20
Three types of negation in Brazilian Portuguese Lilian Teixeira de Sousa * UFBA, Rua Bara ̃ o de Jeremoabo, 147, Campus de Ondina, Cep 40.170-115 Salvador, Bahia, Brazil Received 2 February 2014; received in revised form 1 March 2015; accepted 9 March 2015 Available online 8 April 2015 Abstract Sentential negation in Brazilian Portuguese (BP) may be accomplished through three different kinds of structures: Neg1 [Neg VP], Neg2 [Neg VP Neg] and Neg3 [VP Neg]. This distribution is quite rare in natural languages, which usually feature only one structure to express sentential negation, with an optional structure to convey a discourse function. In the linguistic literature, it is frequently claimed that the Neg1 structure is the only one devoid of syntactic restrictions. This idea has been presented as an argument for its status as the standard negation form in BP. Schwenter (2005) has explained the other two constructions in terms of information structure. However, he also recognizes some differences between the informational status of Neg2 and Neg3. In this study, I aim to describe the occurrence of these structures in BP in an attempt to formulate an analysis that considers both distributive characteristics and possible interpretations of these structures. © 2015 Elsevier B.V. All rights reserved. Keywords: Negation; Proposition; Event; Brazilian Portuguese 1. Introduction Several typological studies (Dahl, 1979; Payne, 1985; Jespersen, 1917) have distinguished three formal possibilities for expressing sentential negation in natural languages: preverbal, Neg1; embracing, Neg2; and post-verbal, Neg3. Brazilian Portuguese (BP), 1 however, exhibits all three structures, 2 all of which are productive and not in competition: (1) a. A Ana na ̃ o/num 3 foi ao teatro (Neg1) the Ana neg/neg-CL went-2PS to the theater Ana didnt go to the theaterb. Eu na ̃ o/num fui no teatro na ̃ o(Neg2) I neg/neg-CL went-1PS to the theater neg I didnt go to the theaterwww.elsevier.com/locate/lingua Available online at www.sciencedirect.com ScienceDirect Lingua 159 (2015) 27--46 * Tel.: +55 7132836207. E-mail address: [email protected]. 1 EP also displays the three types of negation found in BP, but Neg2 and Neg3 are even more restricted in their distribution and licensing conditions. Neg2 in EP, unlike in BP, can only be interpreted as metalinguistic negation, and the final clause na ̃ o is not integrated with the clausal spine, which may be recognized because of the presence of a pause between the sentence and the negative item. According to Pinto (2010), Neg3 in EP expresses metalinguistic negation, but, unlike in BP, it cannot appear in answers to polar questions; the negative item is also separated from the sentence spine by a pause and there is always an explicit rectification. 2 South Brazil is an exception. According to Teixeira de Sousa (2012), people from south Brazil only recognize Neg1 as part of their dialect. 3 In the preverbal position, the negative item nãois frequently reduced to the clitic form num. http://dx.doi.org/10.1016/j.lingua.2015.03.003 0024-3841/© 2015 Elsevier B.V. All rights reserved.

3 Types of Negation in BP

  • Upload
    mb

  • View
    220

  • Download
    0

Embed Size (px)

DESCRIPTION

Negation, minimalist syntax

Citation preview

Page 1: 3 Types of Negation in BP

Available online at www.sciencedirect.com

ScienceDirect

Three types of negation in Brazilian Portuguese*

www.elsevier.com/locate/linguaLingua 159 (2015) 27--46

Lilian Teixeira de SousaUFBA, Rua Barao de Jeremoabo, 147, Campus de Ondina, Cep 40.170-115 Salvador, Bahia, Brazil

Received 2 February 2014; received in revised form 1 March 2015; accepted 9 March 2015Available online 8 April 2015

Abstract

Sentential negation in Brazilian Portuguese (BP) may be accomplished through three different kinds of structures: Neg1 [Neg VP], Neg2[Neg VP Neg] and Neg3 [VP Neg]. This distribution is quite rare in natural languages, which usually feature only one structure to expresssentential negation, with an optional structure to convey a discourse function. In the linguistic literature, it is frequently claimed that the Neg1structure is the only one devoid of syntactic restrictions. This idea has been presented as an argument for its status as the standard negationform in BP. Schwenter (2005) has explained the other two constructions in terms of information structure. However, he also recognizes somedifferences between the informational status of Neg2 and Neg3. In this study, I aim to describe the occurrence of these structures in BP in anattempt to formulate an analysis that considers both distributive characteristics and possible interpretations of these structures.© 2015 Elsevier B.V. All rights reserved.

Keywords: Negation; Proposition; Event; Brazilian Portuguese

1. Introduction

Several typological studies (Dahl, 1979; Payne, 1985; Jespersen, 1917) have distinguished three formal possibilitiesfor expressing sentential negation in natural languages: preverbal, Neg1; embracing, Neg2; and post-verbal, Neg3.Brazilian Portuguese (BP),1 however, exhibits all three structures,2 all of which are productive and not in competition:

(1)

* Tel.:E-m

1 EP

conditiospine, wNeg3 inseparat

2 Sou3 In th

http://dx0024-38

a.

+55 7ail addalso disns. Neghich m

EP exed fromth Braze prev

.doi.org41/© 2

A Ana na o/num3 foi ao teatro (Neg1)

the Ana neg/neg-CL went-2PS to the theater ‘Ana didn’t go to the theater’

b.

Eu na o/num fui no teatro na o(Neg2) I neg/neg-CL went-1PS to the theater neg ‘I didn’t go to the theater’

132836207.ress: [email protected] the three types of negation found in BP, but Neg2 and Neg3 are even more restricted in their distribution and licensing2 in EP, unlike in BP, can only be interpreted as metalinguistic negation, and the final clause nao is not integrated with the clausalay be recognized because of the presence of a pause between the sentence and the negative item. According to Pinto (2010),presses metalinguistic negation, but, unlike in BP, it cannot appear in answers to polar questions; the negative item is also the sentence spine by a pause and there is always an explicit rectification.il is an exception. According to Teixeira de Sousa (2012), people from south Brazil only recognize Neg1 as part of their dialect.erbal position, the negative item ‘na o’ is frequently reduced to the clitic form ‘num’.

/10.1016/j.lingua.2015.03.003015 Elsevier B.V. All rights reserved.

Page 2: 3 Types of Negation in BP

L. Teixeira de Sousa / Lingua 159 (2015) 27--4628

4 Newresulting

5 The

c.

appro in stru

sole o

Fui no teatro na o(Neg3)

went-1PS to the theater neg ‘I didn’t go to the theater’

The possibility of multiple sentential negations is frequently associated with a process of language change calledJespersen’s Cycle. Jespersen noted that the position of the negation may indicate different stages of a particularlanguage. The process of change, according to him, historically begins with preverbal negative elements which arereduced and incorporated into the verb while other post-verbal elements emerge and acquire a negative valence. Dahl(1979) continues Jespersen’s work, suggesting the existence of a negative cycle with the following stages: Neg V > Neg VNeg > V Neg. In this process of language change, the co-occurrence of more than one negative structure is expectedduring a certain period of time, as result of two grammars in competition.4

Despite the Neg3 and Neg2 possibilities, BP is assumed to be a language where negation is expressed in terms of apre-verbal element, since Neg3 has a discursive function and the clause-final nao in Neg2 structures is taken as an actualnegative concord item (see Biberauer and Cyrino, 2009a,b; Vitral, 1999; Cavalcante, 2007). The possibility of negativeconcord, per se, means that BP is a language where the head is lexically realized. According to Wood (1997), if thenegation is in the head position, it is weakened and does not interfere with a second negative. If the negative alternation inBP does not represent a stage of Jespersen’s Cycle, in which the head of the NegP is not lexically realized, what would bethe nature of the negative alternation in BP?

The first studies on this topic (Schwegler, 1991; Roncarati, 1997) associated the alternation of negative structures toemphasis or pragmatic presupposition, i.e., the embracing and post-verbal forms were understood as denials of someoneelse’s belief. Recently, Schwenter (2005) has argued that the presuppositional explanation of the function of thesestructures cannot be supported empirically. According to him, the non-canonical negative structures are related toinformation structure: the proposition denied by embracing negation, for instance, is contextually activated. In this paper,I will argue that the difference between the three structures is not precisely related to emphasis, presupposition or contextactivation but to their interpretation as either event, propositional or metalinguistic negation. According to this approach,each type of formal negation in BP has a different interpretation.

The paper is organized as follow: in section 2, I discuss the two main approaches to sentential negation in BP --emphasis/presupposition (Schwegler, 1991; Cavalcante, 2012) and activation by context (Schwenter, 2005). I examinedata on embracing and post-verbal negation in BP, which cannot be explained by presupposition or context activation. Insection 3, I also present some arguments that show that each type of negation has restrictions due to its interpretation.Section 4 concludes the paper by summarizing the results.

2. Negation in BP

2.1. Negative alternations in BP: emphasis, presupposition and context activation

Analyses of the alternation of negative structures in BP tend to associate the occurrence of the non-canonical formsNeg2 and Neg3 with pragmatic issues. Without effectively distinguishing between the three structures, especiallybetween Neg1 and Neg2, certain authors allude to reinforcement or emphasis, to presuppositional negation or to contentactivated by context. In this study, I will bring new evidence showing that Neg2 cannot be treated the same way as Neg3,given that Neg2 functions as semantic negation, unlike post-verbal negation, which, according to analyses from previousstudies, is strictly discursive. In this section, I will discuss some approaches to alternation of negative structures in BP, withrespect to the following topics: emphasis/presupposition and context activation.

