Cliticization Phenomena in Languages ‘on the Border’

The paper investigates clitic doubling in two non-standard Serbian and Slovenian dialects, Prizren-Timok Serbian and Gorica Slovenian, respectively. These dialects have clitic doubling but lack overt articles, which, prima facie, seems problematic for Bošković\u27s 2008 generalization that clitic doubling is found only in languages with articles. Nevertheless, a thorough analysis of the dialects at stake reveals that not only is doubling limited to pronouns exclusively but also that pronouns in these dialects enjoy both lexical/N and functional/D status. The major evidence for N status is based on the fact that the adjectival modification of personal pronouns is allowed, being banned, however, when the pronoun is doubled, which reveals its D status in the clitic doubling environment. In line with Kroch’s 1994 account of syntactic change, I argue that the presence of the dual pronominal behavior in PTS/GS is the reflection of an ongoing language change, with the transitional stage containing two mutually exclusive systems. Further, several identical phenomena attested in these dialects, such as the impossibility of a verb to intervene between a clitic and its associate, the impossibility of a verb to precede the entire doubling construction, and doubling with full NPs, further demonstrate that the doubling constructions are undergoing a change. In line with Bošković\u27s 2001 approach to cliticization in South Slavic, I argue that the order verb-clitic arises through a lower copy pronunciation, which I claim is blocked in the clitic doubling environment in the dialects in question. Finally, doubling with full NPs, attested with some speakers in the two dialects, yields no specificity/definiteness effects and licenses left branch extraction, which I show lends further credence to Bošković\u27s 2008 claim that languages without overt articles do not project a DP layer on top of NP in the syntax