ABSTRACT
Dual causation does not appear as a concept in most of the literature on causation. One can understand this in the philosophical literature, since few causal theorists rely on a base of causal logic, but it is somewhat surprising that the idea has not made more progress among empirical scientists. The causal diagrams and models in the social sciences are oblivious to dual causation because of the probability models that they have chosen to use. In epidemiology, however, where the dominant trend has been to think in terms of factors (even when there is a better, underlying continuous measurement), the failure to appreciate dual causation is particularly hard to understand.
