In this paper I challenge the accepted view that atypical development necessarily serves as a window on the normal mind/brain. Such a stance ignores the dynamics of development. Rather I submit that the development of many atypical groups cannot be thought of in terms of a normal brain with whole cortical circuits intact or impaired, as in the adult neuropsychological model, because the brains of genetically impaired children develop differently at multiple levels – brain anatomy, brain chemistry, temporal patterns of brain activity – throughout embryogenesis and postnatal development. This is briefly illustrated with reference to studies of Williams syndrome. It is argued that very subtle differences in developmental timing, neuronal density, neuronal efficiency, firing thresholds, transmitter types, synaptogenesis and pruning, may have crucial but very indirect effects on the resulting phenotypes, affecting certain systems more than others. Subtle differences at a much lower level than cognitive modules are likely to explain the range of phenotypical outcomes that humans can display.