Negation in Catalan : some Remarks with regard to no pas

NEGATION IN CATALAN. SOME REMARKS WITH REGARD TO NO PAS M. Te esa Espinal Uni e si a Au bnoma de Ba celona In his pape I shail app oach ce ain ques ions ha can be posed in he s udy o Ca alan nega i e cons uc ions: (a) how many nega i e ma ke s can be dis inguished; (b) how he di e ence be ween no and no pas should be desc ibed in he co e g amma o his language, and (c) wha so o syn ac ic ep esen a ion should be assigned o nega i e sen en iai cons uc ions in Ca alan. The s uc u e o he pape is o ganized as ollows. Sec ion 1 is de o ed o desc ibing he syn ac ic dis ibu ion o he a ious nega i e ma ke s exis ing in Ca alan. I is p oposed ha , apa om no, he e a e wo mo e nega i e ma ke s: no pas I, which is a nega i e head, and no pas 2, which is a nega i e speci ie . Sec ion 2 desc ibes he seman ic p ope ies ha cha ac e ize hese nega i e ma ke s. I is a gued ha he associa ion wi h ocus is no an inhe en p ope y o he logical seman ics o no pas, and ha wha singles ou his ope a o , in con as wi h no, is ha i has a quan i ica ional domain o e a se o concep ual en i ies o which a subdomain is selec ed as he nega i e con as ed i em. Sec ion 3 deals wi h he so o syn ac ic s uc u e ha should be assigned o Ca alan nega i e sen ences and i ini oduces he p oposal ha he ele an s uc u e has a NegP as he oo node and an Ag P as iis complemen . TP is he complemen o he la e unc ional ca ego y. 1. Syn ac ic Dis ibu ion I shall app oach he i s ques ion by b ie ly ske ching ou he basic syn ac ic dis ibu ion o he simple lexical i em no and he compound lexical i em no pas in Cen al Ca alan.1 A close examina ion o he o de in which nega i e s uc u es a e in e nalized in he p ocess o i s language acquisi ion and an inqui y in o he equency o appea ance o he a ious exis ing Ca alan Wo king Pape s in Linguis ics (CWPL), 1991: 33-63 Uni e si a Au hnoma de Ba celona nega i e s uc u es sugges ha i is con enien o dis inguish wi hin he g amma o his language be ween ma ked and unma ked nega i e cons uc ions, wi h he syn ac ic dis ibu ion shown in (1). UNMARKED MARKED no no V no X UNMARKED no pas MARKED I I no pas I: no pas 2: Nego SPEC [ + NEG] no pas X El I is impo an o no ice ha V can be kFIN], and V* is ei he a V o a complex e b. On he o he hand, X canno be a ini e e b o m.2 Some examples which illus a e his syn ac ic dis ibu ion a e gi en in (2). (2) a. La Ma ia no a ap o a he Ma ia no PAST pass 'Ma ia did no pass' b. *La Ma ia a no ap o a c. No o s els es udian s an ap o a no all he s uden s PAST pass 'No all he s uden s passed' d. La Ma ia no a pas ap o a he Ma ia no PAST pas pass 'Ma ia did no pass' (cancelling o con i ming he speake 's expec a ions)3 e. La Ma ia no a ap o a pas (2) . *La Ma ia a pas ap o a g. *LaMa ia aap o a pas h. No pas LA MARIA no a ap o a no pas he Ma ia no PAST pass 'I was no Ma ia who passed' (cancelling o con inning he speake 's expec a ions on Ma ia) i. *No pensa a que la Ma ia ap o a ia pas no hink-PAST ha he Ma ia pass-COND pas 'I didn' hink ha Ma ia would pass' j. *No emia pode pas ap o a la Ma ia no ea -PAST be able pas pass he Ma ia 'I had no ea o being able o pass Ma ia' k. No c ec que ap o i pas la Ma ia no hink ha pass-SUB pas he Ma ia 'I don' hink ha { Ma ia is going o pass, I'm going o pass Ma ia )' Exa nples (2a,b,c) show he dis ibu ion o he simple lexical i em no, ei he p eceding a e b o n o a non- e bal i em. Howe e , in his pape I am mo e conce ned wi h nega i e no pas cons uc ions. No ice ha in (2) no pas appea s in wo di e en syn ac ic en i onmen s (compa e (2d,e) wi h (2h)). Wi h ega d o his ac wo dis inc nega i e ma ke s will be pos ula ed. No pas 1, as illus a ed in (2d,e), is cha ac e ized by coappea ing wi