Explanations of where our behaviour comes from are frequently presented in terms of the exclusive importance of one set of factors, either genetic or environmental. Unravelling the external and internal sources of individual differences is a useful first step in analysing behavioural development. Nevertheless, the analytical method that was well designed for extracting influences from a confusing mass of data was never a substitute for a theory. It was simply a means to an end. Descriptive statements about the genetic and environmental sources of variation in the population do not offer an adequate basis for understanding what happens to individuals. That awareness was an important step in moving towards an adequate theory of behavioural development. As an example of how that may be done, I discuss the interplay between the developing individual and its environment in highly regulated learning processes such as imprinting. Getting the level of explanation right is crucial. A purely molecular or synaptic account of the processes involved in the development of behaviour is inadequate. Nevertheless, those connectionist models that are properly rooted in a thorough knowledge of behaviour and physiology do provide a promising route out of the reductionism and the empty interactionism that characterized the old nature-nurture debates.