A-movement and clitic doubling in Rioplatense Spanish

SAAB, ANDRÉS; PUJALTE, MERCEDES

Utrecht

Congreso; Going Romance XXXII; 2018

Universidad de Utrecht

Goals: In this talk, we first describe the distribution of dative clitic doubling in Rioplatense Spanish (1a) that, as it is already known, is much less restricted than the distribution of accusative clitic doubling (1b) and, then, explore certain interactions between both types of doublings that shed new light on the syntax of objects:(1)a.Le entreguéel libroaMaría. CL.dat.3sggavethebookto M.?I gave the book to María.?b. LaviaMaría.CL.acc.fem.3sgsawDOMM.?I saw María.?Dative doubling and A-movement: We show that dative doubling is essentially mandatory for any type of indirect object (IO). We propose then an analysis in terms of A-movement, according to which doubling is the surface reflex of an A-dependency between the phase head v and the dative DP. In this respect, dative doubling behaves like accusative doubling, which is also a reflex of A-movement to the v edge, as already shown by Di Tullio et al (in press). Among other relevant diagnostics, Di Tullio et al argue that presence of A-movement in accusative doubling can be detected by the particular distribution of Weak Crossover effects (WCO). This distribution is replicated in exactly the same way when it comes to dative doubling. First, as shown in (2), dative doubling repairs WCO in regular wh/focus extraction (both sentences in (2) are degraded without doubling in those dialects/idiolects that accept non-doubling variants): (2) a. ¿A quiénile entrególanota sui profesor?towhoCL.dat.3sggavethe gradeherprofessor b.AMARÍAile entregó la nota sui profesor.toM.CL.dat.3sggavethegrade his professorSecond, this repair is sensitive to the position of the possessive phrase (see Ishii 2006 for English): Whenever the possessive phrase is not in the same clause as the doubling clitic the result is degraded: (3) a. *?¿A quiéni cree sui madre que le entregó el libro Juan? to whobelieves her mother that CL.dat.3sg gave the book J.?b. ¿A quiéni cree Juan que sui madre le entregó el libro? to who believes J. that her mother CL.dat.3sggave the book?(4)a. *?A MARÍAi cree sui madre que le entregó el libro Juan. to M. believes her mother that CL.dat.3sg gave the book J.b. A MARÍAi cree Juan que sui madre le entregó el libro. to M. believes J. that her mother CL.dat.3sg gave the book.Crucially, this particular distribution is not attested in other constructions that also repair WCO (see Di Tullio et al for a discussion on clitic left dislocation). The key for understanding the pattern in (2)-(4) is the type of movement involved in each case. WCO arises whenever an operator crosses the possessive DP. A-movement doesn?t involve any operator-variable chain; then, whenever the IO crosses the subject DP through A-movement, absence of WCO is predicted. This is the situation that obtains in (2), (3b) and (4b). In the sentences in (3a) and (4a), the IO crosses the DP in the main clause through A?-movement giving rise to WCO effects. Thus, we reach the conclusion that dative doubling, like accusative doubling, involves A-movement to the first vP edge. However, given the broad extension of dative doubling in Rioplatense Spanish (i.e., dative doubling is insensitive to the nature of the doubled IO), we conjecture that the motivation for such a movement to the vP edge is different in IOs and DOs. For the latter, Di Tullio et al provide evidence in favor of the hypothesis that clitic doubling is the result of an optional person feature on accusative DPs. Our conjecture is that IOs are forced to move because of a property on the v head and not on the IO itself. Such a property is a purely selectional feature (a D or EPP feature, depending on certain assumptions). On the ban on double doubling: Regardless the particular implementation, both IOs and DOs are in competition for exactly the same position, Spec,vP. This makes a couple of important predictions. First, given that dative doubling is mandatory in Rioplatense Spanish, IOs always win. This explains why we cannot have double doublings like those in (1a), a hitherto unnoticed observation in the literature.(5)a. *A Juan, se la presenté a María. to J.CL.dat.3sgCL.acc.3sg.femintroduce.1sg to M.b. A Juan, le presenté a María. to J. CL.dat.3sgintroduce.1sgto Mary. The contrast in (5) is straightforwardly captured under our analysis. Here is why. In (5a) the IO requires A-movement before topicalization. For this reason, mandatory A-movement of the IO blocks A-movement of the DO. The sentence in (5b) is grammatical just because the DO doesn?t A-move to the vP edge. The second important prediction is that topicalization of the DO should give grammatical results even in presence of a doubled IO, because accusative clitic doubling is optional; i.e., the DO can be topicalized without being subject to A-movement to Spec,vP. This prediction is borne out. Notice that the final result is the opposite of what we observe in (5): absence of dative doubling is ungrammatical in colloquial (oral) Rioplatense Spanish: (6)a. A María, se la presenté a Juan. to M. CL.dat.3sgCL.acc.3sg.fem introduce.1sgto J.b. *?A María, la presenté a Juan.(better in formal register) to M. CL.acc.3sg.fem introduce.1sg to J. Importantly, the contrast between (5) and (6) cannot be accounted in terms of other well-known restrictions involving a-marked DPs. Concretely, distinctiveness in the sense of Richards (2010) (i.e., the ban for two category identical DPs to be linearized within the same phase) cannot be the cause of the ungrammatical results in (5a) and (6b): the two a-DP are in different phases and, consequently, distinctiveness is respected. Evidently, the grammatical counterparts in (5b) and (6a) also preclude any explanation in terms of something like Richards? condition.We?ll also show that no animacy restriction (Ormazabal & Romero 2007, 2013) is behind our basic minimal pairs. Contrasts like those in (7), in which we obtain the same results as in (5) and animacy is not at issue, allow us to reject alternative analyses along these lines: (7) a. *A instituciones prestigiosas, solo se lo recomiendo a Juan.to institutions prestigious only CL.dat.3pl CL.acc.3sg.mas recommend.1sg to J. b. A instituciones prestigiosas, solo les recomiendo a Juan. to institutions prestigious only CL.dat.3pl recommend.1sg to J.In summary, we provide new evidence for the hypothesis that clitic doubling ? independently of its ultimate motivation? is the PF reflex of an A-movement chain. Importantly, accusative and dative doubling are in competition for the same position in the vP-area, a fact that precludes double doubling even in long-distance configurations; i.e., even when IO and DO are not within the same phase domain.