h a e b o m (ei he simple o complex) and by being a discon inuous lexical i em. No e also ha in he dialec ha I desc ibe, pas canno occu alone, wi hou no (see examples (2 ,g)), which sugges s ha -unlike F ench and Occi an- in cu en Cen al Ca alanpas is no an independen lexical uni , bu a mo pheme which is lexically and syn ac ically insepa able om no. Fu he mo e, he ung amma ical sequences gi en in (2ij) e eal ha he se ies o e bs ha in e up he compound lexical i em no pas 1 canno con ain a sen en ial bounda y. In his espec , he g amma icali y o an exa nple such as (2k) seems o be ela ed o he ense-mood o he subo dina e clause. Acco ding o Picallo (1984), he Tense node o a subjunc i e clause lacks he ea u es necessa y o be ully speci ied as [+ TNS], o i is bound o he Tense speci ica ion o he main clause. On he o he hand, pas a S-s uc u e mus always be c-commanded by no (c . May (1985)), which means ha he e canno be any maximal p ojec ion be ween no and pas . The e o e, in he syn ac ic con ex illus a ed in (2k) i seems as i he CP node o he subo dina e clause in he subjunc i e ense is no isible wi h ega d o binding e ec s. One possible in e p e a ion o hese da a is o assume ha no pas 1 is -1ike no - he head o a ~e~p.4 Wha di e s be ween hem is ha no pas 1 equi es an explici complemen a S- s uc u e, and his complemen mus ha e he o m o an in lec ed e b, as he answe s in (3B) e eal. (3) A. Vind iilaMa ia? come-FüT he Ma ia 'Will Ma ia come?' B. a. No b. *Nopas c. No ind h pas F om a syn ac ic poin o iew i is impo an o poin ou ha no pas can also occu a a di e en syn ac ic en i onmen , exempli ied so a in (2h). No pas 2 , as I shall e e o his nega i e ma ke , is in a simila syn ac ic dis ibu ion o o he lexical i ems, such as nome's, solamen 'only', enca a 's ill' and ins i o 'e en'. In ac , all hese ad e bials can occu a se e al syn ac ic posi ions, modi ying ei he he e b o some o he cons i uen o he sen ence. Among hem, no pas in oduces a nega i e pola i y e ec . Mo e exa nples which illus a e he loca ion o hese ad e bials in syn ac ic posi ions o he han a he le o he e b a e gi en in (4), (5) and (6) (4) a. { Només, solamen , ins i o , enca a ) VOSALTRES se eu sanciona s [ only, only, e en, s ill ] you be-FUT punished '{ Only, e en, s ill ) you will be punished' b. No pas VOSALTRES no se eu sanciona s no pas you no be-FUT punished 'I 's no you who will be punished' (5) a. Només una mica de PA ens han dona only a li le o b ead us ha e gi en 'Only a Ii le b ead has been gi en o us' b. No pas gens de PA no ens han dona no pas no hlng o b ead no us ha e gi en 'No b ead a all has been gi en o us' (6) a. La Ma ia comp a pa només al o n he Ma ia buys b ead only a he bake y 'Ma ia buys b ead only a he bake y' b. La Ma ia comp a pa no pas al o n, sinó a la pas isse ia he Ma ia buys b ead no pas a he bake y, bu a he pas y shop 'Ma ia buys b ead no a he bake y bu a he pas y shop' In (2h), as in (4b), (5b) and (6b), he e is no supe icial discon inui y be ween no andpas. This p ope y seems o sugges ha no pas in hese examples is -as be o e- a single lexical i em which occupies a single syn ac ic posi ion. No ice, u he mo e, ha i would be inco ec o claim ha in his so o s uc u e no pas is he head o a NegP, o i would be qui e anomalous o i ems o he ad e bial ca ego y o ha e indis inc ly nominal, quan i ied o p eposi ional complemen s: a speci ic p ope y o no pas 2 is ha i can p ecede any majo lexical ca ego y, wi h he excep ion o ini e e b o ns. I ad e bs which beha e like quan i ie s a e no heads, he only posi ion hey can ill is he SPEC posi ion o a majo syn ac ic ca ego y: NP, QP, PP, e c. A u he dis ibu ional p ope y ela es o he no ions o scope and ocus, due o he ac ha hose lexical i ems which in o he languages beha e like no pas, només, solamen , e c. a e known in he linguis ic li e a u e -a nong o he ems- as " ocus ad e bs", "scopal ad e bs", "scala pa icles" and " ocusing pa icles". Since he cons i uen associa ed wi h his ype o ad e bial has been some imes named ocus and o he imes scope, in he nex sec ion I shall desc ibe he dis ibu ion o no pas wi h ega d o ocus and scope. 