The logic of causation in epidemiology.

The paper attempts to model causality with logical conditionals by way of conditional probability. This provides a broad conceptualisation of causality according to which we merely observe tendencies toward sufficiency or tendencies toward necessity. Cohort studies evaluate the first tendencies, and case-control studies the latter. This conceptual approach also clarifies the logic of what unifies the two methods. Ways to measure causal tendencies are proposed. Some of the consequences of this conceptual analysis are discussed and among them, the growing role of chance instead of determinism in science and, although rarely acknowledged yet, in epidemiology.