Reviewed by Reilly Jones

This is a landmark book, encompassing daring new holistic ideas about living systems. Stuart Kauffman is Professor of Biochemistry and Biophysics at the School of Medicine, Univ. of Pennsylvania, and External Professor at the Santa Fe Institute. The application of the mathematics of complexity theory to specialized branches of natural science is progressing rapidly. Kauffman’s is the first comprehensive effort to apply complexity to the theory of evolution by placing evolution within a larger biophysical framework of potential universal laws.

By looking at evolution in such a new, over-arching way, Kauffman has placed himself at a considerable distance from established reductionists in developmental biology. He will doubtless be proven wrong in many of the details he covers, but by being bold, general and asking questions arising from new graphical computer modelling techniques, he points in many different directions for fruitful research.

The book begins with an ‘Introduction’ containing the contemporary theory of evolution and some of the peripheral challenges to it, along with Kauffman’s effort to place Darwinism within a larger framework of biophysics. Part I is entitled ‘Adaptation on the Edge of Chaos’ and deals with fitness landscapes and adaptation in sequence spaces (protein and DNA). Some bold hypotheses on construction requirements of complex evolving ecosystems are presented based on novel modelling techniques and the technological promise of Adaptive Molecular Evolution is outlined. Part II is entitled ‘The Crystallization of Life’ and deals with the origins of life, metabolism and coding. Chapter 10 on random grammar models is a gem all by itself; it contains open-system analog modelling techniques of biological, economic, technological and cultural systems. Part III is entitled ‘Order and Ontogeny’ and deals with cell differentiation and morphology. This section reads more like a textbook than the earlier sections, it contains much collected research material but no conclusions on the relative influence of spontaneous order in within-cell versus between-cell genetic regulation.

The first two sections of the book are written for general science readership although some familiarity with complexity theory would be of help. The last section is more of interest to the developmental biologists. The combination of complexity and evolution brings forth new concepts such as ecosystem attractors, extinction and speciation power laws, and frequent definitions of ‘spaces’ (what we used to naively call systems). In fact, there is such an abundance of ‘spaces’ throughout the book that Kauffman could be characterized as ‘space’-happy.1 He frequently uses ‘If/then statements’ typical of much of biology (as opposed to empirical laws typical of physics). For example, “If it is the case that systems poised between order and chaos are indeed the natural culmination of selective evolution, we shall have found deep laws indeed.”

He clearly is aware of the new ground he is breaking with his holistic point of view, but throughout the book shows deep respect for the body of knowledge containing reductionist microphysics at the cellular level. In his own words: “The theories presented are merely the beginnings of a new area of thought and investigation in biology, chemistry, and physics-;perhaps even in economics and other areas of social sciences. The spirit of all the ideas discussed... is a kind of unrepentant holism and a sense of synthetic biology rather than the familiar reductionistic analytic mold.”

His search for universals, or what he terms the “physics of biology” leads him to conclude that, “Biology is surely harder than physics.” He proposes some very broad potential universal laws that have direct relevancy to Extropian principles of directed self-transformation and boundless expansion. The broadest of all is of great interest towards development of more complex selectional systems than mere survival; it involves the possibility of evolutionary feedback. “...The capacity to evolve is itself subject to evolution and may have its own lawful properties. The construction principles permitting adaptation, too, may emerge as universals.”

These construction principles will be profound. What theory of morphology would enable us to predict features of organisms that would evolve on any planet, in any environment? What forms of life are highly unlikely to evolve and how does selection work to achieve new families of forms? How does such a universal theory of forms fit the empirical facts of our own past and what part does random drift play in the speed at which forms evolve? These questions are of vital interest to potential development of new biological ecosystems or desired alterations of pre-existing systems.

An important hypothesis that Kauffman reaches, strikes me as a description of self-interested (myopic) individuals interacting in a free market. “In coevolution, organisms adapt under natural selection via a metadynamics where each organism myopically alters the structure of its fitness landscape and the extent to which that landscape is deformed by the adaptive moves of other organisms, such that, as if by an invisible hand, the entire ecosystem coevolves to a poised state at the edge of chaos.” He even treads the dangerous ground of social science when he examines “What is a functional whole and how does it transform when its components are altered?” He finds features in technological, economic and cultural systems that are phase transitions between finite and potentially infinite growth.

Self-transformation is related to his concept of “evolvability” and boundless expansion is related to his concept of “sustained fitness.” There are potential biophysical laws that govern what paths our future evolution can take such that we can choose our destiny and transform ourselves faster in more complex ways than we safely could have without these models. “...Proper evolutionary tuning of mutation rate, population size, and landscape structure might simultaneously optimize both evolvability and sustained fitness.”

