Expletive Negation

Abstract

The aim of this paper is to exptions which correspond to explthat a syntactic operation of logiory of grammar in order to accit will be characterized by meanative marker or a negative indecontext; it will be characterizedin explicit syntax) or feature motax) to an averidical Neg headKey words: syntax, semantics,

Resum. Negaci expletiva, con

Lobjectiu daquest treball s extacticosemntica que corresponPostular que una operaci sintnecessria en la teoria de la gramen un context no negatiu i la cadun marcador negatiu lleuger oun context negatiu i la caracterdespecificador-nucli a la sintapecificador-nucli no es dna a Paraules clau: sintaxi, semnt

* This paper has benefited frosupported by the Spanish Mcollegues M. L. Hernanz, J. vious manuscript of this papSyntax of Central Romance ference (Groningen, Decem

The content of this paper

CatWPL 8, 2000 47-69

Expletive Negation, Negative Concordand Feature Checking*

M. Teresa EspinalUniversitat Autnoma de Barcelona. Departament de Filologia Catalana08193 Bellaterra (Barcelona). [email protected]

CatWPL 8 001-137 29/5/2001 17:29 Pgina 47lore some aspects of the syntax-semantics interface representa-etive negation (EN) and negative concord (NC). I shall postulatecal absorption, conceived as feature checking, is needed in the the-ount for both phenomena. EN instantiates a nonnegative context;s of a covert negative feature movement, from either a light neg-finite, up to a nonveridical X head. NC instantiates a negative

as either category movement (when the Spec-Head relation holdsvement (when the Spec-Head relation does not hold in overt syn-

. negation, Catalan, Spanish.

cordana negativa i comprovaci de trets

plorar alguns aspectes de les representacions de la interfcie sin-en a la negaci expletiva (EN) i a la concordana negativa (NC).ctica dabsorci lgica, concebuda com a verificaci de trets, stica per tal de donar compte dambds fenmens. La EN apareix

racteritzar mitjanant el moviment encobert dun tret negatiu des dun indefinit negatiu a un nucli X0 no verdic. La NC apareix enitzar com el moviment duna categoria (quan hi ha una relacixi explcita) o com el moviment dun tret (quan la relaci des-la sintaxi explcita) a un nucli Neg0 averdic.ica, negaci, catal, espanyol.

m two project grants (PB93-0893-C04-01 and PB96-1199-C04-02)inistery of Education and Culture. I would like to specially thank myMaci, C. Picallo, G. Rigau and J. Sol for their observations to a pre-er. I would like to thank the audiences at the Third Workshop on theLanguages (Girona, December 1996) and at the Going Romance con-ber 1997) for helpful comments.remains basically as was presented to these conferences.

Received: March 3rd 2000Accepted: October 3rd 2000

1. AimIn this paper I shall first pin (1) and (2) (see below),head which can be non-ture (from now on, FNeg).

Second, I shall describtions with EN and configuchecking, feature attractionveridical and averidical negfor the phenomenon of neg

Third, I shall explore sthe theory of grammar.

Some representative exin (1) and (2).(1) Expletive Negation

a. Preferiraprefer+COND.1trabajando todoworking whoI would rather

b. A cuntas pto how many pSo many people

c. (No) senot himDatCLI must have told

(2) Negative Concord (a. Nadie dijo

nobody said.3sNobody said an

b. No me llamnot meCL calleHe never called

48 CatWPL 8, 2000 M. Teresa Espinal

1. I thank M. Llusa Hernanz fo

Table of Contents1. Aim

2. EN. Structural Configuration3. NC in Catalan and Spanish

4. Similarities and Differences betweenEN and NC5. Related IssuesReferences

CatWPL 8 001-137 29/5/2001 17:29 Pgina 48resent a brief grammatical analysis of the sentences by postulating that their syntactic structure contains aovert defined with an inherent negative formal fea-

e some similarities and differences between configura-rations with NC. Theoretical notions such as feature, and feature strength of negation, with regard to non-ative features, will become relevant in order to accountative absorption in natural languages.everal issues related to this analysis of EN and NC in

amples of EN and NC, taken from Spanish, are given

(EN)salir con vosotros que (no) estar

sg go-out with you than not beel fin de semana.

le the end of weekgo out with you than be working the whole weekendersonas (no) habr matado este dictador!eople not have+FUT.3sg killed this dictator must have been killed by this dictator!lo habr dicho veces esto! 1itCL have+FUT.1sg told times this him this so many times!NC)

nada.g anythingything. nunca.

d.3sg never me.

r bringing this example to my attention.

Notice that Spanish, like Catalan, typically shows a light negative marker no inpreverbal position. A light negative marker refers to its head status (Haegeman 1995),and to the fact that it appears preverbally and is always adjacent to the verb. The samenegative marker no, is used for pure sentential negation and nonnegative negation.

2. EN. Structural ConfigurationEN has been characterized as a natural language phenomenon induced by speci-fic lexical items (such as Italian finch, and Catalan ms in (3)) appearing underspecific structural conditio

(3) Italiana. Resto finch

stay.1sg until Ill stay until s

Catalanb. Val ms que

better that sols.2aloneIts better you

These data illustrate noin specific syntactic envireffective contributions to tstituent (see the English gthe fact that a negative itemify the truth value of the p

The relevant syntactic

(4) [ ... [XP X [CP C[NeBesides the classical s

another instance of non-ne

(5) Spanisha. A cuntas

to how manyjuventud!youth(S)he must hav

Expletive Negation, Negative Concord and Feature Checking CatWPL 8, 2000 49

2. The nonoptionality of the ncontiguous que complemen

3. X is the lexical item whoseoccurs when minimality is rwhen there is no logical oper

CatWPL 8 001-137 29/5/2001 17:29 Pgina 49ns.

(non) arriva qualcuno.not arrives somebody

omebody arrives.

vingueu que no que us quedeucome.SUBJ.2pl than not that youCL remain.2pl

come than you stay on your own.

t only the fact that the negative marker (non, no) appearsonments, but also that this negative marker makes no

he interpretation of the whole string containing this con-losses). What is characteristic of expletive negation is, which lexically contributes to negation, does not mod-

roposition in which it occurs. configuration for (1a) and (3) is given in (4):3

gP Neg ... ]]]]et of examples illustrating EN, natural languages offergative negation in degree wh- sentences.

personas no engaara en supersons not deceive+COND.3sg in his/her

e deceived so many people in his/her youth!

egative marker in this particular example must be related to the twotizers at the output string. inherent FF license an expletive reading for the negative word. Thisespected among the contiguous nodes X-que-Neg within the tree, andator intervening between X and Neg at the level of LF (Espinal 1992).

