In this article I discuss three existing complement-of-D analyses as possible solutions to the behaviour of the German determiner d-jenig, which requires the presence of a restrictive relative clause. As it turns out, none of them is without problems. As an alternative I propose a structure without the DP-hypothesis which meets some of the very problems. In the end I provide a number of observations pertaining to its morphology and its discourse behaviour, which may lead to conclude that the determiner in question has a complex internal structure.