how nature works: the science of self-organized criticality [Paperback]

1. There is no deep significance to a power law distribution. All it means is that there is no natural scale of the phenomenon. (No power law runs from zero to infinity, so at best it means that there is no natural scale in the range in which the distribution is a power law.)

2. There is nothing new about this conclusion. It was understood in the 1940's by Kolmogorov in his theory of turbulence and Fermi in his theory of cosmic ray acceleration. The only thing Bak added was the application of these old ideas to the artificial and uninteresting problem of sandpiles.

There is nothing wrong with taking old ideas and applying them to new problems. However, Bak never admitted where the ideas came from. He never cited this earlier work. For example, models identical to his ``Self Organized Criticality'' were published by scientists working on earthquakes (Journal of Geophysical Research 90, 1894 [1985] and 91, 10412 [1986], building on earlier ideas by Knopoff). Although Bak was aware of this work, published before he ever began working on ``SOC'', he didn't acknowledge it. To a scientist this is an unpardonable sin, equivalent to a banker cooking the books.

There is a striking resemblence between Bak's How Nature Works and Wolfram's A New Kind of Science. Each author claimed to have made profound and original insights which explain not just one phenomenon, but almost every aspect of the world around us. In each case these supposed insights are essentially mathematical, based on models which ignore the actual physical, chemical or biological processes involved. And in each case closer examination shows that the work is almost trivial, and irrelevant to real scientific problems. You get out what you put in, and when you don't put any science in you, don't get any science out.

Each of these authors has also claimed credit to which he is not entitled by failing to acknowledge the prior work of others. They suffer from a would-be-genius syndrome, in which someone is so infatuated with his own supposed brilliance that he both exaggerates the slight significance of his work and fails to give credit to those who actually did it earlier....

Although there were portions of Bak's work that were a little belabored-I found my interest in sand piles began to sag after the initial discussion, for instance-much of the rest of the book was enlightening. The discussion in Chapter 1 of the contrast between the clarity and simplicity of the laws of physics and the complexity and unpredictability of nature was particularly interesting as was the discussion of the difference between chaos and complexity. His explanation in Chapter 2 of the theory of self organized criticality and the history of its development is far clearer than I found Stuart Kauffman's to be. It might make a better starting place for anyone wishing to understand the theory a little better before going on to Kauffman's and other books on the subject.

Essentially the theme of the book involves the self organization of much of the universe, from stars and volcanoes to traffic jams and economics, into critical states sustained as stable systems until they evolve through cascade events or what Bak calls avalanches (after his sand pile paradigm) or catastrophes. Bak explains that the system maintains itself along a critical line, above which chaos rules and nothing can be predicted and below which nothing happens so there is nothing to predict!

Chapter 5 which deals with earthquakes and volcanic eruptions interested me in particular because of my own study of geology. Here Bak suggests that geophysicists' attempts at prediction of events is a lost cause. He believes it to be based upon the mistaken human habit of looking at random events for patterns and periodicity where none exists. While the history of a given event can be studied in some detail after the fact, the information derived is useless in predicting the future. In Bak's opinion, the variables involved are so legion and are interrelated in so convoluted a way as to be impossible to monitor before the fact.

In chapters 7, 8, and 9 the author attempts to model Darwin's gradual evolution, Gould's punctuated equilibrium, and the Santa Fe Institute's fitness landscape to see which fits the facts better. In general Darwin's theories are vindicated---no real surprise there---while punctuated equilibrium is also found to have it's place in a complete theory of evolution. Chapter 11 contained a section on the unavoidability of catastrophes and fluctuations---and by their extension, one supposes, biological evolution-which casts light on the boom and bust character of economics among other things. This chapter extends the use of the theory of SOC to human activities as well as to human evolution.

The author's style is very chatty, which makes it readable and personable. By filling in the human details of the discoverers, he makes the book more personal. In all, though I found myself occasionally losing the thread of the author's theme, I nevertheless found the content of each chapter well worth.