He explores the possibility of genetic design rules. One design consideration is the amount of DNA needed to generate novel cell types. For example, if we wanted to add cell types to boost the complexity of our consciousness or to produce regenerative neurons to increase longevity, we would need a hefty increase in the amount of DNA in our chromosomes. ‘Junk’ DNA may support the complexity of cell types in a functionally whole way that will not reduce down to function codon by codon. It is, in fact, possible that much of large-scale order in genetic design “is a direct reflection of fundamental features of polymer chemistry.”

Kauffman discusses an epistemological boundary we should keep in mind when working with complex design considerations in genomic or immune regulatory systems. He points out that these systems are so fluid that they “are dancing away from us faster than we may ever be able to grasp them. ...We may never to be able to carry out the reductionistic dream of complete analysis but will want nevertheless to understand how these systems work.”

While he doesn't use the terminology of memetic evolution, he does interpret results of his models as showing how meaning and learning arise in complex organisms. Meaning does not arise in his digital, Boolean models but does arise in his random grammar models through modular interactions exhibiting their functional couplings within an evolving system. The appearance of meaning in this model is structurally similar to theories of human meaning arising from embodiment of the mind and social interactions.2 We should be able to model how meaning will change in the future with accelerating self-transformation and more complex social interactions. Learning is characterized as “a walk in synaptic weight space seeking good attractors. Learning itself may be the fundamental mechanism which converts chaotic attractors to orderly ones.” This is similar to current theories of memory formation through nitrous oxide cellular diffusion within statistical ensembles of neurons.3 The unit of selection is the individual cell but appears to be a group selection because of the mechanisms of the attractor.

Kauffman identifies two major limitations to selection, what he terms “complexity catastrophes.” In one scenario, as the complexity of the species increases, the fitness landscape it is operating in deforms to lower the overall possible fitness level. In the other scenario, as the complexity of the species increases, the population is unable to hold to the fitness peaks and falls back to a lower average fitness. Much of the discussion in Part I of the book discusses strategies to increase complexity in species while avoiding either of these catastrophes. A very promising possibility is the mapping of complex cost surfaces with the goal of optimizing energy flow to allow for increasing levels of civilization.

While he makes productive use of his NK fitness and Boolean models in many areas, he is careful to point out the inadequacy of digital models to really approximate the analog biological world. He does not hide his excitement over the potential of random grammar models to unite the natural sciences with the social sciences. He hopes to find universal classes of behavior in functionally whole systems through exploration of “grammar space.” “Thereby we may obtain models of functional couplings among biochemical, technological, or ideational elements without first requiring detailed understanding of the physics or true laws governing the couplings.”

In very strong theoretical support for boundless expansion, a sequence is traced from the open thermodynamic system on earth prior to the origin of life, to cascades of catalyzing organic molecules, to the explosion in organic diversity we see today. He says “open chemical systems can be self-extending. The fact that the biosphere as a whole is supracritical serves, I believe, as a fundamental wellspring for a persistent increase in molecular diversity.” As an aside, I could not help but reflect that recent pictures of the large-scale structure of galaxies in the universe look remarkably similar to what Kauffman calls “filagreed fog” random grammar end-states.4 He then makes a random grammar model connection between bounded physical systems such as thermodynamic constraints in chemistry and budget constraints in economics, and “the worlds of ideas, myths, scientific creations, cultural transformations, and so on” that are unbounded.

Two very interesting results from these models are of particular note. The first is that we model each other’s potential behavior (analogous to trust) in such a way that society tends towards a poised state at the edge of chaos. In essence, high degrees of trust (the most complex, discriminating models) will lead to decreased trust while low degrees of trust (the most brute, simple ‘tit-for-tat’ models) will lead to increased trust. While everyone knows that familiarity breeds contempt, it is also true that contempt breeds familiarity. The second result is that: “The extent to which the planner looks into the future governs whether the economy grows at all, slowly, or rapidly. ...Technological growth is strongly correlated with the capacity to see its implications.... If the consumer places little value on the future, diversity of goods and services remains small.” The clear implication of this is that the Extropian principles, if adopted, will by themselves be self-fulfilling. There is good reason for dynamic optimism, it works!

Spontaneous order in the absence of outside work is found throughout biology in the form of small attractors. These attractors represent cell types, immune responses, etc. and are easily attained by natural selection to produce stable structures. However, in a strongly counter-intuitive finding, as the complexity of systems increases, natural “selection cannot avoid the order exhibited by most members of the ensemble. Therefore, such order is present not because of selection but despite it.”