Catalanb. Quin desastre no heu fet!

what disaster not have.2pl madeWhat a mess youve made of everything!

It seems entirely plausible that the syntactic configuration corresponding to thesentences which license nonnegative negation in wh-exclamatives (such as (1b,c)and (5)) should be analysed in similar terms to clausal structures licensing EN(such as (1a) and (3)) (see Espinal 1997). In accordance with these hypotheses Ihave postulated a degree prwith the feature FNeg (besi

The relevant syntactic sgiven in (6).5

This hypothesis is supshow cross-categorial degrdegree) applies over differpreted as an affective quanquantifiers; the verb, altho


4. The basic contrast between of the fact that in the set of elogical content might be saidlicenses EN cannot be the wan affective predicate (Laduwords are not specialized fo

5. The order NegP < AgrP is nomodal interpretation of the f

(6) DegP

Deg

C

CatWPL 8 001-137 29/5/2001 17:29 Pgina 50ojection above CP with an implicit affective Deg( defineddes other FF), since it licenses EN.4tructure of degree wh-exclamatives which license EN is

ported by the fact that sentences such as (1b) and (5)ee, in the sense that gradability (more specifically, highent categories of the sentence: the wh quantifier is inter-tifier, it is non-specific, and has wide scope over otherugh in indicative, has an intensional interpretation (see

classical instances of EN and EN in degree wh- sentences is a resultxamples given in (1b,c) and (5) there is no apparent lexical item whose to be responsible for the phenomenon of EN. The constituent which

h- word, for it is neither a comparative marker denoting inequality norsaw 1980, Haegeman 1995), and it is absent in (1c). Furthermore wh-r exclamatives.t considered as relevant at this point. MP is postulated for the nonfutureuture marker (see examples (1b,c)).

CP

NegP

Neg TP

T (MP)

(M) VP

V

the glosses in (1b) and (5)); and the negative marker does not entail falsity of theproposition.6

In support of postulating a functional category DegP we can adduce variousoutput conditions, since a sentence with an intensifier Deg constituent has specificsuprasegmental and semantic properties which build into particular phonetic andsemantic interpretations. In addition to that, in support of postulating a DegP pro-jection above a CP projection consider the following four arguments (Espinal1997).7

First, the sentences in (1b) and (5), always have a wh- expression in clausalonset position. The only wa CP1 (+ WH( from a CP2may or may not be inheremore there are degree sentheless license a degree intthe best scenario seems to concept of complementizeprojection.

Second, it should be nindependent clauses. If thexpect to find expletive exthe case. Therefore, Deg s

Third, Deg must be hiand unboundedness of whthese wh- expressions areclause, under whose scopjection postulated for ques

Fourth, some overt dedegree position in (6) exitive degree words such aslicense EN.9


6. Furthermore, this hypothesstructural similarities to focand Greek. There exist, howwh-constituents with inheretential initial position and within a sentence, (iii) the itrasted with the degree intetive constructions, and (iv) EN.

7. See Postner Zanuttini (199tive CP higher than the CP p

8. See Bosque - Masullo (19969. See Grevisse (1986) for a d

are claimed to be linguistic ea quality or a property expre

CatWPL 8 001-137 29/5/2001 17:29 Pgina 51ay to avoid postulating a DegP would be to distinguish( + WH, + DEG(. However, since most wh- expressionsntly lexically specified as degree markers, and further-tences which do not have a wh- expression, but never-ensification on a bare plural noun (see the example (1c)),be one which dissociates the concept of degree from ther, and where the Deg projection is separated from the C

oted that exclamative sentences entailing EN are alwayse degree operator was fused with the C, then we wouldclamatives in subordinate clauses, which, in fact, is nothould be separate from C.gher than C in clause structure, since the nonspecificity- expressions in exclamatives can only be explained if affected at the level of LF by some constituent of thee they are licensed, which is distinct from the CP pro-tions.8gree operators which seem to alternate with the covertst in both Catalan and Spanish. I am referring to posi- b, bien and si really, so, well, highly, which do not

is should be related to the observation that wh- expressions showused expressions in a number of languages, among them Hungarianever, some important differences between focalized constituents andnt degree modification: (i) the latter must obligatorily move to sen-

cannot remain in situ, (ii) there can only be one degree constituentdentificational interpretation of focalized constituents should be con-rpretation of wh- constituents at clause onset position of exclama-sentences with a focalized constituent do not license by themselves

8) for an analysis of exclamatives according to which they have a fac-ostulated in questions.) for a study on the relation between gradability and unboundedness.escriptive analysis of bien and si in French as degree adverbs. Theyxpressions which affect the intensity of an action expressed by a V, ofssed by an A or an Adv.

(7) Catalana. B que s ho ha cregut!

really that seReflCL itCL has believed(S)he really did believe that!

Spanishb. Bien me has engaado.

really meCL have.2sg deceivedYou really did deceive me!

c. Si ser atrevido!very be+FUT.3sg daring(S)he is so daring!

These linguistic expresthey do not license EN. Cfeatures which make themdegree markers of the sorhaving negative features o

Let me now briefly con

3. NC in Catalan and SpNC is characterized as a lia sentence. NC is involvedonce, although it is expreinvolves negative markers ative indefinites in generaalthough they need not beCatalan n-words).

Unlike Germanic langthe phenomenon of NC. Tative marker with negativedo not cancel each other oevant data is given in (8) a

(8) Catalana. No ha vist a

not has seen to(S)he has seen

b. Ning (no) hnobody not hNobody has se


CatWPL 8 001-137 29/5/2001 17:29 Pgina 52sions may be taken as evidence for a DegP, even thoughlose to them, other degree expressions contain negative optimal licensers for EN; for example, overt and covertt exemplified in (3b) and (5). My analysis will rely onn these heads.sider the syntactic structure corresponding to NC data.

anishnguistic phenomenon spread over various items within in syntactic contexts where negation is interpreted justssed more than once. In both Catalan and Spanish it

such as no and other constituents (e.g. n-words, and neg-l) defined by means of an abstract formal feature FNeg, morphologically negative (consider the case of most

uages, it is well-known that Romance languages showhey have the possibility of combining a sentential neg- indefinites, in such a way that the negative constituentsut, but they jointly express a single negation. Some rel-nd (9).

ning.anybody

no one.a vist res.as seen anythingen anything.

