Research on spoken languages has shown that response particles may indicate the truth of a previous utterance or the polarity of the response. In responses to negative antecedents, the two functions come apart and particles become ambiguous. We present the first quantitative study on response strategies in sign languages by discussing data from a production experiment in German Sign Language (Deutsche Gebärdensprache; DGS). The results indicate that DGS does not exploit the potential of simultaneous manual and nonmanual strategies to disambiguate responses. Still, the type of articulator influences the choice of response element. We propose an optimality-theoretic model to account for the role of articulator type, the disambiguation potential, and the morphosyntax of response elements in DGS.