In: M.T. Espinal, M. Leonetti & L. McNally (eds.), Proceedings of the IV Nereus International Workshop “Definiteness and DP Structure in Romance Languages”. Arbeitspapier 12x. Fachbereich Sprachwissenschaft, Universität Konstanz 200x, pp-pp.
Remarks on Focus Structure and Non-Specificity
Manuel Leonetti (Universidad de Alcalá)
‘¿Por qué te fuiste, mamá? Poca ropa me lavabas.
¿Por qué te fuiste, mamita? Raras veces te pegaba.’
Les Luthiers, Pieza en forma de tango
1. Verum Focus Fronting
It is usually assumed, at least for most European languages, that two different fronting constructions should be distinguished, besides wh- constructions: (Clitic) Left Dislocation and Contrastive Focalisation (Focus Movement), as in the examples in (1) and (2):
(1) El libro, ya lo he terminado.
The book, already cl have.prs.1sg finished
‘The book, I have already finished (it).’
EL LIBRO he terminado (no el artículo).
The book have.prs.1sg finished not the paper
‘It is the book that I have finished, not the paper.’
Some authors have pointed out that a third kind of fronting operation is available under certain conditions in Romance languages (see Cinque 1986, 1990:74-76, Vallduví 1993, Zubizarreta 1998:102-103, Ambar 1999, Barbosa 2001, 2009 for the basic data). The properties of such a construction have not received much attention, except in Quer (2002) and, under the label of mild focalization, in Gallego (2007), although it provides us a valuable testing ground for the study of definiteness / specificity constraints. More recently, an account of this construction in terms of information structure has been proposed in Leonetti and Escandell-Vidal (2008, 2009), with Quer (2002) as the main source of inspiration. The list of examples gathered in (3) represents the prototypical instances of this construction in Spanish:
(3) (a) Nada tengo que añadir.
Nothing have.prs.1sg to add
‘I have nothing to add.’
(b) Algo debe saber.
Something must.prs.3sg know
‘(S)he must know something.’
(c) Poco más te puedo decir.
Little more you.obl can.prs.1sg say
‘Little more can I say to you.’
(d) Bastante trabajo tengo ya.
Enough work have.prs.1.sg already
‘Enough work I have already.’
(e) A alguien encontrarás que te pueda ayudar.
To someone find.fut.2.sg that you.obl can.prs.3.sg help
‘You’ll find someone that can help you.’
(f) Mucho dinero debe tener.
Much money must.prs.3sg own
‘(S)he must have a lot of money.’
(g) Demasiada concesiones hemos hecho ya.
Too many concessions have.prs.1.pl done already
‘We have already made too many concessions.’
(h) Menos estudiantes teníamos el año pasado.
Fewer students have.pst.1.pl the year past
‘We had fewer students last year.’
(i) Tantas quejas hubo que tuvieron que suspenderlo.
So many complaints there-be.pst.3.sg that have.pst.3.pl to suspend-it
‘There were so many complaints that they had to suspend it.’
The construction, called Quantificational QP-Fronting in Quer (2002), shows the following cluster of grammatical properties:
there is no emphatic stress on the fronted constituent, which can neither be interpreted as a contrastive focus nor as a narrow informative focus;
there is no resumptive clitic, except in cases where it is independently licensed (this is the most salient difference with respect to Clitic Left Dislocation);
the subject is postverbal, by virtue of an adjacency requirement between the finite verb and the fronted constituent, as in Spanish wh-interrogatives and contrastive focalization;
the fronted constituent has been extracted from the clause and moved to some position in the left periphery:1 thus, the construction displays the typical behaviour of operator-variable structures (sensitivity to island constraints, weak cross-over effects);
almost any category can be fronted2 (cf. Leonetti and Escandell-Vidal 2009);
fronting does not affect truth conditions, but has some notable effects: the resulting interpretation is usually emphatic, affective, argumentatively oriented, i.e., in some sense marked with respect to the alternative version with canonical order.
Given this set of properties, the main question raised by the analysis of this kind of QP Fronting or fronting without resumption is this: what is the crucial distinctive factor that defines it among the remaining types of fronting? The answer presented in Leonetti and Escandell-Vidal (2009) is based on information structure, and goes along the following lines. Since the fronted constituent cannot be interpreted as a topic (see §2.1), nor as a contrastive focus, fronting forces an interpretation of the sentence with no informational partition, and focus is thus limited to sentence polarity, i.e. what is usually known as Verum Focus. The construction can thus be named Verum Focus Fronting (from now on VFF). It represents an instance of ‘altruistic movement’, in Erteschik-Shir’s (2007) terms, given that in this case syntactic movement does not obey any feature matching requirements between heads and phrases, but simply blocks the possibility of getting a Topic / Comment split, or a Focus / Background split, and thus induces a Verum Focus reading as a last resort interpretive mechanism —the only way to assign any information structure to the sentence.3 Fronting acts as the syntactic trigger for the Verum Focus reading. Focus falls on sentence polarity only and the rest of the explicit constituents form the background. The emphatic value, the explicit indication that the content is true, and the argumentative orientation that characterize VFF constructions should all be treated as effects of Verum Focus: focus overtly marks the assertion of the propositional content while rejecting any other alternative proposition.
In what follows I will assume this approach to the syntax and semantics of the constructions in (3) in order to concentrate on one particular aspect of their grammar, namely the constraints they impose on definiteness and specificity in the fronted DPs/NPs. Notice that the fronted DPs in (3) are all indefinite. VFF with definite DPs and strong quantifiers gives anomalous results, as shown in (4), unless used under very specific contextual conditions (see §3.1).4
(4) (a) ??El libro he terminado.
The book have.prs.1.sg finished
‘I have finished the book.’
(b) ??Todos los detalles te voy a contar.
All the details you.obl go.prs.1.sg to tell
‘I am going to tell you all the details.’
There are two main reasons for paying attention to this aspect of VFF: one is purely descriptive and simply boils down to the need to obtain an adequate picture of the kinds of nominal expressions that are compatible with VFF, at least in Spanish; the other one is theoretical, and has to do with the possibility of explaining the constraints by resorting to general and well established principles governing DP interpretation, especially the distribution of specific and non-specific readings. As the approach to VFF I just sketched is built on basic notions of information structure and is not tied to any particular hypothesis about the syntax of the left periphery, my main interest will be in the connection between definiteness / specificity and information structure, which I rely on as one of the fundamental notions for understanding the interpretation of DPs in a grammatical context.5
The present paper has three main sections. Section 2 presents and discusses two previous views of specificity / definiteness constraints in fronting constructions: the first one is Cinque’s (1986, 1990), based on Italian data quite similar to the Spanish examples in (3), and the second one is Quer’s (2002), based on data from different Romance languages, where the main issues related to definiteness / specificity constraints were addressed for the first time. Section 3 is devoted to the nature of such constraints and includes a discussion of different facts in Spanish. Finally, section 4 aims at formulating a reasonable explanation for such facts that situates them in the context of the interactions between definiteness / specificity and information structure. I hope to provide a suitable basis for integrating Spanish VFF facts into a more general view of such interactions. The discussion will be predominantly data-oriented. A number of interesting issues will inevitably be left for future research: among them, the place of VFF within a general picture of crosslinguistic variation in information structure, the ways in which the constraints on VFF can be treated in a formal model, and the reason why restrictions on definiteness / specificity disappear when PPs or adjuncts are fronted.