Doubling in Dutch restrictive relative clauses: rethinking the Head External Analysis

In colloquial Dutch the higher and the lower clause of long-distance relative clauses (RCs) can both be introduced by a relative pronoun. I argue that these doubling data are most easily compatible with a (specific implementation of the) head external analysis (HEA) of RCs. I show that although head internal analyses (HIAs) of RCs have gained a lot of ground in recent years, they face a variety of problems that are not encountered by a HEA. I furthermore show that the most prominent argument in favor of HIAs – namely the presence of reconstruction effects – is not very well founded, as reconstruction without movement or copies seems to be needed anyway. This strongly suggests that reconstruction is not a full proof diagnostic for movement, and that the presence or absence of reconstruction effects in RCs thus provides inconclusive evidence to support any analysis of RCs.