2. No pas As a Quan i ie -like Ope a o No pas, jus like only, is always seman ically associa ed wi h some sen en ial cons i uen , bu his seman ic associa ion does no ha e o be e ealed in he syn ax h ough p ecedence s uc u al ela ions. Mos o en, he cons i uen associa ed wi h no pas is ei he iden i ied wi h he ocus o he sen ence (see (4b) and (5b)) o wi h a cons i uen o a con as i e s uc u e (no pas X, sinó Y 'no X, bu Y'), as exempli ied in (6b). In he i s case, he cons i uen associa ed wi h no pas ends o be in ona ionally mo e p ominen .5 Howe e , i is no igh o assume ha no pas is au oma ically associa ed wi h he p osodic ocus o he sen ence, o i is possible o ind examples whe e he ocus o he sen ence does no coincide wi h he cons i uen ha no pas p ecedes.6 Le us conside he pa adigm o sen ences gi en in (7). (7) a. La Ma ia no ha pas ingu he Ma ia no has pas come 'Ma ia has no come' b. LA MARIA no ha pas ingu 'I is Ma ia who has no come' (7) c. La Ma ia no ha pas VINGUT 'Ma ia has no COME' d . No pas LA MARIA no ha ingu 'I is no Ma ia who has come' In (7a) he e is no p osodic ocus, whe eas in he o he examples he cons i uen in capi al le e s is he mos s essed elemen and co esponds o he bea e o p osodic peak. (7a), he e o e, illus a es ha he associa ion wi h p osodic ocus is nei he an essen ial p ope y o he occu ence o no pas, no a necessa y condi ion o licensing his ope a o a LF. Wha should be no iced is ha hese examples show an impo an dis inc ion be ween nega i e s uc u es wi h no and hose wi h no pas , o only no pas con ains wi hin i s scope he con as ocus o he sen ence. Acco ding o B ugman (1986), he con as ocus o a ce ain s uc u e has o be iden i ied wi h he con as ed i em, which need no be iden ical wi h he s essed elemen , bu can p ope ly include he bea e o p osodic peak. In he examples I am desc ibing, he con as ocus does no necessa ily ha e o co espond o he cons i uen ha no pas p ecedes a any le el o syn ac ic ep esen a ion. Ac ually, in (7a-c) he con as ed i em, o quan i ica ional domain (in Roo h's (1985) e ms), o he meaning o no pas can be any sen en ial cons i uen , ei he he VP, o he NP, o he main e b. This is so because a he le el whe e ins uc ions o meaning ep esen a ion a e codi ied (i.e. a LF), he nega i e ope a o modi ies and c-commands e e y cons i uen wi hin i s absolu e scope, which - ollowing May (1985)- is he CP, he mos immedia e maximal p ojec ion domina ing all sen en ial ope a o s. The e o e, he quan i ica ional domain o an ope a o such as no pas can be any cons i uen wi hin i s absolu e domain.7 In acco dance wi h his claim, (7a-c) may be in e p e ed in he e ms exempli ied in he pa aph ases in (8), (9) and (10). (8) a. La Ma ia no ha pas ingu , qui ha ingu ha es a el Joan he Ma ia no has pas come, who has come has been he Joan 'Ma ia has no come, he one who has come is Joan' b. La Ma ia no ha pas ingu , simplemen ha ele ona he Ma ia no has pas come, simply has phoned 'Ma ia has no come, she has simply phoned' (9) a. LA MARIA no ha pas ingu , qui ha ingu ha es a EL JOAN 'Ma ia is he one who has no come, he one who has in ac come is Joan' b. LA MARIA no ha pas ingu , ELLA simplemen ha ele ona 'Ma ia is he one who has no come, in ac she has simply phoned' (10) a. La Ma ia no ha pas VINGUT, en eali a s'ha limi a a TELEFONAR 'Ma ia has no ac ually come, in ac all she has done is