Facets of Systems Science

An inevitable prerequisite for this book, as implied by its title, is a presupposition that systems science is a legitimate field of scientific inquiry. It is self-evident that I, as the author of this book, consider this presupposition valid. Otherwise, clearly, I would not conceive of writing the book in the first place.

The common-sense definition, as expressed by Eq. (1.1), looks overly simple:

Systems science is a phenomenon of the second half of this century. It developed within a movement that is usually referred to as systems movement. In general, systems movement may be characterized as a loose association of people from different disciplines of science, engineering, philosophy, and other areas, who share a common interest in ideas (concepts, principles, methods, etc.) that are applicable to all systems and that, consequently, transcend the boundaries between traditional disciplines.

To characterize the domain of systems science more specifically requires a conceptual framework within which systems are characterized. Each framework determines a scope of systemhood properties that can be described within it and leads to some specific systemhood-based taxonomy of systems. To capture the full scope of systemhood phenomena we are currently able to envision, a comprehensive framework is needed.

Systems methodology is understood in this book as a family of coherent collections of methods for dealing with the various systems problems that emanate from the conceptual framework employed.* Thus, for example, one systems methodology is based upon the GSPS framework outlined in Chap. 4. Systems methodologies based upon different but equally general conceptual frameworks are capable of covering, by and large, the same class of problems. With some adjustment, methods developed under one framework can usually be converted into methods for dealing with comparable systems problems under another framework.

As argued previously, the principal aim of systems science is to understand the phenomenon of systemhood as completely as possible. The first step in achieving this aim is to divide the whole spectrum of conceivable systems into significant categories. The second step is to study the individual categories of systems and their relationship, and to organize the categories into a coherent whole. The third step is to study systems problems that emerge from the underlying set of organized systems categories Finally, we address methodological issues regarding the various types of systems problems.

In every traditional discipline of science, we develop systems models of various phenomena of the real world. Each of these models, when properly validated, represents some specific knowledge regarding the relevant domain of inquiry. In systems science, the domain of inquiry consists of knowledge structures themselves—the various categories of systems that emerge from the conceptual framework employed. That is, the objects of investigation in systems science are not objects of reality, but systems of certain specific types.

Complexity is perhaps as important a concept for systems science as the concept of a system. It is a difficult concept, primarily because it has many possible meanings. While various specific meanings of complexity have been proposed and discussed on many occasions, there is virtually no sufficiently comprehensive study that attempts to capture its general characteristics. The reason for this situation is well expressed by John Casti [1986]:

The notion of system complexity is much like St. Augustine’s description of time: “What then is time [complexity]? If no one asks me, I know; if I wish to explain it to one that asks, I know not.” There seems to be fairly well-developed intuitive ideas about what constitutes a complex system, but attempts to axiomatize and formalize this sense of the complex all leave a vague, uneasy feeling of basic incompleteness, and a sense of failure to grasp important aspects of the essential nature of the problem.

In some contexts, complexity is a desirable property, i.e., we search, within given constraints, for systems with a high degree of complexity. Cryptography and the design of random number generators are two typical examples of such contexts. In some situations, a certain degree of complexity is a necessary condition for obtaining some specific systems properties, usually referred to as emergent properties. Self-reproduction, learning, and evolution are examples of such properties.

Literature dealing with various issues that emanate from recognized categories of goal-oriented systems is voluminous and growing rapidly. The subject of goal-orientation does not always appear in the literature under this general and neutral term. More frequently, it is discussed under other names, which designate special types of goal orientation. Typical examples are: regulation, control, self-organization, learning, autopoiesis, self-reproduction, self-correction, adaptation, evolution. No attempt is made in this chapter to cover this broad subject comprehensively since each of the special types of goal-orientation alone could easily occupy a whole book. Instead, the focus here is on a few key concepts and issues pertaining to goal-oriented systems.

