In reviewing the original introduction for Systems Thinking: Selected Readings in the 1969 Penguin paperback, there’s a few threads that I only recognize, many years later.
The tables of contents (disambiguating various editions) were previously listed as 1969, 1981 Emery, System Thinking: Selected Readings.
— begin paste —
In the selection of papers for this volume, two problems have arisen, namely what constitutes ‘systems thinking’, and what systems thinking is relevant to the thinking required for organizational management? The first problem is obviously critical. Unless there were a meaningful answer, there would be no sense in producing a volume of readings in systems thinking in any subject. A great many writers have manifestly believed that there is a way of considering phenomena which is sufficiently different from the well-established modes of scientific analysis to deserve the particular title of systems thinking. Reasons for believing that the distinction is of value are spelt out in the first selections.
There have in fact been two arguments for a systems approach to the analysis of living phenomena.
-  Throughout the volume we have kept to the strand of thought that runs from theorizing about biological systems in general to social systems. We have practically ignored the strand that arises from the design of complex engineering systems. Through such movements as operations research and cost-benefit analysis this influence is being strongly felt by management but its methods and language are so different as to require separate treatment.
First has been the argument that only such an approach will reveal the ‘Gestalten’ properties that characterize the higher levels of organization which we call ‘living systems’. It will be noticed that we are parenthesizing the key terms. This pretty well indicates the uncertain state of knowledge in this field.
Second has been the argument that many of these Gestalten properties are common to the different levels of organization of living matter (from bacteria to human societies) and hence provide a valid and powerful form of generalization. [p. 7, editorial paragraphing added]
There is at least one further line of argument although it has had little apparent attraction to the main contributors to the systems approach. This is that a systems analysis of living organizations is likely to reveal the ‘general in the particular’. Analysis of part systems in cause-effect terms, for example, of liver disorders, death rates, recruitment, training, or productive efficiency, builds up a certain kind of knowledge. However, the total systems of which they are a part usually offer alternative paths which will minimally meet organizational requirements and/or provide substitute feedback control systems, Analysis of the total system is likely to reveal those properties, general to the species, that have enabled the species to adapt and survive in its typical environment. This line of argument clearly accepts the first argument, namely that there are Gestalten qualities of living organizations that are unlikely to be revealed by the ordinary modes of scientific analysis. It goes beyond this in suggesting that there may be properties that can be generalized to the ‘species’ and yet claim no necessary generalizability to all living systems because systems analysis presupposes a knowledge of what functions the part systems can undertake. Without this knowledge it would be very difficult indeed to determine what the total system was supposed to be coordinating and controlling. [pp. 7-8]
We posed a second question, ‘What systems thinking is relevant to the thinking required for organizational management?’ The editor believes that it has been shown that living systems, whether individuals or populations, have to be analysed as ‘open systems’, i.e. as open to matter-energy exchanges with an environment. Human organizations are living systems and should be analysed accordingly. The fact that it faces us with the task of analysing forbiddingly complex environmental interactions gives us no more of an excuse to isolate organizations conceptually than the proverbial drunk had when searching for his lost watch under the street lamp because there was plenty of light when he knew he had lost it in the dark alley. [p. 8]
A great deal of work has been done in the social sciences to elucidate the properties of social systems considered as isolated entities. We shall not draw on this material. Management is concerned with the control of social systems, technologies, and markets. Our central purpose in selection has been, therefore, to depict the emergence and clarification of the view that living systems are essentially ‘open systems’, not ‘closed systems’. Despite Koehler’s contribution (Reading 3) in 1938 and Angyal’s (Reading 1) in 1941, the major impact came from von Bertalanffy (Reading 4) in 1950. It is perhaps unfortunate that this impact gave rise to a movement for General Systems Theory along with its search for dynamic principles common to all kinds of systems living or mechanical. This movement has been attractive to those with a systems engineering orientation, but has so far failed to further its unifying mission and has tended to overshadow the carly recognition by Ashby and Sommerhoff that if living systems are to be treated as open systems we must be able to characterize their environments. [pp. 8-9]
In Part Three we present a set of papers which have sought to deal specifically with the properties of environments that are relevant to adaptive behaviour.
