Crises & Tipping Points: From Statistical Physics to Social Sciences
As P. W. Anderson wrote in 1972 in his article "More is Different", the behavior of large assemblies of individuals (/molecules) cannot be understood by extrapolating the behavior of isolated individuals (/molecules). On the contrary, completely new behaviors, sometimes spectacular and difficult to anticipate, can appear and require new ideas and methods. The purpose of statistical physics is precisely to try to understand these collective phenomena, which do not belong to any of the underlying elementary constituents. In particular, small changes at the individual level can lead to dramatic effects at the collective level. Several simple examples will be discussed, which demonstrate the need to go beyond the models of classical economics, based on the idea of a "representative agent" (moreover rational), and for which only exogenous events can lead to crises - whereas many socio-economic or financial phenomena seem to be fundamentally endogenous in nature.