A Special Integration Group (SIG) of the
International Society for the Systems Sciences (ISSS)
originally SGSR, Society for General Systems Research.





An activity of the Primer Group


December 1, 1996; to December 31, 1997

Applications of


Systems Theory

James Grier Miller and Jessie L. Miller

Adapted from

Analysis of Dynamic Psychological Systems, Volume 2: Methods and Applications, edited by Ralph L. Levine and Hiram E. Fitzgerald. Plenum Press, New York, 1992.

Living systems theory identifies basic principles that underlie the structure and processes of living things and relates them to the nonliving physical world, integrating and bringing order to the ever-growing mass of empirical data about them. In addition, living systems models and methodology are useful in empirical research on the great variety of systems of interest to psychology and related fields and in study of individual systems at any of the eight levels of living systems.

Research in General Systems Theory

Systems science research, whatever its particular theoretical bias, is concerned with processes of heterogeneous, complex systems. The systems of interest may be (a) living, (b) nonliving or (c) mixed living and nonliving. The last class (c) includes both man-machine and ecological systems. These are complex not only because they have many interacting parts but also because they change over time in ways that are not necessarily predictable from their initial states. What Forrester calls

1973) "the counterintuitive behavior of complex systems" results from multiple feedback loops that connect nonlinear variables in hierarchical structures.

Models in General Systems Theory

Systems scientists use models to display or discover general principles or isomorphisms (formal identities) among systems of different types or levels.

Adequate representation of complex systems, particularly living systems, requires new theoretical models in natural language and in types of mathematics capable of handling nonlinear interactions among a large number of variables.

Prigogine (1947) originated nonlinear thermodynamics, the thermodynamics of irreversible processes, to provide models for systems of this sort. Catastrophe theory, set theory, group theory, topology, fuzzy set theory, bifurcation theory, stability theory, and hierarchy theory are all used in systems models. Other mathematical approaches useful in analyzing data from this type of system or representing system processes include information theory, game theory, queuing theory, statistical decision theory, conceptual structure theory, inventory theory, factor analysis, and cluster analysis.

Computers are of great value in the study and modeling of complex systems. They can process massive amounts of data rapidly. They can also be used to simulate the interactions of thousands of variables as they develop over time from a given starting point. In this way effects of changes in initial conditions and the impact of exogenous variables on processes throughout the system can be traced. Such a simulation can be used to predict future states of systems. In addition, computerized artificial intelligence expert systems can be a part of applied work with systems of various sorts.

Models are also used to study the characteristics of a class of systems or to examine a particular system for classification, study, or diagnosis and correction of pathology.

Cross-Level Models

The following general systems models are of living and nonliving systems of many sorts. They are not based on living systems theory but are compatible with it. Living systems theory stresses the importance of cross-level research to discover general properties of living systems of all kinds. All of these general systems models are applied to systems at different levels, although they are not specifically described as "crosslevel" by their authors.

Nicolis and Prigogine (1977) have used nonlinear thermodynamic models to demonstrate the phenomenon of "self-organization." Their concepts are very complex but so important that they need to be understood by psychologists and other behavioral scientists. In simple terms, they believe self-organization to be a fundamental formal identity of all systems able to maintain over time an internal order that apparently defies the second law of thermodynamics. This law of nature provides, in its broadest form, that the most probable state of systems, toward which all isolated systems tend, is an equilibrium in which molecules are randomly distributed.

Self-organizing systems are able to maintain a nonrandom and, therefore, improbable state because they are open systems that exchange inputs and outputs of matter and energy with their environment. Under proper conditions, inputs or outputs can cause fluctuations that destabilize the system so that "branching" or "bifurcation" of its states occurs. Bifurcation is a transition from the ordinary progression of thermodynamic states to new, stable, and more complex configurations; that is, order can increase in such self-organizing systems. These configurations are "dissipative" structures that can remain ordered only so long as sufficient flows of matter and energy are available.

The models of Prigogine and his colleagues illustrate the operation of this principle in many kinds of systems such as interacting chemical reagents, cells undergoing processes like glycolysis and synthesis of protein, developing embryos, neural nets, the immunity system in organisms, species in evolution, population dynamics in communities, and processes of social systems such as traffic flow in cities.

Conrad's (1983) thermodynamic models were intended to show that adaptability is a general characteristic of biological systems. Adaptability is defined as the use of information to handle environmental uncertainty.

These mathematical models are concerned with evolutionary processes in ecosystems made up of a nonliving part, the physical environment, and a living part, the biota. Conrad considers organisms the most important units in the biota. Above the organism are aggregations of organisms, including social systems and populations. Below the organism are organs, cells, organelles, molecules, and atoms.

Conrad's models describe the contribution to adaptability of several important processes which, in living systems theory, would be called adjustment processes. These include transformability, compensation, cycle formation, and the minimal tendency of adaptability. They operate at multiple levels of living systems.

