The aims of this article are to arguefor an alternative approach to theorizing about social skill and to apply a model developed within this alternative perspective in identifying the mechanisms that give rise to various performance deficits. Theorizing about social skill is currently dominated by models that emphasize the role of knowledge and motivation in skilled behavior. Despite this, it should be clear that people may possess adequate knowledge and motivation and still fail to perform skillfully. For this reason, it is desirable to examine the mechanisms of behavioral production that give rise to nonoptimal behavior even when knowledge and motivation are sufficient. Toward this end, a particular model of the behavioral-production system, action assembly theory, is presented, and its implications for understanding performance deficits explored.It seems an unavoidable and frustrating fact of our social lives that on many occasions we produce messages only, on further reflection, to succeed in formulating a response that would have been more skilled or effective than the message we actually gave. Who among us hasn't come up with an original and incisive response to a critic on a conference panel long after the last group of weary conventioneers has abandoned the hotel lobby for the evening? It seems that we often struggle to comfort a friend, graciously accept a compliment, handle conflict with skill, or write just the right sentence, and yet later figure out a way to have done it "better."One of our aims here is to argue that this common phenomenon suggests the need for a reorientation in our thinking about social skill. Instances in which we succeed in formulating superior messages after the fact indicate that while we may possess appropriate knowledge and motivation to perform well, these factors do not insure skilled behavior. Our point, however, is not simply that there are factors in addition to knowledge and motivation that are relevant to social skill, but instead, that a style of theorizing, distinct from what currently dominates approaches to social performance, is required.The broad-based and sustained nature of research on social skill and competence has given rise to a large number of explanatory and organizational frameworks reflecting an array of specific behavioral and/or cognitive foci, root metaphors, and assumptive foundations (see Curran