Schwegler (1991) was one of the first researchers to address the issue of multiple negations in BP and the first todistinguish Neg2/3 from Neg1. For him, the non-canonical forms are distinguished from the canonical in that they are usedto contradict presuppositions. Though Schwegler does not use the word ‘‘emphasis’’, it is implicit from his indication of thepresence of intonational prominence in these structures. Roncarati (1997) also refers to Neg2 as implicitly emphatic andrestricts Neg3 to the sentence sei nao5 (‘‘I don’t know’’). Furtado da Cunha (1996) refers to ‘‘contrary to expectation’’, andBiberauer and Cyrino (2009a,b), who are the first to distinguish Neg2 from Neg3 as different structures, speak ofreinforcement.

The notion of emphasis, previously underscored by Schwenter (2005), is not clearly specified in studies about thealternation of negative structures in BP. According to the author, the role of emphasis in explaining this alternation is

aches (Biberauer, 2009; Gelderen, 2008; Ingham, 2007) have recognized that the process may stop at one of the steps involved,ctural reorganization. It might have occurred in BP. Such an analysis, however, would need a historical approach.ccurrence in the corpus she collected.

Page 3: 3 Types of Negation in BP

L. Teixeira de Sousa / Lingua 159 (2015) 27--46 29

related to the idea of reinforcement, described as an important step of Jespersen’s Cycle. In this case, the use of anadditional negative element (Neg VP Neg) is interpreted as emphasis, but, as the author himself adds, there is nonecessary relationship, intuitively, between reinforcement and emphasis; despite containing an additional negativeelement, Neg2 seems no more emphatic than Neg1.

According to Schwenter (2005), a distinction between the non-canonical structures Neg2 and Neg3 that focuses onemphasis does not cover certain facts about their use in BP, because only Neg2 has the additional phonological elementto justify the interpretation of these structures as emphasis.

Understanding emphasis as creating prosodic prominence, one can observe that prominence may occur in a pre-verbal na o in Neg1 structures or even in the clitic num, which indicates that, unlike the French ne, the pre-verbal negativemarker in BP maintains the property of emphasis. This characteristic is even clearer in contexts of expletive negation, inwhich the na o/num has no negative value, but rather one of emphasis:

(2)

6 Accextraling

E nao/num cheguei atrasada! (=I was really late!).

and neg/neg-CL arrive late

Thus, I validate Schwenter’s intuition (2005) that the criterion of emphasis cannot distinguish between the threestructures. That is, if the pre-verbal na o in BP, despite having a clitic form, is still able to receive prosodic prominencefluently as well as negate the sentence on its own, it is not logical to consider the co-occurrence of another negativeelement as having the sole function of reinforcement. This leads to the conclusion that the information simply denied, oremphatically denied in the case of a structure using pre-verbal na o, and that of an embracing structure ([Neg VP Neg]),are not effectively the same. Thus, it also does not seem correct to analyze the Neg2 structure as negation with greateremphasis ([Na o+neg VP Na o+emphasis]).

Regarding the emphatic nature attributed to Neg2 by various authors (Biberauer and Cyrino, 2009a,b; Furtado daCunha, 1996; Schwegler, 1991), I propose that this is not sufficient to explain the use of this structure, given that thepreverbal na o in BP, despite having a clitic form, is able to receive prosodic prominence and can itself be emphatic.Although Neg2 may be interpreted as emphatic, as we shall see, this is not a characteristic intrinsic to the structure.Stressing Schwenter’s point, given that negative polarity items have the same emphatic function on negative information --Eu nao tenho um centavo no bolso (‘‘I’m completely broke’’), the label ‘‘emphasis’’ is insufficient to distinguish thephenomenon of two co-occurring negative items in the structure [Neg VP Neg] from other clearly distinct phenomena.

Cavalcante (2012) also rejects the explanation of the final na o structures as emphasis. According to him, the structuresNeg2 and Neg3 do not encode emphasis, given that the structure with the negative marker nada ([VP nada]), in acomplementary distribution with [VP Neg], is the emphatic negative structure of BP. In this proposal, the Neg2 and Neg3structures are related to the negation of presupposed content.

Regarding the description of the Neg2 and Neg3 structures being related to presupposition, Schwenter states that thisterm is also used to describe canonical negation, given that many authors (Horn, [1989]2001; Givon, 1978) argue thatnegatives are more marked in terms of presuppositions when compared to affirmatives -- every negative sentencepresupposes an affirmative one. Thus, according to Schwenter, the presuppositional content would not suffice todistinguish between the canonical and non-canonical forms.

Cavalcante (2012), however, claims that what distinguishes the post-verbal na o is not precisely its presuppositionalnature, but rather the way in which the presupposition is placed in the discourse. According to him, in Neg1 case, thenegative sentence does not require the affirmative proposition to be previously introduced in the discourse, as it is thenegative sentence that draws the interlocutor’s attention to the embedded affirmative proposition. Meanwhile, Neg2 andNeg3 sentences correspond to the negation/rejection of the proposition put forward or made available in the discourse.Thus, for these kinds of sentences to occur, it would be necessary for the presupposed proposition to correspond to aprevious assertion, explicit or inferred from a communicative situation. This interpretation is quite similar to Schwenter’sproposal of context activation (2005); however, as we shall see, there is an important distinction to be made between theinformative nature of the negated element and the use of Neg2 and Neg3.

On the idea of presupposition, Schwenter refers to the work of Stalnaker (1978, 1974) on pragmatic presupposition, inwhich the author argues that presuppositions are taken from propositions located in the interlocutors’ common ground,6

but these presuppositions are not necessarily shared knowledge in the case of negations. Thus, considering the argumentthat Neg2 in BP presupposes a previous affirmation or contradicts an assertion, it is understood that Neg2 contradicts aproposition that is in the common ground, which indicates a pragmatic and not a semantic presupposition. Once again,taking into account Stalnaker’s analysis, this explanation can be extended to Neg1. Schwenter argues, then, that Neg2

ording to Stalnaker (1978), common ground may be defined as information that has been previously given in the discourse oruistic context and which is shared (or assumed by the speaker to be shared) between the listener and the speaker.

Page 4: 3 Types of Negation in BP

L. Teixeira de Sousa / Lingua 159 (2015) 27--4630

cannot be used to deny any pragmatic presupposition7 or proposition in the common ground, but would rather berestricted to the negation of salient propositions.

In his analysis, Schwenter adopts Price’s model (1992), which makes an important distinction between the status of thediscourse and the status of the listener of a referent NP, with two possible values for each: new and given. Neg2 in BPwould be, according to him, sensitive to the status of the discourse, but not of the listener. To justify his analysis, he offersthe following example:

(3)

7 Thebackgro

8 All w

[the speaker is walking down the street and remembers that she has left her oven on]

Nossa! (Eu) nao desliguei o fogao (#na o)! damn (I) neg turned-off-1PS the stove (neg) ‘Damn! I didn’t turn off the stove!’

(Ex. (2a) -- Schwenter, 2005)8

According to Schwenter, the use of Neg2 in the above sentence is inappropriate because the statement is directed atthe speaker herself, which defines her as new in the discourse. If another interlocutor were added to the same scenario in(3), the status of the discourse may be considered given and thus, according to the author, Neg2 becomes appropriate tothe context:

(4)

[the same situation as (3)] A:

speakeund of

ord-by

Você desligou o fogao, né?

you turned-off-1PS the stove, ‘You turned off the stove, right?’

B:

Nossa! Na o desliguei nao! damn neg turned-off-1PS neg ‘Damn! I didn’t turn it off!’

(Ex. (2b) -- Schwenter, 2005).

In addition to the concepts of given and new, Schwenter holds that inferable propositions also have a role in the use ofNeg2. In the sentence below, the speaker infers that the interlocutor believes that it is cold; in this case, the use of Neg2 isappropriate:

(5)

[the speaker sees the interlocutor putting on a coat that is much too heavy for the current weather] Na o tá muito frio na o! neg is very cold neg ‘It’s not very cold out!’

Despite there being a breach of expectation in many data of Neg2, Schwenter argues that the use of Neg2 is not relatedto the breach of expectation, but rather to a proposition activated by some given material. Thus, in an attempt tounderstand the nature of Neg2, Schwenter then analyzes this structure using the typology of negations proposed byGeurts (1998), which includes: (1) negation of proposition (descriptive negation); (2) negation of presupposition; (3)implicature (scalar); and (4) form (pronunciation, lexical choice, etc.).

This typology further extends Horn’s ([1989]2001) already discussed binary distinction between descriptive negationand metalinguistic negation. In this model, metalinguistic negation is divided into negation of the presupposition, theimplicature and the form. Thus, this perspective focuses on the target of the negation’s objection; that is, whether what isbeing denied is a proposition, a presupposition, an implicature, or some aspect of linguistic form.

According to Schwenter, the use of Neg2 as negation of presupposition has already been established and offers, then,some data that enable a judgment of whether this structure can also be used as a negation of presupposition, implicatureor form.

Regarding the presuppositional content, Schwenter offers evidence that Neg2 cannot be used in contexts in which thepresupposed content is not contained in the asserted content, but rather in a previous statement. Observe the contrast inthe following examples:

r’s presuppositions, according to Stalnaker, may be considered those propositions whose truth is assumed as part of thethe conversation.-word glosses in Schwenter examples were included in this paper.

Page 5: 3 Types of Negation in BP

L. Teixeira de Sousa / Lingua 159 (2015) 27--46 31

(6)

a. A: O Joao já deixou de fumar. the Joa o already stopped-3PS of to-smoke ‘J. has stopped smoking.’