o elephone' b. La Ma ia no ha pas VINGUT, pe 6 en Joan si que ha ingu 'Ma ia has no ac ually come, bu Joan has in ac come' One o he hings hese da a sugges is ha , in o de o ep esen adequa ely he seman ic agueness o he sequences in (7a-c), a LF he nega i e ad e b should be in a sen en ial posi ion om which i could c-command e e y cons i uen o he nega i e s uc u e. This posi ion could be an adjunc IP posi ion. (1 1) a. La Ma ia no ha pas ingu LF: [cp [c' [?? no pas [jp la Ma ia ha ingu I]]] b. LA MARIA no ha pas ingu LF: [CP [SPEC~ la Ma ia ] [c' [?? no pas [~p ei ha ingu I]]] (1 1) c. La Ma ia no ha pas VINGUT LF: [cp [c' [ci ingu 1 [?? no pas [~p la Ma ia ha ei ]I]] No ice ha any o hese ep esen a ions in oduce ague in e p e a ions, because he scope o no pas is he se o cons i uen s i c-commands a LF, and his se includes bu is no necessa ily iden i ied wi h he emphasized cons i uen . A u he obse a ion is ha mos o en he unma ked eading which co esponds o (7b) o (7c) is ha in e p e a ion acco ding o which he emphasized cons i uen is aken as p o iding he se o ele an en i ies which bea he quan i ica ional domain o no pas. The eason why his happens should be a ibu ed o he posi ion ha he cons i uen ma ked wi h he abs ac ope a o [+ EMPH] occupies a LF. I assume ha , since he cons i uen iden i ied wi h he p osodic ocus o he sen ence occu s a he mos p ominen LF posi ion, i co esponds o he kind o en i y o e which -o he hings being equal- i is mo e easonable o unde s and as he no pas quan i ica ional domain. O he wise, in any ague example i is going o be he con ex ual in o ma ion which will ci cumsc ibe he a ge o nega ion. Con as ing wi h (7a-c), he LF co esponding o (7d) would ha e he logical ep esen a ion in (12). (12) No pas LA MARIA no ha ingu LR [cp [SPEC~ no pa la Ma ia ] [CI [?? no [~p ei ha ingu I]]] In his s uc u e he absolu e scope o nega ion is s ill he whole CP, because i is he minimal maximal p ojec ion domina ing e e y cons i uen wi hin he nega i e sen en ial s uc u e. Wha is c ucial is ha in (12) a ague in e p e a ion is no licensed, and his ac can only be explained i he posi ion ha no pas has a LF is aken in o conside a ion. No e ha in his s uc u e no pa is placed wi hin he SPEC o he CP, oge he wi h he emphasized cons i uen .8 In conclusion, he ac ha no pas migh be associa ed wi h a p osodic ocus does no seem o be pa o i s logical seman ics. Wha eally de ines his ad e b seman ically is ha i has a quan i ica ional domain o con as ocus. The e o e, he seman ic dis inc ion be ween no and Whe eas cli ics ha e o p ecede he ini e o ms o he e b and ollow he non- ini e o ms, no mus always p ecede bo h he e b -ei he ini e o non- ini e- and he cli ics adjoined o i , and pas mus always ollow bo h he e b o n -ei he ini e o non- ini e- and he cli ics adjoined o i . This syn ac ic dis ibu ion is schema ized in (21). (21) a. no + (cl) + V b. V+(cl)+pm As al eady no iced by Zanu ini ( 1990), i he syn ac ic s uc u e o languages like I alian and Ca alan was he one in (18a) -pos ula ed by Pollock (1989)-, hen we should expec ha he mo emen o he nega i e ma ke om he head posi ion o NegP o he head o TP, oge he wi h he p onominal cli ics, migh lead o some so o eo de ing p ocess among hese i ems. Howe e , his eo de ing -a leas ega ding no -, ne e akes place. On he o he hand, al hough I assume ha pas mus be eo de ed co ec ly a S-s uc u e, i clea ly di e s om he cli ic p onominal sys em as a as he second mo pheme o no pas I is no es ic ed o ollow non- ini e o ms o he e b. 1 1 F om hese commen s we can conclude he ollowing: i he NegP is assumed o be a syn ac ic p ojec ion highe han TP and Ag P, hen he e is no eason o suppose ha Ca alan nega i e ma ke s a e cli ics and ha hey mus be aised up o a highe