On the left periphery of Spanish indirect interrogatives

Abstract

Abstract Spanish doubly filled complementizer (DFComp) clauses differ from plain embedded questions in a number of respects (availability of discourse-related projections, islandhood, sequence of tenses, licensing of discourse particles). I argue that the contrast is caused by the presence in the left periphery of these clauses of an illocutionary projection (Haegeman 2004, 2006; Coniglio and Zegrean 2012; Woods 2016b) between the leftmost projection, here identified as Haegeman’s (2004) SubP, and the criterial interrogative projections (InterP and QembP). This illocutionary projection prevents syncretism of the clause-typing and the criterial projections, the default option in plain embedded clauses. This not only explains the range of structural phenomena differentiating DFComp clauses and embedded questions, but also a key semantic property of the former, namely their speech-act denotation. Finally, DFComp clauses are compared with plain embedded questions displaying root behavior under first-person matrix subjects and with English inverted embedded questions. Both are shown to pose minimal variants of the structural pattern proposed for DFComp clauses.