Bulgarian relative and factive clauses with an invariant complementizer

The paper offers an analysis of Bulgarian relative clauses introduced by the invariant complementizer deto ‘that’, whose distribution also extends to factive contexts. Using reconstruction as primary evidence for movement, I review the basic facts for its presence (amount readings, idiom interpretation, binding and scope) and absence (Principle C) and argue that both a raising and a matching analysis must be available for the derivation of deto-relatives. I also discuss the distribution and structural properties of resumptive clitics which are shown to block reconstruction in all types of contexts and hence to be compatible with a matching derivation only. Given the structural ambiguity in the derivation of Bulgarian complementizer relatives and in search of a unified treatment of their potentially ambivalent behavior, the paper applies Cinque’s (2003, 2008) analysis of relative clauses, which postulates two identical copies of the relativized Head (internal and external) and exploits different movement options to account for the raising and the matching derivations. It is then argued that such a proposal, which also exploits distinct CP positions, can successfully accommodate all of the observed reconstruction effects (or lack thereof), including the option of resumption. The paper also offers a discussion of factive clauses introduced by the same complementizer and proposes that they are best treated as hidden relatives embedded under a more complex structure involving a PP projection and a silent D head