This talk aims to provide a panorama of how the discrete and the continuous are dealt with in medicine. Modern Medicine is oriented towards diagnosis, and treatment or prevention of discernable diseases. Medical practice, research, education, as well as the regulation of health care draw on more or less clear-cut disease classifications and more or less precise circumscriptions of patient populations.

This talk addresses theoretical and practical challenges in demarcating the normal from the pathological. These include (i) the possibility of an ontological continuity between health and disease or between (co-)morbidities, (ii) the abundance of epistemic uncertainty in diagnosis, and (iii) the diversity of interests in defining relevant outcome parameters within the various arenas of the generation, evaluation, and application of medical knowledge. The talk proceeds in discussing four strategies that deal with the challenges of blurred boundaries in medicine: (1) the definition of threshold values, (2) the introduction of sub-types, degrees of severity, or prodromal stages, (3) the use of vague terms, and (4) the dissolution of disease concepts. It shows that all of these strategies, individually and in combination with each other, have merits and vices.

The presentation concludes with a reflection on what it means to observe at the margins. It argues that borderline cases exist in a dusky light where specific aspects of the phenomenon of interest become particularly evident – just like, as the saying goes, the contours of a tree in l'heure entre chien et loup.