ON COMPLEX PREDICATES IN BRAZILIAN PORTUGUESE
Sonia Cyrino (Unicamp, Brazil)
Clitic climbing in Romance, triggered by restructuring verbs (cf. Rizzi 1982), has been recently
studied as part of a more general phenomena, known as complex predicates (cf. Wurmbrand
2001, 2004; Cinque 2004; Gonçalves, 1999, 2001, a.o.). Although considered a Romance
language, Brazilian Portuguese (BP) presents a challenge for these current analyses.
Some studies (Pagotto 1992; Cyrino 1993) have shown that BP has lost clitic climbing, and
this was related, in the 90s, to the concomitant loss of null subjects in the language (cf. Kayne
1989), both changes being comparable to the losses French has undergone.
However, the fact that the loss of clitic climbing in BP has also affected other verb groups
has not yet, to my knowledge, received a proper analysis. Unlike French, BP has the syllabic
clitic in the pre lower verb position not only in a “restructuring configuration” (1), but also in any
compound tense sequence (2a-b), as well as in passive constructions (2c):
(1) João pode/quer/vai te ver.
João can/ wants/goes you-CL see-INF
`João can/wants to/is going to see you´.
(2) a. João está te vendo. b. João tinha me visto.
João is you-CL see-PresPart João had I-CL see-PstPart
`João is looking at you.´ `João had seen me´.
c. O livro foi nos enviado ontem.
The book was we-CL send-PassPart yesterday
`The book was sent to us yesterday´.
In this paper, I claim that complex predicates in BP have the lower verb in a nondefective
functional projection below the higher V, in both the compound tense and the restructuring
context, in contrast to what has been proposed for other Romance languages. Gonçalves
(1999,2001), for example, argues for a biclausal structure in restructuring contexts in European
Portuguese (EP), with a defective T projection in the lower clause, inert for checking the features
of the lower V, forcing this V to incorporate into the higher V, as in Roberts (1997). As for the
compound tense sequence, Gonçalves & Costa (2002) argue for a monoclausal structure.
I show that, in BP, the fact that the lower verb is in this nondefective projection explains the
possibility for intervening subjects in constructions where we would expect clause union effects.
Thus, BP lacks both faire+infinitive constructions (3a) and long object movement (3b), as
compared to EP (cf. Gonçalves 1999); therefore, a nominative argument may intervene (3c, 3d):
(3) a. João fez comer o bolo ao Pedro. (*BP/√EP) b. Hoje, querem-se ter regalias. (*BP/√EP)
João made eat the cake to Pedro. Today, want-PL-se-CL-ACC have advantages.
`João had Pedro eat the cake´. `Today, advantages are wanted´
c. João fez ela comer o bolo (√BP/*EP) d.. Hoje, quer se ter regalias. (√BP/ *EP)
João made she-NOM eat the cake. Today, want-SG se-CL-NOM have advantages.
`João made her eat the cake´ `Today, one wants to have advantages´.
Another piece of evidence comes from the fact that BP lacks ECM constructions. Sentences
with causative and perception verbs in BP (4) show that the higher verb does not have a bearing
on the Case of the argument of the lower verb, contrary to what is usual. Example (4b) shows
that we cannot rely on the loss of 3rd person clitics in BP (cf. Cyrino 1993) to explain the
phenomenon, since we may have other pronouns morphologically marked as nominative in that
(4) a. Maria viu ele passar. b. João mandou eu comer o bolo.
Maria saw he-NOM pass João ordered I-NOM eat-INF the cake
`Maria saw him pass by.´ `João ordered me to eat the cake.´
The possibility for similar constructions with a clitic, as in (5a), could point to counterevidence
for my claim that we do not have a complex predicate structure in BP here. However, alongside
(5a), we can also have (5b), which shows that, in fact, the clitic in (5a,b) is only a reduplication
of sorts, as has been proposed for BP clitics by some scholars (Cyrino 1990, Kato 1993, a.o):
(5) a. Maria me mandou comer o bolo. b. Maria me mandou eu comer o bolo.
