I analyze the process by which we react cognitively to information that contradicts our culturally held sentiments in the context of affect control theory. When bizarre, unanticipated events come to our attention and we have no opportunity to act so as to alter them, we must reidentify at least one event component: the actor, the behavior, or the person who is the object of the action. Currently, however, affect control theory includes no formal model predicting which concept will undergo reidentification. I consider theoretical reasons why actors, behaviors, and object persons each might be predicted to undergo reidentification more than the others. I also present an alternative prediction, consistent with affect control theory logic, based on the contribution made by each concept to the total deflection of the event; I label these discrete contributions "component deflections." I then test these possibilities in two separate experimental studies, using event vignettes. By far, the most commonly reidentified event component is the behavior; this finding suggests a dynamism in the conceptualization of behaviors which has been underappreciated in a social psychological literature that focuses on self-identities. (Contains 1 figure, 10 footnotes and 9 tables.)