Resumo

This paper  aims  at investigating  Brazilian   Portuguese ‘lá’  (‘there’) in structures with rhetorical question, imperative, directive, emphatic assertion, and predicative. We argue that, in these constructions,  ‘lá’  is  merged in the  specifier of   FocusP  and  ForceP  in the  CP- cartography. This  proposal is based both in the F-Spec Theory (CINQUE, 1999) and in the cartographic project (RIZZI, 1997; CINQUE & RIZZI, 2008).  The former claims that AdvPs are merged in the specifier of  functional  categories. The  latter identifies a domain to the left of  the IP  which is made up by a range of  functional  categories facing both discourse and syntax.

Introduction

Several researchers have observed that the adverb , both in BP (MARTELOTTA & REGO, 1996) and in EP (MARTINS, 2010), has shown non deictic properties which allow it to perform an “emphatic marker” function in many different syntactic structures. For a matter of clarification, we may divide these researches in at least two groups. Roughly, the first one is concerned with the semantic and discursive values conveyed by , without developing a syntactic analysis. The second one provides these realizations of with a syntactic analysis, but does not offer means of examining the restrictions and identity of each realization of in BP.

For instance, MARTINS (2010) claims that Spec,TP is the position where is merged in all the structures that she took into account in EP. Though this proposal is meant to be comprehensive, it seems to lack principles for explaining the singular syntactic and semantic features of each structure where occurs. Actually, this analysis puts together, in a sole category (Spec,TP), a bunch of different interface features, which seems to be problematic.

Therefore, a unified analysis is needed, but it also has to fit with the peculiar properties of in its many realizations. In order to meet these requirements, which are not mutually exclusive, the ensuing analysis will pursue the following assumptions. Firstly, AdvPs are functional specifiers (CINQUE, 1999). As such, they are expected to have a rigid order determined by the Universal Grammar and to check head features. Secondly, the IP (RIZZI, 1997) projects a domain dedicated to elements of interface, that is, the ones which have syntactic and discourse import. In this way, the left periphery or CP bears a space rich in functional categories, such as focus, force, finiteness, and so forth. Thirdly, syntax is governed by elementary mechanisms, such as merge and move (CHOMSKY, 1995), that can generate highly complex hierarchical blocks.

.

Minimalism focuses on the elementary mechanisms which are involved in syntactic computations […] and cartography focuses on the fine details of the generated structures, two research topics which can be pursued in parallel in a fully consistent manner (CINQUE; RIZZI, 2008: 49).

.

Based in these assumptions, we hypothesize that checks functional features in the IP periphery1. Therefore, this paper examines in rhetorical questions (1.1), imperatives (1.2), directives (2.1), emphatic assertions (2.2), and predicatives (2.3), arguing that, in these structures, would be merged in the specifier position of FocusP and ForceP.

1 Lá in Spec,FocusP

This section investigates in rhetorical questions (1.1) and in imperatives (1.2).

1.1 Rhetorical questions

To study in these structures, we first give a brief explanation on the properties of rhetorical questions. Then, we argue for an analysis of as Spec,FocusP in the IP-periphery. Finally, we discuss MARTINS’s (2010) proposal for EP. In the meantime, we also intend to clarify the difference between as a rhetorical question marker, on the one hand, and as a negative marker, on the other hand.

Let’s observe (1a).

(1)

a. Isso é atitude de um homem?

This lá is attitude of a man

‘Is this a man’s behavior?’

b. This is not a man’s behavior.

.

In (1a), belongs to a rhetorical question which, by definition, (i) “does not expect to elicit an answer” (HAN, 1998: 1) and (ii) has the capacity of inverting the polarity of a sentence. That is why (1b), a negative sentence, may be a paraphrase of (1a).

In rhetorical questions, can be either pre-verbal (1c) or post-verbal (1d) and can be omitted, without changing the propositional content of a sentence, as follows:

(1)

c. Isso () é atitude de um homem?

d. Isso é () atitude de um homem?

This is lá attitude of a man

‘Is this a man’s behavior?’

.

Due to the fact that can be pre-verbal, it is situated higher than the position where the verb is located, that is, higher than IP, probably in the left periphery because conveys information with emphatic import. As a consequence, post-verbal position of would be explained by V- raising to a position higher than the one where is located.

