Resumen de Subordinación y determinación: completivas precedidas de artículo definido en español
This dissertation deals with a case where subordination and determination converge: embedded subject and complement clauses preceded by a definite article (henceforth el que clauses) in Spanish. A priori, the article, which appears in a default masculine singular form, seems optional. However, el que clauses have a quite restricted distribution so the alleged optionality of the determiner is questioned. On the one hand, the predicates allowing these sentential complements are reduced. Actually, it has been claimed in the literature that only factive verbs can select them. On the other hand, the positions they can occupy also show some restrictions. In particular, el que clauses cannot be combined with prepositional verbs. Surprisingly, Spanish is the only Romance language that has this type of sentence: they cannot be found in French, Portuguese, Romanian, Italian, Galician or Catalan. To provide an explanation for these puzzling properties is the purpose of this thesis.
Chapter 1 describes the most relevant properties of el que clauses. Furthermore, I present some notes about the diachrony of these sentences, which first appeared in Spanish in the XVI century. Finally, I compare them with similar structures, i.e. embedded clauses preceded by a definite article, in other languages, such as Modern Greek, Persian and Norwegian.
In chapter 2, I describe the distribution of el que clauses, especially when they occupy the object position. I depart from the generalization established in the literature, according to which el que clauses only appear as complements of factive verbs. However, I observe that it is not empirically appropriate since it cannot account for some new data I provide. Consequently, I give a more complete list of verbs which can select el que clauses. Finally, I observe that the possibility of taking these sentential complements cannot be reduced to a matter of simple lexical selection since some verbal classes do not behave homogeneously regarding this issue. This determines that we pursue the grammatical properties the clauses per se has to have in order to allow the insertion of the article.
In chapter 3 I begin to investigate those properties. For this purpose, it is unavoidable to study the so called main clause phenomena since at first sight these phenomena and el que clauses are in complementary distribution. Consequently, I explain what main clause phenomena are, what languages they can be found in and what contexts allow them. After that, I justify with the relevant data the parallelism suggested above. Finally, I hypothesize that both phenomena are related to the same grammatical property which is materialized in an opposite way: on the one hand, it triggers V2 and, on the other hand, it blocks the article¿s insertion.
In chapter 4, I focus on verb second (henceforth V2), the most studied main clause phenomena in the literature, especially V2 in Mainland Scandinavian. After revising some influential proposals about this phenomenon, I assume that V2 occurs in ¿assertive¿ clauses. As a consequence of the generalization established in chapter 3, el que clauses should appear in ¿non assertive¿ clauses (i.e. the opposite contexts to V2). In order to characterize what non assertive clauses are, I examine some proposals which tried to define them. As it will be seen, some additional concepts, as projective content and at-issueness (Roberts et al. 2009, Simons et al. 2011, Tonhauser et al. 2013), will be incorporated into the discussion. Those concepts will allow to narrow down the set of non assertive clauses. Finally, I analyze what kind of information (in terms of new versus old) el que clauses express.
After studying the dicursive properties el que clauses have, I devote chapter 5 to the semantics of these sentences. I depart from a fact observed in the lexical selection of those verbs allowing el que clauses when they take DPs. In particular, I have noticed that those predicates only selects referential complements since they can combine with definite DPs, shows a clear preference for a specific interpretation of indefinite DPs and do not take bare singulars or plurals. On the contrary, the predicates not allowing el que clauses admit both referential and non referential DPs. This observation leads us to pay attention to some proposals in the literature according to which complement clauses have referential properties (De Cuba y Ürögdi 2009, 2010; Haegeman y Ürögdi 2010a). Assuming this idea, we can maintain that the embedded sentences selected by those predicates compatible with el que clauses are referential. On the contrary, the embedded sentences selected by those verbs which reject el que clauses should be non referential.
Faced with the difficulty of defining a clauses as (non) referential, I will conciliate the hypothesis above with Zucchi¿s (1993), according to which embedded complement clauses can denote different semantic objects in the ontology: propositions, events and states of affairs. Specifically, I will propose that Zucchi¿s distinction between propositions and state of affairs can be implemented in a different way if we combine it with both kratzerian situation semantics and the dichotomy type (or kind) and token due to Carlson (1977). As it will be shown, these ingredients will shed light on the definition of referentiality applied to embedded sentences.
In chapter 6 I refine the semantics of complement clauses, both those admitting the article and those not doing it. For this purpose, I take into account some proposals about the configuration of the NP and, finally, I assume Borik and Espinal¿s (2012, in press) approach. According to the authors, a common noun denotes a set of properties of kinds. In combination with the functional node Num(ber), it becomes to denote a set of individuals. When later on the definite article is merged in the structure, the DP becomes to denote a unique individual which have the property expressed by the noun. In this chapter, I apply Borik and Espinal¿s proposal to embedded clauses (i.e. to the C domain). Consequently, I propose that clauses begin denoting a set of kind situations. When the relevant functional category ¿whose nature will be discussed¿ merges, they come to denote a set of individualized situations. Furthermore, I account for the grammatical role that the article plays and it is shown that, although it can be inserted or not in the structure, its presence or absence has both syntactic and semantic effects. Finally, I provide some predictions which follow from our proposal and I show that it also accounts for some data not previously analyzed.
To finish, in chapter 7 I summarize the main conclusions to the questions raised in chapter 1. In addition, I present some open issues for future research.
