원출처: Systems Research and Behavioral Science, March-April, 1997
by J.P. van Gigch
- The Quark and the Jaguar: Adventures in the Simple and the Complex.
- By Murray Gell-Mann (Part XIII of Design of the Modern Inquiring System(1)).
- Published by W.H. Freeman, New York, 1994,
- ISBN 0 7167 2725 0
(Note: the French translation of this review will appear in Revue Internationale de Systemique, Dunod / AFCET, Paris)
In this section, Gell-Mann's extraordinarily rich book is reviewed and commented. Through generous quotes and summaries, his theory of the behaviour of complex adaptive systems is presented in condensed form. Implications of this theory for system science and for the sciences of complexity, in particular, and for system scientists, in general, are discussed.
Complexity has been called a `new field of scientific research' (Ferris, 1995). Of course, for those of us who have worked in system science for a relatively long time, the subject is far from new. In his book Gell-Mann summarizes the research results that he and his colleagues, at the Santa Fe Institute, have reached since the Institute's relatively recent inception. These results must be placed in the perspective of previous work in the field. I am reminded of the work of Boulding (Systems Research, 1995), Simon (1962) and Klir (1985) among many others (for a broad survey of these ideas, see van Gigch, 1991, Chapter 7). There is no question that, in his book, Gell-Mann gives us a new perspective on the subject of complexity which is absent from all previous treatments of the subject. In this sense, there is much to learn from this book.
Gell-Mann's book has four parts. [:]
- Part I (approximately one-third) covers `The Simple and the Complex'.
- Part II (another one-third) covers `The Quantum Universe'.
- Parts III and IV cover diverse subjects such as evolution, adaptive systems, diversity and sustainability.
In Part I, Gell-Mann analyses the behaviour of complex adaptive systems (CAS) which manifest complexity, diversity and individuality:
The common feature of [CAS] is that each one acquires information about its environment and its own interaction with that environment, identifying regularities in that information, condensing those regularities into a kind of schema or model, and acting in the real world on the basis of that schema ...andSince a complex adaptive system separates regularities from randomness, it affords the possibility of defining complexity in terms of the length of schema used by a complex adaptive system to describe and predict the properties of an incoming data stream. (pp. 17 and 55, respectively.)
The above quotations are important because they embody two of the many `sound bites' of Gell-Mann's book. His definition of effective complexity is based on separating randomness from regularity. It is also important to note that he is defining an entity's effective complexity as observed by that which is external to it. Definitions of complexity very much depend on a description of one system by another (p. 33). The CAS is observing and constructing a schema to encompass the other entity's regularities.
Effective complexity is defined as the required minimum length of the schema which is required to encompass all of the entity's regularities. Another way of stating it is, `how long a message would be required to describe certain properties of the system in question' (p. 28). A good example of a schema is given by the some 10 million bits of a genotype's string which identify and compress its particular regularities. Gell-Mann explains that the concept of effective complexity is not sufficient as a `diagnostic of regularities'. We also need what he calls algorithmic information content (AIC), which is the length of the schema to describe and compress an entity's regularity, plus its randomness. For a given message length, AIC is at its minimum, i.e. very nearly zero, when the message exhibits the maximum number of regularities. Contrariwise, AIC is at its maximum when no regularities exist, randomness is maximum and we cannot find a schema to encompass either or both (p. 59). According to Gell-Mann, effective complexity reaches its largest value in between the two extremes of the AIC's values. Armed with the two concepts of AIC and effective complexity, the author proceeds to develop his ideas concerning CAS.
What about simplicity? In Gell-Mann's view, simplicity is not the opposite of any of the concepts of complexity defined above. Rather, sciences can be ordered in a hierarchy which features particle physics, as the most fundamental science, in terms of which all the other sciences can be expressed or derived. In the hierarchy, and beyond particle physics, we find, in this order, chemistry, biology, biochemistry and life sciences. In the same vein, systems of particle physics are called `simple' compared with living systems. The effective complexity of the latter is relatively high compared with that of the former. Because a unified theory of physics can be expressed mathematically by a compact, short equation, its effective complexity is low because the length of the schema to describe it and encompass its regularities is brief.
Gell-Mann contrasts his hierarchy where conditions of life represent high effective complexity and situations between order and disorder to what he calls the `Caltech approach' or `bottom-up approach'. He recognizes that the latter has produced spectacular technological and scientific achievements but he deplores its reductionist bias, a veiw that most of us in the systems movement will find rather dated. He states the `Caltech is making a mistake' by neglecting most of the `sciences of complexity' and praises the founding of the Santa Fe Institute for representing `the rebellion against the excesses of reductionism' (p. 119). Welcome to the club! While we greet Gell-Mann among our ranks, we cannot help reminidng him that, way back in 1947, von Bertalanffy wrote his own declaration against reductionism. It is obvious that Gell-Mann is not writing for the audience of system scientists but rather for the public at large, which unfortunately remains unconvinced.
