This thesis adopts a complex systems stance towards the origin and growth of intelligence in biological and artificial systems at both reactive and reflective levels of adaptation. A synthetic strategy is employed dually informed by characterisations of complex biological systems together with the engineering of embodied artificially intelligent systems. Following a review of the two dominant, and often construed as mutually exclusive, "rule-based" and "behaviour-based" characterisations of systems it will be argued that neither provide a comprehensive account of the growth and organisation of complex biological systems, and that both result in brittle, hand-crafted and, critically, pre-interpreted artificial systems. Furthermore, it will be suggested that the growing consensus that this impasse can be resolved by conjoining conventional reactive and reflexive components in a single architecture should be rejected. Such considerations indicate that a new agenda is required. A synthetic intelligence stance is adopted herein, motivated by convergent biological and engineering approaches within a dynamic systems framework which emphasises system self-organisation. In this context some new experiments on mobile robot navigation are reported which demonstrate the feasibility of the synthetic approach. The implications of this novel stance are discussed with particular emphasis on the future of self-organising artificial systems.