posi ion. 12 The hypo hesis ha he NegP is he highes node would also p edic he possibili y ha bo h no and pas can bea s ess (c . Zanu ini (1990)), as well as he ac ha in Ca alan he e b can ne e occu o he le o no. (22) a. *Vaig no eu e'l (pas) PAST no see-him pas 'I didn' see him' (22) b. * Vindh no (pas) come-FUT no pas 'He will no come' c. *Se no (pas) pun ual és un mal cos um be-INF no pas punc ual is a bad habi 'No o be punc ual is a bad habi ' In addi ion o accoun ing o he non-cli ic na u e o nega ion and he ela i e posi ion o nega ion wi h ega d o he e b, he analysis o sen en ial nega ion mus also p edic he mo phological cha ac e is ics o he Ca alan e b sys em. Acco ding o he Mi o P inciple (c . Bake (1985)), i is p edic ed ha mo phological de i a ions mus di ec ly e lec syn ac ic de i a ions (and ice e sa). This means ha , aking in o conside a ion he mo pheme s uc u e o he e b o ms and, mo e speci ically, he ela i e o de o ense and ag eemen mo phemes wi h espec o he e b, he p ojec ion o Tense canno be highe han he p ojec ion o Ag eemen (c . Belle i (1988, 1990b), Sola (1989)). (23) a. pin a en /pín +á+ba+n/ pain + lconj.+ impe ec i e ense + 3p.pl. They pain ed' b. do mi íem /dó m+i+ ía+m/ sleep + 3conj. + condi ional ense + 1p.pl. 'We would sleep' As illus a ed in (23), i he syn ac ic and mo phological o de o Ca alan e b o ms is V+T+Ag , hen a D-s uc u e Ag P mus be highe han TP; which is qui e he e e se o wha Pollock (1989) pos ula es. Paying a en ion o his ema k, and assuming ha in Ca alan -as in F ench, I alian and Spanish- a ixa ion is done h ough a head mo emen ule which aises all e bal heads o highe hie a chic posi ions (in o de no o iola e he Head Mo emen Cons ain and he Emp y Ca ego y P inciple; c . Chomsky (1981), T a is (1984)), hen he p oposals (18b) and (18d) a e he only ones ha can co ec ly p edic he igh o de o e bal su ixes in his and simila languages. Bu , i a he same ime we a emp o explain he non-cli ic na u e o Ca alan nega i e ma ke s, hen we a e le wi h he syn ac ic ela ion be ween unc ional ca ego ies pos ula ed in (18d). In summa y, he answe o he ques ion o wha ac s as complemen o nega ion seems o poin o he unc ional ca ego y Ag P, which in u n would ha e TP as i s complemen .13 The ad an age o (18d) is ha i p o ides an accoun o he so o da a gi en in (24). (24) a. No ob i (pas) la po a quan uquen po se p uden no open-INF pas he doo when call may be sensible 'No o open he doo when someone is calling may be sensible' b. No ob in (pas) la po a, és di ícil que e passi es no open-GER pas he doo is di icul ha you happens some hing 'No opening he doo , i is unlikely ha any hing will happen o you' c. *No obe a (pas) la po a, hau is d'en a pel ga a ge no open-PP pas he doo ha e-FUT o go-in h ough he ga age The doo no (being) open, you will ha e o go in h ough he ga age' d. *No ob e (pas) la po a no open-IMP pas he doo 'Do no open he doo ' e. No ob i (pas) la po a no open-SUB pas he doo 'Do no open he doo ' (2 p. sg.) (24) . No ob iu (pas) la po a no open-1ND pas he doo 'Do no open he doo ' (2 p. pl.) 14 In o de ha a nega i e i em migh be licensed as he bea e o sen en ial nega ion, a condi ion on he assignmen o sen en ial scope mus be sa is ied. (25) Cons ain on he Assignmen o Sen en ial Scope o Nega ion Nega ion can ake sen en ial scope only i -a S-s uc u e - i is in a posi ion om which i c-commands bo h he TP and he Ag P (c . Zanu ini (1989: 14)) In acco dance wi h his condi ion, he nega i e ad e bs which occu in (24c,d) canno be licensed as p oposi ional ope a o s, o in hese examples hey coappea wi h e b o ms which a e no ma ked o Tense. Wi h ega d o pas pa icipial clauses, I ollow he idea, al eady poin ed ou by Belle i (1989, 1990b), ha hey a e Ag Ps which