Parts III and IV of the book will be treated together. One chapter is devoted to the process of biological evolution. Gell-Mann introduces the concept of 'fitness' which he explains as follows: `Differential rates of survival and reproduction can often be described in terms of a fitness quantity, defined so there is a general tendency for organisms with higher fitness to propagate their genes more successfully than those with lower fitness' (p. 249). He coins the idea of 'fitness landscapes' and 'basins of attractions' to give a graphical representation of movements of the genotypes towards regions of higher fitness in their biological evolution (pp. 249-266). The author also applies the concept to advantage to explain the fitness of a theoretical idea in science as `a measure of the extent to which it improves existing theory, ... by explaining new observations while maintaining or increasing the coherence and explanatory power of that theory' (p. 266).
Gell-Mann applies his views of CAS [complex adaptive system] to individuals and organizations. According to him, `an organization behaves as a complex adaptive system, with schemata and selection pressures' (p. 297). He contends that models, goals, plans, practices, procedures and rules act as schemata. Profits and survival in the competitive market place exert selection pressures on firms. The pressures filter right down to the individual managers who fight for their survival. The manager and the corporation can be considered separate CAS in their own right. On the other hand, the corporation with its managers can be considered another CAS where each problem can be viewed as a situation offering sets of choices subject to variation. A search process takes place over the sets of choices to find the one which will ensure the survival of the system. The marketplace and the economic system apply the selection pressures which result in the adoption of certain policies and actions while others are discarded. In the larger setting, certain firms will thrive while others will wither. Gell-Mann argues that individuals can become part of higher-order CAS in which the `character of the human direction can itself evolve'. Presumably, the same can be said of the `character' of the economic enterprise. When dealing at the scale of whole societies or the entire humanity, Gell-Mann advises the use of computers and simulation to be able to analyse the great quantitites of relevant information to study the 'choices offered by the branching alternative histories of the future'. `Computers acting as CAS can serve us both by learning or adapting themselves and by modelling or simulating systems in the real world that learn or adapt or evolve' (p. 305). One chapter is devoted to the subject of machines that learn or simulate learning, another to the subject of diversities under threat.
Moving on to Chapters 21 and 22, the author examines how humanity could function as `a composite, richly diverse complex adaptive system'. To accomplish this feat, Gell-Mann suggests that humanity needs to equip itself with a measure of collective foresight--some degree of understanding of the branching histories of the future ...'. Humanity needs to undergo a `highly adaptive change', an `ideological transition', `a major step towards planetary consciousness' (p. 366). The author envisages this scenario by applying his theory, of how complexity evolves, to the operation of complex ecosystems. He draws an optimistic picture of how, over long periods of time, CAS `establish themselves, operate through the cycle of variable schemata, accidental circumstances, phenotypic consequences and feedback of selection pressures to the competition among schemata' (p. 329). `Life forms contain an extraordinary amount of information.' He asks whether the destruction `in the course of a few decades of a significant fraction of the complexity that evolution has built over such a long time' make sense. He states that, as the twentieth century nears its end, `the conservation of biological diversity is one of the most important tasks facing humanity' (p. 330). The final chapter is a recapitulation of the main points in the book.
IMPLICATIONS OF THE BOOK FOR SYSTEMS SCIENCE
The New View of Complexity
Gell-Mann provides us with a new view of complexity based on the length of the schemata required to describe a system's regularities and/or randomness. One may well ask whether this new measure is useful and practical to assess the complexity of systems with which we interact on a daily basis. In other words, `Is Gell-Mann's measure "relevant" to decision makers?' or `does our knowledge of the new measure matter?' In my personal opinion, all views are important, no matter what their ultimate applicability may be. To the extent that Gell-Mann's contribution changes the focus of the discussion on the subject makes it invaluable. Except for the work of the Santa Fe Institute in the USA and that of the European Association of the Modelization of Complexity (GRASCE, 1995), including the contribution of Le Moignae (1994, 1995) (van Gigch, 1996b), the sciences of complexity have remained mostly static after Simon's and Klir's contributions. System science needs more pioneering work of the kind reviewed here.
The Diffusion of Scientific Knowledge
Will Gell-Mann's work have significant repercussions among lay people? I seriously doubt it. Without a doubt, this is an important book for scientists, but due to its erudition it will go vastly ignored by the general public. We know from Ferris (1995) and from Gell-Mann himself that writing this sort of book was not easy for him. Atomic physicists and Nobel prize winners have better things to do. On the other hand, we are indebted to Dr Gell-Mann to have persevered and given us the fruit of his own knowledge and that of his colleagues at the Santa Fe Institute. However, this book is not for everyone and its reading and interpretation require serious study. Gell-Mann's attempt to bring difficult concepts `to the masses' brings up interesting questions: How does knowledge trickle down from the halls of science to the average reader? Is this kind of knowledge important to the public at large? How do we ensure that new ideas about important subjects such as complex adaptive systems, evolution, biodiversity, sustainability, the future of complex ecosystems, etc., are filtered, digested and understood by more people? Who needs to know? What is the role of technical journals and the media in this process of dissemination and diffusion? How does this process work and how can be, the scientific community, improve it? There are no easy answers to any of these questions. Journals like Systems Research play an invaluable role in this process. However, the systems community knows only too well that, outside, its work is mostly ignored. Even Gell-Mann's book attests to this sad situation. He hardly acknowledges that any previous or current work on the subject exists. (중략)
글
댓글 없음:
댓글 쓰기