(9) Spanisha. No ha visto a nadie.

not has seen to anybody(S)he has seen no one / (S)he has not seen anybody.

b. Nadie (*no) ha visto nada.nobody not has seen anythingNobody has seen anything.

The postverbal negative constituent a ning, a nadie no one is obligatorilypreceded either by a negawith no and postverbal neexpresses sentential negaNotice the lack of parenth

Examples (8b) and (9verbal no in Catalan and aHowever, it should be note(consider the data in (10position of negative elemeform negative polarity ite

(10) Old Spanish (Llora. Ninguno non

nobody not Nobody dares

Dialectal Spanishb. para que ya

so that ana arreglarse ato get-ready iso that nobomore

These examples haveing to one of two perspe(Zanuttini 1991, Haegemaversal well-formedness conative meaning, and predichave the properties of univtion operation when a negsecond hypothesis conceiLaka 1990, 1993), and pronly under an operator-va


10. See Espinal (2000) for a receit is argued that n-words aretifiers, underspecified for q

CatWPL 8 001-137 29/5/2001 17:29 Pgina 53tive head no or by another negative item. The examplesgative constituents are similar to sentences in which notion, in that in both cases pre-verbal no is obligatory.eses on no in (8a) and (9a).

b) illustrate a preverbal negative item with optional pre- non-overt negative marker in modern standard Spanish.d that both in Old Spanish and in some dialectal variants

)) no is overt, regardless of the preverbal or postverbalnts, which Suer (1995) takes as an argument for a uni-

m status of negative elements in Spanish.ens 1929)los ose defender.themCL dares defend to defend them.(Snchez Ferlosios El Jarama)

nunca nadie no venga jamsymore never nobody not come+SUBJ.3sg ever

mi casan my housedy never ever would come to get ready in my home any-

been accounted for in the literature by basically adher-ctives. A first hypothesis postulates a NEG-criterionn-Zanuttini 1991, Haegeman 1995), conceived as a uni-dition on syntactic representations, which cancels the neg-

ts that n-words are inherently negative expressions whichersal quantifiers (i.e. n-words are involved in an absorp-ative operator has scope over a number of variables). Aves NC as a polarity construction (Bosque 1980, 1994;edicts that n-words are polarity items which are licensedriable structure.10

nt analysis on the status of n-words in Catalan and Spanish. In this paper indefinites incorporated into a numeral meaning (0(, and weak quan-uantificational force.

In the next section I shall compare EN with NC and will show that movementof negative features (sometimes strong negative features) to specific targets pro-vide important insights into the relation between the syntactic structure and thesemantic interpretation of these representations.

4. Similarities and Differences between EN and NCThe question I would now like to approach is what sort of relation, if any, existsbetween EN and NC. I will claim that both these constructions share: clause-boundedness and strict structural locality to an X/Neg head with spe-

cific morphosyntactic negative absorption, c sensitivity to nonverid

4.1. Clause-boundednessFrom a structural perspecfact that both these phenolocality to some X F Negcal operator in EN structtures).

Consider the example

(11) Catalana. Abans que pa

before that han aniria.enCL go+COBefore anythi

b. Em temomeCL am-afraiIm afraid tha

c. *Abans [quebefore thatme n anmeCL enCL go

d. *Em temo [director]]meCL afraid tdirector


CatWPL 8 001-137 29/5/2001 17:29 Pgina 54 features,onceived as feature checking, andicality, of which a subset is averidicality.

and strict localitytive the similarity between EN and NC stems from themena require not only clause-boundedness but also strict(which in section 4.3 will be claimed to be a nonveridi-ures and an averidical operator in the case of NC struc-

s in (11) and (12).

ssi (*gaireb/absolutament) res, jo meppens almost/absolutely anything I meCL(EN)ND.1sg

ng happens, I would leave.que no escullin nou director. (EN)

d that not elect+SUBJ.3pl new directort a new director would be elected.

contesti [que passa res, joanswer+SUBJ.3sg that happens anything, I

iria]

que diguin [que no escolliran nou

hat say+SUBJ.3pl that not elect+FUT.3pl new

(ONLY GRAMMATICAL AS NEGATIVE)

(12) a. No funciona (gaireb/absolutament) res com haurianot works almost/absolutely anything as have+COND.3sgde funcionar. (NC)of work(Almost/absolutely) Nothing works as it should.

b. Res no funciona com hauria de funcionar. (NC)anything not works as have.COND.3sg of workNothing works as it should

c. *Res s vernothing is trude funcionar.of work

At his point it is imporLatin noun, can be interptext. Similarly, no may alscontext; in fact it is nonne

Furthermore, in (11a) existential nonnegative radverbs and has a negativ

Examples (11c,d) are unegative indefinite and theordinate clause and are no

In minimalist terms Imovement has applied (seture checking, that is, covetural closeness.11 See (13)


11. In Espinal (1992) it is arguand categorial (i.e. ( N, ( V)

(13) XP

X

Fneg

CatWPL 8 001-137 29/5/2001 17:29 Pgina 55itat [que no funciona com hauriae that not works as have.COND.3sg

tant to notice that res, although it derives from a positivereted as either anything or nothing, depending on con-o get a nonnegative reading depending on the syntacticgative in (11b) but fully negative in (12a,b).res cannot be preceded by almost/absolutely and has aneading, whereas in (12a) res can be modified by thesee reading.ngrammatical under an expletive reading because both the light negative marker are embedded within a second sub-t adjacent to X (abans in (11c) and temo in (11d)). would like to claim that in EN constructions, once Ce a lexical form such as finch in (3a)), EN involves fea-rt movement of a FNeg to check another FNeg, under struc-.

ed that the CP projection, whose head que lacks inherent referential features, in addition to not blocking government due to its intrinsic

CP

ti NegP

Neg

Fneg

This movement takes place at the level of LF because in this configuration X,specified from the lexicon with a particular FNeg that identifies it as the target of adislocation property, attracts the FNeg that identifies the thing that is to be dislo-cated: either a Neg head or an n-word, under the Last Resort and the Minimal Linkconditions (Chomsky 1995:280,297,311).

(14) a. K attracts F if F is the closest feature that can enter into a checking relation with a sublabel of K

b. Last Resort CoMove F raises Fwith a sublabel

c. Minimal Link CK attracts ( onlattracts (

According to these defture attraction operation mthe feature checker in theerty: nonveridicality. Theing the polarity items resX head (abans, temo). O(see section 4.2).