For these reasons, we presume that in (1a) is merged in the CP- domain, specifically in Spec,FocusP. This position seems to be suitable because is usually pronounced with an emphatic intonation, is not compatible with focalized items (2), and allows topic recursion higher (3b, d) and lower (3c) than its position.

(2)

a. TUDO, o João comprou. Não faltou nada.

Everything, the João bought. Not lacked nothing

‘João bought EVERYTHING. Nothing is missing’.

b. *TUDO, o João lá comprou?

EVERYTHING, the João lá bought

.

(3)

A: O João comprou um carro recentemente.

‘João has bought a car recently’.

B:

a. Você quis dizer: O PAI DO JOÃO comprou um carro recentemente.

‘You mean: João’s father has bought a car recently’.

b. O João lá comprou um carro? Foi o pai dele.

The João lá bought a car? Was the father of-him

‘Has João bought a car? It was his father who did it’.

c. Lá um carro o João comprou? Aquilo é uma lata velha.

Lá a car the João bought? That is an iron old.

‘Has João bought a car? That looks like scrap iron’.

d. O João comprou lá um carro? Aquilo é uma lata velha.

The João bought lá a car? That is an iron old.

‘Has João bought a car? That looks like scrap iron’.

.

The underlined items, in contrast to what happens to o pai do João in (3Ba), carry given information, which means that they are available in the discourse. That is why they can be interpreted as topics. We claim that o João (3b), um carro (3c), o João (3d) and comprou2 (3d) move to topic positions either higher or lower than the one where is situated, according to the derivations shown below.

Figure 1.

According to MARTINS (2010), in European Portuguese (EP), is post-verbal, as given in the examples (4a, b).

(4)

a. “Eu doente? Ora essa! Eu sou criatura que adoeça!”

Me sick? Now that! I am lá creature that gets-sick!

‘Me sick? What a silly idea! I’m not someone to fall sick’ (MARTINS, 2010: 12).

b. “Há coisa melhor que estar na praia?”

‘Is lá thing better than be-INFIN in-the beach?

‘Could anything be better than staying in the beach?’

(MARTINS, 2010: 16).

.

Surfacing in this position, is lower than the verb: while the latter is situated in ΣP, the former is in TP.

.

Spec,TP in European Portuguese is a dedicated Utterance Time position […] non-argumental deictic locatives may give content to Spec,TP by external merge, in which case they act as emphatic markers devoid of locative meaning (MARTINS, 2010: 18).

.

Nevertheless, in BP, in rhetorical questions can be pre-verbal, as seen in the example (1a), repeated below, which means that it is in fact higher than the verb.

(1)

a. Isso lá é atitude de um homem?

This lá is attitude of a man

‘Is this a man’s behavior?’

.

Besides, according to the author, in (4a) is different from in (4b). The former is a negation marker while the latter is a rhetorical question marker. This difference, however, does not apply to BP. In BP, (4a) is grammatical only if it is pronounced with an intonation of questions. It means that (4a) would be in fact a rhetorical question just like (4b).

We still need to know, however, why both (1a) and (4a, b) are understood as negation in BP. It follows probably from a peculiar property of rhetorical questions which is the reversal polarity, that is, “a rhetorical positive question has the illocutionary force of a negative assertion” (HAN, 1998: 1) and other way round. As a matter of fact, if is left out, the sentence will still keep its negative interpretation. Therefore, a straightforward conclusion from this fact is that , in rhetorical questions, is not a negation marker. Negation derives purely from illocutionary force.

Nonetheless, it turns out that, in sentences like (5a) and (5b) below, can be analyzed as a negation marker. In this case, there is no question intonation, is post-verbal3 only and can not be left out otherwise the negative interpretation is lost. In addition, , as a negation marker in BP, has some restrictions to be met, for instance, it has to appear with the verbs saber (‘to know’), as in (5a), and importar (‘to mind’), as in (5b). When it occurs with saber, it is also restricted either to 1st singular person (5a) or to 3rd singular plus arbitrary -se, as in (5c).

(5)

a. Sei lá!

Know lá

‘I don’t know!’

b. Importa-me lá!

Mind me lá!

I don’t care!

c. Sabe-se lá se ela casou.

Know-se lá if she married

It is not known if she has got married.

.

To sum up, we have pointed out that is not a negation marker in (1a), but only in specific contexts where the sentence does not have interrogative intonation and shows lexical and grammatical restrictions. In addition, at least in BP rhetorical questions, can be preverbal, which means that it is probably higher than Spec,IP.