Odum's (1983) ecological models also describe systems in thermodynamic terms of energy and entropy. He analyzes the processes of a variety of living systems, including social and economic systems as well as nonliving systems and ecological systems, to support the proposition that systems prevail that maximize the flow of useful energy (the maximum power principle). They ordinarily operate at intermediate rates to conserve energy but exercise maximum power when that is necessary to compete against other systems. He considers the struggle for survival to be essentially competition for free energy, that is, energy that is available for work.

The "dynamic" models of Forrester and his associates are systems of equations that are programmed in a computer language, DYNAMO, developed for the purpose. His models of a factory (1961) and an urban community (1969) apply not to any particular factory or community but to classes of systems with similar characteristics.

He and his colleagues have also developed global models that incorporate variables such as food supplies, industrial production, agriculture, pollution, and population (Forrester, 1973; Meadows, Meadows, Randers, & Behrens, 1972).

By running a model with different initial conditions and assigning different values to variables, the effects of changes in such things as prices and policies can be traced through a system. The simulation can be extrapolated into the future, and predictions of future system behavior can be made. Such models are useful to the extent that the selection of variables, processes, and data accurately reflect the real systems they represent.

Of course it is not possible to include all variables and interactions of such complex systems in even the largest simulation. Those used in a particular simulation are selected as typical of the aspects of real systems with which the modeler is concerned. Data are drawn from statistical or other sources.

Leontief and his colleagues (1953) have developed an input-output model for analyzing the flows of goods and services through an economy. He identifies various producing and consuming sectors that can be found in the economies of regions of a nation, an entire nation, regions of the world, and the world as a whole. The sectors he identifies are commonly used in economic analysis and are compatible with the Standard Industrial Classification Manual. The number of sectors in a model depends upon the degree and type of aggregation of data required.

Leontief's input-output models consist of sets of differential equations that represent the producing and consuming sectors of an economy. These are connected by flows of goods and services. Input and consumption coefficients for each sector describe the combination of goods and services it needs in order to produce a unit of its output, or, in the case of households, a unit of their expenditures and income. Stocks of capital goods and natural resources are also represented.

The large number of equations in his models are solved by computer. The resulting data are presented in tables that give a detailed picture of the economic system being studied.

In addition to its use as a general-purpose economic model, inputoutput analysis has been employed to study the possible futures of developing countries as well as world military spending. A global model (1977) has been run with alternative projections for the years 1980, 1990, and 2000. Eight sets of assumptions about, for example, growth rates of populations, growth rates of per capita incomes, and pollution abatement created eight "scenarios" for the future of the world economy with outcomes that range from bleak to optimistic.

Living Systems Theory: Basic Research and Applications

Basic Research

Any scientific theory derives its credibility and eventual validation from its correspondence to the real phenomena to which it is applied, its usefulness in solving problems and answering questions, and the degree to which it contributes to science in general. It is, therefore, critical that research be undertaken on testable hypotheses derived from it.

Basic living systems research is concerned with intersystem generalization, the search for common aspects of structures or processes among living systems of different kinds. Generalization may be among individuals of the same species or type, among systems of different species or types at the same level, or among systems at different levels.

Cross-level research is the most powerful of these approaches, although all have their place in the study of living systems. It seeks to discover isomorphisms among systems at two or more levels and to apply models based upon them to a variety of systems. This is a useful approach because it can illuminate previously undetected regularities in nature and can result in deeper understanding of basic characteristics of living systems at two or more levels. Scientists must always be alert to the fact that greater generality is a major goal of science.

Cross-Level Tests of Living Systems Hypotheses

A list of 173 testable cross-level hypotheses appeared in Living Systems (Miller, 1978), and others have been stated since. Some of these apply to all levels, others to two Or more. Several have been tested empirically.

Formal identities across levels cannot be rigorously demonstrated unless the dimensions and units used in measurements at various evels are compatible. If this were not so, curves that plot data at different levels might appear to be identical when they were not. Conversely, such curves might appear to be of entirely different shapes and yet actually be the same. For instance, a curve with equal intervals along the abscissa and logarithmic intervals along the ordinate looks entirely different from a curve derived from the same data with logarithmic units on both coordinates.

Information Input Overload. The first multilevel experimental test of a hypothesis derived from living systems theory was research on information input overload (Miller, 1960, 19 78). The hypothesis, which was tested at levels of cell, organ, organism, group, and organization was the following:

Hypothesis 5.1-1: As the information input to a single channel of a living system-measured in bits per second-increases, the information output-measured similarly-increases almost identically at first but gradually falls behind as it approaches a certain output rate, the channel capacity, which cannot be exceeded in the channel. The output then levels off at that rate, and finally, as the information input rate continues to go up, the output decreases gradually toward zero as breakdown or the confusional state occurs under overload.

Experiments at each level were conducted by specialists in relevant fields. Measurements of information input and output rates were made on single fibers (from the sciatic nerves of frogs), optic tracts of white rats (retina or optic nerve to optic cortex), human subjects working alone, human subjects in groups of three, and laboratory "organizations" made up of nine subjects.

In one form of experimental organization, two groups of three people simultaneously received information inputs. In each of these, two members served as input transducers, each member receiving a different sequence of visual inputs. The input transducer components sent electronic signals to the third member of the group who output them to a display before one member of a three-person group in another room. Each of these forwarded the signals he received to the final person in the second room, who acted as decider and output transducer. He compared the signals, made a decision about them for the total organization, and output the decision to a recording device.

