As the title suggests, this chapter attempts a synthesis of the mechanisms by which installations evolve in time. I must admit I feel dwarfed by the problem. The sophistication of the mechanisms and their number makes biological evolution (of which the principles are reminded at the beginning of this chapter) look almost simple in comparison.

Cultural evolution draws on a series of different mechanisms for producing new variants and selecting them. It is complex because these mechanisms not only apply locally to the existing situation (as in biology), but take into account the (simulated) anticipation of possible consequences and the results of past experience. These are made possible by symbolic representation systems and the use of ‘external’ representation techniques (language, documents, etc.)

There is a ‘dual selection’ making concurrent use of ‘real’ and ‘simulated’ experiments and tests for fitness; it is extremely efficient and reduces risk and waste. This dual selection is monitored by institutions. This monitoring ensures that local changes do not provoke negative externalities in the larger social system.

Finally, although installations’ local functional processes evolve through the betterment loop, their component layers evolve also with their own rationale, partly independently of local installations. This results in installations evolving not only under their local conditions of fitness and efficiency, but also under the crossed impact of other installations.

Each of these mechanisms is easy to understand and to recognize in actual installations. It is their combination that makes the evolution of installations complex. The chapter ends with a simplified diagram that I tried to keep minimal. I apologize to the reader if this chapter is not always straightforward: as for Simon’s ant, the complexity is in the ground I tour, and I still have not found the best path.