French Relative Qui

Starting from Kayne's (1976) motivation for the existence of the French que/qui rule based on the complementizer system of French relative clauses, I show that French in fact has a double paradigm of wh-elements, a weak one and a strong one, much like what is found in the (strong/weak) pronominal system. Although only French is discussed here in any detail, such a split seems to have much wider relevance, in other Romance languages, in some Germanic and Scandinavian languages, and beyond (Wolof). This split in turn shows that the que/qui rule (and its cognates) should be looked at differently—in particular, that they should be uncoupled from constraints on subject extraction.