CLITIC DOUBLING Elena Anagnostopoulou elena@phl.uoc.gr , elena.anagnostopoulou01@gmail.com 1. Introduction 1 This is an updated version of my 2006 contribution to the Blackwell Companion to Syntax (Anagnostopoulou 2006). I investigate the syntax of clitic doubling constructions focusing on the question of variation within and across languages. In the course of the discussion, two different patterns of clitic doubling will emerge, with different factors causing variation across languages in each case: a) In the case of direct object clitic doubling, languages vary with respect to whether they permit it (i) only with pronouns or also with DPs, (ii) with humans, animates or also with inanimates and (iii) with specific indefinites, partitives or only definites. Direct object clitic doubling of DPs is generally optional (doubling of pronouns is in many cases obligatory). b) In the case of indirect object doubling, languages vary w.r.t. whether (i) doubling is obligatory in the double object construction or (ii) doubling is optional in the double object construction. The research on clitic doubling will be presented from a historical perspective focusing on the nature of the Clitic Doubling Parameter and its potential implications for the syntax of cliticization. I will identify four periods in the study of doubling which coincided with different stages of the Principles and Parameters framework and Minimalism: (i) Early (GB) studies (e.g. Jaeggli 1982; 1986, Borer 1984) take clitic doubling as an argument for a base-generation analysis of clitics and the movement-properties of cliticization as properties of Chains rather than Movement (contra Kayne 1975 who argued for a movement analysis of French clitics, a language lacking clitic doubling). In their attempt to isolate the factor underlying the property that permits formation of clitic-argument pairs in some languages (Spanish, Romanian) but not in others (French, Italian), these accounts capitalize on Romance and Semitic clitic doubling which is limited to DPs that are preceded by special prepositions (a in Spanish, pe in Romanian, šel in Hebrew). (ii) Late GB studies maintain the base-generation syntax of clitics and concentrate on certain interpretive effects associated with direct object clitic doubling, which will be referred to by the term “specificity” (see Enç 1991 and Diesing 199β for discussion and references). This change of perspective is initiated by Suñer (1988), who argues on the basis of data from Argentinean Spanish that direct object clitic doubling does not rely on the presence of special prepositions, contrary to what has been previously thought. 1 Parts of the material included in this paper has been presented in guest classes taught at the Universities of Stuttgart (June 2009, Erasmus exchange class) and the Leiden (September 2012, mini-course “Meaning Domains, Phases and Agree”). I would like to thank the organizers and the audiences, especially Artemis Alexiadou, Gianina Iordăchioaia, Florian Schäfer, Roberta D’Alessandro, Irene Franco, Laura Migliori and Giuseppe Torcolacci. I would also like to thank Winfried Lechner, Dominique Sportiche,Vina Tsakali and an anonymous reviewer for their feed-back. This work has been partially supported by an Alexander von Humboldt Friedrich Wilhelm Bessel Research Award (2013).