Data from all the levels-cell, organ, organism, group, and organization-yielded information input-output curves alike in form. Each curve rose sharply at first, leveled off, and then, as the channel capacity of the systems was exceeded, fell toward zero. Transmission rates were lower at each higher level.

These results confirmed the hypothesis of a formal identity in this aspect of information processing at these five levels.

Nonrandom Nets. Rapoport and a group working with him did experiments on communication networks in elementary and junior high schools (Rapoport & Horvath, 1961). Rapoport (1957) had previously investigated models of neural nets through which information in the form of bioelectric pulses is propagated and conjectured that a similar model might apply to communications networks in systems at higher levels, such as organizations. They tested the following hypotheses:

The structure of the communication networks of living systems at various levels are so comparable that they can be described by similar mathematical models of nonrandom nets.

Nonrandom or "biased" nets are those in which communication over some channels is more probable than over others. They were able to predict with reasonable accuracy some aspects of the friendship patterns in schools.

Effects of Conflicting Signals on Decision Making. An experimental study by Lewis (1981) on individual human organisms and on groups tested an hypothesis that relates to the decider subsystem at those levels:

When a system is receiving conflicting command signals from several suprasystems, it intermittently being a component of all of them, the more different the signals are, the slower is its decision making.

Subjects played a computerized game in which they had to defend their own ships and destroy the enemy in war in space. A computer presented "commands" from five officers, all of the same rank and all superiors of the subjects. The subjects had to decide which command to obey. They could "fire," "warp...... scan...... dock," or "wait." Eight patterns of commands with differing amounts of conflict were used, ranging from total agreement among the officers to total disagreement. Subjects could act only on command of at least one officer. The game was scored so that subjects were more likely to succeed when more officers agreed upon a command.

The research design was the same for both groups and single subjects. In the group experiment, three subjects worked together to arrive at a joint decision. After 40 turns, a final score was calculated, based on the subject's success in defending his own ships and destroying the enemy. The results were highly significant in both the individual and the group experiments, an outcome that supported the hypothesis.

The preceding three sets of basic researches are examples of crosslevel hypotheses based on living systems theory. Other cross-level hypotheses have been tested by other scientists.

Living Systems Studies of Individual Systems

Living systems theory can be used to analyze and work with individual living systems at all eight levels. It has been applied to selected systems in diagnosis of pathology, in treatment, and in efforts to improve both efficiency and effectiveness of system processes.

When a physician examines a patient, he or she uses diagnostic tests and instruments to check the structure and processes of organ systems. The living systems method for studying systems at other levels is similar. It involves observing and measuring important relationships between inputs and outputs of the total system and identifying the structures that perform each of the 19 subsystem processes discussed in Chapter 2 (Volume 1). The flows of relevant matter, energy, and information through the system and the adjustment processes of subsystems and the total system are also examined. The status and function of the system are analyzed and compared with what is average or normal for that type of system. If the system is experiencing a disturbance in some steady state, an effort is made to discover the source of the strain and correct it.

A set of symbols, shown in Figure 1, has been designed to represent the levels, subsystems, and major flows in living systems. They are intended for use in simulations and diagrams and are compatible with the standard symbols of electrical engineering and computer science. They can be used in graphics and flow charts like those used in several applications discussed in this chapter (e.g., see Figure 2, p. 177).

It is of great importance that measures or indicators be developed for structures and processes of systems at all levels and that normal values and ranges of critical variables be determined. This has been achieved for thousands of physiological variables of human beings. Normal values and ranges of many variables of interest to psychologists are also known, particularly in areas like sensation, perception, learning, and child behavior, but many are still to be established. For some other types of systems, such as organizations, such information could be obtained, but efforts to do so are rarely made.

Some of these studies consist of subsystem reviews only with little use of other living systems concepts. Others are sophisticated efforts to analyze systems and make recommendations for their improvement.

The Level of the Cell

Conceptual Structures

A new method for collecting and integrating facts has been described and computer software developed by Sowa (1983). It can be used in any scientific discipline, profession, or other area of interest in which knowledge accumulates over time. His method uses logic, mathematics, and/or simulation to display the conceptual structure of a field and show the relationships among the concepts of different authors. As data from more and more sources are added, the growing conceptual structure is amplified, refined, and made more precise. Sowa's procedures also specify the logical relationship of each new fact to preceding facts and alters the conceptual structure as necessary.

To perform an integration that reflects the current state of knowledge of a subject, it is necessary to provide the computer with a database. Parameters, variables, and specific values can be found in published research in the field of interest or related fields. Qualitative as well as quantitative statements can be included in this sort of conceptual structure and, as the field advances and knowledge increases, less precise qualitative statements can be replaced by quantitative expressions.

Cellular biology, cellular genetics, and neuropsychopharmacology have generated a vast literature that includes results of millions of experimental and clinical studies. Thousands of articles in each of these areas are published annually. As a result, no individual person can remember all the things that have been learned in his area.