1.2 Imperatives

This subsection is concerned with in imperatives (6).

(6)

a. Olha lá, hein? Pare de acusar os outros.

Look lá, huh? Stop of accuse-INF the others

‘Be aware of it, huh? Stop accusing people.’

.

In (6), can only be post-verbal and, as expected for imperatives, it is incompatible with conditionals (6b), embedded clauses (6c) and non- finite clauses (6d).

(6)

b. *Se você olha lá, as pessoas não vão ficar bravas.

If you look lá, the people not go-FUT stay-INF angry

c. *Eu disse que olha lá.

I said that look lá

d. *Olhar lá é a chave para as pessoas não ficarem bravas.

To-look lá is the key to the people not stay angry

.

According to MARTINS (2010: 13), expresses “vehement requests, by which the speaker intends to grant a positive response from the interlocutor”, as in (7).

(7) A: Dá-me um beijo.

‘Give me a Kiss!’

B: Não. ‘No.’

A: Dá .

‘Please!’ (MARTINS, 2010, p. 13-14).

.

Contrastively, in BP imperatives4, , besides indicating a kind request, as given in (8a), may also be used to indicate a threat, as given in (6a).

(8)

a. “Avisa lá que eu vou chegar mais tarde”

Tell lá that I will arrive more late

‘Tell them please that I will arrive later’.

.

According to MARTINS (2010), is situated in Spec,TP due to its post-verbal position. In our view, however, belongs to the IP- periphery, being merged probably in Spec,FocusP. This hypothesis takes into account that: firstly, does not seem to be compatible with focalized items (8b); secondly, is post-verbal (8c), which follows from the fact that V raises to Force in imperatives (PLATZACK & ROSENGREN, 1998); thirdly, is not only post-verbal but also adjacent to the verb (8d). That is why may be merged immediately below ForceP.

(8)

b. *Avisa VOCÊ lá.

Tell YOU lá

c. *Lá avisa!

Lá tell

d. *Avisa com atenção lá.

Tell with attention lá

.

Therefore, even though is merged in Spec, FocusP and act as a kind of ‘emphatic’ marker both in rhetorical questions and in imperatives, it belongs to two different structures. In rhetorical questions, there is no V-raising to Force, contrary to what happens in imperatives. Furthermore, Force, in rhetorical questions, bears [+Interrogative] features, while in imperatives, Force is [+Imperative].

Considering these facts, a derivation for (6a) would be as follows:

Figure 2.

FIGURE 2: Lá in Spec,FocusP of imperatives

2 Lá in Spec,ForceP

This section deals with sentences having the following types of illocutionary force: directive (2.1), assertive (2.2) and conditional (2.3). The hypothesis we have for them is that is merged in Spec,ForceP.

2.1 Directives

In order to analyze directive sentences, this subsection comprises a description of expressions such as French voilà/ci and English t/here you go/are, comparing them with BP lá vai (9).

(9) Lá vai!

Lá goes!

‘There you go!’

.

Comparing voilà and lá vai, we will be interested in what they are similar and what they are different in relation to the following properties: (i) syntactic function of the post-verbal NP; (ii) possibility to appear in embedded clauses; (iii) replacement of the NP for an embedded clause and (iv) replacement of by aí.

Firstly, starting from the syntactic function of the post-verbal NP, in French (10a), the NP is an internal argument of voir (‘to see’), while in BP, the NP is an external argument of ir (‘to go’) (10b).

(10)

a. Voison sac (internal argument).

‘There is his bag’.

b. Lá vai a bolsa (external argument).

‘There is the bag’.

.

Secondly, while voilà may appear in embedded clauses, as in relatives (11a), lá vai seems to be restricted to main clauses (11b).

(11)

a. “L’homme que voilà est mon amant” (BERGEN; PLAUCHÉ, 2001: 7).

‘The man (who is) there is my lover’ (BERGEN; PLAUCHÉ, 2001: 7).

b. *Esta é a bolsa que lá vai.

This is the bag that lá goes

.

Thirdly, the realization of an embedded clause in the position of the NP is allowed with voilà (12a), but not with lá vai (12b).

