STRATEGIES FOR WINNING ACCEPTANCE OF SECOND ORDER CYBERNETICS by Stuart A. Umpleby Department of Management Science George Washington University Washington, DC 20052 USA Guest Scholar, Summer 1991 Institute for Medical Cybernetics and Artificial Intelligence University of Vienna Freyung 6, A-1010 Vienna, Austria August 20, 1991 Presented at the International Symposium on Systems Research, Informatics and Cybernetics Baden-Baden, Germany, August 12-18, 1991 ABSTRACT In the last two decades an intellectual movement associated with the term "second order cybernetics" has taken shape within the systems theory and cybernetics community. The ideas associated with this movement promise to have an important impact on the philosophy of science. However, the ideas have spread slowly. This paper reviews the strategies that have so far been employed in attempts to win acceptance for second order cybernetics. Four strategies are discussed: develop the new point of view; develop applications of the new ideas to show their usefulness; reconstruct the philosophy of science incorporating the new ideas; and contrast the philosophical origins of cybernetics with that of the dominant scientific world view. KEYWORDS Second order cybernetics; epistemology; knowledge; constructivism; scientific revolution. INTRODUCTION In the United States cybernetics as a scientific field is usually said to have started in 1943 with the publication of two papers. In that year Arturo Rosenblueth, Norbert Wiener, and Julian Bigelow published their paper "Behavior, Purpose, and Teleology." In this paper they suggested that goal-seeking and purposeful activities could be performed by suitably designed machines. In the same year Warren McCulloch and Walter Pitts published an even bolder article, "A Logical Calculus of the Ideas Immanent in Nervous Activity." In this article McCulloch and Pitts proposed a theory of how ideas arise in the brain as a result of the activity of a network of neurons. These articles initiated a research program whose purpose was to develop machines that could replicate at least some of the intellectual activities of human beings. The strategy of the research program was to study the nervous system, create a formal theory of human cognition, and then use this theory to design intelligent machines. During the 1950s and 1960s most of the work which was called cybernetics tended to focus on control systems in engineering or on applications of the concept of feedback in fields ranging from mathematics to sociology. At the 1970 meeting of the American Society for Cybernetics in Philadelphia Heinz von Foerster sought to redirect attention to the original interests which had led to the founding of the field of cybernetics. In a paper titled "Cybernetics of Cybernetics" he made a distinction between first order cybernetics, the cybernetics of observed systems, and second order cybernetics, the cybernetics of observing systems (von Foerster, 1974). Others have distinguished first and second order cybernetics in a number of ways, see Table 1. Table 1 DESCRIPTIONS OF FIRST AND SECOND ORDER CYBERNETICS First Order Second Order Author Cybernetics Cybernetics ----------------------------------------------------------------- von Foerster observed systems observing systems Pask purpose of a model purpose of the modeller Varela controlled systems autonomous systems Umpleby interaction among interaction between variables in system observer and system observed Umpleby theories of social a theory of the systems interaction between ideas and society In the years since 1974 von Foerster and his colleagues and the members of the American Society for Cybernetics have sought to establish this new, or renewed, conception of cybernetics and to work out its implications for a variety of fields. They have met with some success. Most members of the cybernetics and systems science communities have at least heard of second order cybernetics and many have been won over. However, in the larger scientific community there has been little impact so far. This situation is in marked contrast with the rapid and widespread acceptance of other scientific ideas, for example Thomas Kuhn's conception of scientific revolutions or more recently chaos theory. What have the second order cyberneticians done to present their case? What resistance have they encountered? And what steps might be taken to dissolve the resistance and to win the acceptance of the scientific community? STRATEGIES TO DATE This article briefly reviews the strategies employed to date. For a more complete discussion of the arguments in behalf of second order cybernetics, the reader is referred to the references listed at the end of the paper. 