Maria CL-I ordered eat the cake Maria CL-I ordered I eat the cake
`Maria ordered me to eat the cake.´ `Maria ordered me to eat the cake.´
If my claim is right, BP should allow intervening subjects in compound tense structures as well,
on a par with the possibility for anaphoric pronouns in restructuring contexts. Indeed, it does, see
(6), in contrast to what has been proposed for EP (6b,c) (cf. Gonçalves & Costa 2002:25 and 67):
(6) a. João quer ele ganhar a loteria. (√BP/√EP) b. João tinha ele visto o jornal. (√BP/*EP)
João wants he to win the lottery. João had he seen the newspaper.
`João wants himself to win the lottery´ `João himself had seen the newspaper´
c. Já que está todo mundo ocupado, João pode/vai ele fazer o jantar. (√BP/*EP).
Since that is every body busy, João can/is going he to make the dinner.
`Since everybody is busy, João himself can/is going to make dinner´.
Besides that, we can explain one possible interpretation available in easy-to-please
constructions in BP. Roberts (1997) assumes the adjective triggers restructuring, thus creating an
extended projection involving the two clauses and accounting for the clause union effect. The
object of the lower verb moves over the empty subject (ec) of the lower clause:
(7) Johni is pleaseV+easy [ec tV ti]
In BP, unlike other Romance languages, the sentence is ambiguous, and John can be
interpreted as the subject of the lower clause. This can be explained if we assume BP allows for a
structure with no complex predicate. Here, again, I compare BP to EP, for illustration:
(8) Joãoi é fácil de [ eci agradar] (√BP/*EP)
João is easy of please
`João is easy to please people´
In all these cases, the unity of the sequences found in complex predicates is broken. My
analysis, which relies on recent discussion on shared subjects (cf. Hornstein, 2003), is able to
encompass different phenomena of BP which have been explained case-by-case in the literature.
Cinque, G. 2004. ‘Restructuring’ and Functional Structure. In A. Belletti (ed.) Structures and beyond: the
cartography of syntactic structures. New York: OUP. Cyrino, S. 1990. O objeto nulo no português do Brasil: uma
mudança paramétrica? ms, Unicamp. Cyrino, S. 1993. Observações sobre a mudança diacrónica no português do
Brasil: objeto nulo e clíticos. In I. Roberts & M. Kato (eds.) Português brasileiro. Campinas: EdUnicamp. Cyrino,
S. & Matos, G. 2005. Local licensers and recovering in VP ellipsis. Journal of Portuguese Linguistics 4(2): 79-
112. Gonçalves, A. 1999. Predicados complexos verbais em contextos de infinitivo não preposicionado do
português europeu. PhD diss. University of Lisboa. Gonçalves, A. 2001. Defectividade functional e predicados
complexos em estruturas de controlo do português, ms. Gonçalves, A. & Costa, T. 2002. Auxiliar a compreender
os verbos auxiliares. Lisboa, Colibri. Hornstein, N. 2003. On control. In R. Hendrick (ed) Minimalist syntax.
Malden, Blackwell. Kato, M. 1993. A theory of null objects and the development of a Brazilian child grammar. In
R. Tracy & E. Latley (eds.) How tolerant is Universal Grammar. Tübingen: Newmeyer. Kayne, R. 1989. Null
subjects and clitic climbing. In O. Jaeggli. & K. Safir (eds.) The null subject parameter. Dordrecht, Kluwer.
Pagotto, E. 1992 A posição dos clíticos em português: um estudo diacrônico. MA thesis, Unicamp. Roberts, I.
1997 Restructuring, head movement, and locality. Linguistic Inquiry 28:423-460. Wurmbrand, S. 2001. Infinitives:
restructuring and clause structure. Berlin, Mouton de Gruyter. Wurmbrand, S. 2004. Two types of
restructuring—Lexical vs. functional. Lingua 114,8: 991-1014.