These efforts are all concerned with global properties of organizational environments. In this capacity these theories seem to be far removed from the specificity of organization-environment relation that Sommerhoff shows to be necessary for a science of living systems. The gap may well prove to be more apparent than real. As Gibson (1966) and Tomkins (1953) have argued, living systems probably learn and hence adapt because of their ability to react to the general and less variable properties of the environment, rather than because of their sensitivity to the concrete events and objects which do after all yield a constant flux of stimulation. Measuring organizational environments along the dimensions suggested by Simon, Ashby, and the others, may be all that is required to realize Sommerhoff’s general theory.
Part Four brings together papers that concern the extension of these notions to social systems. It will be clear that these efforts have barely begun to encompass the richness of thought exhibited in the preceding sections. There is no reason why the serious student of management should not regard this as a challenge to join in the bridging operation. To encourage this we present six principles which reach back as far as the work of Koehler in order to identify lessons of value to systems management:
1. The primary task of management is to manage the boundary conditions of the enterprise. The boundaries of an enterprise are those levels of exchange with the environment which allow it to survive and grow. They can be managed only by managing the co-variation of internal and external processes. In so far as a manager has to co-ordinate or otherwise resolve internal variances then he is distracted from his task. [p. 9]
2. The goals or purposes of an enterprise can be understood only as special forms of interdependence between an enterprise and its environment, They cannot be identified with the state of equilibrium that is the end-state for closed systems. The state of equilibrium represents a minimum level of potential energy or capacity for work. The enterprise seeks to establish and maintain those forms of interdependence that enable it to maximize its potential energy or capacity for work. As in Koehler’s example of the flame (pp. 63-5), a steady state is achieved only at the level of maximum potential energy. The form of this potential capacity and the exchanges are determined by the special forms of interdependence into which the enterprise enters, but achievement of a steady state is the most general dynamic trend in an open system. [pp. 9-10]
3. An enterprise can achieve a steady state only when there is (a) constancy of direction, i.e. despite changes in the environment or in the enterprise, the same outcomes or focal conditions are achieved. Put another way, the system remains oriented to the same end; (b) that with respect to that end, the system maintains a rate of progress toward it which is within limits defined as tolerable. A more precise statement of ‘rate of progress’ might be that the enterprise achieves the required focal condition with lesser effort, with greater precision for relatively no more effort, or under conditions of greater variability. In any of these cases, the level of exchange would be more favourable to the enterprise. One implication of this proposition is that an enterprise can have no equilibrium state such as can be found in physical systems (because in the former case, the relevant internal and external variables – are capable of independent variation)– the state of one does not automatically determine the other). A positive implication is that an enterprise cannot hope to achieve steady state (except accidentally) unless it sets a mission for itself in terms of outcomes that are capable of achievement and yet are sufficiently beyond present performance to allow for some measurable degree of progress.
4. Given the last two propositions, the task of management is governed by the need to match constantly the actual and potential capacities of the enterprise to the actual and potential requirements of the environment. Only in this way can a mission be defined that may enable an enterprise to achieve a steady state. However, the actions of management cannot in themselves constitute a logically sufficient condition for achievement of a steady state. [p. 10]
5. A ‘steady state’ for the system cannot be achieved by any finite combination of regulatory devices or mechanisms that are aimed at achieving a steady state for some partial aspect of the system such as input-output rates, internal change, or environmental contact. In a human organization, the two requirements for a steady state, unidirectionality and progress, can be achieved only by leadership and commitment. The end-state of the system must be clearly enough defined and agreed upon to enable the system to be oriented toward it regardless of a wide range of changes in their relations. Secondly, the members of the organization must be so committed to the end-state that they will respond to emergencies calling for greater efforts. The basic regulation of open systems is thus self-regulation — regulation that arises from the nature of the constituent parts of the system. [pp. 10-11]
One corollary is that it is only within this framework that regulatory mechanisms, such as cost controls, can make an effective contribution. In creating these mechanisms it is essential to ensure that they do not run counter to, or undermine the requirements for self-regulation, and to remember that mechanisms which are appropriate in one phase of a system’s existence may, with a change in location with respect to the mission, become inappropriate.
The measure of whether these processes of self-regulation are operating effectively, of whether the system is healthy and maturing, is to be found in the steady growth of potential capability with respect to the mission. In the case of any enterprise, the critical question at any time is whether it is more capable than before of fulfilling the tasks arising from its mission. A good record of recent performance, e.g. high profit yield, would not in itself exclude the possibility that potential capacity had in fact been reduced.
