Faragó et al (hereafter FMS&G) draw attention to an important issue for researchers of human-robot interaction (HRI): can we conceive a scheme for making social robot behaviour both comprehensible and appropriate in human social settings? We agree with the authors concerning the potential utility of drawing on the example of domestic animals-particularly dogs, the species with which we have the longest history of co-evolution as social interactors. Here we seek to extend from the authors emphasis on the detail of species-specific interaction to a general blueprint for robot action selection. We particularly emphasise the integral role of emotions in facilitating social interaction selection as social signals of internal agent states that are relevant to joint action. Our research questions concern the general abilities of artificial agents, particularly robots, to express their current, transient internal states in ways that people find comprehensible and acceptable. This requires that researchers consider not only the potential communicative value of a social signal but also the validity or utility of the internal state which it describes. Were it the case that the role of the human was to correctly identify a signal, as a passive observer of the robot, it would be a simple matter to construct a repertoire of discriminable social actions. However, this leads us to an important issue in HRI research not emphasised by FMS&G : the nature of interaction itself as a concept that requires simultaneous consideration of the actors and the acted-upon. Dynamic selection decisions for emotional signalling must depend on considerations that span human and robot: a question of emotional interaction selection.