No historical reflection upon systems science and its impact on other areas of human endeavor can be definitive at this time since systems science is currently still in the process of forming. It is by far not established as yet to a degree comparable with traditional disciplines of science such as, e.g., physics, chemistry, psychology, or economics. One of the difficulties in examining systems science in this formative stage is the lack of unified terminology. Thus, the very notion of systems science, as conceived in this book, is often discussed in the literature under the names systems research or systems theory, sometimes with the adjective general.

For a long time, people have been trying to characterize or define the notion of system. After all, “systems” are supposed to be what System Theory is about. The results so far have been contradictory and unsatisfactory. This confusion at the foundations has led many to conclude that there is no such thing as a “system” and hence to deny that System Theory is about anything.1,3 Even those most sympathetic to the notion have difficulties at this level. The very founders of System Theory did not try to say what a system was; and as for System Theory, they characterized it only obliquely, by saying it comprised all studies of interest to more than one discipline.5 They thereby begged the entire question.

It has increasingly been recognized that a number of countries, primarily the United States and other countries in the West, are at some unique historical crossroad of great significance. This crossroad is usually described as a transition from an industrial into postindustrial phase of society. It is compared in its significance with the previous major societal transition—the change from the pre-industrial into industrial society. Although the transition into the industrial society occurred in most Western countries in the nineteenth and early twentieth centuries, most of the world today is still characterized by the pre-industrial society.

When the Neapolitan philosopher Giambattista Vico published his treatise on the construction of knowledge,† it triggered quite a controversy in the Giornale de’Letterati d’Italia, one of the most prestigious scholarly journals at the time. This was in the years 1710–1712. The first reviewer, who remained anonymous, had carefully read the treatise and was obviously shocked by the implications it had for traditional epistemology—all the more so because, as he conceded, the arguments showed great learning and were presented with elegance. He was therefore impelled to question Vico’s position, and he very politely suggested that one thing was lacking in the treatise: the proof that what it asserted was true.

General Systems Theory is a name which has come into use to describe a level of theoretical model-building which lies somewhere between the highly generalized constructions of pure mathematics and the specific theories of the specialized disciplines. Mathematics attempts to organize highly general relationships into a coherent system, a system however which does not have any necessary connections with the “real” world around us. It studies all thinkable relationships abstracted from any concrete situation or body of empirical knowledge. It is not even confined to “quantitative” relationships narrowly defined—indeed, the developments of a mathematics of quality and structure is already on the way, even though it is not as far advanced as the “classical” mathematics of quantity and number. Nevertheless because in a sense mathematics contains all theories it contains none; it is the language of theory, but it does not give us the content. At the other extreme we have the separate disciplines and sciences, with their separate bodies of theory. Each discipline corresponds to a certain segment of the empirical world, and each develops theories which have particular applicability to its own empirical segment. Physics, Chemistry, Biology, Psychology, Sociology, Economics and so on all carve out for themselves certain elements of the experience of man and develop theories and patterns of activity (research) which yield satisfaction in understanding, and which are appropriate to their special segments.

The emergence of general system theory is symptomatic of a new movement that has been developing in science during the past decade: Science is at last giving serious attention to systems that are intrinsically complex. This statement may seem somewhat surprising. Are not chemical molecules complex? Is not the living organism complex? And has not science studied them from its earliest days? Let me explain what I mean.

We live in a largely artificial world, one made by man as a result of the most powerful activity man has discovered: the activity of science. The intellectual and practical adventure which began in Western Europe in the 16th and 17th Centuries with Copernicus, Kepler, Galileo, Newton, has made our world. Only 100 years ago there was little doubt that the application of science, leading to the creation of wealth and the elimination of much disease had shown the way to a happier future. Today, noting the manifest inability of the most scientifically advanced countries to solve the problems of the real world (as opposed to the self-defined, artificial problems of the laboratory) we wonder whether the fragmentation of science into its many separate disciplines is not a significant weakness.