On the other hand, in the negative marker and thative features of these consappropriately for the paraddency that holds betweenthey entail the falsity of p

Concerning NC structby the strength of the FNeg oWatanabe (1997), since in in NC must take place in ture of the phrasal expressstrong FNeg need a head forative marker no in Catalan a


properties, does not block mentizer bears temporal inis bound to the referential feaed and is submitted to a proical head which governs it,the input to a logical absorp

12. This claim can be made com1993 and Giannakidou 1997allow long-distance licensin

CatWPL 8 001-137 29/5/2001 17:29 Pgina 56ndition to target K only if F enters into a checking relation

of Kondition

y if there is no (, ( closer to K than (, such that K

initions, EN can be said to illustrate the relevance of a fea-otivated by the properties of the target. My claim is that

target category is a FNeg with a specific semantic prop-refore, EN involves dislocation of the FNeg characteriz-and no in (11a,b), and adjunction of this feature to thence this movement has taken place, absorption applies

the case of NC structures, clause-boundedness betweene n-word at syntax, and strict locality between the neg-tituents at the level of LF are required in order to accountigm in (12), as well as for the averidical semantic depen- the two constituents in italics (that is, for the fact that).12ures I am going to assume that feature checking is forcedf specific negative items in the checking domain. Following

NC structures FNeg is interpretable, the movement involvedovert syntax, and is triggered by the strong negative fea-ions, not by the Neg head. Phrasal expressions with the the purpose of checking, but not viceversa, since the neg-nd Spanish can occur without phrasal negative expressions.

absorption of negation, and seems to be invisible at LF. The comple-formation, since the TP of the subordinate clause (in the subjunctive)tures of the main clause. Therefore, it is claimed that CT cannot be delet-cess of raising and incorporation (i.e. category movement) to the lex-

either in the syntax or at LF. The output of complementizer raising istion operation.patible with the observation already described in the literature (Quer

) that certain contexts (such as mood and specific selecting predicates)g of negative items.

Elimination of superfluous FNeg is part of the interpretive processes contingentupon movement to a Neg head. In Catalan, movement involving NC items makesuse of either adjunction through pure F movement (see (12a)), or an Spec-Headrelation with a Neg head which can be full via category movement (as illus-trated in (12b)).

In the Spanish example (9b), like Italian, the null Neg head checks the fea-tures of the negative item through category movement, for the Spec-Head relationholds in overt syntax. In (9a) the Neg head, represented by no, checks the featuresof the negative polarity item through feature checking, since the Spec-Head relationdoes not hold in overt syn

In sum, NC in Romanvated by feature strength. (as in (12a)) or to the maxieither F movement or catetural representation in (15

To summarize, the phetarget holding a particulaa case of F movement in ceither category or F moveFNeg .13

4.2. Negative Absorption In (16) I quote Giannakidapproach to absorption, wNC postulated here.


13. I leave for future research thment in overt syntax.

(15) NegP

Neg

weak Fneg

CatWPL 8 001-137 29/5/2001 17:29 Pgina 57tax.ce languages illustrates the relevance of movement moti-The output of this movement is adjunction to either Negmal projection of Neg (as in (12b)), depending on whethergory movement is the relevant operation. See the struc-).

nomenon of EN involves the process of F attraction to ar set of semantic features, and can be accounted for asovert syntax, whereas the phenomenon of NC involvesment in overt syntax, motivated by the strength of the

as Feature Checkingous conception of the NEG-criterion and Watanabeshich seem to be both relevant to the analysis of EN and

e conditions that control the application of either category or F move-

...

NegItem

strong Fneg

(16) a. Giannakidou (1997:182) reinterprets the NEG criterion as an agree-ment requirement with respect to averidicality.

b. Watanabe (1997) argues that absorption can be regarded as a conse-quence of overt movement of pure features (p.19). Movement involv-ing NC items makes use of either adjunction through pure feature move-ment or the Spec-head relation via category movement (p.13). Themovement involved in NC takes place in overt syntax and it is triggeredby the strong feature of the phrasal expressions, not by the Neg head(p.14).

In the light of the precEspinal 1992) might be renation of superfluous negafeature checking, and feaveridicality, in the case ofof LF. This logical deletionupon movement.

(17) absorbs , = Xiffwith regard to a con(i) [XP Spec [X X(ii) has been Attr

In the output LF confthere are no maximal projtive absorption correspondsor of negation (in terms pof natural language configuwith a nonveridical negatirations licensing NC is asnegative feature (FAver ).

Thus, logical absorptiofeature or category movemcertain negative constituependent negative conceptsto provide an interesting apparently different synta

4.3. Sensitivity to Nonveri

The next hypothesis I wotions of different kinds ofmeans of the same theore


CatWPL 8 001-137 29/5/2001 17:29 Pgina 58eding discussion, absorption of negation (as defined ininterpreted in feature checking terms, that is, as elimi-tive features. Logical absorption is to be understood as

ture deletion of either sensitive FNeg (sensitive to non- EN) or redundant FNeg (in the case of NC) at the level operation is part of the interpretive processes contingent

FNeg and = FNeg of a negative item

figuration such as

FNeg ]]acted/Moved to either Spec,XP or to X

iguration ( and ( mutually c-command one another, forection boundaries between them. The process of nega-s to the factorization of one single and abstract expres-ut forward by Ladusaw 1992, 1996), which in the caserations licensing EN is associated with a head X( defined

ve feature (FNonver), and in the case of syntactic configu-sociated with the Neg head defined with an averidical

n of negation, conceived as feature checking, after eitherent has applied, provides an explanation of the fact thatnts in specific configurations are not licensed as inde-. Furthermore, the operation of feature checking seemsline of research on the syntactic relation between twoctic structures: EN and NC.

dicalityuld like to point out is that EN and NC are manifesta- sensitivity to negation, which can be accounted for bytical tools.