This finding of such an inseperable relationship between self-organization and selection that varies with the scale of the parts and the whole is typical of the holism found throughout this book. Other major examples of functional wholes include the idea of autocatalytic polymers being ‘chicken-and-egg.’ There is a lengthy outline of ‘knower-and-known’ systems where representation of and interaction between entities in their environment depends on stability of both the entities and the environment. “In a phrase, organisms have internal models of their worlds which compress information and allow action.” Also, proper growth of organisms depends on a control system of ‘map-and-interpretation.’ “...The entire genomic system is, in reality, a single coupled system whose attractors constitute both map and interpretation at once.” This holism of Kauffman’s seems akin to the ‘undivided universe’ ontological interpretation of quantum mechanics by physicist David Bohm based on experimental results of non-locality.5 I am also reminded of the position of the English philosopher Frances Bradley: “And what I repudiate is the separation of feeling from the felt, or of the desired from desire, or of what is thought from thinking, or the division of anything from anything else. For judgment is the differentiation of a complex whole, and hence always is analysis and synthesis in one.”6

Technology is being developed that will allow experimentation in areas that have the potential of transforming society. Kauffman proposes that we create life anew. He notes that function must be extremely redundant in DNA and protein sequence space, and therefore, life is created far more easily than we have previously thought. “Life is an expected, collectively self-organized property of catalytic polymers. ...Self-reproduction and homeostasis, basic features of organisms, are natural collective expressions of polymer chemistry.” The experiments must risk a complexity sufficient to achieve catalytic closure, but once accomplished, the path is open to make empirical tests on coding mechanisms to help understand why DNA coding is so prevalent today.

Kauffman is seeking patents in a field of molecular nanotechnology that he calls “Applied Molecular Evolution.” There is a finite number of enzymes that will catalyze all polymer reactions. The drive is on to explore DNA, RNA and protein sequence space to custom design for any function desired at all. The potential for individualized drug treatment is promising in immunology and cancer research. “...Using antibodies from an infected individual, it becomes possible, in principle, to find vaccines for diseases where the pathogen is not yet even known!” Such a powerful technology must be available on the private market for commercial uses and consumer benefits. The military potential for rapid development of customized, biological offensive weapons and equally rapid defenses against them is enormous. Biological warfare may not be plague versus plague, but highly accurate, targeted strikes and selected defense responses.

Finally, Kauffman uses artificial life researcher Thomas Ray’s Tierra model ecosystem to show how closely artificial extinction patterns obey the same power law that has been recorded in earth’s fossil record. The artificially-generated graph is a close match with the actual graph. I bring this up because Ray’s latest paper references Kauffman and discusses “ecological attractors” at length.7 Ray also confirms the superiority of analog models to digital models for realism and even references Hans Moravecs’ article “Pigs in Cyberspace” in Extropy #10.

Kauffman's deepest insight is a direct challenge to the current view of our lives as being merely the result of a series of frozen accidents. “I have made bold to suggest that much of the order seen in organisms is precisely the spontaneous order in the systems of which we are composed. Such order has beauty and elegance, casting an image of permanence and underlying law over biology. Evolution is not just ‘chance caught on the wing.’ It is not just a tinkering of the ad hoc, of bricolage, of contraption. It is emergent order honored and honed by selection.”

This book is a challenging read for those interested in shaping spontaneously ordered living systems towards increased complexity and meaning. The search for a “physics of biology” to help minimize tragic and time-consuming trial-and-error methods of human-directed evolution is brought to the forefront of scientific priorities by Kauffman's bold thinking.

Notes:

1. Phenotypic space, morphospace, protein space, sequence space, trait spaces, genotype space, complex fitness spaces, RNA space, catalytic task space, shape space, space of biological systems, state space, synaptic weight space, local strategy space, action space, space of symbol strings, peptide space, space of possible polymers, open state space, fixed state space, grammar space, composition space, parameter space.
2. Lakoff, G. Women, Fire, and Dangerous Things: What Categories Reveal About the Mind. Chicago: Univ. of Chicago Press, 1987. Johnson, M. The Body in the Mind: The Bodily Basis of Meaning, Imagination, and Reason. Chicago: Univ. of Chicago Press, 1987.

3. Edelman, G. Bright Air, Brilliant Fire: On the Matter of the Mind. New York: BasicBooks, 1992. Schuman, E. & Madison, D. “Locally Distributed Synaptic Potentiation in the Hippocampus.” Science 28 January 1994: 532.

4. Travis, J. “Cosmic Structures Fill Southern Sky.” Science 25 March 1994: 1684.

5. Bohm, D. & Hiley, B.J. The Undivided Universe: An Ontological Interpretation of Quantum Theory. London: Routledge, 1993.

6. Bradley, F.H. Appearance and Reality. 2nd ed., Oxford, 1897.

7. Ray, T. In press. “An evolutionary approach to synthetic biology, Zen and the art of creating life.” Artificial Life 1(1): xx--xx. MIT Press. I found this paper in the AI Expert Forum Library on Compuserve, dated 21 October 1993.

Home Page