It is a great honor for me to have been invited to present this year’s Ludwig von Bertalanffy Memorial Lecture. I was privileged to have known Ludwig von Bertalanffy personally in the last few years of his life, when he came to the State University of New York at Buffalo in the late 1960s. Although his appointment was in the Faculty of Social Sciences, his offices were located in the Center for Theoretical Biology, of which I was then Assistant Director. I cannot imagine a more fitting arrangement; von Bertalanffy’s roots were solidly anchored in the theory of biological systems, of which he initiated some of the deepest pioneering studies; from these roots grew unique insights into the character of social and behavioral systems, as well as the creation of a General Systems Theory which bound all these activities together into an integrated intellectual edifice of unique amplitude and power.

It may be useful to describe some of the peculiarities of cybernetic explanation.

The world does not present itself to us neatly divided into systems, subsystems, environments, and so on. These are divisions which we make ourselves, for various purposes, often subsumed under the general purpose evoked by saying “for convenience.” It is evident that different people find it convenient to divide the world in different ways, and even one person will be interested in different systems at different times, for example, now a cell, with the rest of the world its environment, and later the postal system, or the economic system, or the atmospheric system.

We all know, I think, that system theory is a revolutionary development in scientific thought. The blossoming of system theory over the past three or four decades betokens a massive paradigm shift, on a scale which has not been seen since the publication of Newton’s Principia. However, whereas previous paradigm shifts have resulted in the relinquishing of one paradigm in favor of another, ultimately equally restrictive one, system theory offers us not one alternate paradigm, but many; it has shown us that there exists many primary modes of system analysis, each one independent of the others, and thus that there exists no single over-arching analytic procedure which encompasses all others. The mode we choose in dealing with a particular problem must be determined by the nature of the question we are asking, and must be guided by intelligence and insight. These last are qualities which have been excluded from science for too long by slavish adherence to a single mode of system analysis; system theory has re-introduced them, and indeed established them as the first prerequisite in the generation of scientific theory.

The past two decades have witnessed an evolution of classical circuit theory into a field of science whose domain of application far transcends the analysis and synthesis of RLC networks. The invention of the transistor, followed by the development of a variety of other solid-state devices, the trend toward microminiaturization and integrated electronics, the problems arising out of the analysis and design of large-scale communication networks, the increasingly important role played by time-varying, nonlinear and probabilistic circuits, the development of theories of neuroelectric networks, automata and finite state machines, the progress in our understanding of the processes of learning and adaptation, the advent of information theory, game theory and dynamic programming, and the formulation of the maximum principle by Pontryagin, have all combined to relegate classical circuit theory to the status of a specialized branch of a much broader scientific discipline—system theory—which, as the name implies, is concerned with all types of systems and not just electrical networks.

I believe we are leaving one cultural and technological age and entering another; that we are in the early stages of a change in our conception of the world, a change in our way of thinking about it, and a change in the technology with which we try to make it serve our purposes. These changes, I believe, are as fundamental and pervasive as were those associated with the Renaissance, the Age of the Machine that it introduced, and the Industrial Revolution that was its principal product. The socio-technical revolution we have entered may well come to be known as the Resurrection.

In retrospect, it seems that the general shape of my systems profile was determined early in my professional career and under rather special circumstances. I refer here to the period 1952–1964: the 1950s were the years of my university studies (undergraduate studies in electrical engineering and graduate studies in computer science) and the early 1960s were the years during which I was primarily involved in industrial computer research. The place was Prague, the capital of Czechoslovakia.

