Several theories of the stress-disease link have now incorporated prolonged activation. This article argues that these theories still lack an important element, that is, the cognitive nature of the mechanism that causes stress responses to be sustained. The perception of stress and the initial response to it do not automatically lead to prolonged activation. The active cognitive representations of stressors need to be prolonged in order to extend their physiological concomitants. We call this mediating process perseverative cognition, and it is manifested in phenomena such as worry, rumination, and anticipatory stress. We summarize evidence suggesting that these phenomena are indeed associated with physiological activation, including cardiovascular, endocrinological and immunological parameters. This evidence is still far from sufficient, due to the many methodological insufficiencies in the studies involved. Nevertheless, it makes clear that cognitive phenomena characterized by perseverative cognition may be likely candidates to mediate the effects of stress sources on somatic disease. We also argue that there is a dearth of evidence supporting the role of prolonged activation. There are a limited number of studies demonstrating prolonged activity related to stressors and emotional episodes, and their methodologies often do not allow unambiguous conclusions. Even more important, the crucial assumption that prolonged activation actually leads to pathogenic states and disease has received hardly any attention yet and therefore is still largely unsupported. There are only a few studies that showed that anticipatory responses and slow recovery from stress predicted disease states.