Networks represent the spaces of economic (inter)actions, which transform along emergence and exit of institutional forms. Connections between institutions link between social practices and enforce social learning across the whole population, as we have explored in the previously presented agent-based model of institutional change. But these networks are not constant over time, they are intrinsically different from our geographical space, which is rather fixed. Socioeconomic networks constitute systems which are entities for themselves in other systems of higher hierarchy, i.e. ‘system-element duality’ (compare Potts 2000, p. 68). Potts (2000, p. 69) articulates a ‘hyperstructure’ as a system of systems and ‘combines the notions of emergence and hierarchy into a single construct.’ The important point is that Potts (2000) follows already the notion of synthetic programming by using the language structure of algorithms instead of closed form mathematics, as discussed in Chapter 12. Potts (2000, p. 116) defines the ‘hetero economicus’ as the ‘algorithmic man’ and adopts thereby the idea of evolution as computation. Hyperstructures are formalized with graph theory to in his evolutionary microeconomic theory. They represent the complex universe of economic interactions, spaces where heterogeneous agents are engaged in a ‘circular flow from knowledge to capital’ (Potts 2000, p. 119). Capital is conceived as the by-product of knowledge evolution in this context, feeding then back again as another source for novelty; reaching transitional states of super-criticality. Conclusively the complexity and evolution of a hyperstructure is strictly dependent on the weak connections between the hierarchical systems and their corresponding knowledge repositories, i.e. their institutions. Potts (2000, p. 149) outlines according to Herbert Simon that it is therefore necessary to decompose the hyperstructure in order to understand the systematics of its weak connectivity.