... in the three decades between now and the twenty-first century, millions of ordinary, psychologically normal people will face an abrupt collision with the future. (Ref. 58, p. 18)

Man not only exists but knows that he exists. In full awareness he studied his world and changes it to suit his purposes. He has learned how to interfere with ‘natural causation,’ insofar as this is merely the unconscious repetition of immutable similars. He is not merely cognizable as extant, but himself freely decides what shall exist. Man is mind, and the situation of man as man is a mental situation. (Ref. 31, p. 11)

The quantification of nature, which led to its explication in terms of mathematical structures, separated reality from all inherent ends and, consequently, separated the true from the good, science from ethics. (Ref. 38, p. 122)

In April 1982 a conference concerned with Model Realism took place at Bad Honnef in Germany. It was remarkable because the organizer, Horst Wedde, attempted to force comparability between the different approaches to modeling proposed by asking all participants to illustrate their methodologies applied to one of three well-defined case histories. The papers and commentaries given at the conference are available in the book

Adequate Modeling of Systems

.

61

This paper is based on an evening address given by the author in a wine cellar as part of the lighter side of the conference. It attempts to put our endeavours to create increasingly

real

global models in the wider perspective of man’s search for meaning, illustrating the general points made by quotations from the book

Groping in the Dark

40

which is based on comparisons of seven global models at the Sixth IIASA Symposium on Global Modelling.

M. E. Van Valkenburg writes in the foreword to Steiglitz’s Introduction to Discrete Systems,1 that “Given the widespread availability of computers, there seems little doubt that the teaching of electrical engineering should undergo an evolution ... In the emerging pedagogical approach, equations should be written in discrete form as difference equations, instead of in continuous form as differential equations. Indeed, equations should seldom be used, since principles should be stated directly in algorithmic form.” Approaches having this character are increasingly being proposed not only in electrical engineering but in other fields where continuous mathematical methods have been traditional, and not only are pedagogical changes being suggested. Digital computing and discrete models are influencing our conception of real world systems and the role classical mathematical methods are to play in modelling them.

It is a great pleasure and a privilege for me to deliver this lecture and pay thus a tribute to W. Ross Ashby, a brilliant scholar who contributed so much to systems research and cybernetics.

Recent work on the fundamental processes of regulation in biology (Ashby, 1956) has shown the importance of a certain quantitative relation called the law of requisite variety. After this relation had been found, we appreciated that it was related to a theorem in a world far removed from the biological—that of Shannon on the quantity of noise or error that could be removed through a correction-channel (Shannon and Weaver, 1949; theorem 10). In this paper I propose to show the relationship between the two theorems, and to indicate something of their implications for regulation, in the cybernetic sense, when the system to be regulated is extremely complex.

Information theory was created for the purpose of studying the communication of messages from one point to another, and since its appearance,14 its focus has remained on the question, “how can the constraint between the two variables X (message sent) and Y (message received) be measured and maximized”? Although the theory was generalized to N dimensions,10,2 and its relation to the analysis of variance noted,9 not much use seems to have been made of the result, perhaps in part because the descriptors “N-dimensional Information Theory” or “Uncertainty Analysis” did not adequately represent what can actually be seen as the analysis of constraints in multivariable systems. In any statistically-analyzable system of several variables interacting in a lively way, some variables (or sets of them) exert effects on others. These effects are reflected statistically as non-independence of the variables involved, and it is this deviation from independence which we indicate by the term “constraint.” We prefer this term to the term “dependence” because the latter suggests dependence of XonY while the former is neutral as to direction. To the extent that the variables are not independent, they are “in communication” with one another, and information theory can be used to analyze the non-independence. In addition, the fluctuation of values taken by any variable can be viewed as a message it sends, a flow of information about itself to all other parts of the system which are “listening.” The view of systems as networks of information transfer leads to quantitative conclusions about system behavior and structure which are somewhat novel and of wide applicability.

Science has led to a multitude of results that affect men’s lives. Some of these results are embodied in mere conveniences of a relatively trivial sort. Many of them, based on science and developed through technology, are essential to the machinery of modern life. Many other results, especially those associated with the biological and medical sciences, are of unquestioned benefit and comfort. Certain aspects of science have profoundly influenced men’s ideas and even their ideals. Still other aspects of science are thoroughly awesome.