1. Develop the new point of view The primary strategy for advancing second order cybernetics has been to develop and refine the basic idea. This work has been done primarily by von Foerster (1981), von Glasersfeld (1987), and Maturana and Varela (1987). Very briefly, cyberneticians studied the nervous system in order to understand human cognition. They came to the conclusion that observations independent of the observer are not physically possible. To illustrate, as I write in English, I am using a structure in my nervous system which is the result of language interactions which I have had since I was a child. Furthermore, what each of us chooses to focus our attention on is a function of our cultural upbringing, our scientific backgrounds, and our personal interests, all of which are embedded in the current structure of our nervous systems. But recall that the classical conception of science was that scientific theories are objective, that is, the characteristics of the observer should not enter into the observation. Hence, by studying the nervous system, cyberneticians came to reject a central tenet of the philosophy of science. Because observations independent of the characteristics of the observer are not possible, one is forced to the conclusion that all knowledge is subjective. "Objectivity" then becomes a matter of shared subjectivity. By the way, it is not the case, as some have suggested, that widespread adoption of this view would lead to chaos. Rather, adoption of this view would mean that people would seek confirmation of their views by others. One can claim, in fact, that this is what already occurs. 2. Develop applications of the new ideas to show their usefulness One strategy for advancing second order cybernetics has been to develop applications of the new ideas in order to demonstrate their usefulness. This strategy has been particularly important in the United States where ideas are evaluated almost solely on the basis of their utility. In psychology the field of family therapy traces its origins directly to cybernetics (Watzlawick, 1967). In family therapy the assumption is that bizarre behavior can be the result of interactions within the family rather than, as in psychoanalysis, the history of the individual and traumatic incidents in childhood. For artificial intelligence cybernetics offers an alternative philosophical tradition and suggests that language is more usefully thought of as a means of preserving and enhancing the autonomy of the individual within a social setting rather than as strings of symbols which represent objects and processes (Winograd and Flores, 1986). In management the contribution of cybernetics has been to focus attention on communication and control processes within firms (Beer, 1972). Recently the emphasis has been on developing methods for interactive planning among groups of people within organizations (Umpleby, 1986). For epistemology cyberneticians have proposed that cognition is a biological phenomenon and that speculations about the nature of knowledge should have a biological foundation (Segal, 1986). In education cyberneticians have joined with those who are advocating methods based on a constructivist approach to learning. For example, rather than simply teaching physics as is usually done, the idea behind the new approach is to identify the beliefs in the student's existing, naive view of physics and to help the student learn how formal physics is superior. For the field of ethics cyberneticians have made the radical suggestion that if people construct their own "realities," then people are responsible not only for their actions but also for the "world" which they perceive. Just recently a connection between second order cybernetics and economics has been made with the suggestion that social systems can be thought of as being composed of biased, thinking participants rather than assuming that people are perfectly-informed, rational actors (Soros, 1987). This work has also led to a connection to political science through a discussion of the gap between description and experience (Soros, 1990). This theoretical perspective offers the best explanation I have seen of why the change in thinking in the Soviet Union and Eastern Europe occurred so rapidly. 3. Reconstruct the philosophy of science incorporating the new ideas Another strategy for winning acceptance for second order cybernetics has been to look at the history of science and seek to understand how previous scientific ideas came to be widely accepted. Using Thomas Kuhn's (1970) idea that science advances through a series of periods of normal science separated by scientific revolutions, the claim is that second order cybernetics constitutes a scientific revolution. The transition from a period of normal science to a scientific revolution is marked, according to Kuhn, by incommensurable definitions. However, Kuhn has less to say about how a scientific revolution turns into a new period of normal science. Kuhn implies that the new paradigm is simply accepted (see Figure 1). A clue about how a scientific revolution changes into a new period of normal science is provided by those scholars, usually philosophers of science rather than historians of science, who have been critical of Kuhn. They suggest that major changes in science can be described using the