(12)

a. Voilà que Marie part. (BERGEN; PLAUCHÉ, 2001: 8). ‘There is Marie leaving’ (BERGEN; PLAUCHÉ, 2001: 8).

b. *Lá vai que a Maria desaparece.

Lá goes that the Maria disappears

.

Having pointed out the features which distinguish voilà and lá vai, we will show now the features which make them alike. Firstly, locative may be replaced with ci (13a) in voilà and with (13b) in lá vai without changing the propositional content of the sentence. According to Bergen & Plauché (2001: 2), “voilà and voici were historically used to differentiate between proximal and distal relations, as ci and still do […]. At present, voilà and voici are mostly interchangeable without semantic effect”.

(13)

a. Voilà/ci son sac. (Bergen; Plauché, 2001: 1).

‘There/Here is his bag’.

b. Lá/aí vai a chave que você pediu.

There/Here goes the key that you asked

‘There/Here is the key that you asked me’.

.

Another common feature between voilà and lá vai is the realization of a speech act which results in an action from the listener. Following Bergen & Plauché (2001: 2), “Voilà and voici derive historically from imperative forms of the verb ‘to see’”, i.e., there is a request saying “look at that thing there” (BERGEN; PLAUCHÉ, 2001: 6). Similarly, directive lá vai is usually uttered in a context where the speaker gives or sends something to his interlocutor, as in (14).

(14) Scene: a girl replies to an e-mail message of her brother.

a. Lá vai: Rua da Bahia, n. 16, CEP ....

There goes: Street of Bahia, number 16, Postal Code …

‘T/here is my address: 16, Bahia Street, Postal Code …’

Scene: a woman throws a key to her husband.

b. Lá vai!

There goes

‘T/here is the key!’

.

To sum up, on the one hand, (i) the syntactic function of the post- verbal NP, (ii) the insertion into an embedded clause; and (iii) the replacement of the NP for an embedded clause establish a contrast between French voilà and BP lá vai. On the other hand, (iv) locative interchangeability makes them look alike. The comparison between BP lá vai and French voilà is aimed at showing that like voilà, lá vai performs a presentative function.

So far, we investigated the similarities and differences between French voilà and BP lá vai. From now on, we will try to make a parallel between BP lá vai and English t/here you go/are. These expressions are “used when you are giving something to someone, or showing something to them”5, as in (15) and (16).

(15)

a. “There you are. I’ll just wrap it up for you”6.

b. “Here you are. A box full of tools”7.

.

(16)

a. A cashier gives a customer the shopping already packed, saying: “T/here you go!”.

b. “‘Here you go’. Callum handed her a glass of orange

juice”8.

.

We may observe that, in each of the examples above, lá vai may be replaced with a verb in the imperative form, such as tome (‘take’), receba (‘receive’) or pegue (‘hold’). Therefore, lá vai, just like voilà, conveys a directive speech act, because it elicits a reaction from the speaker. For example, in (14b), the husband is expected to be prepared to hold the key which will be thrown in his direction. This is the most striking similarity between there you go and vai: both has what we may call directive illocutionary force. In the examples above, we may also notice that there, as in (15a) and (16a), and here, as in (15b) and (16b), are sometimes interchangeable like and ci in voilà.

All of this means that, far beyond a locative import, the comparison with voilà, on the one hand, allows us to identify a presentative function in lá vai. On the other hand, the comparison with there you go allows us to identify a directive illocucionary force in the Brazilian Portuguese expression. Therefore, because there is a relevant matter of illocutionary force in the expression lá vai, and because is pre-verbal, we hypothesize that this adverb is merged in Spec,ForceP, in order to check [+directive] feature in Force, according to the following derivation.

Figure 3.

FIGURE 3: Directive lá in Spec,ForceP

2.2 Emphatic assertions

In this subsection, we will examine the properties of in emphatic assertions (17B).

(17)

A: _A vida não tem sentido sem trabalho e fé. ‘Life is not worthy without work and faith’.

B: _Lá isso é verdade.

There this is true

‘This is definitely true’.

.

Firstly, the fact that is fully compatible with aqui (‘here’), as showed in (18a), supports the idea that it is not deictic locative. Secondly, is rigidly pre-sentential (18b), which results in its high position in the syntactic hierarchy. Thirdly, is restricted to root clauses, which might be determined by the illocutionary force of emphatic assertions. As such, tests seem to confirm that is prevented from occurring in conditional (18c), embedded (18d) and non-finite (18e) clauses. Fourthly, may co-occur with positive polarity items (18f), which indicates that it does not belong to PolP. In addition, is not allowed to follow these items (18g), which confirms that this adverb is high, specifically, higher than PolP and, hence, situated in the CP-domain. Fifthly, may occur with topicalized (ficamos) and focalized (nós) items. In this case, precedes them (18h).

