This article recounts the background to the development of the Elston-Stewart algorithm for the genetic analysis of pedigree data. The algorithm can be used in the context of two contrasting research programmes. The first approach is analytical and reductionist. Traditional biological disciplines, such as physiology, embryology and comparative biology, describe phenomenological domains in which biological organisms are treated as complex systems with emergent properties. In the reductionist approach, these biological domains are largely abandoned in favour of descriptions in the domain of the component elements of these systems, resulting in a concentration on molecular genetics as the central object of biological science. In the second approach, it is explicitly recognized that genes and molecules, per se, cannot be explanatory with respect to emergent systemic properties. However, since the genetic variation which occurs in natural populations can influence virtually every aspect and level of biological organization, from cell lineages to ecology, the study of such variation opens new possibilities for establishing meaningful articulations between the various biological disciplines. The author’s expressed preference is for the second alternativ