We propose to attempt to achieve integration of knowledge at the level of the cell by building a conceptual structure that will include all relevant information. A small group of computer scientists, mathematicians, and cellular biologists will work together in the project. Data about cells will be drawn from the literature of several relevant fields. The levels, subsystems, and subsystem variables of living systems theory can constitute a set of parameters on which to base such a conceptual structure. It could be developed much as biologists' knowledge of cells has accumulated over the last few centuries. At the same time, the components of cells could be related to the subsystem processes of living systems theory (see Table 1).

The structure would first include facts discovered soon after microscopes were first used, such as the shape and range of sizes of cells and the presence of a nucleus in many cells. Other discoveries, like the chromosomes and other nuclear structures, the behavior of the genetic material in mitosis, and the fact that chromosomes are the sites of genes that specify the characteristics of descendants would be added successively. Thus a description of the cellular reproducer subsystem would begin to arise. It would become more precise as the structure and function of DNA and RNA and the reproducer components in the cytoplasm were included in the conceptual structure.

Descriptions of the other subsystems of the cell would be created in the same way, and ultimately the relationships among subsystems would be included in the growing representation of cellular structure and process. As it developed, the conceptual structure would become more and more precise, detailed, and congruent with current knowledge in cellular biology. Also it would be increasingly useful as a reference source for scientists and a source of suggestions as to what future studies were needed.

It seems possible that conceptual structures could be developed at all the other levels of living systems by an analysis of the relevant literature using the same method we propose to use on cells. Because the parameters used at each level would be those of living systems theory, it would be most interesting to find out whether the conceptual structures were similar at the different levels. To the extent that this was true, it would tend to confirm the evolutionary cross-level aspects of living systems theory.

The Level of the Organ

No research at this level is planned at present, but it is obvious that particular organ components, like the liver or the bladder, could be analyzed in terms of living systems theory. In addition, a living systems conceptual structure of any organ could be created using Sowa's methods.

The Level of the Organism

Expert Systems

Artificial intelligence expert systems are used as research tools or as practical devices to solve problems or replace a human interviewer. For instance, such systems have proved to be effective for securing admissions information in hospitals. Two diagnostic programs, Mycin (Buchanan & Shortliffe, 1984) and Internist I (Miller, Pople, & Myers, 1982) are available to suggest probable diagnoses in clinical situations.

Expert systems are of two n s: (a) those t at model the way human beings behave or solve problems and (b) those that solve problems in the optimal way regardless of whether their logical processes are like human thinking.

Living systems theory probably has little relevance to the second sort of artificial intelligence system, but it can be usefully applied to the first, the systems that are designed to operate as much as possible like human problem solvers. At present, no unifying conceptual system is in use by the developers of expert systems. Each selects her or his own space and dimensions. The living systems subsystem analysis, the designation of flows as matter-energy, information, or both, and the use of comparable units and common dimensions could integrate the field and make it possible to relate the various expert systems.

Almost all expert systems today are at the level of the organism, but the cross-level emphasis of living systems theory leads us to ask whether expert systems might some day carry out some functions of work groups on assembly lines, of certain decision-making organizations like stock markets, and even of higher level systems. The increasing availability of very large computer chips makes such developments appear possible.

Diagnosis and Treatment

Living systems theory is of potential value to physicians and mental health professionals. Although not widely accepted, it has been used as a model for diagnosing a patient's physical and psychological state (Kolouch, 1970); applied in the diagnosis and treatment of asthma (Kluger, 1969); and used in analysis of school phobia (Bolman, 1970).

A subsystem view of the sort described was used by members of a clinical seminar in an analysis of a teenage boy with clinical symptoms of anorexia nervosa. The analysis was carried out at both organism and group levels. This case illustrates how important it is to consider not only the subsystems of a system in which pathology is diagnosed but also the suprasystem, in this case, the family of the patient.

The patient was brought to the hospital by his parents, who were concerned about a major loss of weight and other symptoms brought about by his refusal to eat or drink. Physical and neurological examination showed him to have no structural abnormalities in his matter- energy-processing subsystems, although there were functional abnormalities in the ingestor (failure to input matter-energy). Functional problems were also observed in the distributor (abnormal heart action) and extruder (urinary system and bowel malfunction). All these symptoms disappeared when he was given nourishment in the hospital.

Neurological examination disclosed no disease or structural abnormality in his nervous system (channel and net and other informationprocessing subsystems). Psychological examination showed normal associator and memory function for his age. Encoding and output transducing were also normal.

Certain of the information-processing subsystems were, however, functioning abnormally. He was extremely sensitive to noise (input transducer) and, unlike most patients, remained so when starvation ended. He insistently denied feeling hunger and thirst and reported feeling cold when others were uncomfortably hot (decoder). Like other anorexic patients, he perceived his body to be too fat when he was, in fact, very thin (decoder).

The major pathology in this boy appeared to be in the decider subsystem. Unlike other teenage boys, he seemed to have no personal goals but was apparently motivated entirely by reward and punishment. There was no sign of the desire for emancipation that is common among American boys of his age. He also appeared unable to control his behavior by setting limits for himself. He had become interested in collecting some time before and had filled his room with so many things that he had to move out. When he started to draw or write, he continued as long as he was rewarded for doing so. Further study of this case centered on his family.

