CONICET | Buscador de Institutos y Recursos Humanos

1. Introduction. The present paper explores the phenomenon of clitic doubling (CD) in Argentinean Spanish. Our main hypothesis is that clitic doubling is a PF effect that expresses an agreement relation of the object with the verb. This effect appears as a consequence of an A-movement to Spec,vP triggered by the [PERSON] feature of the object. We propose that variation facts regarding CD can be explained by how the [PERSON] feature is grammaticalized in lexical DPs. 2. Background. It is a well-known fact that the clitic doubling phenomenon presents an important degree of variation across dialects of Spanish. In General Spanish, the phenomenon is restricted to personal pronouns (1). In turn, in Argentinean Spanish, it extends to other classes of nominal expressions, such as proper names and definite descriptions (2).(1)Juan *(la) vio a ella.[All dialects]Juan CL saw ACC her ?Juan saw her.?(2)a. Juan (la) vio a María.[Argentinean Spanish] Juan CLsaw ACC María ?Juan saw María.?b. Juan(la) vio ala enfermera. Juan CL saw ACC the nurse ?Juan saw the nurse.?The literature on this topic has been misguided with respect to the morphosyntatic conditions that trigger this phenomenon, at least, in Spanish. The basic problem is that they didn?t provide accurate diagnostics to tease apart bona fide cases of CD from other kinds of doubling phenomena, such as Clitic Right Dislocation (CLRD) (3a) and Clitic Left Dislocation (CLLD) (3b). (3)a. Lo compré ayer, el auto.[CLRD] CL bought yesterday the carb. El auto, lo compré ayer.[CLLD] the car CL bought yesterday?I bought the car yesterday.?As a consequence of this short sight on the phenomenon, there seems to be no clue regarding the source of variation across Spanish dialects. 3. Proposal. The purpose of this paper is threefold: first, to provide reliable tests to distinguish CD from other kinds of pronominal duplications; second, to stablish the morphosyntactic conditions that triggers CD; and third, to propose a hypothesis explaining the variation facts. We put foward different tests that show that CDed DPs can be interpreted in the domain of the focus, whereas CLRDed ones cannot. These tests let us to conclude that in Argentinean Spanish, clitic doubling requires the object DP to be differentially case marked, a fact that has involved a great deal of discussion in the literature (Jaeggli 1982, Suñer 1988). As can be seen in (4), in a constext of association with focus, a differentially marked object can be clitic doubled, but the same result is ungrammatical if the object lacks the differential case marker /a/. So, CD in Argentinean Spanish must obey Kayne?s Generalization (Jaeggli 1982, Zdrojewski 2008, Saab & Zdrojewski 2012, pace Suñer 1988 a.o).(4)a. Solo la vi a María only CL saw ACC María ?I only saw María.?b. Solo (*lo) compré el auto.only CL bought the car.?I only bought the car.?Following the reasoning Bobaljik (2008), regarding the relation between agreement and case, we can conclude that if the accusative marker /a/ is introduced at PF, and the clitic depends on the accusative marker, the clitic has to be introduced at PF as well.On top of this, we contrast CD with CLLD constructions. Argentinean Spanish allows for Focus Fronting plus CD (Di Tullio & Zdrojewski 2006, Saab & Zdrojewski 2012). This structure results superficially similar to CLLD structures, as can be seen in (5):(5)a. A MARÍA(la) critiqué.[Focus Fronting + CD]ACCMaría CLcriticized ?I criticized MARÍA.?b.A María, la critique ayer.