lack a empo al speci ica ion. This Ag P co esponds o he p ojec ion which in he ull clausal s uc u e ep esen s he pas pa icipial complemen o Aux: Ag P0bjec (c . Chomsky (1989)). In his syn ac ic con ex nega ion is no allowed, because NegP ills a posi ion ou side he pas pa icipial Ag P. Lack o Tense and Ag eemen ea u es wi h he subjec can also be pos ula ed in ela ion o ue impe a i e o ms. No ice ha his mood is ully de ec i e: in mos Ca alan e bs he impe a i e is educed o a single o m, which is used in associa ion wi h he second pe son jus because he impe a i e is supposed o appeal o he add essee's a en ion, bu no e ha a e b onn such as he one in (24d) (ob e ) does no bea any second pe son ag eemen ea u es.15 Consequen ly, (24d) is ung amma ical, because a nega i e ad e b occu s wi h an impe a i e o m which is cha ac e ized by i s lack o Tense and subjec Ag eemen ea u es. The examples in (24e, ) illus a e wha a e ac ually known as impe a i e supple i e o ns, ei he aken om he p esen indica i e ense o he p esen subjunc i e ense. The e b o m in (24e) (ob i ), aken om he subjunc i e, may be de ined by means o he ea u es [-T, +Ag ] (c . Picallo (1984)). Being posi i ely de ined o a leas one o hese ea u es, he ad e bs no and no pas 1 can occu wi h his e b o m wi hou leading o any so o ill- o medness. Likewise, he e b o m in (24 ) (ob iu ), aken om he indica i e, is posi i ely de ined o bo h Tense and Ag eemen and, he e o e, i can eely appea wi h ei he nega i e ad e b. Wi h ega d o he in ini i e and ge und o ms, exempli ied in (24a,b) (ob i , ob in ), no ice ha in Ca alan hey do no show any so o Ag eemen ea u es wi h he subjec o he sen ence. Ne e heless, i seems easonable o assume ha hese o ms hold some aspec ual ea u es (impe ec i e in he in ini i e o ms, and p og essi e in he ge und o ms) which a e insepa able om hei empo al speci ica ions. In acco dance wi h his assump ion in ini i es and ge unds may be said o ha e a leas a Tense p ojec ion and, as a consequence, can eely coappea wi h a nega i e ad e b. As a conclusion, he se o da a which has been desc ibed in his sec ion p o ides se e a1 a gumen s o pos ula ing ha he NegP mus be in a s uc u al posi ion highe han Ag P and TP. In he inal sec ion I shall add ess he ques ion o wha so o o malism could allow he loca ion o he nega i e mo pheme pas in he app op ia e syn ac ic posi ion wi hin a sen en ial s uc u e. I a n going o assume he exis ence o a p ocess o syn ac ic inco po a ion (c . Bake (1988)) acco ding o which he e b complex is adjoined o he head o he NegP. My p oposal is ha only a e he applica ion o his e b mo emen , which also a ec s in ini i e o ms, pas can ind an app op ia e posi ion wi hin he new syn ac ic cons i uen buil h ough adjunc ion. 3.3. Inco po a ion o Nega ion I has al eady been a gued (c . Llinis (1990)) ha in Ca alan a complex e b Aux + V is an Xo syn ac ic uni , which implies ha his e b sequence canno be in e mp ed by any independen syn ac ic cons i uen , o example an Ad P o a NP (c . Espinal (1983)). (26) a. *Ha so in ele ona ? has o en phoned 'Has he o en phoned?' b. *Va en Joan ele ona ? PAST he Joan phone 'Did Joan phone?' The only cons i uen s ha can in e up he componen s o his e b complex a e a g oup o i ems s ongly ela ed o he e b: ei he he cli ic p onouns o he nega i e mo phemepas, as he exa nples in (27) illus a e. (27) a. Quan a ha e -la is a, se'n a ana when PAST ha e-he seen, le 'Once he had seen he , he le ' b. No ha pas ingu no has pas come 'He has no come' In o de o explain his dis ibu ion, I am going o ha e ecou se o a p ocess o syn ac ic inco po a ion, which esul s in he con igu a ion speci ied in (28).16 (28) P ocess o Syn ac ic Inco po a ion (c . Lliniu (1990)) I has been obse ed ha , once his