Since what is characteristic of EN is the presence of a negative expression(either a negative marker no or an n-word) which is absorbed by an attracting head,my claim is that EN is obtained when the negative marker is interpreted as a polar-ity item. Let us consider the following definitions:

(18) a. Polarity item (Giannakidou 1997:14,16)(i) A polarity item a is an expression whose distribution is limited bysensitivity to some semantic property b of the context of appearance;(ii) The semantic property bneed not be polar

b. Condition on P(i) A polarity itesion z which sudepends on, and(ii) z is said to b

Accordingly, to claimitem means that it is a sensto the definitions in (18)text of grammaticality, annegative marker is license

Following Montague (1(nonveridicality and averidi

(19) (Non)veridicality (ZLet Obe a monadic Op p is logicallynonveridical operat

Following Giannakidodependent definition given

(20) Context dependent (In a context c,(i) A propositional orequires that p be tr(ii) An operator Op require that p be tru(iii) A nonveridical requires that p be fa

These definitions view(AVERIDICAL ( NONV


14. For connections between m15. M(x) stands for some individ

CatWPL 8 001-137 29/5/2001 17:29 Pgina 59I-licensingm a is licensed iff (a) the context provides some expres-

pplies the semantic property b as proper interpretation (b) a is found in the semantic scope of z

e the trigger of a

that the negative marker in EN structures is a polarityitive expression, that is, an expression that according can only be licensed by a property bpresent in the con-d this property is nonveridicality.14 That is, in EN thed under the scope of a nonveridical operator.969) and Zwarts (1995), veridicality and the related notionscality) are viewed as properties of propositional operators.

warts 1995:287)sentential operator. O is said to be veridical just in case valid. If O is not veridical, then O is nonveridical. A

or O is called averidical iff Op p is logically valid

u (1997), (non)veridicality can be assigned the context in (20).

non)veridicality (Giannakidou (1997:110))

perator Op is veridical iff the truth of Op p in cue in some model M(x) in cis nonveridical iff the truth of Op p in c does note in any model M(x) in coperator Op is averidical iff the truth of Op p in clse in any model M(x) in c15

averidical operators as a subset of the nonveridicalERIDICAL), so every averidical operator is also non-

ood and non-veridicality, see Quer (1998).uals worldview and represents the epistemic status of that individual.

veridical, but not viceversa. Furthermore, they relativize (non)veridicality withrespect to context. We can infer from them that a propositional operator is veridi-cal iff Op entails p, that is, whenever Op p is true, p is true too. Op is nonveridicaliff Op does not entail p, that is, whenever Op p is true, p may or may not be true.Nonveridical operators do not entail the falsity of p. Entailing the falsity of p is thedefining property of averidical operators.

In EN the negative marker is to be thought of as a polarity item sensitive to a spe-cific semantic property: the nonveridicality supplied by before, until, comparativeoperators denoting inequality, a specific high degree operator, and some negativeand adversative predicates.in the semantic scope of anthe negative marker is licmanifest a polarity item st

(21) Catalana. Abans que (no

before than noBefore anybod

b. Abans que etbefore than yoBefore anyboy

Notice, furthermore, ththe proposition expressed ing you may or may not b


16. See Espinal (2000) for an aallow the occurrence of n-wtive concord structures, areproperty and a variable undtions on the LF building pro

17. EN is also allowed in the abexample (1b)), but it is not (i d)).

(i) a. En Joan mor aDet Joan died bJoan died before h

b. ?En Joan mor aDet Joan died bJoan died before h

c. En Joan revisDet Joan checkedJoan checked the

d. * En Joan revisDet Joan checke

CatWPL 8 001-137 29/5/2001 17:29 Pgina 60 When either a light negative marker or an n-word appears expression which supplies a nonveridical property, then

ensed as nonnegative (or expletive) and n-words makeatus.16 This is illustrated in (21a-b).

) et vegin, vs- te n.t youCL see+SUBJ.3pl go+IMP youCL enCLy sees you, you should go. vegi ning, vs- te n.uCL see+SUBJ.3sg anybody go+IMP youCL enCL sees you, you should go.

at in (21) abans before is nonveridical with respect toin the subordinate clause, in the sense that anybody see-e true.17

nalysis predicting that languages (such Catalan and Spanish) whichords in polar and negative contexts, as well as in expletive and nega- those that have n-words lexically defined with a specified negativeerspecified quantificational force, subject to various licensing condi-cess.ans before clause in those contexts where it is inferred that ~ q (seelicensed in those contexts where it is inferred that q (as illustrated in

bans de conixer els seus nts. ( ~ q)efore of get-to-know the his grandchildrene got to know his grandchildren.

bans que no conegu els seus nts. ( ~ q)efore that not got-to-know+SUBJ.3sg the his grandchildrene got to know his grandchildren.

el cotxe abans de comprar-lo. ( q)the car before of buying itCL

car before he bought it.el cotxe abans que no el comprs. ( q)

d the car before that not itCL bought+SUBJ.3sg

Lack of veridicality is also illustrated in (22), since this example shows thatSpanish hasta until is nonveridical with refard to its internal argument: the factthat the speaker be thrown out may or may not be true.

(22) SpanishMe quedar hasta que (no) me echen.meCL stay+FUT.1sg until that not meCL throw.3plI will stay until they throw me out.

Similarly, expletive neas Catalan tmer fear) anto the proposition expresthat it will rain may or mbe false.

(23) Catalana. Em temo

meCL am-afraiIm afraid that

Spanish (R. Carnicb. Hay quin

is there whohistorias?storiesDoes anybody

Finally, the high degrplement clause, for the pr


18. The linguistic literature onexclamative clauses are facis interesting to notice thatoccur under the scope of a

(i) Catalana. s inslit com

is unusual howIt is unusual how

b. *s increble qis incredible w

(ii) Spanisha. Es extraordinario

is outstandingIt is outstanding h

b. *Me sorprendemeCL surprises

CatWPL 8 001-137 29/5/2001 17:29 Pgina 61gation is licensed under the scope of verbs of fear (suchd verbs of doubt, which are all nonveridical with respect

sed in the subordinate clause. This is illustrated in (23):ay not be true, and the referred stories may or may not

que (no) plogui.d that not rain+SUBJ.3sg it will rain.er, La Vanguardia)dude que (no) son falsas las talesdoubt+SUBJ.3sg that not are false the such

doubt the falseness of such stories?

ee operator is also nonveridical with regard to the com-oposition expressed may or may not be true.18

exclamatives (Elliott 1974, Grimshaw 1979) shows that positive wh-tives, that is, they presuppose the truth of the proposition. However, it neither Catalan nor Spanish wh- exclamatives which license EN canfactive verb, as the following contrasts make explicit:

es porta de malament!esReflCl behaves of bad

bad (s)he behaves!uines mentides no diu en Joan!hich lies not says Det Joan

cuntas tonteras habr dicho este estudiante!how many stupid-remarks have+FUT.3sg said this student

ow many stupid remarks may have said this student!cunto no pesarn estos paquetes!how much not weigh+FUT.3sg these parcels

(24) CatalanQuantes mentides (no) deu haver dit des que estem casats!how many lies not must.3sg have told since that are.1pl married(S)he must have told so many lies since we got married!