(18)

a. isso é bem verdade aqui na região.

Lá this is well true here in-the region

‘This is definitely true here in this area’.

b. ≠Isso lá é bem verdade.

This lá is well true

c. *Se lá isso é verdade, ...

If lá this is true, …

d. *Eu disse que lá isso é verdade. I said that lá this is true.

e. *Lá isso ser verdade é a condição para o acordo.

Lá this to-be true is the condition to the agreement

f. Lá isso sim é verdade.

Lá this yes is true.

‘Yes, this is definitely true’.

g. ≠Isso sim lá (em São Paulo) é verdade. This yes there (in São Paulo) is true. Yes, this is true there (in São Paulo).

h. Lá ficamos NÓS sem almoço.

Lá stayed we without lunch

‘We finished by not having lunch’.

.

In sum, precedes the whole sentence, positive polarity items, topic and focus and is restricted to root clauses. Therefore, there are at least five reasons to support the analysis of in Spec,ForceP. We claim that , in emphatic assertions, is directly merged in Spec,ForceP in order to check [+Declarative] features, according to the derivation below.

Figure 4.

FIGURE 4: Lá in Spec,ForceP of emphatic assertions

2.3 Predicatives

The data in 19 (a - d) show that may appear in a predicative structure which is made up by different heads.

(19)

a. Seja lá qualquer pessoa que for, comporte-se.

Be-PRES.SUBJ. lá any person that be-FUT.SUBJ., behave- yourself

‘No matter who s/he is, behave yourself ’.

b. Seja lá que pessoa/quem/qual livro for, aceite.

Be-PRES.SUBJ. lá any person/who/any book be-FUT.SUBJ, accept-IMPER.

‘No matter who s/he is, accept her/him’.

‘No matter which book it is, accept it’.

c. Seja lá como/onde/por que razão/de que direção for, prossiga.

Be-PRES.SUBJ. lá how/where/for any reason/from any direction be-FUT.SUBJ, go-IMPER.

‘Never mind how/why/where/which direction they will travel, just go’.

d. Seja lá bonito como for, não compre.

Be-PRES.SUBJ. lá beautiful how be-FUT.SUBJ, not buy

‘No matter how beautiful it is, don’t buy it’.

d’. Seja lá quão bonito for, não compre.

Be-PRES.SUBJ. lá how beautiful be-FUT.SUBJ, not buy

‘No matter how beautiful it is, don’t buy it’.

.

The predicative structures in (19) belong to a complex wh-item made up by X + be-SUBJUNCTIVE such that X may be a DP, PP, NP, AP, QP, AdvP, and so forth. An evidence for claiming that this phrase is a functional wh-item comes from the fact that quem for, o que for, por que for, como for, onde for and bonito como for may be translated into English by wh-items, such as whoever, whatever/whichever, why, however, wherever and how beautiful. Moreover, even in BP, (19d) may be paraphrased by (19d’) with a wh-item (quão bonito), though in a very literary style.

Additionally, the predicative clauses in (19) can be compared with a subordinate clause like (20a).

(20)

a. Diga quemi ele é ti.

Say-IMPER. who he is

Tell me who he is.

b. *Diga ele é quem.

Say-IMPER he is who

c. *Seja lá for quem, aceite.

Be-PRES.SUBJ. lá be-FUT.SUBJ who, accept-IMPER.

.

In (20a), “quem ele é” is a subordinate clause. It has the feature [+interrogative] in Forceº and requires wh-raising (20b). In (19), although there is also a requirement for wh-raising (20c), Force is [+conditional]. In order to illustrate this assumption, we may take a look in the example (19b). In this case, the root clause seja and the embedded one lá quem for together may be paraphrased by conditional clauses with se (‘if ’) (21), a conjunction which is usually described in ForceP. Of course, lá quem for does not have se, but it also bears conditional force because seja and for are in the subjunctive, which is a Mood dedicated to hypothetical situations.

(21) Se for a Maria/o João/uma empregada/um palhaço, cumprimente.