B:

Ele na o deixou de fumar (#na o). Ele nunca fumou. he neg stopped of to-smoke (neg) I never smoked-3PS ‘He hasn’t stopped smoking, he never smoked.’

b.

A: O Joao já deixou de fumar. the Joa o already stopped-3PS of to-smoke ‘J. has stopped smoking.’

B:

Ele na o deixou de fumar (nao), ele ainda fuma. he neg stopped-3PS of to-smoke (neg) he still smokes ‘He hasn’t stopped smoking, he still smokes.’

(Ex. (14a--b) Schwenter, 2005).

In the examples presented in (6), the presupposition of statement A is that Joa o has smoked in the past, and it is thispresupposition that is denied in (6a), considering what follows. Meanwhile in (6b), the interpretation made evident by thecontinuation of the statement is that the negation applies to the asserted content and not to the presupposition. As the datashow, Neg2 is not appropriate when applied to presuppositional elements.

Also in cases of denying the implicature, Schwenter offers data that make the incompatibility of Neg2 with expressionsof scalar value evident:

(7)

a. Eu na o gosto do meu professor (#na o). Eu adoro ele! I neg like-1PS of-the my professor (neg) I adore-1PS he ‘I don’t like my professor, I adore him!’

b.

Eu na o gosto do meu professor (nao). Eu odeio ele! I neg like-1PS of-the my professor (neg) I hate-1PS he ‘I don’t like my professor, I hate him!’

(Ex. (15a--b) Schwenter, 2005).

While in (7a) it is the scalar value of the verb gostar (like) that is denied, as shown by the subsequent use of the verbadorar (adore), in (7b) it is the truth of the proposition Eu gosto do meu professor that is denied, as occurs in a descriptivenegation. Thus, it is clear that Neg2 is also not appropriate when applied to scalar values.

Analyzing the fourth type of negation, negation of form, Schwenter once again demonstrates the distinction betweenNeg1 and Neg2:

(8)

a. A: Ele trouxe (trúsi) feija o pra festa. he brought-3PS beans to-the party ‘He brought [‘‘incorrect’’ pronunciation] beans to the party’

B:

Ele na o trouxe (trúsi) feijao (#nao), ele trouxe (trowsi) feija o. he neg brought-3PS beans (neg) he brought-3PS beans ‘He dind’t (trúsi) beans, he (trowsi) beans.’ (=metalinguistic form denial)

b.

A: Ele trouxe (‘trusi) feija o pra festa. he brought-3PS beans to-the party ‘He brought [‘‘incorrect’’ pronunciation] beans to the party’

B:

Ele na o trouxe feijao (nao), trouxe arroz. he neg brought-3PS beans (neg) brought-3PS rice ‘He didn’t bring beans, he brought rice.’ (=propositional negation)

(Ex. (16a--b) Schwenter, 2005).

In (8a), the target of the negation and subsequent correction of statement A is the phonetic realization of the verb formtrouxe, which is pronounced as (‘trusi). Meanwhile, in (8b), the target of the negation is not the pronunciation of the verb,but rather the truth of the proposition Ele trouxe feijao pra festa. Thus, Neg2 is also inappropriate when applied to lexical or

Page 6: 3 Types of Negation in BP

L. Teixeira de Sousa / Lingua 159 (2015) 27--4632

register choice. Therefore, Schwenter concludes that Neg2 in declarative sentences is strongly restricted to propositions,so that it corresponds to descriptive negation.

According to Schwenter, this result is surprising, in a way, given that Neg2 is intuitively considered emphatic, which is acommon aspect of metalinguistic negation, given that they are more costly to process and have an effect of temporaryambiguity (Garden-path).

To bolster his hypothesis, Schwenter offers examples in which Neg2 and Neg1 have different interpretations:

(9)

A: O Joa o votou no Lula? the Joa o voted-3PS for-the Lula ‘Did Joao vote for Lula?’

B1:

(Na o.) Na o votou na o. neg neg voted-3PS neg ‘(No.) He didn’t vote (not)’

B2:

(Na o.) Ele nao votou. neg he neg voted-3PS ‘(No.) He didn’t vote (for anyone)’

(Ex. (17) Schwenter, 2005).

According to Schwenter, responses B1 and B2 are not equivalent. The interpretation of B1 is that Joao did not vote forLula, but voted for another candidate; meanwhile, the interpretation of B2 is that Joao did not vote for anyone. Thus,Schwenter holds that, apparently, the function of the final nao in Neg2 structures is to index the response with the mostaccessible given proposition. Testing this same context among native speakers of BP, I notice that none of the informantshad any doubt about the interpretation of B1 as ‘‘Joao did not vote for Lula’’; sentence B2, meanwhile, was considered to beambiguous by most of the speakers, possibly meaning either ‘‘Joao did not vote for Lula’’, or ‘‘Joao did not vote for anybody’’.The justification for the ambiguity, however, does not seem to be related to the type of negation, but rather to the presence ofthe subject in B2, which indicates that there was no movement of the verb to a higher category (∑P) (see Oliveira, 2000).

In certain contexts, according to Schwenter, both Neg2 and Neg3 can be used. A typical example would be responsesto yes-no questions:

(10)

A: Você gostou da palestra da Maria? you liked-2PS of-the talk of-the Maria ‘Did you like Maria’s talk?’

B:

Gostei na o. liked-1PS neg ‘I didn’t’

(Ex. (24) Schwenter, 2005).

In the above data, Neg3 could just as easily be substituted by Neg2. However, as Schwenter points out, both of theseforms are not always possible in the same context:

(11)

A: Você gostou da palestra da Maria? you liked-2PS of-the talk of-the Maria ‘Did you like Maria’s talk?’

B1: #

fui na o. went-1PS neg

B2:

Eu na o fui na o. I neg went-1PS neg ‘I didn’t go’

(Ex. (25) Schwenter, 2005).

The data above brings an important piece to our puzzle. The sentence in B2 introduces a new proposition rather thananswers the question ‘‘Did you like Maria’s talk?’’. The comparison between (10) and (11) exhibits the key differencebetween Neg2 and Neg3; while Neg2 can introduce new propositions, Neg3 is an actual answer, incapable of introducinga negative proposition.

As previously argued, the contrast evident in the above example makes it clear that Neg3 is more restricted in itsdiscursive-pragmatic distribution than Neg2. It seems that Neg3 requires the negation of a proposition that has beendirectly activated by the context of the question, as occurs in (10). As the negated proposition was not explicitly activated in(11), only Neg2 was possible.

Page 7: 3 Types of Negation in BP

L. Teixeira de Sousa / Lingua 159 (2015) 27--46 33

Schwenter further presents some data on Neg3 from the PEUL corpus9 in which an explicitly activated proposition isalso negated. This may be concluded, according to him, from the use of the same verb as the interviewer in the responsein every one of the examples analyzed:

(12)

9 This c

E:

orpus

Você tem vontade de mudar um dia?

you have-2PS desire of to-move one day ‘Do you have a desire move someday?’

F:

Tenho na o. have-1PS neg ‘I don’t’

(Ex. (26) Schwenter, 2005).

(13)

E: Mas você cozinha. E você deve ter algum prato que seus fregueses gostam mais. Qual é? but you cook-2PS. And you must-2PS have some dish that your clients like-3PP more. what is ‘But you cook. And you must have some dish that your clients like most. What is it?’

F:

Ah, eu cozinho na o. Minha tia é que cozinha! ah I cook-1PS neg. My ant is that cooks-3PS ‘Ah, I don’t cook, my aunt is the one that cooks!’

(Ex. (27) Schwenter, 2005).

(14)

E: . . . Você pode comparar isso. . . sensaçao que você tem, quando está desfilando na escola de samba? you can-2PS compare that sensation that you have-2PS when is-2PS parade-GER with-the school of-the samba ‘You can compare that. . . sensation that you have, when you’re parading with the samba school’

F:

Posso na o, duas coisas diferente. can-1PS neg, two things different ‘I can’t, two different things.’

(Ex. (28) Schwenter, 2005).

Schwenter holds, however, that the number of occurrences of this kind of structure in the corpus is too small for there to

be a concrete generalization. On the other hand, the work of Roncarati (1997) seems to include the same kind ofoccurrence, that is, the repetition of the same verb present in a preceding question. Most of the data found by Roncaratiwas the sei nao (I don’t know) construction in response to yes/no questions.

Without elaborating the analysis of Neg3, Schwenter argues that this kind of structure may be understood as a subsetof Neg2, given that both require a proposition to be activated by content given in the discourse. The difference between thetwo structures is in the restriction of Neg3 to contexts in which the proposition is directly activated in the current discourse.

As seen in Schwenter’s framework (2005), the preference for the type of negative structure in BP is strictly related to itsdiscursive status, whether it is new, inferable or explicitly activated.

The data presented by Schwenter on the negative structures Neg2 and Neg3 are extremely important for understandingthis construction in BP, given that they demystify the relationship between these structures and the pragmatic notions ofemphasis and presupposition. Two other issues brought up by the author are the indexical nature of Neg2 and, moreimportantly, the use of this structure as a negation of a proposition. This last point has been essential to the development ofthe present research. There are, however, relevant facts about negative structures in BP that were not discussed bySchwenter and that are critical to understanding the simultaneous occurrence of three different negative structures in thelanguage.