A number of proposals have been advanced in recent years for the development of “general systems theory” which, abstracting from properties peculiar to physical, biological, or social systems, would be applicable to all of them. We might well feel that, while the goal is laudable, systems of such diverse kinds could hardly be expected to have any nontrivial properties in common. Metaphor and analogy can be helpful, or they can be misleading. All depends on whether the similarities the metaphor captures are significant or superficial.

There is an enormous literature on the complexity of systems, and with attempts at specifying intrinsic measures of complexity. In this note, we will take an opposite view; that complexity is not an intrinsic property of a system, but rather manifests our capabilities to interact with the system. Since our capabilities to interact with systems around us are continually changing, so too do their apparent complexities. Some consequences of this viewpoint, bearing on system descriptions, and on the capacity of systems to make errors, will be dealt with below; a fuller discussion of these matters will be presented in another place.

I would like in this presentation to explain why the study of complexity has become so interesting today. The idea of studying complex systems is, of course, by no means new. Some 2,000 years ago, Aristotle had already studied domains as varied as marine biology and the political organization of towns. But, it is our present situation, rather, which is, in a sense, unique. When one thinks of complex phenomena, one immediately thinks of biology, society, economics, or areas of this kind, and when one thinks about simple phenomena, the repeatable experiments of physics and chemistry and the domain of planetary motion are what spring naturally to mind. The remarkable feature of our time is that the gap between these two sets of phenomena has narrowed dramatically.

I was privileged to have known W. Ross Ashby personally, albeit briefly. We had the opportunity to interact rather intensively over a six-week period in the mid-1960s, when we both participated in a Summer Colloquium on Theoretical Biology, sponsored by NASA, which then had an interest in such things. I have very vivid memories of those days, and of Ashby himself, and accordingly I am most honored to be invited to present this Ashby Memorial Lecture.

In order to understand the successes of science, we can do no better than to examine physics—and particularly mechanics—for these sciences are often taken to be ideal models. The beauty of the mechanical model of the world was well expressed by Deutsch,

1

who said that mechanism

... implied the notion of a whole which was completely equal to the sum of its parts; which could be run in reverse; and which would behave in exactly identical fashion no matter how often these parts were disassembled and put together again, and irrespective of the sequence in which the disassembling or reassembling would take place. It implied consequently that the parts were never significantly modified by each other, nor by their own past, and that each part once placed in its appropriate position with its appropriate momentum, would stay exactly there and continue to fulfill its completely and uniquely determined function.

What is systems theory? Today the answer is becoming clear, and I would like to describe how the subject seems to be taking form and developing logical coherence.

Today, as a step towards the control of complex dynamic systems, models are being used ubiquitously. Being modelled, for instance, are the air traffic flows around New York, the endocrine balances of the pregnant sheep, and the flows of money among the banking centres.

Questions of principle are sometimes regarded as too unpractical to be important, but I suggest that that is certainly not the case in our subject. The range of phenomena that we have to deal with is so broad that, were it to be dealt with wholly at the technological or practical level, we would be defeated by the sheer quantity and complexity of it. The total range can be handled only piecemeal; among the pieces are those homomorphisms of the complex whole that we call “abstract theory” or “general principles.” They alone give the bird’s-eye view that enables us to move about in this vast field without losing our bearings. I propose, then, to attempt such a bird’s-eye survey.

I have come to believe that an understanding of anticipatory systems is crucial, not only for biology, but also for any sphere in which decision making based on planning is involved. These are systems which contain predictive models of themselves and their environment, and employ these models to control their present activities.

