Relationship processes in adult romantic relationships have been vastly studied under the aegis of attachment theory. Attachment theory is interactionist in nature, proposing that individual differences in levels of both attachment avoidance and anxiety predict an individual's thoughts, feelings, and behaviors in contexts that activate attachment concerns. A growing body of research is beginning to systematically test the conditions in which individual differences in attachment orientations both predict and fail to predict relationship processes. In this article, we focus on anxious attachment and review a program of research showing that the potentially destructive relationship processes typically observed in highly anxious individuals do not always appear in neutral or benign contexts, or when security needs are met. We argue that research needs to more thoroughly investigate the conditions that should, or should not, activate attachment concerns and thus result in links between individual differences in attachment orientations and relationship processes.