Biberauer and Cyrino (2009a,b) also contrast Neg2 and Neg3 and present important contributions. These authorsclearly distinguish the two structures with respect to omission (ungrammaticality/change in meaning), modifiability,stressability and increase in meaning (presupposition or reinforcement). According to them, Neg2 and Neg3 do not exhibitidentical behavior in terms of omissibility, nao3 (clause-final nao in Neg3) exhibits behavior very similar to the ‘‘real’’ negator,although nao3 clearly does not work in the same range of negation contexts as the real negator. They also observe that onlynao3 is restricted to matrix clauses in the way anaphoric negators are in a more general sense as a consequence of theiroccupying CP-structure not being projected in embedded clauses (cf. Poletto, 2008). In conclusion, they argue that nao2and nao3 in fact derive from different sources: nao2 (clause-final nao in Neg2) is a genuine concord element, lexicalizing aPol-head, i.e. it is fully integrated with the clausal spine (cf. Laka, 1994); nao3 is not a concord element; it is related to theanaphoric negator, which is not fully integrated with the clausal spine (cf. Zanuttini’s (1997) NO).

consists of sociolinguistic interviews with Rio de Janeiro residents recorded in the 1980s.

Page 8: 3 Types of Negation in BP

L. Teixeira de Sousa / Lingua 159 (2015) 27--4634

An important issue considered by Biberauer and Cyrino (2009a,b) is that there are not only discursive, but syntacticrestrictions as well, on the occurrence of Neg2 and Neg3. As discussed, the use of Neg3 is not possible in any kind ofembedded clause, but according to the Teixeira de Sousa (2012), Neg2 is also incompatible with infinitive or temporalembedded clauses.

(15)

*

*

*

*

*

[Complement clauses]

A Maria acha que o Pedro na o comprou o carro. the Maria thinks-3PS that the Pedro neg bought-3PS the car ‘M. thinks P. didn’t buy the car.’

b.

A Maria acha que o Pedro na o comprou o carro na o. the Maria thinks that the Pedro neg bought-3PS the car neg

c.

A Maria acha que o Pedro comprou o carro na o. the Maria thinks that the Pedro bought-3PS the car neg

[Subject clauses]

d. É melhor na o ficar acordado até tarde.

is better neg stay awake until late

‘It is better not to stay up late.’

e.

É melhor na o ficar acordado até tarde na o. is better neg stay awake until late neg

f.

É melhor ficar acordado até tarde na o. is better stay awake until late neg

[Conditional clauses]

g. Se você na o for viajar mesmo, passa lá em casa no fim de semana.

if you neg go-FUT-2PS travel surely go-by-2PS there in home at-the end of week

‘If you are not going traveling, come visit me at the weekend’

h.

Se você na o for viajar mesmo na o, passa lá em casa no fim de semana if you neg go-FUT-2PS travel surely neg, go-by-2PS there in home at-the end of week

i.

Se você for viajar mesmo na o, passa lá em casa no fim de semana. if you go-FUT-2PS travel surely neg, go-by-2PS there in home at-the end of week

[Non-defining relative clauses]

j. Tem político que na o rouba.

have-3PS politician that neg steal-3PS

‘There are politicians who do not steal.’

k.

Tem político que na o rouba na o. have-3PS politician that neg steal-3PS neg

l.

Tem político que rouba na o. *have-3PS politician that steal-3PS neg

(16)

[Infinitive clauses] a. Na o fumar, faz bem à saúde.

neg to-smoke, do-3PS well to-the health

‘Not smoking is better for one’s health.’

b.

Na o fumar nao, faz bem à saúde. neg to-smoke neg, do-3PS well to-the health

c.

Fumar na o, faz bem à saúde. *to-smoke neg, do-3PS well to-the health

(17)

[Temporal clauses] a. Eu na o durmo, enquanto minha filha nao chega em casa.

I neg sleep-1PS until my daughter neg comes in home

‘I can’t sleep until my daughter comes home.’

b.

Eu na o durmo, enquanto minha filha na o chega em casa nao. I neg sleep-1PS until my daughter neg comes in home neg

c.

Eu na o durmo, enquanto minha filha chega em casa nao. *I neg sleep-1PS until my daughter comes in home neg
Page 9: 3 Types of Negation in BP

L. Teixeira de Sousa / Lingua 159 (2015) 27--46 35

Moreover, regarding discursive restrictions, and in spite of Schwenter’s position, one can observe that, although theNeg2 structure occurs more frequently with given information, the structure is also possible in contexts of new information.Observe the following data:

(18)

A: Você sabia que o professor Pedro foi assaltado? you knew-2PS that the professor Pedro was robbed ‘Did you know that professor P. was robbed?’

B:

Nossa! Falando do Pedro, eu na o entreguei o trabalho dele na o. damn speaking of Pedro, I neg gave-3PS the homework him neg ‘Damn! Speaking of P., I didn’t give him my homework.’

In the case above, only the subject, the teacher Pedro, is given information; the proposition, contrary to Schwenter’sprediction, is new information and takes the Neg2 structure. Thus, it seems that it is the presence of propositions in thediscourse that leads to the occurrence of Neg2. The salience mentioned by Schwenter may come from the understandingof Neg2 as propositional, which goes beyond a situation being denied.

Another characteristic of Neg2, which will be an important point in this analysis, concerns the incompatibility of thisstructure in narrative contexts:

(19)

Maria acordou pela manha . Olhou pela Janela. Na o viu sinal de chuva (#nao). Saiu sem seu guarda-chuva. Maria woke-up-3PS in-the morning. looked-out-3PS at-the window. neg saw-3PS sign of rain (neg). left-3PSwithout her umbrella ‘M. woke up in the morning, looked out at the window. She didn’t see a sign of rain. She left without her umbrella.’

Concerning Neg3, Schwenter’s analysis points in an interesting direction: the need for content explicitly activated in thecontext. However, perhaps due to the limitations of the data to which he had access, some important questions on themetalinguistic use of the structure went unraised. Applying the same tests used by the author on Neg2, it is evident that Neg3is possible in every case: negation of a pragmatic presupposition, negation of a scalar value expression and negation of anexpression:

(20)

A: Eu vi a Maria com o pé enfaixado. Ela quebrou o pé? I saw-1PS the Maria with the foot bandaged. she broke-3PS the foot ‘I saw M. with a bandaged foot. Did she break it?

B:

Quebrou na o. broke-3PS neg ‘No, she didn’t.’

(21)

A: Isso vale a pena. it worth-3PS the penalty ‘It’s worth it.’

B:

Vale a pena na o. Vale a galinha inteira! worth-3PS the feather neg. worth-3PS the chicken whole ‘It’s really worth it.’

(22)

A: O Joa o bateu as botas he Joao beat-3PS the boots ‘J. kicked the bucket.’

B:

Bateu as botas na o, faleceu. beat-3PS the boots neg, passed-away-3PS ‘He didn’t kick the bucket, he passed away.’

Biberauer and Cyrino (2009a,b) do not mention the possibility of Neg3 being metalinguistic negation. According to theiranalysis, this structure originates from short answers, which parallel those available in a positive context. This is also animportant context for the occurrence of Neg3:

Positive answers:

(23)

Q: Você tem muitas dívidas? you have many debts ‘Do you have many debts?’
Page 10: 3 Types of Negation in BP

L. Teixeira de Sousa / Lingua 159 (2015) 27--4636

Table 1Distributio

Context

1. Seman2. Metalin3. New in4. Inferab5. Compl6. Condit7. Relativ8. Infinitiv9. Tempo10. Narra

A:

n of Ne

tic negguistic

formatile/expliement cional clae clause clausral clautive

a.

*

g1, Ne

ation

negatioon

citly aclausesuses

es

es

ses

Tenho.

have.1SG ‘Yes’

b.

Tenho muitas dívidas have many debts ‘I have many debts’

c.

Sim, tenho. Yes have.1SG. ‘Yes, I do’.

d.

Tenho, sim have.1SG yes ‘Yes, I do.’

(Ex. (28) Biberauer and Cyrino, 2009a,b).

Negative answers:

(24)

Q: Você tem muitas dívidas? you have many debts ‘Do you have many debts?’

A:

a. Na o. no ‘No.’

b.

Na o, na o/num tenho. No, not have.1SG ‘No, I don’t’.

c.

Tenho na o have.1SG not ‘I do not’

d.

Tenho, na o. *(Ex. (29) Biberauer and Cyrino, 2009a,b)

The distribution of the three negative structures as presented in this section is summarized in Table 1, which takes intoaccount both new observations and facts reported in the literature:

Given the characteristics of Neg1, Neg2 and Neg3 presented above, it is clear that the three structures are notequivalent; there are syntactic restrictions that, as we have seen, apply to Neg3 and not to Neg2. Also, in matters ofinterpretation, it is evident that Neg2 may be used to negate new information, while Neg3 cannot. Furthermore, as hasbeen pointed out by Schwenter (2005), Neg2 denies only propositions, which qualifies it as a semantic negation; Neg3,meanwhile, as shown in the data (20--22), is used as a metalinguistic negation. Neg1, on the other hand, is understood assemantic negation over events, but it retains the possibility of metalinguistic negation expression. Thus, it is clear that thethree structures cannot be treated as a single phenomenon. Based on these conclusions, the next section will deal withthe nature of these structures.

Taking as a salient point the restriction against Neg2 in narrative contexts in both temporal and infinitival embeddedclauses, I propose that this structure corresponds, as does Neg1, to a semantic negation. The difference between Neg1

g2 and Neg3 in BP.

Neg1 Neg2 Neg3

U U *n (presupposition, scalar value, expression) U U U

U U *tivated information U U U

U U *U U *U U *U * *U * *U * *

Page 11: 3 Types of Negation in BP

L. Teixeira de Sousa / Lingua 159 (2015) 27--46 37

and Neg2 is in the scope of what is denied: while Neg1 denies events, the scope of Neg2 includes propositions. Given thepragmatic nature of Neg3, this structure is defined as an external or metalinguistic negation (Horn, [1989]2001).