inco po a ion p ocess has been ca ied ou , a cli ic may change i s p ecedence s uc u al ela ions wi h ega d o he complex e b (ei he be o e he ini e e b, o a e he non- ini e e b), wi hou a ec ing he g a nma icali y o he sequence. (29) a. La a eu e he PAST see 'He saw he ' b. Va eu e-la In a pa allel way, my p oposal is ha a e he e b inco po a ion p ocess has applied, he e b complex is successi ely adjoined i s o T, hen o Ag and inally o Neg. 17 Along he same lines as i is cu en ly assumed ha T and Ag a ac V only i hey a e p o ided wi h a ce ain mo phological s eng h, I assume ha , when nega ion is p esen , Neg a ac s V. Fu he inco po a ion o his complex cons i uen o he head o CP is uled ou because no mo phological o syn ac ic in onna ion would be a ached o V. The esul o his head- o-head inco po a ion p ocess is illus a ed in (30). Nego I A no pas A Ag V 1 T No ice ha his p ocess o syn ac ic inco po a ion applies bo h o he igh and o he le . Mo e speci ically, when a e b is mo ed in o de o be adjoined o ano he lexical head (ei he an auxilia y e b o a nega i e ad e b), i inco po a es o he igh . Howe e , when a e b head (e en a complex e b head) is adjoined o a non-lexical head ( o example, Tense and Ag eemen ), hen i inco po a es o he le . Since he e b o ms in V2 (pa iciples, ge unds and in ini i es) ha e al eady unde gone ag eemen a achmen (c . Belle i (1990a,b)), no mo e a ixes can a ach o hem, and he e o e he Tense and Ag eemen ea u es speci ied in (30) can only be a ached o he head o V1. Once his mo phological p ocess has applied, a syn ac ic s uc u e such as (3 1) is ob ained. NegP I is p ecisely a his poin whenpas, ollowing lexical equi emen s peculia o no pas 1, is eo de ed wi hin he new Nego cons i uen and, as a esul , is placed in any pos e bal posi ion ha can be ound wi hin his complex cons i uen . The da a in (32) p o ide e idence in suppo o his claim. (32) a. No ha pas es a ucan no has pas been inging 'I has no been inging' b. No ha es a pas ucan c. No ha es a ucan pas No ice ha wha e e he syn ac ic posi ion o pas may be, no is s ill c-commanding bo h Tense and Ag eemen and, he e o e, he sen en ial scope o his ope a o is adequa ely licensed a he le el o LF. OThis pape was deli e ed a he P ime Coloquio de G amá ica Gene a i a (Mad id, 1991). Many hanks o C. Plcallo o he commen s. This wo k has been pa ly suppo ed by a g an om he DGICYT (numbe PB89-0323). No es 1 This is he dialec b oadly spoken in he p o inces o Ba celona, Gi ona and Ta agona. 2 Ac ually, he s uc u e no X seems o be lexically and syn ac ically es ic ed o ce ain quan i ie s: no o (s) 'no all', no gai e(s) 'no many', o ce ain quan i ie -like ad e bials: no nom3, no solamen 'no only' and o o he speci ic lexical i ems. (i) a. No o s els es udian s ani an a la uni e si a 'No all he s uden s will go o he uni e si y' b. No gai es es udian s ani an a la uni e si a 'No many s uden s will go o he uni e si y' c. No només els millo s es udian s ani an a la uni e si a 'No only he bes s uden s will go o he uni e si y' d. Se celeb a & una con e kncia in e nacional dels paisos no alinea s 'An in e na ional con e ence o he non-aligned coun ies will be held' 3 F om now on each example con aining no pas can be in e p e ed ei he as a p esupposi ion- cancelling ype o nega ion ( ejec ing some o he speake 's assump ions wi h ega d o wha is being said), o as a p esupposi ion-con i ma ion ype o nega ion ( ein o cing hese assump ions) (c . Espinal ( 1990), Wilson-Spe be ( 1988)). Analysing pas as a speci ie (c . Pollock (1989)), o he same so as cli ics, a icles and weak possessi e o ms would inco ec ly p edic ha pas a S-s uc u e should be able o eme ge a he le o no. Howe e , his possibili y only yields ung amma ical sequences.