correspondence principle -- any new theory should reduce to the old theory, to which it corresponds, for those cases in which the old theory is known to hold (Krajewski, 1975). That is, science progresses by adding a new dimension which had previously been neglected. The efforts of second order cyberneticians to distinguish their ideas from previous ideas by making tables comparing two or more points of view is evidence that incommensurable definitions are involved and that a scientific revolution is in the making. But since second order cyberneticians have been making these lists for about fifteen years, I think the time has come to shift to the second part of the loop, that is, to enter a new period of normal science by reconstructing science using the correspondence principle. In this effort the crucial point is to define a new dimension which scientists had previously neglected. Based on the literature in cybernetics, an obvious candidate for the new dimension is "amount of attention paid to the observer." NORMAL SCIENCE | /\ | | incommensurable | | the correspondence definitions | | principle | | \/ | SCIENTIFIC REVOLUTION Figure 1. The Cybernetics of Science But if the new dimension is "amount of attention paid to the observer," what aspects of the observer are we interested in? I suggest that there are at least three aspects of the observer of interest to scientists. Each of these aspects entails self-reference. a. The second order cyberneticians, who are interested in biology and epistemology, are concerned with the nature of knowledge, language, cognition, and communication. b. Scholars such as Thomas Kuhn, who are interested in the sociology of the scientific community, point out that a scientist is a participant in the scientific community. Hence, a scientist's judgment is influenced by his or her role in that community. c. Scholars who are interested in political and economic systems are aware that the behavior of these systems changes as a result of people acting on new knowledge. In the social sciences, unlike the physical sciences, theories change the way that systems operate. When physicists changed their thinking from classical, Newtonian mechanics to modern, quantum mechanics, the behavior of atoms did not change. But theories developed by Adam Smith, Karl Marx, John Maynard Keynes, Milton Friedman and others have changed the way that economic systems operate. In my opinion if we are to say that the new dimension is "amount of attention paid to the observer," then we need to distinguish at least two instances of this attention -- attention paid to the observer as a biological entity, in order to understand the nature of knowledge, and attention paid to the observer as a participant in a social system. 4. Contrast the philosophical origins of cybernetics with that of the dominant scientific world view A further strategy for attempting to explain the significance of cybernetics for the scientific community is to note how the development of cybernetics has followed a different direction from that of the philosophy of science in general. "The received view on theories" is to a large extent the product of philosophers who were connected with the Vienna Circle in the early days of this century. Austrians also are influential members of the group of second order cyberneticians. Von Foerster, von Glasersfeld, and Watzlawick are all Austrian in origin. Perhaps the second order cyberneticians have chosen a different path away from an earlier German/Austrian point of view. The two axes in Figure 2 are similar to the two dimensions discussed in the context of the correspondence principle. That is, paying attention to the observer as a biological entity is similar to the change from a representation conception of knowledge to a coherence conception of knowledge. Likewise, paying attention to the observer as a social participant is similar to the distinction between the construction of knowledge as an individual activity and the construction of knowledge as a social activity. a coherence conception of knowledge ------------------------------------------ |3 | 4| | second order | social construction | knowledge | cybernetics | of reality | knowledge as an | | | as a individual ------------------------------------------ group activity | | | activity | German | Vienna | | Idealism | Circle | |1 | 2| ------------------------------------------ a representation conception of knowledge Figure 2. Two Paths to a Similar Outcome German idealism implied that through knowledge it is possible to impose order on experience. The German word for education or instruction, Ausbildung, means to "build out" into the world from an origin within the individual. By analogy the purpose of government is to impose order on a chaotic society. In contrast British empiricism and American pragmatism assume that the world has its own identity. The purpose of exploration is to discover the nature of the patterns of organization in the world. By analogy the purpose of government is to make