To conclude, EN instantiates a nonnegative context, in which the negative mark-er is a polarity item sensitive to some specific nonveridical semantic dependen-cies. The target FNeg is associated with some lexical head (other than Neg) whichhas nonveridical properties. EN has been formulated as a case of covert F move-ment, from either the lighFNeg of the target nonveridtional operators, and averiaveridical FNeg of a light item can be cancelled undlexical heads.

By contrast, NC instanis an averidical operator wassociated with the Neg hminimalist program, sincechecking, triggered by theing domain, we have claimthe Spec-Head relation hoSpec-Head relation does nmovement has applied, a r

In Section 4 I have shoboth syntactically and semedness, strict locality andconclusion in a study on tinterface of negative structuer no (Meibauer 1990) invdifference between EN andis associated with the NegEN the target FNeg is assohas nonveridical propertieattraction, whereas in NC syntax.

Accordingly, expletivetion is cancelled, in such anegative operator. This apptational function: to checkthe local domain. This doeno can be licensed differeexpressed by various mean

This analysis has the arelations hold in complex


CatWPL 8 001-137 29/5/2001 17:29 Pgina 62Negt negative marker or the negative indefinite, up to theical operator. Since nonveridicality concerns proposi-

dical expressions form a subset of the nonveridical, thenegative marker or the nonveridical FNeg of a negativeer feature checking with a nonveridical FNeg of specific

tiates a negative context, in which the negative markerhich licenses negative polarity items. The target FNeg isead of the clause, which has averidical properties. In the negative agreement has been reinterpreted as feature

strength of the FNeg of the negative items in the check-ed that NC requires either category movement (when

lds in explicit syntax) or feature movement (when theot hold in overt syntax) to an averidical Neg. Once thisedundant FNeg is absorbed.wn that the similarity between EN and NC is important,antically. Both these phenomena require clause-bound- negative absorption, which, I think, is a significant

he configuration of negation and the syntax-semanticsres. There is only one NegP and a single negative mark-olved in different licensing conditions. The structural NC is due to the fact that whereas in NC the target FNeg head of the clause, which has averidical properties, inciated with some lexical head (other than Neg) whichs. In addition to that, in EN F movement is a result of Fcategory or F movement is forced by F strength in overt

no is a negative lexical item whose semantic contribu- way that this marker is not licensed as an independentarent imperfection of language has of course a compu-

the morphosyntactic and semantic properties present ins lead to an interesting conclusion: the negative markerntly depending on the fact that negation itself can bes in natural languages.

dditional advantage that it can account for the meaningexamples such as those in (25).

(25) Catalana. Abans que ning digui res, deixeu -me

before that anybody say+SUBJ.3sg anything let meCLdonar -vos la benvinguda. (EN)give youCL the welcomeBefore anybody says anything, let me welcome you.

b. Abans que ning no digui res, anullarem (NC)before that anybody not say+SUBJ.3sg anything cancel+FUT.1plla reuni.the meetingBefore nobody

As said with regard toand is nonveridical with clause. Therefore, in (25a) ning and res are sensitiveabans, defined with a nonvand res. On the other hanwords, there is a Neg heaveridical FNeg. In this conand F movement of the FNture (with F checking of clause), thus entailing an

5. Related IssuesIn this section I shall mentmy analysis of EN and NC

5.1. The relevance of the dand C/Q (see the spli

Notice the contrast betweeator (such as si, and bien (26) Spanish

a. Cuntas vechow many timI must have to

b. Si sereally him.DatCI must have to

c. Bien de vecesso of timesI have told him


CatWPL 8 001-137 29/5/2001 17:29 Pgina 63 says anything, well cancel the meeting.

(21) Catalan abans is a lexical item which licenses PIregard to the proposition expressed in the subordinateanybody says anything may or may not be true, and both to the nonveridical semantic dependency introduced byeridical FNeg which attracts and checks the FNeg of ning

d, in (25b), in addition to the presence of abans and n-ad in the subordinate clause defined with an inherenttext, both category movement of the negative item ningeg of res to the Neg head apply, thus building a NC struc-negative features with regard to the Neg head of theaveridical interpretation.

ion some issues, independently motivated, which support.

istinction made in linguistic theory between Degt degree system hypothesis postulated by Corver 1997)n Spanish sentences which contain an overt degree oper-really, so, highly), and sentence (1c).

es se lo habr dicho esto!es him.DatCL itCL have+FUT.1sg told thisld him this so many times!

lo habr dicho veces esto!L itCL have+FUT.1sg told times thisld him this so many times!le he advertido. (Moliner 1982)himDatCL have.1sg noticed so many times!

Example (26a), as said before, has a covert degree operator, whereas (26b-c)have an overt degree operator in italics. (26c) shows that, when the degree spreadsnot over the whole sentence but over a noun, bien is next to an explicit QP.

On the other hand, if we consider an example such as (1c) (repeated here for con-venience),

(27) SpanishNo se lo habr dicho veces esto!not him.DatCL itCLI must have told him

we shall be able to notice (26b), and the semantic paoperator (inherently definno.

This paradigm suggestsentence such as (26b), whshould look like (28),

(28) [DegP Si F Pos [CP/QP C

whereas (27) is assumed t

(29) [DegP Deg F Neg [CP/Q

5.2. The relevance of the ditems (Bosque 1996, G

Notice the ungramaticatain either bien or si and a

(30) a. *Bien no sequite not him

b. *Si no sequite not him.

This phenomenon wathat bien is an adverb of emphatic P, thus in compsis put forward in this papcome from the lexicon wiLF these lexical items arepolarity.

By contrast, a sentenceat the level of LF must not


CatWPL 8 001-137 29/5/2001 17:29 Pgina 64have+ FUT.1sg said times this this so many times!

that, in spite of the syntactic superficial similarity withraphrase relation with (26a,b), (27) has a covert degreeed with a nonveridical property) which licenses the PI

s that the syntactic representation that corresponds to aich has a degree head defined with the feature [+ POS],

/Qi [ eci[ veces ]]]]

o have the syntactic structure in (29).

P C/Qi[ NegP no F Neg [ eci[ veces ]]]]]

istinction between triggers of polarity and polarityiannakidou 1997, Progovac 1994).lity, illustrated in (30), of Spanish sentences which con-

negative marker.

lo puedo haber dicho veces esto..DatCL itCL may.1sg have told times this

lo habr dicho veces esto.DatCL itCL have+FUT.1sg told times this

s first described by Hernanz (1995), who postulatedpositive polarity generated in the Spec position of anlementary distribution with no. According to the analy-er, I shall claim that bien and si are degree heads whichth the FF [+ DEG, + Q, + POS], so that at the level of interpreted as veridical operators that trigger positive

such as (27) contains a light negative marker no, whichbe interpreted as a trigger of negative polarity but, rather,

as a polarity item with regard to a nonveridical Deg operator, as has been repre-sented in (29).

