Puzzling response particles: An experimental study on the German answering system

Abstract

This paper addresses the use and interpretation of the German response particles ja, nein and doch. It presents a series of four acceptability-judgment experiments that collected preference data for the full paradigm of standard German particles in responses to positive and negative assertions and were designed to test the empirical validity of two recent accounts of response particles, Roelofsen & Farkas (2015) and Krifka (2013), which view response particles as propositional anaphors, and which we refer to as feature model and saliency account, respectively. The results for responses to negative antecedents were unpredicted and inconsistent with either account. A further unexpected finding was that there were two groups of participants that differed in their preference pattern for affirming responses to negative antecedents. We discuss possible revisions of the feature model and the saliency account to account for the findings, and explore in how far the findings can be accounted for in alternative, ellipsis accounts of response particles.