The Level of the Group

Diagnosis and Treatment

During the course of diagnosis and treatment of the anorexic boy discussed above, a social worker had several interviews with the parents together and, infrequently, saw each parent separately. A living systems analysis of the family was based upon her report.

This family differed in many ways from others of similar cultural and educational backgrounds. The deviation was sufficiently great and the effects on family members so profound that several family processes appeared to be pathological. The primary pathology at the group level, as it had been at the organism level, was in the decider subsystem. Pathologies in several other subsystems appeared to be secondary.

Although families differ greatly in the way they make decisions, this family was clearly unusual. All decisions, including those ordinarily made by mothers, were made by the father. The mother was allowed no more independence than the children. The father insisted that the family be as isolated as possible from the community and that the parents be separated from the children only when it was unavoidable (matter-energy boundary). He censored incoming information (information boundary) by deciding what family members were allowed to read or watch on television and by prohibiting friendships. All information that entered the family was interpreted by the father in terms of his own fears and suspicions (decoder). No one could discuss personal feelings and attitudes (internal transducer). Communication among family members was discouraged (channel and net).

The results of the analyses at organism and group levels were used to make the diagnosis, decide what further treatment would be offered, and set treatment goals.

The pathological adjustment processes of this patient and his family appeared to be consequences of the father's aberrant information processing. The primary pathology was at the organism level, but in the father rather than the patient. In his interviews, the father had revealed fear that he would lose his family as he had lost his parents early in childhood.

The seminar group believed that the patient was unlikely to adjust normally unless his father was helped to gain insight into the reasons for his excessive control. It also seemed probable that the other children would experience difficulties when they reached adolescence. Treatment, therefore, would focus on him. Both individual treatment of the patient and treatment of the family appeared necessary.

Goals for treatment were (1) to help the father gain the insight that he could allow his family increased independence and still retain their love, and (2) to work with the patient toward an adjustment more normal for this age.

Family Interviews

One description of how any family could be analyzed dealt with the structure, processes, and pathologies of each subsystem as well as feedbacks and other adjustment processes (Miller & Miller, 1980).

A subsystem review of a real family (Bell, 1986) was carried out in a videotaped interview that followed a schedule designed to discover what family members were included in each of several subsystems, how the family decided who would carry out each process, how much time was spent in each, and what problems the family perceived in each process. In addition, sections of the schedule concerned how rules are made in families with children (a decider subsystem process) and how the rules are enforced and children disciplined (an adjustment process). We consider rule making to be a decider function, while enforcement and discipline are adjustment processes.

Ant Nests

For decades, analogies have been made between animal organisms and the nests of ants, termites, and other social insects. At times there was widespread skepticism among biologists as to whether such cross-level analogies are valid. In recent years, as more evidence has accumulated, the organismlike character of social insect colonies has been widely accepted. It therefore seems reasonable that the input-output flows of matter, energy, and information through the 20 critical subsystems can be observed and measured in the nests of ants and other social insects. We recognize the cross-level formal identity between organisms and insect nests, but we maintain that those systems are groups rather than societies, as they are often called.

We are planning to videotape an ant nest for several hours continuously and to make such observations and measurements from the tapes. We intend to compare our observations of these subsystem processes with the 26 kinds of behavioral acts recognized by Oster and Wilson (1978) in ant nests.

Level of the Organization

Living systems theory offers a means for analyzing the structure, function, and processes of organizations and finding disfunctions that reduce a system's effectiveness in achieving its purposes. It has been used in studies of several sorts of organizations including hospitals (Merker & Lusher, 1987). A psychiatric ward (see page 174), several public schools in a community (Banathy & Mills, 1985), and a public transportation system (Bryant & Merker, 1987) among others.

Measures of effectiveness may be related to financial variables such as total profit, profit per share of stock, or cost per unit of goods or services output. They may also relate to less tangible variables like mortality rates of hospital patients, average attendance at performances or meetings, or number of applicants for admission to universities. The factors that contribute to effectiveness in an organization can be revealed by monitoring flows of matter-energy and information through components or subsystems of the system rather than using only inputoutput measures of the total system.

Each type of organization is specialized for certain subsystem processes of the society to which it contributes and may be quite unlike organizations of other types in its subsystem structure and in the type of living and machine components that carry out its processes. Organizations of the same type may also differ in many aspects of structure and process. It is, therefore, necessary to observe how the individual system under consideration is structured and what living and nonliving components are involved in each process.

United States Army Project

The first large-scale application of living systems theory was a 3-year study of 41 U.S. Army battalions (Ruscoe et a]., 1985). The Army's own assessments had revealed problems that affected its capacity to achieve and maintain optimal levels of training efficiency and the ability of units to accomplish their assigned missions. There was no consensus about the methods of evaluating battalion effectiveness that the Army has used for many years. These consist of many independent objective and subjective measures that do not relate to any integrated conceptual system. Although these may distinguish good and less good battalions, they do not reveal the basis for differentiating them.