‘If Mary/John/a servant/a clown appears, greet her/him!’

.

In sum, we suggest that is merged in Spec,ForceP, higher than wh-items. The latter raises from IP/VP to Spec,FocusP. Therefore, in these structures, wh-items, constituted by a diverse range of categories such as APs, AdvPs, PPs, and so forth, are moved to Spec,FocusP. Also, regarding , instead of being in the Spec of each one of these projections, it is in fact the Spec of a sole category, as follows:

Figure 5.

FIGURE 5: Lá in Spec,ForceP [+conditional]

Final remarks

In this paper, we worked on a formal analysis of BP sentences with . Following the cartographic approach, this research made it possible to recognize and to identify functional projections in the left periphery. Accordingly, the properties of so far considered as an “emphatic” marker were “syntacticized” (CINQUE; RIZZI, 2008: 52) in a way that this adverb was classified as specifier of FocusP and ForceP.

In Spec,FocusP, belongs to rhetorical questions and imperatives. In this case, is usually high pitch accented and is prevented from occurring with focalized items, which indicates a possible dispute for the same position. However, these structures are clearly different. In rhetorical questions, Forceº is [+interrogative] and there is no V-raising to Forceº. In imperatives, V raises to Forceº which is [+imperative]. Moreover, in contrast to what is observed in EP, in BP there is a striking distinction between as a rhetorical question marker and as a negation marker. The latter can be neither pre-verbal nor omitted, besides being restricted in many other ways.

In Spec,ForceP, belongs to directives, emphatic assertions and predicatives. In directives, lá vai looks like voilà and there you go, in that while may be replaced by , lá vai may be replaced by a verb in the imperative. In emphatic assertions, precedes focus, topic and positive polarity items. That is why it is situated in a high position in the hierarchical structure. In predicatives, we compared the complex structure lá X for with English wh-items. From this comparison, we concluded that, even though X may be categorically diverse (D, A, Adv, N, Q, etc.), it moves to Spec,FocusP. As precedes these items, we hypothesized that is merged in Spec,ForceP just above them and that Force would be [+conditional] for two reasons mainly: on the one hand, seja lá X for is paraphrased by a subordinate clause with se (‘if ’); on the other hand, seja and for in the subjunctive, like conditionals, represent irrealis Mood.

In sum, this research has shown that it is possible to provide different realizations of with a relatively unified analysis, as this adverb seems to match properties of left periphery projections.

Referências

  1. Voilà voilà: Extensions of deictic constructions in French BERGEN Benjamin, PLAUCHÉ Madelaine. In: Cienki Alan, Luka Barbara, Smith Michael , eds. Conceptual and Discourse Factors in Linguistic Structure. 2001 .
  2. The Minimalist Program CHOMSKY Noam. Cambridge, Mass: MIT Press; 1995.
  3. Adverbs and Functional Heads: A Cross- Linguistic Perspective CINQUE Guglielmo. New York: Oxford University Press; 1999.
  4. The cartography of syntactic structures CINQUE Guglielmo, RIZZI Luigi. CISCL Working Papers.2008;v. 2:42-58.
  5. Deriving the interpretation of rhetorical questions HAN Chung-Hye. Proceedings....1998;:237-253.
  6. Gramaticalização de lá MARTELOTTA Mário, RÊGO Lana. In: MARTELOTTA M, VOTRE S, CEZARIO M, eds. Gramaticalização no português do Brasil: uma abordagem funcional. Rio de Janeiro: Tempo Brasileiro; 1996 .
  7. Deictic locatives, emphasis and metalinguistic negation MARTINS A. In: GALVES Charlotte , et al . Diachronic Syntax: Parameter Theory and Dynamics of Change. Oxford University Press; 2010. (Forthcoming). .
  8. A sintaxe cartográfica de lá no português brasileiro: um estudo da periferia esquerda PEREIRA Bruna. 2011.
  9. On the subject of imperatives: a minimalist account of the imperative clause PLATZACK Christer, ROSENGREN Inger. The Journal of Comparative Germanic Linguistics.1998;v. 1(n. 2):177-224.
  10. The fine structure of the left periphery RIZZI Luigi. In: HAEGEMAN Liliane, ed. Elements of grammar: handbook of gererative syntax. Dordrecht, Boston, London: Kluwer Academic Publishers; 1997 .