The approach advocated here differs from most in the emphasis it places on the constructive capacity of the nervous system, and the advantages of flexibility that results from the course of development not being laid down in detail but discovered by trial and error. On the other hand, it insists that development is not a process of blind trial and error: it is desirable for particular processes to be performed, and advantageous to have mechanisms designed especially for them. However, it avoids postulating ill-specified mechanisms such as a `language acquisition device' by insisting (in line with the analysis in [5]) that these complex developments be broken down into their elementary steps of progress, and that these be understood individually. In this way we may hope to have arrived at a very practical way of probing the extreme complexities of the development of human cognitive abilities.