3. Three structures, three negations

As previously mentioned, explanations of the negative alternation in BP are frequently related to pragmatic categoriessuch as presupposition or context activation. However, as I have claimed, many important syntactic restrictions on thethree negative structures are ignored by the majority of authors. In this section, I will argue that the constraints observed inthese structures are related to their interpretation as event, propositional or metalinguistic.

In the previous section, it was shown that Neg2 and Neg3 cannot be considered equivalent phenomena, as therespective functions of these structures are not effectively equal; while Neg2 is close to Neg1 with regards to its semanticinterpretation, Neg3 is clearly discursively marked. With respect to the observed syntactic restrictions, it has beendemonstrated that Neg3 is far more restrictive than Neg2, the latter being excluded only in narrative contexts and intemporal and infinitive subordinate clauses. According to Schwenter (2005), the Neg2 structures work precisely asdescriptive negation, since it is the proposition that is negated in these structures. However, if both Neg2 and Neg1function as descriptive negations, how does one explain both the coexistence of the two forms in the language and therestrictions associated with Neg2? In an attempt to account for this question, I will utilize the ternary division of the types ofnegation described by Dahl (1979) and seek to associate them with different scopes.

In logical-semantic studies, negation is understood as an operation in which a sentence (S1) is converted into another(S2), so that S2 is true if and only if S1 is false. Many authors (Dahl, 1979; Rajagopalan, 1982; Horn 1989[2001]) claim,however, that this definition is insufficient, as it does not distinguish between types of negation. Dahl (1979), using Englishsentences, asserts that it is possible to distinguish three types of negation:

(25)

10 It is w11 One oby Davidsthat can qand, of cresistancmight pos

It is not raining.

(26) It is false that it is raining. (27) It is not the case that it is raining.

(ex. (1--3) Dahl, 1979)

All of the above sentences fit the same definition of negation. However, (25) is distinct from (26) and (27) in that it doesnot contain an embedded clause, such that, from a grammatical point of view, it is a single sentence. The other twosentences are better understood as denials of a previous statement or assumption made by an interlocutor, being thusdefined as external negation. In this type of negation, the negative sentence ‘not S’ does not comment on the state ofaffairs, but rather denies the truth of judgment S, which was previously stated or implied.

Negation is, thus, broken up into two basic kinds: internal and external. The distinction between these two types ofnegation is oftentimes one of scope. Thus, the scope of a negation may be over the predicate, resulting in ‘A is not B’,or over the statement, resulting in ‘A is B -- is false.’ This distinction between internal and external negation may alsobe called descriptive negation and metalinguistic negation, respectively. Therefore, although Dahl points to threedifferent structures in order to account for the different interpretations of a negation, it is normally divided into onlytwo types, namely the internal and the external. However, upon close analysis of sentence (26), one can observe thatit is the truth value that is at stake, unlike (25), which simply describes a state of affairs, and (27), which deniesthe assertability of an utterance. The concept of a proposition, in semantic theory, is precisely that which can beeither true or false; thus, sentence (26) contains an explicit proposition negation structure. For Kratzer (1989),a proposition is a set of possible situations10; propositions, then, would classify situations into either true orfalse. Thus, if the structure in (26) negates a proposition, then the best interpretation of (25) is as event negation.11

Thus, three types of negation are distinguished: (i) negation of events, (ii) negation of propositions and (iii) metalinguisticnegation.

orth recalling that Kratzer prefers the term ‘‘situation’’ to ‘‘event’’.f the most relevant works on the notion of event in formal semantics is, without a doubt, the article The Logical Form of Action Sentenceson, published in 1967. In his article, Davidson holds that many phenomena of natural language can be explained through logical formsuantify over events. Among the phenomena cited by Davidson are nominalization, adverbial modification, factives, anaphora, pluralsourse, tense and aspect. Given its explanatory power, this model has gained many adherents over the years, but there is still somee precisely because, as Kamp and Reyle (1993) point out, it is difficult to determine what events are and what general properties theysess.

Page 12: 3 Types of Negation in BP

L. Teixeira de Sousa / Lingua 159 (2015) 27--4638

3.1. Proposition, event and negation in BP

In order to better distinguish between Neg1 and Neg2, it is necessary to define both event and proposition. InDavidson’s view, events are entities about which an indefinite number of statements can be made. Thus, the relationshipbetween a sentence and its logical form is not univocal, but rather one with possible variation. This makes the introductionof an existential quantifier critical. According to Davidson, a sentence like Amundsen flew to the North Pole in May 1926does not describe an event, but if it is true, there must be an event that will make it true.

Among the various theoretical approaches to events (Kamp and Reyle, 1993; Kratzer, 1989, among others), the mostdecisive criterion in the line of reasoning that will be followed is the recognition of events as either universal (things thatoccur in different places and times) or specific (things that occur in a specific place and time). For this study, the approachto events as specific is the most relevant. According to this perspective, events are indexically constructed to fill a spatio-temporal location. Thus, tense is understood as an extension composed of instants. These instants are what are termedan event or eventuality. This relationship between time interval and instants is an important point in establishing theconcepts of event and proposition.

According to semantic theories, events speak to a state of affairs and are limited to instants along a time interval, whilethe proposition is always defined in relation to a specific time, which comprises every instant within a time period (Kampand Reyle, 1993; Kratzer, 1995). Thus, a sentence such as Peter works in the garden does not make clear reference to atime interval and, consequently, is interpreted as an event.

For Kratzer (1989), a proposition is a set of possible situations. This means that propositions classify situations intoeither true or false. According to her model, a proposition p is true in a situation s if and only if s is contained in p. If p is nottrue in s, then p is not necessarily false in s. The proposition p may not yet be true in s, but may become true in somesituation which contains s. Thus, it is observed that there is always an event associated with a proposition, but an eventmay only be termed a proposition if it has a value of either true or false over a period of time x. Based on this, it isunderstood that the davidsonian event in Kratzer’s model works as an argument for spatio-temporal situations.

Kratzer argues that propositions in natural languages follow certain restrictions. If a proposition is true in a situation s,then it is true in any situation in which s takes part. Thus, the meaning of a (timeless) sentence like Socrates is in prison isthe set of all the possible situations in which some temporal state of Socrates is in prison. The meaning of the sentence‘‘Socrates is in prison now’’ is the set of possible situations in which the present temporal state of Socrates is in prison.

The tense of a statement appears as an interval and not as an instant. In order to assess the truth value of an assertion,it is necessary to assert that something is true in any instant within a period of time. A sentence is true at the time ofstatement t only if it is true at every instant included within t.

Some data that is revealing of the distinction between the interpretations of Neg1 and Neg2 may be obtained from theinteraction of these structures with distributive phrases introduced by cada (each). According to Negra o (2002),predicates containing a distributive quantified phrase introduced by cada as the subject are only acceptable if the contentof the argument for spatio-temporal localizations is explicitly modified:

*

(28) a.

*

Cada funcionária está grávida.

each employee is pregnant

b.

Cada funcionária está grávida num período do ano. each employee is pregnant at-a period of-the year ‘Each employee is pregnant at a certain time of year.’

(Ex. (9--10b) Negra o, 2002)

In sentence (28b), the event argument of the stage level predicate is modified and thus the sentence is acceptable.With this, the author concludes that in BP the distributive quantifier phrases introduced by cada require a modification ofthis argument so that they may function as a domain of distribution; in these cases, then, for the sentences to beacceptable, there must be a predicate with an argument structure containing an event argument. Observe in the data in(29) the interaction between the distributive quantifier cada and the negative structures Neg1 and Neg2:

(29)

a. Cada aluno leu um livro. each student read-3PS a book ‘Each student read a book.’

b.

Cada aluno na o leu um livro. each student neg read a book

c.

Cada aluno na o leu um livro na o. each student neg read a book neg
Page 13: 3 Types of Negation in BP

L. Teixeira de Sousa / Lingua 159 (2015) 27--46 39

In the sentences above, it is evident that the structure with the distributive quantifier cannot interact with the Neg1structure, but is acceptable in Neg2 sentences. As we have seen, cada, a strongly distributive quantifier, is acceptable inpredicates containing a temporal sequence of events, and the infelicity of sentence (29b) may be perfectly explained if it isunderstood that the Neg1 structures associate the subject to only one event, precluding the distributive interpretation.When cada appears in a Neg2 structure, the sentence becomes more acceptable, as the proposition is within the scope ofthe negation, and thus over the entire temporal sequence. Therefore, there are arguments that are independent of theinterpretation of Neg1 as negation of events and Neg2 as negation of propositions.

As I have attempted to demonstrate in the beginning of this study, Neg2 distinguishes itself from Neg1 in threeadditional essential contexts: narrative, infinitival and temporal embedded clauses. My proposed explanation of thisdistinction is that the scope of Neg1 extends over events while the scope of Neg2 extends over propositions. If theincompatibility of Neg2 with narratives, infinitival and temporal embedded clauses already indicates an event/propositiondistinction, its interpretation as referring to a complete time interval confers greater weight on our hypothesis.

Infinitives such as ‘Nao fumar faz bem à saúde’ (Not smoking is better for one’s health) are timeless sentences, which canbe interpreted as a general (usual, recurrent, predictable) truth, but not as a truth in a specific time. Since timeless sentencesdo not make clear reference to a time interval, they are consequently interpreted as events. If Neg2 is negation overpropositions, while Neg1 is negation over events, the preference for Neg1 in infinitives rather than Neg2 (see examples (16a--c)) can be clearly associated to the Neg2 restriction to propositions, which are always defined in relation to a specific time.

