Hidden Order: How adaptation builds complexity
by John H Holland
Helix Books,
Addison Wesley
ISBN 0-201-40793-0
1995; $24.00 US
Reviewed by Roger F Malina
This book is about how complex adaptive systems (economies, cities, diseases, ecologies, civilisations, the nervous system) emerge, survive, evolve. With remarkable clarity Holland outlines a research agenda that began over the last twenty years and will become one of the main new sciences of the next to the invention of genetic algorithms, and his own road map (a class of models called "echo" models) of the fundamental principles that he thinks will underpin successful theories of complex adaptive systems. Upon re-reading, the word clarity summarizes best Holland's achievement in this book. He clearly defines the problem set as he sees it, clearly identifying the underlying principles, and succinctly describes the necessary interaction between observed phenomena in diverse domains, the models and simulations, and the resulting rigorous mathematically-based theories. Like the Lectures on Physics of Richard Feynman, this book and Holland's work will have wide influence.
In his closing chapter Holland states boldly his beliefs about the broad requirements of a successful approach to theory of complex adaptive systems. Each one of these are rich areas for discussion in themselves:
1. Interdisciplinarity: the theory must apply to very different domains - from biology, medicine and ecology to sociology, anthropology and history. A good theory of complexity must apply equally well to the emergence of AIDS and the fall of Central American civilizations, as well as the evolution of the Internet.
2. Computer-based thought experiments: Computer-based models allow complex explorations not possible with real systems. These can guide theoretical thinking, but Holland is insistent that computer models that happen to match certain characteristics of real systems should not be mistaken for a deeper understanding of underlying principles and predictive theoretical constructs.
3. A correspondence principle: Holland insists that a successful theory of complex adaptive systems must encompass standard models from prior theories (just as Bohr's famous correspondence principle required consistence between quantum theory and classical physics). Thus a successful theory of complex adaptive systems must map cleanly to well established theories of microeconomics or virology.
4. A mathematics of competitive processes based on recombination, Holland insists on the need for rigorous mathematical generalizations that will define the trajectory of the evolution of a system based on the interaction of competition and recombination- something that computer based experiments cannot provide on their own.
Holland's book is almost breathless with the excitement and optimism of the founder of a new scientific discipline. It speaks with joy about the way that the scientific method is being enlarged to deal with phenomena previously thought to be resistant to scientific inquiry. Just as our concept of science was enlarged with the advent of quantum theory at the turn of this century, so the theory of complex adaptive systems promises to redefine the scientific landscape of the next century. I highly recommend this book, particularly to artists and others developing interactive art works. If Holland is right, there is no reason that a good theory of complex adaptive systems should not also apply to the creation of the interactive art systems of the future.
