Playing along: When do interlocutors recognize ostensible refusals?

This paper investigates the key defining feature of ostensible speech acts, namely pretense, by which a speaker pretends to be sincere. For ostensible speech acts to function as the initiating speaker intends, the interlocutor must recognize the pretense. This study explores the timing of such recognition. Two novel discourse gating tasks probe listeners' ability to distinguish pretense from sincerity, ostensible from genuine refusals, and the point in the conversation at which recognition takes place. Both tasks use turns as gates, presenting the conversational stimuli aurally to listeners turn by turn and asking them to make predictions about the outcome of each conversation. The gating tasks differ in terms of how predictions are elicited. Listeners volunteer their predictions as soon as they can make them (open-prediction) or respond to prediction prompts to make a choice (fixed-prediction). Responses elicited from 60 L1 Chinese speakers and 47 L2 Chinese learners are compared across acceptances, genuine refusals, and ostensible refusals. Results show that listeners do not immediately recognize pretense, and thus may not immediately collude or act on mutual recognition, but L1 speakers tend to recognize pretense faster and more accurately than L2 learners.