Predicate doubling asymmetries with factive clauses

VERDECCHIA LEANDRO, MATÍAS NICOLÁS

São Carlos

Congreso; III Colóquio de Semântica Referencial; 2019

Universidade Federal de São Carlos

1. Overview. In this paper I explore a pattern that has not been previously discussed in the literature: an asymmetry in predicate doubling construction with clauses embedded under cognitive and emotive factives. In short, I argue that emotive factives prevent predicate doubling because they always lead to a presupposition failure. Regarding cognitive factives, I show that they allow predicate doubling because they present non-factive readings.2. The puzzle. Spanish exhibits predicate doubling, i.e., a construction in which an infinitive verb appears dislocated in the left periphery of the sentence, doubled by an inflected form of the same lexical verb that occupies a clause-internal position (1a). This construction is also acceptable if the inflected verb is in a complement clause (1b). Importantly, as Vicente (2007) shows, predicate doubling in Spanish triggers a verum focus interpretation (i.e. an emphasis on the positive polarity of the proposition) of the doubled clause.(1)a. Leer, leyó.read.INF read.3SG ?As for reading, he did read.?b. Leer, creo que leyó.read.INF think.1SG that read.3SG?As for reading, I think he did read.?However, while predicate doubling is possible with complement clauses embedded under cognitive factive predicates (2a), it is unacceptable with clauses selected by emotive factives (2b).(2)a. Leer, sé que leyó. read.INF know.1SG that read.3SG ?As for reading, I know that he read.?b. *Leer, lamento que haya leído. read.INF regret.1SG that have.SUBJ.3SG read ?As for reading, I regret that he read.?Predicate doubling has received several analysis in terms of movement (Vicente 2007, Landau 2006, a.o.). Nevertheless, this asymmetry is unexpected under a movement-based analysis of predicate doubling: if it involves A?-movement, we would not expect any contrast between clauses embedded under cognitive and emotive factives, since both classes of factives give rise to weak island effects (3).(3)a. *¿Cómo sabés que bailó Juan? how know.2SG that danced.3SG J. ?How do you know that Juan danced??b. *¿Cómo lamentás que haya bailado Juan? how regret.2SG that have.SUBJ.3SG danced.PP J. ?How do you regret that Juan danced??3. Analysis. I claim that this asymmetry arises due to presuppositional differences between cognitive and emotive factives. First, following Karttunen (1971) and Hooper & Thompson (1973), I assume that cognitive factives can present non-factive readings (i.e. parenthetical readings). Thus, I argue that in predicate doubling constructions, cognitive factive predicates introduce assertions (i.e. their complement is not presupposed). Therefore, they behave exactly as non-factive predicates and, consequently, they allow predicate doubling construction, as in (1b).(4)A. ¿Leyó el libro? read.3SG the book ?Did he read the book??B. Leer, sé que lo leyó (pero no lo entendió) read.INF know.1SG that it read.3SG but not it understood.3SG ?As for reading, I know that he read it (but he didn?t understand it)?Regarding emotive factives, I claim that these predicates prevent predicate doubling because they always lead to a presupposition failure. Adopting a classical Hamblin-Rooth semantics for focus, I assume that a predicate doubling construction with verum focus interpretation (5a) has the focus value of (5b): the proposition it expresses and its negation (for sake of simplicity, I don?t consider the semantics of the bare infinitive topic).(5)a. Leer, [Verum leyó]F.b. {that he read, that que didn?t read}I also assume that focus value is linked to the Question Under Discussion (QUD) in the following way:(6)Context condition (Büring 1997, adapted by Gutzmann et al. 2017)An utterance of sentence S is felicitous in a context c if ⟦S⟧f = QUD(c)That means that (5) is only felicitous in a context in which the QUD is the polar question of whether he read (7a). Adopting a Hamblin-Karttunen semantics for questions, (7a) denotes the set of (7b):(7)a. ¿Leyó?b. {that he read, that que didn?t read}Regarding predicate doubling with emotive factives (8), the embedded clause must be presupposed because of the factive predicate (9a). As I showed before, predicate doubling triggers verum focus, which is linked to a QUD (9b). Therefore, predicate doubling with emotive factives will always lead to impossible presuppositions: the denotation of the QUD triggered by the verum focus (9c) cannot satisfy the presupposition of the embedded clause. Consequently, predicate doubling with emotive factives is unacceptable.(8)*Leer, lamento que [Verum haya leído]F.(9)a. Presupposition triggered by the factive predicate: He read.b. QUD linked to the Verum Focus: Did he read?c. Hamblin-denotation of the QUD: {that he read, that he didn?t read}