The goal of this review was to provide a brief overview of recent developments in the domain of emotional mimicry research. We argue that emotional signals are intrinsically meaningful within a social relationship, which is crucial for understanding the functionality and boundary conditions of emotional mimicry. On the basis of a review of the literature on facial mimicry of emotion displays, we conclude that the classic matched motor hypothesis does not hold for emotional mimicry. We alternatively propose a contextual view of emotional mimicry, which states that emotional mimicry depends on the social context: we only mimic emotional signals that are interpreted to promote affiliation goals and not necessarily what we see. As a further consequence, we are less likely to mimic strangers and we do not mimic people we do not like nor emotions that signal antagonism.