Sign languages apply different means to mark focus and other information structural notions. So-called question-answer pairs are one of these options to highlight constituents. Based on a recent debate on the syntactic analysis of question-answer pairs as a sentential unit in American Sign Language, we discuss their syntactic structure and present two empirical studies comparing possible positions of wh-words in wh-questions and question-answer pairs in German Sign Language (DGS). In wh-questions of DGS, wh-words may appear sentence initially, sentence finally, and doubled in sentence initial and final position, with a tendency of an unmarked final position. Question-answer pairs in DGS exhibit a specific nonmanual marking and show no clear restrictions with respect to the choice of wh-words. Wh-doubling, which is generally used for emphasis in regular wh-questions, is also possible in question-answer pairs in DGS, but more marked than the initial and the default final wh-word position. On these grounds, we argue against Wilbur’s (1996) wh-cleft account and follow a pragmatically modified version of Caponigro & Davidson’s (2011) analysis as complex declarative clauses to account for question-answer pairs in DGS.