Cuantitative clitics in Romance and Slavic

Laburpena

In this article I argue that quantitative clitics in Romance (Italian and French) and Slavic (Czech) perform the task of identifying a quantitative pro inside an internal argument, since they carry the feature [+ quantitative] . In these cases, the clitic, together with its maximal projection, is generated in the Spec of VP. Romance quantitative clitics cannot be related to an external argument, whereas Slavic clitics can. The clitic should be generated in the Spec of AgrP in order to do that. So, the contrast between Slavic and Romance is due to the fact that those clitics have to incorporate into a higher head in order to become visible, but this is not possible unless another head is above Agr. In Slavic, this head is T. The fact that Agr is not enough for pro to be identified by Spec-head agreement (a strategy available in Romance, and incompatible with the presence of the clitic) is due to the fact that Slavic clitics also identify the +gender, +number pro features, which cannot be transmitted by Agr through Spec-head agreement. By contrast, Romance D's governing pro (which carry gender-number morphemes) perform this task.