On Wh-Islands and Obligatory Wh-Movement Contexts in South Slavic

The goal of this paper is to examine Wh-islands effects in Bulgarian and Serbo-Croatian (SC) and constructions in which SC must have Wh-movement. The starting point will be Rudin’s (1988) seminal paper on the structure of multiple Wh-fronting (MWF) constructions in Bulgarian and SC. In section 1 of the paper I briefly summarize Rudin’s (1988) analysis as well as a modification of Rudin’s analysis of SC from my previous work (see Bošković 1997b, 1999, 2002a). In section 2 I turn my attention to Wh-islands in Bulgarian and SC. Although I will argue against Rudin’s analysis of Wh-islands in the languages in question, which has served as a springboard for much recent work (see, for example, Richards 1997), I will end up endorsing the gist of Rudin’s analysis of MWF, namely, that there are two basic patterns of MWF constructions — one where all Wh-phrases are located in SpecCP and one where at most one Wh-phrase is located in SpecCP. However, the distribution of the two patterns will be shown to be quite different from that put forward by Rudin. In particular, the former pattern will be shown to be available in SC in certain contexts, contrary to what Rudin argues. All the differences between Bulgarian and SC MWF constructions will be traced to a single difference in the lexical properties of the interrogative C, more precisely, the PF affix status of the Bulgarian interrogative C. Furthermore, it will be shown that Wh-islands do not provide evidence for the dichotomy of MWF constructions, as argued in Rudin (1988). In this respect, I will demonstrate that the resistance of Bulgarian to the Wh-Island Constraint should not be related to the possibility of MWF in this language. More generally, I will show that the possibility of MWF itself cannot provide an escape hatch from Wh-island effects, contrary to what is standardly assumed. The claim will be situated in a broader linguistic context and shown to receive ample empirical and theoretical support from a variety of sources, including quantifier raising and Superiority.