The objective of this study was to use living systems process analysis to explain how battalions function and relate the quality and quantity of flows of matter-energy and information to battalion effectiveness.

In this study, three types of data were collected: (a) findings of the Army's traditional evaluation methods, which combine performance indicators, command indicators, and perceptions of personnel; (b) process perception data, which includes opinions of unit personnel on how well each matter-energy and information process was being handled, the time spent on each of the members of the unit and the unit as a whole, the importance of each process, and how well the process was being performed in terms of several process variables; and (c) process objective data that includes such things as the percentage of the unit's vehicles that are operational at a particular time.

Data sources were mass administration of standardized questionnaires, interviews with key management personnel, records, and reports available within each unit, surveys from brigade-level managers, and check sheets completed by surveyors. In all, more than 5,000 officers, noncommissioned officers, and other enlisted personnel from battalions in both the continental United States and Europe participated.

Two sets of effectiveness criteria were employed: the traditional Army criteria and a new set based on living systems theory. Although both the living systems criteria and our data collection were derived from living systems theory, we were able to avoid contamination between the two. The living systems criteria led to similar conclusions about distinctions among battalions as traditional measures of unit effectiveness, but they revealed much more about the dynamics of the units.

Among the many discoveries from the study (Ruscoe et a]., 1985, pp. 45-49) was the relationship between information processing and effectiveness. The greater the unit personnel's appreciation of and skill in information processing, the more effective it was. The information variables of meaning, lag, volume, cost, and distortion were repeatedly shown to be good indicators of unit effectiveness.

IBM Study

Civilian corporations and army battalions are different in many ways, but they have many similarities in addition to carrying out the same life processes. This study is planned to use many of the same procedures as the Army study but will include several significant improvements.

1. In the IBM research, to be consistent with conventions used in studies of organizational behavior for about 25 years (Forrester, 1961), we measured five rather than three flows: materials (MATFLOW), energy (ENFLOW), communications (COMFLOW), money and money equivalents (MONFLOW), and personnel (PERSFLOW). These are basically the same matter-energy and information flows used in all other living systems research. We divided information into two classes (communications and money) and added one other flow, personnel, which are lower-level groups and individual persons, each made up of matter-energy and information.

2. We collected only subjective data in the Army study, but in this study we intend to employ objective data in addition. The objective measures will be used to check the reliability of the subjective data.

3. Subjects will give their responses to computers rather than to human interviewers in standardized interviews.

4. We shall observe the location of members of the organization and of equipment objectively, using badges that will send electronic signals to a central computer. Such data will supplement the estimates of time spent in each activity by the people wearing the badges. Use of such badges will be voluntary.

5. Data analysis will be accomplished by computerized expert systems. It is planned to use two types: (1) ALIGN (the interactions among components within the system) and (2) IMPACT (interactions between the total corporation and the local, national, or supranational market in which it competes.

6. Three separate criteria of internal organizational effectiveness will be used: (a) Independent opinions of several judges on the best strategy in different situations. These judges will be chief executive officers. (b) Effectiveness criteria derived from the previous Army study. (c) Effectiveness and productivity data from the extensive literature on organizations.

Hospital Studies

Numerous studies have applied living systems concepts to clinical psy-

chology and to psychiatry. Several of these specifically apply living

systems theory.

1. The components, variables, and possible pathologies of the sub-

systems of the inpatient psychiatric unit of a university hospital were

described by Chase and his colleagues (Chase, Wright, & Ragade,


2. The same group analyzed the decision process in the case of an adult male alcoholic patient (1981b). The process included decisions by the patient's employers, his family, the patient himself, and his family physician before he reached the hospital. From his first appearance at the admission desk to his discharge to outpatient treatment 16 days later, a series of 38 decisions determined he would be admitted, the course of treatment, and the treatment plan after discharge.

The writers concluded that decision analysis of this sort can be of practical clinical significance. It can show what components are involved in the decision process, find significant feedback loops, and reveal otherwise unknown problems in a treatment situation. Superimposing decision analyses for several patients can uncover significant problems within the patient milieu.

3. Following upon two unpublished studies by Whitehead, Goldberg, and others, plans have been made for a full-scale study of hospital cost-effectiveness using methods developed in the IBM Study.

Few, if any, adequate procedures exist for objective evaluation of performance in hospitals, clinics, and health maintenance organizations. Our methods should make it possible to compare alternative pproaches to cost reduction and suggest ways to improve cost-effectiveness without threatening the quality of health care.

We plan to concentrate on information processing in a hospital. There is abundant evidence that information processing in many organizations is unsatisfactory. It may be incomplete; distorted by misinterpretations, opinions, or emotions; untimely; or presented from incompatible viewpoints or in different formats that are difficult to integrate. It may also be wrongly interpreted by the receiver.

A Living Systems Theory of Accounting

Accounting is an internal transducer process that takes place at the levels of the organization and above. Living systems concepts have been used as a basis for a general theory of accounting, and a set of testable hypotheses have been generated (Swanson & Miller, 1986). This approach has been used in a quantitative analysis of the accounts of W. T. Grant, Inc., a large retail store (Swanson & Miller, 1986). Living systems theory has also been applied to marketing, an output transducer process (Reidenbach & Oliva, 1981).