6. An enterprise can only achieve the conditions for a steady state if it allows to its human members a measure of autonomy and selective interdependence. This proposition is clear enough when applied to organizations composed of professionals. It is less clear that it applies to enterprises in general because it introduces an assumption which is new in this context, namely that individuals themselves have open-system characteristics and can be related to each other or to organizations only in ways that are appropriate to such systems. In particular, commitment presupposes that the individual has sufficient autonomy to exercise choice. The requirement that the co-ordination of components be maximally brought about by themselves (proposition five) requires some sacrifice of autonomy and to that extent threatens commitment. This threat can be lessened by allowing selective interdependence.
These principles barely draw upon the potential of systems theory for management theory. They also show little evidence of a coherent and comprehensive theory from which such principles could be rigorously deduced. Nevertheless it is in finding new ways of looking at things that men have managed in the past to advance scientific understanding. [p. 11]
We have tried to select readings that would excite readers to a new way of looking at mundane realities of human organization. They have been arranged to lead from the earlier to the later papers and from consideration of systems in general to social systems in particular.
For Part One we have selected the earlier papers that explicitly or implicitly argue for a new logic in the study of complex systems that display purposive or adaptive behaviour. What these authors have in mind is something more than the model of causal analysis usually associated with the physical sciences; it is also something less than a rigorous predictive theory of some restricted area of behaviour. This latter feature has occasioned some misguided criticism to the effect that theories that cannot predict and hence cannot be experimentally confirmed or disconfirmed are not scientific theories. Such criticism overlooks the all-important role in scientific development of our guiding metaphors and our principles for mapping the real world (Kuhn, 1962). If systems theorizing improves the ‘internal coherence, implicative structure, freedom from the clutter and comprehensiveness’ (Chein, 1967) of the maps we make of human organizations, then we must judge it as a scientific advance.
The selections in Part One only partly argue their case by pointing to principles of system action that can be inferred from their definition of a system. However, the papers in the remaining sections were specifically selected to sample the variety of efforts that have been made to meet scientific requirements of coherence, mplicative structure, etc. These efforts have been very varied. It is almost as if the pioneers, while respectfully noting each other’s existence, have felt it incumbent upon themselves to work out their intuitions in their own language, for fear of what might be Jost in trying to work through the language of another. Whatever the reason, the results seem to justify this stand-offishness. In a short space of time there has been a considerable accumulation of insights into system dynamics that are readily translatable into the differing languages and with, as yet, little sign of the divisive schools of thought that for instance marred psychology during the 1920s and 1930s. Perhaps this might still happen if some influential group of scholars prematurely decide that the time has come for a common conceptual framework. [p. 12]
To give ready access to this variety we decided to select in terms of individual theorists instead of conceptual areas such as coordination, control, regulation, temporal integration. Likewise, to allow space for each theorist to develop his body of concepts and principles we have had to restrict the range of the sample and to leave out articles of great merit as well as those which are essentially polished rehashes or reviews. We can only hope that the heaviness of the resulting selection will be found justified by the number of unmined diamonds that exists in even the earliest.
The importance of a solid ground in this new and explosive field is such that we have not tried to depict the frontiers of systems theorizing, except in the last section introducing notions of systems management. For those wishing to extend their range of coverage in the areas represented here they could do no better than turn to the General Systems Yearbooks. For those eager to see what is happening at the frontiers the best single source is probably the journal Artorga.
The editor is grateful to the Center for Advance Studies in the Behavioral Sciences, Stanford, for the peace and quiet he needed to prepare these Readings.
CHEIN, I. (1967), “Versity vs. truth in the scientific enterprise’, address to Division of Philosophical Psychology, American Psychological Association, 3 September.
GIBSON, J. J. (1966), The Senses Considered as Perceptual Systems, Houghton Mifflin.
KUHN, T. (1962), Structure of Scientific Revolutions, University of Chicago Press.
TOMKINS, S. S. (1953), Affect, Imagery and Consciousness, Springer, vol. 1.
— end paste —
The diligent reader may also be interested in “World Hypotheses, Contextualism, Systems Methods“, with an excerpt from “Part One: Precedents to Systems Theory”, that highlights the omission of Stephen C. Pepper, for reasons of space.
Emery, Fred E. 1969. “Introduction.” In Systems Thinking: Selected Readings, edited by Fred E. Emery, 1st ed., 1:7–13. Penguin. http://books.google.ca/books?id=G2tHAAAAMAAJ.