We are faced with a fundamental problem for any science of complexity. To render the complex world understandable it is necessary to simplify it, but how can we render the complex simple without losing that which is distinctive and interesting about it? If complex systems are indeed ‘irreducible’ and therefore cannot be represented in ‘simple’ analytical models (Cilliers, 1998:9), is scientific reflection on complex systems doomed to be uninteresting reductions that can never grasp what is truly important? And if we insist on the irreducibility of complexity, can we avoid superfi- cial accounts that are prone to relativist interpretations? This is a problem with which Paul was well-acquainted, which is why he dealt extensively with models o...