I argue that the constraints on the distribution and interpretation of coindexed strong and weak (clitic/pro)) pronouns in Serbo-Croatian (SC) are best captured in terms of specific discourse conditions under which the two types of pronouns are licensed. I show that the licensing of coindexed weak and strong pronouns crucially depends on the notions of topic and focus such that the former pick out topic antecedents, while the latter require new information focus antecedents. I argue that the information status of the antecedent can explain the well-known competition effects between different pronoun types (see Cardinaletti & Starke 1999, a.o.). I also demonstrate how the above discourse conditions allow us to unify the behavior of pronouns in forward and backward pronominalization.
Sent from Mail for Windows (https://go.microsoft.com/fwlink/?LinkId=550986)