Key points are not available for this paper at this time.
We present a model of compliance, for domains in which a dialogue agent may become adversarial.This model includes a set of emotions and a set of levels of compliance, and strategies for changing these.
Roque et al. (Mon,) studied this question.