The Level of the Community

A City

Communities are complex systems with many problems that could well be approached from the point of view of living systems theory, but so far few studies have been made at this level. An analysis of the city of Louisville, Kentucky (Vandervelde & Miller, J. L., 1975), was the first attempt to describe the subsystems of a community in living systems terms.

Community Health Care

An analysis of the different levels of living systems involved in community health care and the subsystems of the health care system of a community, which is a part of the producer subsystem, addressed some of the problems involved in health care delivery (Miller, 1970a).

A Community in Space

It appears likely that within a few decades a settlement will be established on a satellite in earth orbit or on the moon. In 1984, a study of uch a settlement was conducted by the California Space Institute (Calspace). Representatives of NASA, including former astronauts, met with members of the Calspace staff and outside consultants to discuss the feasibility and possible program of such a project. Although it would be largely a macroengineering project, the needs of the living systems that would inhabit the space station were also considered (Miller, 1987, pp. 202-203).

Because a community of this sort would necessarily be more isolated and therefore more totipotential than a community on earth, a living systems analysis was used to determine what components would comprise each subsystem, and how flows of essential matter-energy and information to all parts of the system would be assured.

A community on the moon, set up in a location favorable for mining and extraction of the abundant useful minerals of the moon's surface, was envisioned. It would receive regular deliveries of essential inputs by shuttle from earth and send its mineral output to earth in the same way. Figure 2 is a diagram of the chief subsystem processes and flows in such a community. The symbols shown in Figure 1 are used to identify components of each subsystem.

The Level of the Society

The Administrative Side of a Government

A study for the General Accounting Office (GAO) of the United States is currently in the planning stage. It would employ the techniques used in the Army and IBM researches and the accounting concepts mentioned above. Originally the GAO was given the responsibility of making traditional accounting audits of each agency in the executive and judicial branches for Congress. A few years ago it was given the additional responsibility of performing "management audits." It is required to report to Congress how cost-effective each agency is. The GAO has been asked if management audits based on living systems theory could be conducted that might improve current methods used by that agency.

The Supranational System

In the latter half of the twentieth century a worldwide channel and net subsystem has developed that connects all but the most remote human settlements. Its components are human beings working with telephone networks, cables under bodies of water, electronic networks using radio, television, and satellites in earth-stationary orbit. Ours is the first generation in which information, traveling more than seven times around the world in a second, can enable billions of people worldwide to communicate with one another.

Networks that use such virtually instantaneous transmission are in use by military organizations, banks, and other business organizations. A network for medical and health information (MEDLINE) connects the United States to many other countries. In each of the past several years, Christmas color television programs have connected religious worshipers in several countries on three continents.

A Canadian/American Transboundary Monitoring Network

The International joint Commission of Canada and the United States has been employing living systems theory as a conceptual framework for exploring the creation of an electronic network to monitor the region surrounding the border separating those two countries (Miller, J. G., 1986). The proposed network is conceived of as a channel and net subsystem for the boundary region system. It is hoped that communication across the border on such a network can improve the adjustment processes to stresses like acid rain and provide for more effective cooperation between the two nations. A similar approach could be used to decrease tensions across many other international borders throughout the world.

University of the World

Use of international networks for educational purposes was foreseen 20 years ago (Brown, Miller, & Keenan, 1967). This vision arose from a recognition of the potential of a supranational channel and net subsystem, a concept derived from living systems theory. Conviction that such networks are possible or needed has been slow to develop, but interest in establishing such a system is growing. Plans are well advanced for such a supranational network under the auspices of a University of the World. It would use satellite transmission to participating countries and radio or television as well as other electronic technologies within countries to transmit courses taught by experts. They would travel worldwide and would particularly benefit areas that have little access to high-quality education or research. Courses would be offered from the preliterate level to a master's-degree level. Examinations would be given at central educational institutions, and degrees might be granted based on such instruction. Interest in the University of the World has been expressed by numerous countries throughout the world.


This has been a description of some of the uses of living systems theory at seven of the eight levels of living systems. These researches and applications should interest psychologists from a number of subspecialties, including physiological, experimental, clinical, social, community, industrial, and educational psychologists.

Psychology today is so fractionated that it is in danger of losing its intellectual cohesiveness. Living systems theory may point the way toward reunification and integration.

James Grier Miller - Departments of Psychiatry and Psychology, University of California at Los Angeles, Los Angeles, California 90024, and Department of Psychiatry, University of California at San Diego, San Diego, California 92093. Pines Road, Suite 203, La Jolla, California 92037.


Jessie L. Miller -


Banathy, B., & Mills, S. R. (1985). The application of living systems process analysis in education. ISI Monograph, 85-7.

Bell, R. A. (1986). Videotape script. (personal communcation)

Bolman, W. M. (1970). Systems theory, psychiatry, and school phobia. American journal of Psychiatry, 127, 65-72.