Ramchand (2005) suggests that there is an event variable that assigns a referential nature to time. Thus, adoptingReichenbach’s (1947) and Giorgi and Pianesi’s (1997) approaches to temporal relations, she distinguishes two necessaryrelations to a predication: (1) a relation between an event (E) and its reference time (R) and (2) a relation between thereference time and speech time (S). Relation (2), S--R, distinguishes the notions of present, past and future, while relation(1), E--R, distinguishes the aspectual notions of perfect, prospective and neutral.

In a matrix clause, the reference time is determined by adverbs or context, possibly generating a narrative reading.When it comes to embedded clauses, it is important to understand the event-temporal relationship that is established withthe event expressed in the clauses. In this type of structure, the reference time in the embedded clause is established withrespect to the matrix clause; thus, the reference time in embedded clause is not determined by the context:

(30)

a. O Pedro acordou. (Depois) olhou pela janela. the Pedro woke-up-3PS. (Then) looked-3PS out-the window ‘P. woke up. (Then) He looked out the window.’

b.

A Maria disse que se sentiu mal. the Maria said-3PS that herself felt bad ‘M. said she felt bad.’

In example (30b), the embedded clause se sentiu mal is linked to a time previous to the time of the matrix clause AMaria disse. It means that the reference time of the embedded clause is fixed with respect to the time of speech expressedin the matrix, before (past), after (future):

(31)

a. A Maria [disse+pas] que o Joao na o [foi+pas] à festa. the Maria said-3PS that the Joao neg went-3PS to-the party ‘M. said that J. didn’t go to the party.’

b.

A Maria [disse+pas] que o Joao na o [vai+fut] à festa. the Maria said-3PS that the Joao neg go-3PS to-the party ‘M. said that J. isn’t going to the party’

On the other hand, one can observe that temporal clauses display an important characteristic with respect to thedetermination of the time of the event.

(32)

a. O Joao [chegou+pas] quando a polícia [estava+pas] (*estiver+fut) aqui. the Joa o arrived-3PS when the police were-3PS (will be) here ‘J. arrived when the police were still here.’

b.

O Joao [vai chegar+fut] quando a polícia [estiver+fut] (*estava+pas) aqui. the Joa o will arrive when the police are (were) here ‘J. will arrive when the police are here.’

As the above examples demonstrate, temporal notions such as future, past and present (relation 2) have to be thesame in matrix and embedded clauses. This leads us to conclude that unlike other kinds of embedded sentences,temporal ones lack a reference time and the event time of the embedded clause is indexed to the reference time of the

Page 14: 3 Types of Negation in BP

L. Teixeira de Sousa / Lingua 159 (2015) 27--4640

matrix clause. As previously stated, Neg2 structures are possible in embedded clauses with the exception of embeddedtemporal clauses. Given that the particular characteristic of temporal embedded clauses in relation to other kinds ofembedded clauses is the linking of temporal reference, one can say that Neg2 structures are not possible in clauses inwhich there is no reference time.

As has been indicated, Neg2 is also incompatible with narrative texts. Thus, if the reference time is established by thecontext, the Neg2 structure is not permissible. It means that this structure does not permit an anaphoric reading of time,because it links itself to the reference time.

Note the following context of coordination:

(33)

Fui almoçar ao meio dia e já na o tinha arroz (*na o) went to-lunch at-the midday and already neg had rice (*neg) ‘I went to have lunch at midday and there was no rice left’

In the above example, já (already) in the coordinate sentence refers to an instant, which is conveyed in the first clause,‘‘ao meio dia’’ (midday). In this case, as well, the Neg2 structure is not acceptable. It seems that this structure may neverrefer to a specific point of a time extension. Given that events are considered an instant in a time interval, it is understoodthat this structure has scope over something other than events.

The existence of different structures indicating different interpretations of negation does not seem to be specific to BP.Ramchand (2005) was the first to recognize the possibility of different morphosyntactic markers for negation with scopeover both events and propositions. According to Ramchand, Bengali contains two negative items (na, ni), occurring indifferent morphosyntactic contexts and with different implications of aspect. Based on the interpretation and distribution ofthese items in certain temporal contexts, she argues that these two elements do not correspond to different forms of thesame functional heads, but rather to two distinct strategies of negation in semantics.

Ramchand’s main argument rests on the complementary distribution depending on the nature of the negated verbform: Na is grammatical with verbs in the simple present, progressive, simple past and future tense, but ungrammatical inthe perfective aspect:

(34)

a.

*

ami am-Ta kha-cch-i na

I-NOM mango-CLASS eat-PROG/PRES-1 NEG ‘I am not eating the mango’

(Ex. (4) Ramchand, 2005)

b.

ami am-Ta khel-am na I-NOM mango-CLASS eat-PAS-1 NEG ‘I did not eat the mango’

(Ex. (5) Ramchand, 2005)

c.

ami am-Ta kheye-ch-i eu manga-CLASS comer-PERF-PRES-1 ‘I have eaten the mango’

(Ex. (6) Ramchand, 2005)

d.

ami am-Ta kheye-ch-i na I mango-CLASS eat-PERF-PRES-1 NEG ‘I have not eaten the mango’

(Ex. (7) Ramchand, 2005)

e.

ami am-Ta kha-i ni I mango-CLASS eat-1 NEG ‘I did not eat the mango’ (have not eaten?)

(Ex. (8) Ramchand, 2005)

The ni, conversely, is grammatical in the past tense and perfective aspect. The two negative markers would thus becomplementary. Based on her analysis of the interaction between the perfective and negation in Bengali, Ramchandargues that negation may be understood as a connector of the time variable or the event variable, and that Bengalicontains both kinds of negation with differences in discourse and morphological implications.

According to Ramchand, the effects of the two types of negation in Bengali would be frequently equivalent, and thedifference is discursive in nature. While na is a simple negation of the event, with the time variable regularly linked via

Page 15: 3 Types of Negation in BP

L. Teixeira de Sousa / Lingua 159 (2015) 27--46 41

context to a specific time in the past,12 ni, which is a quantifier that is directly linked to a time variable, denies that the eventhas happened at any point (in the discursive context).

As she points out, when the use of the two markers is equivalent, native speakers tend to identify the ni structure asmore emphatic, the same interpretation given by Brazilian native speakers with respect to Neg2. Furthermore, there arecontexts in which one form is clearly preferable to another. According to Ramchand, in narrative discourse, in which anevent in the past is related to another in a chronological sequence, the form ni is considered infelicitous. Also, if the time ofa sentence is necessarily related to the time of an attached coordinate clause, the ni is excluded by a felicitous condition.Thus, she argues that the negative marker ni links a time variable, while na links an event variable. The argument is that, inthe case of a narrative, the time variable is emphatically linked to the context in which each sentence moves forward indiscursive time; thus, the form na is superior, as it allows the time variable freedom to be maneuvered over the course ofthe discourse. Meanwhile, ni is inferior, as it links the time variable directly, making it unavailable for discursive anaphora.This observation serves as an additional argument for interpreting Neg2 structures in BP as negating a proposition, giventhat, just like the ni in Bengali, this structure does not occur in narrative contexts.

It is also worth mentioning that although Neg1 and Neg2 may occur in the same context, the interpretations of thesestructures are not equivalent. Neg1 is interpreted as the negation of an event, while Neg2 is essentially propositional. As theevaluation of a proposition as either true or false demonstrates the commitment of the speaker with respect to a state ofaffairs, the emphatic interpretation attributed to this structure may be derived from the commitment of the speaker to theassertion.

3.2. Neg3 as metalinguistic negation

In Teixeira de Sousa (2011, 2012), I have proposed that the Neg3 structures correspond to a metalinguistic negation. Todemonstrate this, I considered the approach to metalinguistic negation proposed by Horn ([1989]2001). According to Horn,metalinguistic negation has the function of negating the assertability of a proposition conveyed in the context of a statement.He refers to three tests that can distinguish it from descriptive negation: (i) metalinguistic negation is legitimized by thediscursive context contradicting a previous assertion; (ii) metalinguistic negation does not legitimize negative polarity items(NPIs) and (iii) metalinguistic negation is compatible with positive polarity items (PPIs). Based on these tests, I will attempt, inthe following lines, to show that the nao present in Neg3 structures has all the characteristics of metalinguistic negation.

With respect to the first test in Teixeira de Sousa (2011), examples (35) and (36) below clearly demonstrate thecontradiction of a preceding discourse, as they express the refusal of the speaker to consider a piece of informationexpressed by the interlocutor valid.

(35)

12 Here,

A:

it is im

Tá chovendo o dia todo!

is raining the day whole ‘It’s been raining the whole day!’

B:

Tá chovendo agora na o! Is raining now neg ‘Now it’s not raining!’

(36)

A: Você cortou o cabelo, na o cortou? you cut-2PS the hair, neg cut ‘You had your hair cut, didn’t you?’

B:

Cortei na o3. cut-1PS neg ‘No, I didn’t.’

The impossibility of Neg3 in wide focus sentences, those that consist entirely of new information, demonstrates thediscursive restriction of this item based on preceding context:

(37)

A: O que aconteceu? the-what happened ‘What happened?‘

B: #

tô achando minha carteira na o3. am finding my wallet neg ‘I can’t find my wallet.’

portant to keep in mind that ni occurs only in the past tense, which is defined by the author as [+past], [+telic] and [+negative].