Notwithstanding their diversity, all living systems must share a common organization which we implicitly recognize calling them “living.” At present there is no formulation of this organization, mainly because the great developments of molecular, genetic and evolutionary notions in contemporary biology have led to the overemphasis of isolated components, e.g., to consider reproduction as a necessary feature of the living organization and, hence, not to ask about the organization which makes a living system a whole, autonomous unity that is alive regardless of whether it reproduces or not. As a result, processes that are history dependent (evolution, ontogenesis) and history independent (individual organization) have been confused in the attempt to provide a single mechanistic explanation for phenomena which, although related, are fundamentally distinct.

High among the interesting phenomena of organization shown by life is that of reproduction. We are naturally led to ask. How can a system reproduce itself? And we go headlong into a semantic trap unless we proceed cautiously. In fact, the answer to the question, “How does the living organism reproduce itself?” is “It doesn’t.”

The ability to construct measuring devices and to predict the results of measurements using models expressed in formal mathematical language is now generally accepted as the minimum requirement for any form of scientific theory. The modern cultural development of these skills is usually credited to the Newtonian epoch, although traces go back at least 2000 years to the Milesian philosophers. In any case, from the enormously broader evolutionary perspective, covering well over three billion years, the inventions of measurement and language are commonly regarded as only the most recent and elaborate form of intelligent activity of the most recent and elaborate species.

The great majority of philosophers of science have ignored general systems theories (henceforth GSTs). And those few who have taken notice of GSTs have too often drawn on popularizations and on careless philosophical formulations, and as a result have come to the conclusion that GSTs constitute a new version of the old holistic metaphysics and the old antianalytic epistemology associated with that metaphysics.

Biology begins with the recognition of what we call living organisms as a separate class of entities, distinguished in structure and properties from the rest of the natural world. The intuitions on which this recognition is based are a mixture of introspections and experience, which despite great effort have never been completely formalized; that is, no one has ever been able to put forward a finite set of structural propositions that are satisfied by exactly those physical systems which our intuition tells us are organisms. Nevertheless, most of us take our intuitions on this matter seriously enough to believe that we can make a useful, scientifically significant distinction between living and nonliving, organic and inorganic. The absence of formalization means, however, that we cannot sharply specify the boundaries which separate the living from the nonliving. We encounter such boundaries when we ask, as some people do, whether viruses are alive, or whether it is possible to construct machines which can “live” in some sense, or whether there are other kinds of physicochemical systems (e.g., on the planet Jupiter) which we would want to classify as “living systems.”

In my own recollections the Society for General Systems Research, as it later came to be called, originated in a conversation around the lunch table at the Center for Advanced Study in the Behavioral Sciences in Palo Alto, California, in the fall of 1954. The four men sitting around the table who became the founding fathers of the Society were Ludwig von Bertalanffy, Anatol Rapoport, Ralph Gerard, and myself—a biologist, an applied mathematician and philosopher, a physiologist, and an economist. Economics, therefore, can certainly claim to have been in at the beginning of that enterprise, although this may have been largely an accident of my own personal interests. Certainly one cannot claim that the interaction between general systems and economics has been very extensive since that date, though the contributions of each to the other may be more than many people recognize. In the intervening years, however, the social sciences in general systems have been represented more by sociologists, such as Buckley,1 and psychologists, such as the late Kenneth Berrien.2 Almost the only other economist I can think of who has played much of a role in the development of general systems is Alfred Kuhn,3 whose interest, like my own, has been primarily in going beyond economics to developing an integrated social science.

Talcott Parsons and I first met in 1939 when we attended clinical conferences presided over by Dr. Stanley Cobb at the Department of Psychiatry of the Massachusetts General Hospital. We remained friends throughout our lives, meeting fairly often both professionally and socially. We liked each other and always maintained a cordial relationship. As he reminded me from time to time, in our younger years we were both greatly influenced by the ideas of Walter B. Cannon, who was my Professor of Physiology in medical school, and Lawrence J. Henderson, who had a direct personal influence on both of us. Henderson’s espousal of Pareto, both in the courses he taught and in his conversation, had a particularly vigorous impact.