Equilibrium phenomena are simpler to analyze than time-dependent phenomena. The fact that non-linear time-dependent systems may display perplexing behaviors is illustrated in Fig. 4.1. It does not come as a surprise, therefore, that the equilibrium requirement is an essential assumption in statistical mechanics. For time-dependent systems the notions of statistical ensemble, temperature, entropy as well as several other fundamental concepts are no longer clearly defined and many standard results of statistical thermodynamics, such as the equipartition law, are no longer satisfied. In a general way, whereas equilibrium statistical mechanics provides a systematic and unified theoretical framework, non-equilibrium statistical mechanics is rather a collection of loosely connected techniques and equations.

It is certainly important to know whether or not a social system can be considered as being in equilibrium. What criteria do we have in this respect? This is the first question that we consider. For a system to remain in equilibrium there must be an equilibrium restoring mechanism, therefore one is naturally led to examine these mechanisms. This is done in the first part of this chapter. In the second part we investigate metastable systems. From a common sense perspective a diamond would seem to be a perfect example of a system in equilibrium; the fact that diamonds are in fact metastable shows that common sense can be fairly misleading. By discussing several examples of metastable systems in physics and in chemistry we try to develop a qualitative understanding of metastability; in particular we emphasize the role of facilitator played by “seeds” or “precursors” of the new organization.