Page 16: 3 Types of Negation in BP

L. Teixeira de Sousa / Lingua 159 (2015) 27--4642

With respect to the second test, Biberauer and Cyrino (2009a,b) have already pointed out the incompatibility of Neg3with NPIs, as the following examples demonstrate:

(38)

a. A:

*

*

O Joao é rico!

the Joa o is rich ‘J. is rich!’

B:

O que?! Ele tem um tostao furado na o what. he has a cent with-a-whole ‘What?! He doesn’t have a red cent!’

(Ex. (16) Biberauer & Cyrino, 2009)

b.

A: vai na festa comigo hoje, né? you go-FUT-2PS to-the party with-me today ‘you’re coming You going to the party with me today, aren’t you?’

B:

vou na festa nem morta na o! go-FUT-1PS to-the party even dead neg ‘By no means will I go this party!’

Furthermore, the above examples illustrate that, as is characteristic of metalinguistic negation, the na o is external tothe clause, given that, with a negative value, it should be able to license negative polarity items. This fact suggests twopossible alternatives: either the item has no scope over the negative polarity item, or this element is not a negative marker,and is thus not able to license a polarity item.

The pattern also applies to Neg3 within strong PPIs:

(39)

A: Você fala pra burro! you talk-2PS to-the donkey ‘You talk like hell!’

B: #

Eu na o falo pra burro. I neg talk to the donkey

B0:

Falo pra burro na o. talk-1PS to the donkey neg

Contrary to what one might expect, clause-final na o does not have scope over the proposition as it is an externalnegation. Observe in the following example that the element denied by na o3 is not the propositional content, but rather anexpression of scalar value:

(40)

A: Isso vale a pena. it worth-3PS the penalty ‘It’s worth it’

B:

Vale a pena na o3. Vale a galinha inteira! worth-3PSthe feather neg. worth-3PS the chicken whole ‘It’s really worth it!’

An additional test to identify metalinguistic negation, referenced by Martins (2010), is that, unlike descriptive negation,metalinguistic negation does not happen in embedded clauses. Na o3 does not occur in these clauses either:

(41)

A: O Pedro disse que vendeu o carro. the Pedro said-3PS that sold-3PS the car ‘P. said that he sold the car.’

B:

O Pedro disse que na o vendeu o carro. the Pedro said-3PS that neg sold-3PS the car

B’:

O Pedro disse que vendeu o carro na o3. *the Pedro said-3PS that sold-3PS the car neg

As we have seen so far, the negative item in Neg3 structures may function as a metalinguistic negation (external) butnot as a negation of a proposition, as is possible with the negative items in Neg2.

Page 17: 3 Types of Negation in BP

L. Teixeira de Sousa / Lingua 159 (2015) 27--46 43

Metalinguistic negation has been addressed in the literature as a pragmatic matter and, for most authors, negativestructures are ambiguous between semantic and metalinguistic readings. However, as pointed by Martins (2014), thereare important syntactic issues to be considered. Drozd (2001), for instance, claims that sentence-peripherical idiomaticexpressions, such as no, appear to be cross-linguistically available as a way to express metalinguistic negation.

For Martins (2014), there are exclusive markers of metalinguistic negation in natural language; she identifies two types:sentence-internal or sentence-peripherical markers. Peripherical negative metalinguistic (NM) markers, unlike theinternal ones, are available in isolation and in nominal fragments, have the ability to deny a negative proposition and arecompatible with contrastive/emphatic constituents, idiomatic sentences and VP ellipsis.

Neg3 distribution, as previously illustrated, is quite similar to the peripheral NM markers described by Martins (2014).Neg3 is also available in nominal fragments, with contrastive constituents, idiomatic sentences and VP ellipsis. Thisconstitutes an additional argument in favor of its interpretation as metalinguistic negation. However, the possibility of Neg3as an answer to polar questions is still an open issue.

One way of interpreting the difference between semantic negation and metalinguistic negation is understanding that, inthe latter case, the truth value of a proposition is not reversed. This interpretation is clear when scalar expressions areanalyzed. In (40), for example, the speaker says vale a pena nao, and the meaning given in the subsequent sentence isthat it is worth it (vale a pena). Thus, the sentence is not contradictory because the na o at the end of the sentence does notdeny the truth value of the proposition. In another instance, however, this relationship is not as obvious -- the context of thepolar question. Observe the following example:

(42)

13 Accorinvolved.

A:

ding to

Você comprou arroz?

you bought-2PS rice ‘Did you buy rice?’

B:

Comprei na o. bought not ‘No, I didn’t.’

In this case, the interpretation of whether Neg3 interferes in the truth value of the proposition is not so clear. Thedifficulty in interpreting this context is due to the fact that it is a polar question, which may or may not contain an implicature.By considering a possible context for that instance, however, it is perfectly possible to recover some implicature, sincethese kinds of questions indicate shared knowledge between the speaker and the interlocutor, and they are not askedwithout a legitimate question. For a question like Você comprou arroz? (‘Did you buy rice?’), it is expected that theinterlocutor knows that he or she should or could have bought the rice. Thus, it is possible to understand that what is beingdenied in the sentence is not the truth value, but the implicature contained in the question. Another example of a similarnature is given below. In this case, as well, a question like A Maria quebrou o pé? (‘Did Maria break her foot?’) can onlyoccur in a context in which there is a Maria who may or may not have hurt her foot. Once again, it can be said that what isdenied is not its truth value, but rather its assertability (ex. (20)). This question, however, requires further investigation.

Once it has been demonstrated that Neg3 in BP may be interpreted as metalinguistic negation, it is necessary toidentify, within a syntactic theory, to which linguistic category it belongs. Once the interpretation of the correction isconsidered metalinguistic, the occurrence of the contrastive accent associated with this kind of structure13 appears to becoherent, considering that metalinguistic negation involves the projection of contrastive or exhaustive projection.

The relationship between metalinguistic negation and contrastive focus may also be observed in an analysis of theassertive structure proposed by Zubizarreta (1998). Zubizarreta does not use the term metalinguistic negation, but ratherrefers to the ‘‘correction’’ effect of the contrastive focus.

According to Zubizarreta (1998), the context of contrastive focus is given by the preceding context in the discourse(context statement). Contrastive focus has two effects: one is to deny the value attributed to the variable in the assertion ofthe preceding context, indicated by an explicit or implicit tag; the other is to introduce an alternative value for the variable.Zubizarreta illustrates these effects with the following example:

(43)

John is wearing a RED shirt today (not a blue shirt). (ex. (14) Zubizarreta, 1998)

To Zubizarreta, the above sentence is composed of two propositions, John is not wearing a blue shirt today and John iswearing a red shirt today, which leads to a conjunction of two main assertions.

Horn (2001), a contrastive accent is more or less mandatory because of the fluctuation of the polar sensibility of some of the items

Page 18: 3 Types of Negation in BP

L. Teixeira de Sousa / Lingua 159 (2015) 27--4644

(44)

14 A: TB: ItC: It

Accordhowever,

A1: There is an x, such that John is wearing x

A2: It is not the case that x (such that John is wearing x) = a blue shirt & the x (such that John is wearing x)= the red shirt

With this, Zubizarreta holds that contrastive focus contains a judgment about the ‘‘truth’’ or correction of the assertionintroduced in the context. Thus, contrastive focus denies certain aspects of the assertion introduced by the context.

The example-sentence used by Zubizarreta corresponds to the data presented by Rajagopalan (1982), having anegative effect over propositions14 and also presenting the same absence of exhaustivity which, according to Horn([1989]2001), characterizes metalinguistic negation.

Whether it is due to the need for a preceding assertion, the expression of a correction or a copula with two assertions, itseems clear that contrastive focus is an effect of the elocutionary act of denial, thus being correct to treat the occurrence of[VP Neg] as an instance of contrastive focus.

A similar analysis is found in Martins (2014). By taking into account the concept of responding assertion (Farkas andBruce, 2010), Martins argues that metalinguistic negation does not bear the [reverse] feature, since it does not reverse thesentence polarity, and proposes that metalinguistic negation bears the feature [objection], which is classified as a relativepolarity feature being grammatically encoded in the CP domain.

In this and the previous section, I have presented data that demonstrate the relationship between the Neg2 structureand negative assertion, on the one hand, and Neg3 and metalinguistic negation, on the other. We have seen that Neg2 issemantic negation which refers to the truth value of a proposition. Meanwhile, Neg3 is pragmatic negation which refers toassertability. As I have previously claimed, some authors recognize two kinds of external negation (Dahl, 1979;Rajagopalan, 1982) which can be translated through the embedding of a proposition in the assertions It is not the casethat. . . or It is false that. . . In the first case, it is evidently a judgment of the assertability of a proposition. In the second case,it seems that what is at stake is the truth value of a proposition. Thus, we can associate Neg3 with It is not the case that. . .and Neg2 with It is false that. . . However, the two types of negation are treated as external negation and, as I have referredto Neg3 as external negation in opposition to Neg2, it bears distinguishing what is commonly called external negation andwhat I consider external negation.

The studies that reference external negation versus internal negation are often based on philosophical studies thatdistinguish between predicate negation and proposition negation. Although there is some disagreement in the field, inthese studies, internal negation is characterized by the application of a predicate to something: ‘A is not B’. Thus, it ispossible to say that internal negation, under this perspective, refers to the negation of a state of affairs. Meanwhile,external negation is propositional negation; propositions are assessed as either true or false---essential in the constructionof arguments, which is what really interests philosophers. If we translate internal and external negation in these terms,Neg1 would correspond to internal negation, while Neg2 would correspond to external negation. In our study, however, Idistinguish between semantic negation as internal negation, and pragmatic negation as external negation. Thus, wewould have, in fact, two internal negations and one external. It is possible, however, to draw a relationship between theaforementioned types of negation and the types of negation found in BP.