Brown, G. W., Miller, J. G., & Keenan, T. A. (1967). EDUNET. New York: Wiley. Bryant, D., & Merker, S. L. (1987). A living systems process analysis of a public transit system. Behavioral Science, 32, 292-303.

Buchanan, B. G., & Shortliffe, E. H. (Eds.). (1984). Rule-based expert systems: The Mycin experiments of the Stanford heuristic programming project. Reading, MA: AddisonWesley.

Chase, S., Wright, J. H., & Ragade, R. (1981a). The inpatient psychiatric unit as a system.

Behavioral Science, 26, 197-205.

Chase, S., Wright, J. H., & Ragade, R. (1981b). Decision making in an interdisciplinary team. Behavioral Science, 26, 206-215.

Conrad, M. (1983). Adaptability. New York: Plenum Press.

Forrester, J. W. (1961). Industrial dynamics. Cambridge, MA: M.I.T. Press. Forrester, J. W. (1969). Urban dynamics. Cambridge, MA: M.I.T. Press. Forrester, J. W (1973). World dynamics (2nd ed.). Cambridge, MA: Wright-Allen Press. Kluger, J. M. (1969). Childhood asthma and the social milieu. Journal of the American Academy of Child Psychiatry, 8, 353-366.

Kolouch, F. T. (1970). Hypnosis in living systems theory: A living systems autopsy in a polysurgical, polymedical, polypsychiatric patient addicted to Talwin. American Journal of Hypnosis, 13(l), 22-34.

Leontief, W., Chenery, H. B., Clark, P. G., Duesenberry, J. S., Gerguson, A. R., Grosse, A. P.,

Grosse, R. N., Holzman, M., Isard, W, & Kistin, H. (1953). Studies in the structure of the American economy. New York: Oxford University Press.

Leontief, W (1977). Thefuture of the world economy. New York: Oxford University Press. Lewis, F. L. II. (1981). Conflicting commands versus decision time: A cross-level experiment. Behavioral Science, 26, 79-84.

Meadows, D. H., Meadows, D. L., Randers, J., & Behrens, W. W III. (1972). The limits to

growth. New York: Universe Books.

Merker, S. -L., & Lusher, C. (1987). A living systems process analysis of an urban hospital.

Behavioral Science, 32, 304-314.

Miller, J. G. (1960). Information input overload and psychopathology. American journal of Psychiatry, 116, 695-704.

Miller, J. G. (1978).-Living systems. New York: McGraw-Hill.

Miller, J. G., & Miller, J. L. (1980). The family as a system. In C. K. Hoffling & J. M. Lewis (Eds.), The family: Evaluation and treatment (pp. 141-184). New York: Brunner/Mazel.

Miller, J. G. (1986). A living systems analysis of a Canada/U.S. boundary region. In P. T.

Haug, B. L. Bandufski, & A. L. Hamilton (Eds.), Toward a transboundary monitoringnetwork: A continuing binational exploration (Vol. 1). Washington, DC: U.S. State

Department, International joint Commission, U.S.A. and Canada.

Miller, J. G. (1987). The study of living systems: A macroengineering perspective. Technology in Society, 9, 191-210.

Miller, R. A., Pople, H. E. Jr., & Myers, J. D. (1982). INTERNIST 1, An experimental computer-based diagnostic consultant for general internal medicine. New England Journal of Medicine, 307, 468-476.

Nicolis, G., & Prigogine, 1. (1977). Self-organization in non-equilibrium systems: Fromdissipative structures to order through fluctuations. New York: John Wiley &

Odum, H. T. (19831. Systems ecology. New York: John Wiley & Sons.

Oster, G. F., & Wilson, E. 0. (1978). Caste and ecology in the social insects. Princeton, NJ:

Princeton, NJ: Princeton University Press.

Prigogine, 1. (1947). Etude thermodynamique des processus irreversibles. Liege: Desoer. Rapoport, A. (1957). Contribution to the theory of random and biased nets. Bulletin.of Mathematical Biophysics, 19, 257-278.

Rapoport, A., & Horvath, W..J. (1961). A study of a large sociogram. Behavioral Science, 6,279-291.

Reidenbach, R., & Oliva, T. A. (1981). A framework for analysis. Journal of Marketing, Fall, 42-52.

Ruscoe, G. C., Fell, R. L. I Hunt, K. T., Merkpr, S.'L., Peter, L. R., Cary, Maj. J. S., Miller, J. G.,,

Loo, Cpt. B. G., Reed, Cpt. Ri W., & Sturm, Cpt. M. 1. (1985). The application of living

systems theory to 41 U.S. Army battalions. Behavioral Science, 30, 7-50.

Sowa, J. S. (1983). Conceptual structures. Reading, MA: Addison-Wesley.

Swanson, G. A., & Miller, J. G. (1986). Accounting information systems in the framework

of living systems theory and research. Systems.Research, 4, 253-265.

Swanson, G. A., & Miller, J. G. (198,9). Measurement and interpretation in accounting: A

living systems approach. New York: Quantum Books.

Vandervelde, K..J., & Miller, J. L. (1975). The urban grant university concept: A systems analysis. Behavioral Science, 20, 273-295.

Send e-mail to

Return to home page

Return to map page