[CLLD]ACC María CLcriticized yesterday ?Mary, I criticized her yesterday.?It is a known fact that Focus Fronting (without CD) displays WCO effects. However, if a Focused DP is CDed, these effects disappear. In this sense, Focus Fronting plus CD and CLLD behave on par; both kinds of structures avoid WCO effects in simple contexts. However, interesting contrasts arise in complex contexts. As can be seen in (6), WCO effects are avoided in CLLD structures either if the possessive phrase is in the embedded clause or the main clause. (6)a. A Juani, sui profesor cree que loi criticó María.ACC Juan POSSprofessor thinksthatCLcriticized María?His professor thinks that María criticized Juan.?b. A Juan i, María cree que loi criticó sui profesor.ACC Juan María thinksthat CLcriticizedPOSS professor?María thinks that his professor criticized Juan.?Crucially, Focus Fronting plus CD patterns with CLLD if the possessive phrase is in the embedded clause (7b), but differs if it is in the main clause, as in (7a).(7)a. ?*A JUANi cree sui profesorque loi criticó María (no a Pedro)ACC Juan thinks POSSprofessor that CLcriticized María (not ACC Pedro)?His professor thinks that María criticized JUAN (no Pedro).?b. A JUANi cree María que loi criticó sui profesor (no a Pedro).ACC Juan thinks María that CLcriticized POSSprofessor (not ACC Pedro)?María thinks that his professor criticized JUAN, (not Pedro).?We claim that this difference is a consequence of the A/A?-movement distinction. As it is well-known, WCO effects appear only if a pronoun-like element intervenes in an Operator-Variable chain. This kind of chains are created by A?-dependencies, but not by A-dependencies. We propose, then, that CD involves an A-movement to Spec,vP triggered by the [PERSON] feature of object. This means that focused CDed DPs never lead to WCO effects in simplex contexts, which is not the case of normal cases of Focus Fronting (without CD). For the cases in (7), then, we assume the reasoning of Ishii (2006) for D-linked wh-extraction. The first step is an A-movement to Spec,vP followed by cyclic A?-movement to its final position. So, (7b) doesn?t display WCO effects, because the movement of the object across the possessive phrase in the subject doesn?t create an Operator-variable chain. (7a) is ungrammatical because the possessive phrase in the main clause intervenes in an Operator-variable chain created by the subsequent A?-movement of the objects to Spec,CP. The lack of WCO effects in (6) is explained because dislocated topics don?t move cyclically, as originally proposed by Cinque (1990)This contrast leads us to a second condition: Clitic doubling is triggered by person features on pronominal/DP objects. These features are the responsible of the A-movement to Spec,vP.If this idea is correct, we can approximate to a hypothesis for the variation facts regarding clitic doubling in Spanish: clitic doubling varies according with the degree of grammaticalization of [PERSON] features on Lexical DPs. 4. Referencies. Bobaljik, J. 2008. Where?s Phi? Agreement as a Post-Syntactic Operation. In D. Harbour, David A. and S. Béjar (eds). Phi-Theory: Phi features across interfaces and modules, Oxford University Press, 295-328. / Cinque, G. 1990. Types of A?-dependencies. Cambridge, Mass.: MIT press. / Di Tullio, Á. & P. Zdrojewski. 2006.Notas sobre el doblado de clíticos en el español rioplatense: asimetrías entre objetos humanos y no humanos. Filología XXXVIII-XXXIX.13-44. / Ishii, T.. 2006. A nonuniform analysis of overt wh-movement. Linguistic Inquiry 37:155?167. / Jaeggli, O. 1982. Topics in Romance Syntax. Dordrecht: Foris / Saab, A. & P. Zdrojewski. 2012. Anti-repair effects under ellipsis: Diagnosing (post-) syntactic clitics in Spanish. In I.Franco, S. Lusini & A. Saab, (Eds.) Romance Languages and Linguistic Theory 2010, Amsterdam, John Benjamins, 177-202. / Suñer, M. 1988.The Role of Agreement in Clitic Doubled Constructions. Natural Language & Linguistic Theory 6:3.391-434. /Zdrojewski. P. 2008. ¿Por quién doblan los clíticos? Restricciones morfosintácticas sobre la duplicación pronominal en el español rioplatense. MA dissertation, Universidad Nacional del Comahue.

enviar mensaje