5.3. The relevance of F movement at LF (Chomsky 1995)Consider the data in (31):

(31) Spanisha. A cuntas personas no matar este dictador!

to how many pHow many peo

b. Qu de tonterwhat of nonseThis student ta

c. * Qu de tontwhat of nons

The analysis of degre[Spec,DegP] (either coverand no + V movement intomeaning of the future tensnonspecific reading of thexpressions that reject the1997).

5.4. The relevance of F mo(Watanabe 1997)

Although raising without pthe minimalist program it ition of movement feeds th(1995), a set of FF is carriethat in covert raising F movical item. It is not movemenframework, since this mov

However, F movementtax. Thus, the cooccurrencexclamative wh- sentenceof the FF [+DEG, -PLU] w(32).20


19. See En (1996)20. I am most grateful to Jaume

CatWPL 8 001-137 29/5/2001 17:29 Pgina 65ersons not kill+FUT.3sg this dictatorple this dictator must be killing!

as no dir este estudiante!nse not say+FUT.3sg this studentlks such nonsense!

eras en concreto no dir este estudiante!ense specifically not say+FUT.3sg this student

e wh-exclamatives requires both wh movement intot, in the case of cuntas, or overt in the case of qu de), Deg, in order to account for the non-referential modale19, the expletive reading of the negative marker and thee wh-expressions (i.e. wh-phrases in exclamatives are specific interpretation, as illustrated in (31c); Espinal

vement and F pied-piping in overt syntax

ied-piping is more economical in some natural sense, ins assumed that FF raise along with F because the opera-e feature checking operation. According to Chomsky

d as a unit when only one of them is attracted. This meansement automatically pied-pipes the rest of FF of the lex-t of the whole category, as understood in the preminimalistement is not required for convergence at PF. and F pied-piping seem also to be relevant in overt syn-e of the Dutch quantifier een a with a plural noun in

s can be accounted for by postulating overt movementhich define een,in this specific configuration. Consider

Sol for bringing these examples to my attention.

(32) Dutcha. Wat heeft hij niet een vragen gesteld!

what has he not a questions raisedHe raised so many questions!

b. Wat heeft hij niet een ellende veroorzaakt!what has he not a mess causedHe created such a mess!

c. Wat een problemen!what a proHow many pr

It is interesting to notiniet + een is not fused intsupport of Watanabe (199I would like to suggest tha

The singular determinexclamative wh- sentencetures licensing EN. A natuis an optional and interpreLF. [PLU] of een mismatcancel out the derivation.take with it raising of the checking configuration wition with [+ DEG], since t

5.5. The relevance of bothnon-derivational op-Ctheory

Notice that examples sucderivational chain formedstrength of FWh.

(33) [wati [[ ti een ] vr

Consider now the Spa

(34) Spanisha. Cuntas vec

how many timI must have to

b. No senot him.DatCL


CatWPL 8 001-137 29/5/2001 17:29 Pgina 66blemsoblems!

ce that in Dutch this is one of the very few cases whereo geen, which actually seems to provide an argument in7)s hypothesis. Accordingly, in this particular structuret F movement takes place in overt syntax.er een a can occur with a plural noun exclusively ins, which in Modern Dutch are the only possible struc-ral way to account for this could go as follows: [PLU]table feature, and because of this it cannot be deleted atches [+ PLU] of vragen, and mismatch of features should Therefore, raising of the [+ DEG] feature of een must[PLU] feature, so that [ PLU] would be adjoined in ath [+ DEG], although it would not be in a checking rela-hese features do not match.

derivational chains (Chomsky 1995) andHAIN representations (Brody 1993) in linguistic

h as the one in (32a) make explicit the existence of a by movement, this movement being motivated by the

agen] ]

nish data in (34).

es no se lo habr dicho esto!es not him.DatCL itCL have+FUT.1sg said thisld him this so many times!lo habr dicho veces esto! (= 1c)itCL have+ FUT.1sg said times this

I must have told him this so many times!c. *Cunta vez no se lo habr dicho esto!

how many time not him.DatCL itCL have+ FUT.1sg said thisd. *No se lo habr dicho vez esto!

not him.DatCL itCL have+ FUT.1sg said time this

Examples (34a,b) are important because they both show EN in a degree sen-tence; but whereas (34a) makes explicit a chain formed by wh- movement, (34b)makes explicit a chain conin degree exclamatives the only with mass nouns andmassness or plurality.

The analysis of Spanis(expletive) operator-CHAImar in order to account forin (32a,b) and the Spanish eifier of either a plural nounconnection to an antecedethe form in (35).

(35) [ Opi [[ eci] veces

Evidence for this assumed in the [Spec,NP] positioveces times requires a quanegative reading corresponhave a nonveridical degree

(36) a. *Esto no sethis not him

b. Esto no sethis not him.DI may not have

The chain representatioabstract operator, which hathere would be an operatoSpanish example in (34b),(32a,b)). The initial expletmine a specific phonetic ingenerated in [Spec,NP] wou


21. I thank M. Llusa Hernanz fo

CatWPL 8 001-137 29/5/2001 17:29 Pgina 67figuration without overt movement. (34c,d) show thatdegree marker cannot combine with singular nouns, but plural nouns, which are intensified for their feature of

h degree sentences makes explicit that representationalNS la Brody (1993) are needed in the theory of gram- the syntactic parallelism between the Dutch examplesxample in (34b), since in both structures there is a spec- or a mass noun that must be identified through chain-nt. The structure corresponding to (34b) can be given

] ]

ption and specifically for the empty category postulat-n stems from the following contrast, which shows thatntifier.21 Consider the ungrammaticality of (36a) and theding to (36b), whose syntactic representation does not operator at head initial position.

lo habr dicho veces..DatCL itCL have+ FUT.1sg said times

lo habr dicho muchas veces.atCL itCL have+ FUT.1sg said many times said this to him many times.

ns postulated in (33) and (35) would have at the edge ans a suprasegmental correlate, and at the end of the chainr in [Spec,NP] (which could be either empty, as in the or filled by a singular Det, as in the Dutch examplesive operator would be a scope marker and would deter-terpretation, licensed through intonation. The operatorld be marked with FDeg and this feature would be attract-

r bringing this point to my attention.

ed to the higher Deg operator at LF, which would determine its appropriate seman-tic interpretation.

