e si y o
Padua. Unpublished disse a ion
276 Ca WPL 7, 1999 Ma ia Luisa Zubiza e a
Ca WPL 7 253-277 13/6/00 12:35 Página 276
Ouhalla, J. (1993). «Subjec -ex ac ion, nega ion and he an i-ag eemen e ec ». Na u al
Language and Linguis ic Theo y 11.
Pole o, C. (in p ess). The Highe Func ional Field in he No he n I alian Dialec s.
Ox o d S udies in Compa a i e Syn ax. Ox o d Uni e si y P ess.
Pollock, J.Y.; Mu ano, N.; Pole o, C. (1998). «Eppu si muo e! On compa ing F ench,
Po uguese and Bellunese Wh-mo emen ». CNRS and Uni e si y o Padua.
Unpublished manusc ip .
Rizzi, L.; Robe s, I. (1989). «Complex In e sion in F ench». P obus 1: 1-30.
Rizzi, L. (1990). Rela i ized Minimali y. Camb idge, Mass.: MIT P ess.
—(1991). «Residual e b second and he Wh-C i e ion». Facul é des Le es, Uni e si é
de Genè e.Technical Repo s in Fo mal and Compu a ional Linguis ics 2.
Speas, M. (1994). Ph ase S uc u e in Na u al Language. Do d ech : Kluwe .
Spo iche, D. (1998). Pa i ions and A oms o Clause S uc u e. Rou ledge.
Suñe , M. (1988). «The ole o ag eemen in cli ic-doubled cons uc ions». Na u al
Language and Linguis ic Theo y 6: 391-434.
—(1992). «Subjec cli ics in he No he n I alian e nacula s and he Ma ching
Hypo hesis». Na u al Language and Linguis ic Theo y 10: 596-640.
—(1994). «V-mo emen and he licensing o a gumen al wh-ph ases in Spanish».
Na u al Language and Linguis ic Theo y 12: 335-372.
Zubiza e a, M.L. (1993). «The g amma ical ep esen a ion o opic and ocus:
Implica ions o he s uc u e o he clause». Cuade nos de Lingüís ica 2. Ins i u o
Uni e si a io O ega y Gasse .
—(1998). P osody, Focus, and Wo d O de . Camb idge, Mass.: MIT P ess.
The Cl(i ic) P ojec ion in Ques ions Ca WPL 7, 1999 277
Ca WPL 7 253-277 13/6/00 12:35 Página 277