Once I have argued that Neg1 is related to the negation of events or situations, we can then relate it to descriptivenegation, that which denies a situation or state of affairs. This negation would, thus, hold scope over the nuclear scope ofthe sentence (vP). Neg2, on the other hand, as I have argued, is a negation of a proposition, linked to the truth value of thesentence and, therefore, to a reference time. That is, it would be related to the inflectional part of the sentence (IP), but stillinternal. Neg3, unlike the other two, would be related to the discursive part of the sentence (CP), defined, for this reason,as external negation. This is, however, just an observation about the relationship between the three types of negation andtense, which must be taken into account in order to propose a derivation for these three types of negation in BP.

4. Final remarks

In this paper, I demonstrated that Brazilian Portuguese exhibits three structures which correspond to differentinterpretations of negation: over events, propositional or assertability. This analysis takes into account that Neg2 andNeg3 forms have strong syntactic restrictions.

he Sky is overcast. is NOT true (that the sky is overcast).

IS true (that the sky is overcast)

ing to Rajagopalan (1982), in the above examples, the response in A counts as an act of denial (the denial of a negation). In both cases, it is true. . . has a metalinguistic function.

Page 19: 3 Types of Negation in BP

L. Teixeira de Sousa / Lingua 159 (2015) 27--46 45

I have noted that, in addition to the differences between the standard and nonstandard forms, there are syntactic andsemantic/pragmatic distinctions between Neg2 and Neg3. Therefore, I have considered them to be separate phenomena:Neg3, as I have observed, is impossible in all kinds of embedded clauses or even when conveying new information, as it isrestricted to response contexts. Neg2, on the other hand, is only incompatible with infinitive or embedded temporalclauses or narrative contexts, and is unrestricted as to the type of information conveyed, that is, whether new or given.

Despite the Neg2 restrictions, this structure and the standard Neg1 are very similar, which lead me to treat them assemantic negations. Since Neg3, unlike the other two structures, occurs solely in the context of response and does notallow for polarity items, I have defined it as a structure marking pragmatic focus.

Acknowledgements

I am grateful to everyone who contributed to this paper in different and important ways, namely, the three anonymousreviewers for Lingua, Sonia Cyrino, Esmeralda Negra o, Mary Kato, Charlotte Galves, and the audience in the Negationand polarity: interfaces and cognition Workshop (Geneva, July 2013).

References

Biberauer, T., 2009. Jespersen of course? The case of contemporary Afrikaans negation. In: Gelderen, E. (Ed.), Cyclical Changes, vol. 146. JohnBenjamin, Amsterdam, pp. 91.

Biberauer, T., Cyrino, S., 2009a. Appearances are Deceptive: Jespersen’s Cycle from the Perspective of the Romania Nova and Romance-basedCreoles. Paper presented at Going Romance 23. University of Nice, Nice.

Biberauer, T., Cyrino, S., 2009b. Negative Developments in Afrikaans and Brazilian Portuguese. Paper presented at the 19th Colloquium onGenerative Grammar. University of the Basque Country, Vitoria-Gasteiz.

Cavalcante, R., 2007. A negaça o pos-verbal no Português Brasileiro: Análise descritiva e teorica de dialetos rurais de afro-descendentes (M.A.thesis). State University of Bahia, Salvador.

Cavalcante, R., 2012. Negaça o anaforica no Português Brasileiro: Negaça o sentencial, negaçao enfática e negaçao de constituinte (Ph.D.dissertation). State University of Sa o Paulo, Sao Paulo.

Dahl, Ӧ., 1979. Typology of sentence negation. Linguistics 17 (1--2), 79--106.Drozd, K.F., 2001. Metalinguistic sentence negation in child English. In: Hoeksema, J., Rullman, H., Sanchez-Valencia, V., van der Wouden, T.

(Eds.), Perspectives on Negation and Polarity Items. John Benjamins, Amsterdam/Philadelphia, pp. 49--78.Farkas, D.F., Bruce, K.B., 2010. On reacting to assertions and polar questions. J. Semant. 27, 81--118.Furtado da Cunha, M.A., 1996. Gramaticalizaçao dos mecanismos de negaçao em Natal. In: Martelotta, V., Cezário, M.M. (Org.), Gramati-

calizaça o no português do Brasil: uma abordagem funcional. Tempo Brasileiro, Rio de Janeiro, pp. 167--189.Gelderen, E.van, 2008. Negative cycles. Linguist. Typol. 12, 195--243.Geurts, B., 1998. The mechanisms of denial. Language 74, 274--307.Giorgi, A., Pianesi, F., 1997. Tense and Aspect. From Semantics to Morphosyntax. Oxford University Press, New York.Givon, T., 1978. Negation in language: pragmatics, function, ontology. In: Cole, P. (Ed.), Syntax and Semantics: Pragmatics, vol. 9. Academic

Press, New York, pp. 69--112.Horn, L., [1989]2001. A Natural History of Negation. Stanford, CSLI.Ingham, R.S., 2007. Ne-drop and indefinites in Anglo-Norman and Middle English. In: Larrivée, P., Ingham, R. (Eds.), The Evolution of Negation:

Beyond the Jespersen Cycle. Mouton de Gruyter, Berlin, pp. 1--27.Jespersen, J.O., 1917. Negation in English and Other Languages. Hist-Fil Meddelelser, vol. 1, No. 5. Det. Kgl. Videnskabernes Selskab,

Copenhagem.Kamp, H., Reyle, U., 1993. From Discourse to Logic: Introduction to a Model-Theoretic Semantics of Natural Language, Formal Logic and

Discourse Representation Theory. Kluwer, Dordrecht.Kratzer, A., 1989. An investigation of the lumps of thought. Linguist. Philos. 12 (5), 607--653.Kratzer, A., 1995. Stage--level and individual level predicates. In: Carlson, G.N., Pelletier, J. (Eds.), The Generic Book. Chicago University Press,

Chicago.Laka, I., 1994. On the Syntax of Negation. Garland, New York.Martins, A.M., 2010. Negaça o metalinguística (lá, cá e agora). In: Brito, A.M. (Ed.), Atas do XXV Encontro da Associaça o Portuguesa de

Linguística. Associaça o Portuguesa de Linguística, Lisboa, p. 20.Martins, A.M., 2014. How much syntax is there in Metalinguistic Negation? Nat. Lang. Linguist. Theory 32, 635--672.Negra o, E.V., 2002. Distributividade e Genericidade nos sintagmas introduzidos por CADA e TODOS. Rev. Gel 1, 87--89.Oliveira, M.de, 2000. Frases assertivas e sua variaçao nas línguas românicas: o seu papel na aquisiçao. Humanitas, Sa o Paulo.Payne, J.R., 1985. Negation. In: Shopen (Eds.), Language Typology and Syntactic Description, vol. 1. Cambridge University Press, Cambridge,

pp. 197--242.Pinto, C., 2010. Negaça o Metalinguística e Estruturas com ‘nada’ no Português Europeu (M.A. thesis). Universidade de Lisboa.Poletto, C., 2008. On negative doubling. In: Cognola, F., Pescarini, D. (Eds.), Quaderni di Lavoro ASlt, vol. 8. pp. 57--84.Rajagopalan, K., 1982. Negation and Denial. A Study in the Theory of Speech Acts (Ph.D. dissertation). Pontifical Catholic University of Sao

Paulo, Sa o Paulo.Ramchand, G., 2005. Two types of negation in Bengali. In: Dayal, V., Mahajan, A. (Eds.), Clause Structure in South Asian Languages. Kluwer,

Dordrecht.Reichenbach, H., 1947. Elements of Symbolic Logic. Macmillan & Co., New York.

Page 20: 3 Types of Negation in BP

L. Teixeira de Sousa / Lingua 159 (2015) 27--4646

Roncarati, C., 1997. A negaça o no português falado. In: Macedo, A., Roncarati, C., Mollica, C. (Orgs.), Variaça o e discurso. Tempo Brasileiro, Riode Janeiro, pp. 65--102.

Schwegler, A., 1991. Predicate negation in contemporary Brazilian Portuguese: a linguistic change in progress. Orbis 34 (1--2), 187--214.Schwenter, S., 2005. The pragmatics of negation in Brazilian Portuguese. Lingua 115 (10), 1427--1456.Stalnaker, R., 1974. Pragmatic presuppositions. In: Munitz, M., Unger, P. (Eds.), Semantics and Philosophy. New York University Press, New

York, pp. 197--230.Stalnaker, R., 1978. Assertion. In: Cole, P. (Ed.), Syntax and Semantics: Pragmatics, vol. 9. Academic Press, New York.Teixeira de Sousa, L., 2011. Sentential negation in Brazilian Portuguese: Pragmatics and syntax. JournaLipp 1, 89--103.Teixeira de Sousa, L., 2012. Syntax and interpretation of sentential negation in Brazilian Portuguese. Thesis (PhD in Linguistics). Unicamp.Vitral, L., 1999. A negaça o: Teoria da checagem e mudança linguística. D.E.L.T.A. 15 (1), 57--84.Wood, J., 1997. Negation in the Paston Letters (M.A. thesis). Arizona State University.Zanuttini, R., 1997. Negation and Clausal Structure. Oxford University Press, Oxford.Zubizarreta, M.L., 1998. Prosody, Focus and Word Order. The MIT Press, Cambridge.