Criticality as a Statistical Process P er Bak's discovery of self-organized criticality in the dynamics of sandpiles provided an explanation for the long-standing puzzle of flicker noise. It also explained the source of some fractal physical patterns because the process produces power functions of complexity cascades [1][2][3]. Subsequently, its patterns have been found in many varied, and sometimes unexpected, places. The most obvious of these include cellular automata and similar sorts of models [4-7], geophysical phenomena [8][9][10][11], ecological systems [12][13][14][15][16], and forest fires and erosion [17,18]. Nevertheless, outside of the physical sciences, it is an understatement to say that self-organized criticality has not been universally embraced. In fact, it has been largely ignored, particularly in the social sciences, where those interested in nonlinear dynamics have focused their attention, instead, on chaos [19].
This seems to be a major mistake because the behaviors of social, political, and economic systems are inherently probabalistic, and because self-organized criticality offers a better way to model their dynamics than does a deterministic approach. Indeed, self-organized criticality is the sort of process that should have great intuitive appeal to social scientists. Like earthquakes, the precise prediction of riots; wars; and the collapse of economies, governments, and societies is not possible in either geophysical or human systems in which catastrophic events can be triggered by minuscule causes.
A number of factors have contributed to the unfortunate lack of enthusiasm for self-organized criticality. Among them, flicker noise was recognized as a major intellectual problem only in the hard sciences, and the importance of its resolution is not apparent to others. In contrast, power functions are an alien concept to most social scientists, who often believe that the behavior of all important variables can be modeled by using normal, binomial, Poisson, or some other simple probability distribution. So even the most obvious examples of the complexity cascades of