The theoretical implication of this argument is that the grammar of natural lan-guages seem to require non-derivational op-CHAIN representations, generated byprinciples of grammar and restricted by conditions on well-formedness, as well asderivational chains created by feature checking needs.

To conclude, there are several independent issues which support some of theassumptions made in the analysis of EN and NC put forward in this paper: the pos-tulation of a DegP (distinct from CP/QP), the distinction between triggers of polar-ity and polarity items, thesyntax, besides F movemeop-CHAIN representation

ReferencesBosque, I. (1980). Sobre la (1994). La negacin y

Gramtica del espaol. Mxico: El Colegio de

(1996). La polaridad mmanuscript.

Bosque, I.; Masullo, P. (19Complutense de Madridin Fullana, O.; Roca, F. Girona: Universitat de G

Brody, M. (1993). Lexico-College London, unpub

Chomsky, N. (1995). The mCorver, N. (1997). Much suElliott, D.E. (1974). Toward

p. 231-246.En, M. (1996). Tense and

porary semantic theoryEspinal, M.T. (1992). Ex

Review 9.4, p. 333-358. (1997). Non-negative n

P.; Martineau, F.; RiveroAmsterdam: John Benja

(2000). On the semantp. 557-580.

Giannakidou, A. (1997). Thtion.

Grevisse, M. (1986). Le bonGrimshaw, J. (1979). Com

10.2, p. 279-326.Haegeman, L. (1995). The s


CatWPL 8 001-137 29/5/2001 17:29 Pgina 68 postulation of F movement and F pied-piping in overtnt at LF, and, finally, the postulation of nonderivationals, besides derivational chains.

negacin. Madrid: Ctedra.el principio de las categoras vacas. In Demonte, V. (ed.)Publicaciones de la Nueva Revista de Filologa Hispnica.Mxico, p. 167-199.odal. Universidad Complutense de Madrid, unpublished

96). On verbal quantification in Spanish. Universidad and Universidad Nacional del Comahue, manuscript. Also(eds.). Studies on the syntax of central Romance languages.

irona, 1998.logical form. A radically minimalist approach. Universitylished manuscript.inimalist program. Cambridge, MA: MIT Press.pport as a Last Resort. Linguistic Inquiry, 28.1, p. 119-164. a grammar of exclamations. Foundations of Language 11,

modality. In Lappin, S. (ed.). The handbook of contem-. Oxford: Blackwell, p. 345-358.pletive negation and logical absorption. The Linguistic

egation and wh- exclamatives. In Forget, D.; Hirschbhler,, M.L. (eds.). Negation and polarity. Syntax and Semantics.mins, p.75-93.ic status of n-words in Catalan and Spanish. Lingua 110,

e landscape of polarity items. Groningen, Ph.D. disserta-

usage. Paris: ditions Duculot.plement selection and the lexicon. Linguistic Inquiry

yntax of negation. Cambridge: Cambridge University Press.

Haegeman, L.; Zanuttini, R. (1991). Negative heads and the Neg Criterion. TheLinguistic Review 8, p. 233-252.

Hernanz, M.L. (1995). Bien y la polaridad positiva en espaol. Ms. Universitat Autnomade Barcelona. Also in Polaridad y modalidad en espaol: en torno a la gramtica debien. Bellaterra: Universitat Autnoma de Barcelona, Research Report GGT-99-6.

Ladusaw, W.A. (1980). On the notion affective in the analysis of negative-polarityitems. Journal of Linguistic Research 1.2, p. 1-16.

(1992). Expressing negation. University of California Santa Cruz, Syntax ResearchCenter 92-03.

(1996). Negation and polarity items. In Lappin, S. (ed.). The handbook of con-temporary semantic theory. Oxford: Blackwell, p. 321-341.

Laka, I. (1990). Negation in syntax: on the nature of functional categories and projec-tions. MIT, Ph.D. dissertation.

(1993). Negative fronting in Romance: movement to (. In Ashby, W.J. (ed.).Linguistic perspectives on the Romance languages. Amsterdam: John Benjamins,p. 315-333.

Llorens, E. (1929). La negacin en espaol antiguo con referencia a otros idiomas.Anejo de la Revista de

Meibauer, J. (1990). Sentenicht in German. Ling

Moliner, M. (1982). DiccioMontague, R. (1969). On

son, R.H. (ed.). FormaHave: Yale University P

Postner, P.; Zanuttini, R. (19rogatives. Georgetown

Progovac, L. (1994). NegatiCambridge University P

Quer, J. (1993). The syntacBarcelona, MA disserta

(1998). Mood at the inGraphics.

Suer, M. (1995). NegativeReview 12, p. 233-273.

Watanabe, A. (1997). AbsInternational Studies, u

Zanuttini, R. (1991). Syntacof Romance languages.

Zwarts, F. (1995). Nonver


CatWPL 8 001-137 29/5/2001 17:29 Pgina 69Filologa Espaola. Madrid.nce mood, lexical category filling, and non-propositional

uistische Berichte 130, p. 441-463.nario de uso del espaol. Madrid: Gredos. the nature of certain philosophical entities. In Thoma-l philosophy. Selected papers of Richard Montague. Newress, 1974, p. 148-187.98). The force of negation in wh- exclamatives and inter- University, manuscript.on and positive polarity: A binding approach. Cambridge:ress.tic licensing of negative items. Universitat Autnoma detion.terface. Ph.D. Thesis. Amsterdam: Holland Academic

elements, island effects and resumptive no. The Linguistic

orption as feature checking. Tokyo: Kanda University ofnpublished manuscript.tic properties of sentential negation: a comparative study

University of Pennsylvania,Ph.D. dissertation.idical contexts. Linguistic Analysis 25, p. 286-312.

Expletive Negation, Negative Concord and Feature CheckingAbstractResum. Negaci expletiva, concordana negativa i comprovaci de trets

Table of Contents1. Aim2. EN. Structural Configuration3. NC in Catalan and Spanish4. Similarities and Differences between EN and NC4.1. Clause-boundedness and strict locality4.2. Negative Absorption as Feature Checking4.3. Sensitivity to Nonveridicality

5. Related Issues5.1. The relevance of the distinction made in linguistic theory between Deg and C/Q5.2. The relevance of the distinction between triggers of polarity and polarity items5.3. The relevance of F movement at LF (Chomsky 1995)5.4. The relevance of F movement and F pied-piping in overt syntax (Watanabe 1997)5.5. The relevance of both derivational chains (Chomsky 1995) and non-derivational op-CHAIN representations

References

Documents

Expletive Negation