This contribution deals with verb second (V2) in subordinative contexts, in particular in complementiserless complement clauses. As V2 is considered as a main clause phenomenon in German (and several other Germanic languages) the question arises how much alike to main clauses such embedded V2 clauses actually are. Focussing on German, I will address the syntactic problem whether the domain in front of the finite verb (the “prefield”) in V2 complement clauses exhibits the same properties and the same structural layout as in V2 main clauses. Evidently, the prefield of V2 complement clauses is restricted in certain ways, for dislocation processes to the left are very limited. This can be seen as indication of a structurally reduced left periphery comparable to the reduced size of subordinate clauses in general. As a first step to address these issues I discuss a piece of indirect evidence in the present paper: complement clauses with a doubled, or resumptive, complementiser dass ‘that’ as in er meint, dass, wenn er das erreicht, dass sich dann auch die Erfolge wieder einstellen ‘he thinks that if he manages this the successes will ensue again’. Since such clauses display a structurally richer left periphery than complement clauses with a single complementiser they shed light on the syntactic structure of the left edge of complement clauses in general and, by implication, on V2 complement clauses. Relying on data from large written and spoken text corpora I argue that the doubling of complementisers involves recursion of FinP (containing the complementisers) together with the top-most topic position(s) of the middle-field. This analysis is supported by the fact that complementiser doubling is not always a repair strategy to solve processing difficulties. It typically involves a specific type of sandwiched material in-between the two complementisers, namely delimitative expressions, or more precisely: framesetting and contrastive topics. Hence, the findings with respect to complementiser doubling support the analysis that subordinate clauses do not possess an articulated C-domain above FinP. Therefore, the assumption that V2 complement clauses have a reduced C-domain - just like any other subordinate clause - should be maintained.