experiments (or reforms) to see how well they work in a world which has a "separate reality." The Vienna Circle moved German philosophy away from idealism and toward a conception of science as a group effort to formulate and test falsifiable theories. Cybernetics has chosen to retain the idea of the construction of a reality by an individual but to reject the assumption that such constructions are universally valid for everyone. Having made these choices, cyberneticians (in quadrant 3 in Figure 2) find themselves diametrically opposed to the received view on theories (quadrant 2). One way to resolve the disagreement between cyberneticians and the scientific community at large might be to have the two groups converge on the view that science is a group activity and that knowledge need not "match" an external world but rather need only be coherent with experience. Recall the earlier analysis where the suggestion was made that the way to move from a period of revolution to a new period of normal science is to reconstruct science in terms of a new dimension, or in this case two dimensions. Thomas Kuhn's sociological view of science has persuaded many American social scientists to think of science as an activity in which acceptance of a theory is a matter of persuading a group of peers rather than simply receiving a favorable verdict from nature. The results of experiments are, after all, subject to interpretation. Cyberneticians also seem to be moving in this direction. Recent work on constructivist epistemologies has emphasized the social character of language and has suggested that human beings need to construct images of others in order to confirm their perceptions. Constructivism is in part a way of incorporating trial and error, what "fits," into the idealist notion that "reality" is created in the mind of the observer. The analysis shown in Figure 2 suggests at least two consequences. a. Whereas some of the arguments in behalf of second order cybernetics are based on studies of the nervous system (see particularly the work of Maturana) other authors (for example von Glasersfeld) emphasize the philosophical tradition of idealism. Figure 2 suggests that it is easier to get to second order cybernetics (quadrant 3) if one is familiar with both idealism (quadrant 1) and empiricism (quadrant 2) than if one is only familiar with empiricism. b. If the current debate is resolved by convergence on the social construction of knowledge (quadrant 4), it is not necessary to go through second order cybernetics (quadrant 3) to get there. Most scientists today are moving from quadrant 2 to quadrant 4. This is a shorter intellectual journey than the trip from 2 to 1 to 3 to 4. However, the shorter journey does not challenge so fundamentally the previous conception of knowledge. Continued development of these and other arguments in behalf of second order cybernetics will probably eventually bring the desired results. REFERENCES Beer, Stafford (1972); The Brain of the Firm; McGraw-Hill Krajewski, Wladyslaw (1975); Correspondence Principle and Growth of Science; Reidel Kuhn, Thomas S. (1970); The Structure of Scientific Revolutions; Second Edition, University of Chicago Press Maturana, Humberto and Francisco Varela (1987); The Tree of Knowledge; New Science Library McCulloch, Warren and Walter Pitts (1943); A Logical Calculus of the Ideas Immanent in Nervous Activity; Bulletin of Mathematical Biophysics, Vol.5 (pp. 115-133). Reprinted in Embodiments of Mind; MIT Press Rosenblueth, Arturo, Norbert Wiener, and Julian Bigelow (1943); Behavior, Purpose, and Teleology; Philosophy of Science, Vol. 10 (pp. 18-24). Reprinted in Modern Systems Theory for the Behavioral Scientist (ed. W. Buckley), Aldine, 1968 Segal, Lynn (1986); The Dream of Reality: Heinz Von Foerster's Constructivism; Norton Soros, George (1987); The Alchemy of Finance; Simon and Schuster Soros, George (1990); Opening the Soviet System; Weidenfeld and Nicolson Umpleby, Stuart A. (1986); Methods for Making Social Organizations Adaptive; Power, Autonomy, Utopia: New Approaches toward Complex Systems (ed. R. Trappl), Plenum Publishing (pp. 133-138) Umpleby, Stuart A. (1990); The Science of Cybernetics and the Cybernetics of Science; Cybernetics and Systems, Vol. 21, No. 1 (pp. 109-121) Von Foerster, Heinz (1974); Cybernetics of Cybernetics; Communication and Control in Society (ed. K. Krippendorff), Gordon and Breach Von Foerster, Heinz (1981); Observing Systems; Intersystems Von Glasersfeld, Ernst (1987); The Construction of Knowledge; Intersystems Watzlawick, Paul, Janet Beavin and Don Jackson (1967); Pragmatics of Human Communication; Norton Winograd, Terry and Fernando Flores (1986); Understanding Computers and Cognition; Ablex Publishing ------------------------------------------------------------------------ S.A. Umpleby, Dept. of Mgt. Science, GWU, Wash. DC 20052 USA tel: 202/994-7530, fax: 202/994-4930, e-mail: umpleby@gwis2.circ.gwu.edu ------------------------------------------------------------------------