Knowledge : Learning Organizations ( Part 3 )

by Kai Larsen, Claire McInerney, Corinne Nyquist, Aldo Santos, Donna Silsbee and Dr. Sue Faerman, May 13, 1996, University of Albany.

The original material is at home.nycap.rr.com/klarsen/learnorg/ 

 

Systems Thinking

In the October 17, 1994 issue of Fortune magazine, Brian Dumaine named Peter M. Senge: "MR. LEARNING ORGANIZATION." ( Dumaine, 1992 ) Why is it that in a field with so many distinguished contributors, Peter Senge was referred to as the "intellectual and spiritual champion ?" ( Dumaine, 1992, p. 147 ) The reason is probably because Senge injected into this field an original and powerful paradigm called ‘systems thinking,’ a paradigm premised upon the primacy of the whole --the antithesis of the traditional evolution of the concept of learning in western cultures.

Humankind has succeeded over time in conquering the physical world and in developing scientific knowledge by adopting an analytical method to understand problems. This method involves breaking a problem into components, studying each part in isolation, and then drawing conclusions about the whole. According to Senge, this sort of linear and mechanistic thinking is becoming increasingly ineffective to address modern problems. ( Kofman and Senge, 1993, p. 18 ) This is because, today, most important issues are interrelated in ways that defy linear causation.

Alternatively, circular causation-where a variable is both the cause and effect of another-has become the norm, rather than the exception. Truly exogenous forces are rare. For example, the state of the economy affects unemployment, which in turn affects the economy. The world has become increasingly interconnected, and endogenous feedback causal loops now dominate the behavior of the important variables in our social and economic systems.

Thus, fragmentation is now a distinctive cultural dysfunction of society.4 ( Kofman and Senge, p. 17 ) In order to understand the source and the solutions to modern problems, linear and mechanistic thinking must give way to non-linear and organic thinking, more commonly referred to as systems thinking-a way of thinking where the primacy of the whole is acknowledged.

THE PRIMACY OF THE WHOLE

David Bohm compares the attempt to understand the whole by putting the pieces together with trying to assemble the fragments of a shattered mirror. It is simply not possible. Kofman & Senge add:

"The defining characteristic of a system is that it cannot be understood as a function of its isolated components. First, the behavior of the system doesn't depend on what each part is doing but on how each part is interacting with the rest ... Second, to understand a system we need to understand how it fits into the larger system of which it is a part ... Third, and most important, what we call the parts need not be taken as primary. In fact, how we define the parts is fundamentally a matter of perspective and purpose, not intrinsic in the nature of the 'real thing' we are looking at." ( Kofman and Senge, 1993, p. 27 )

In his prominent book, The Fifth Discipline, Senge identified some learning disabilities associated with the failure to think systemically. He classified them under the following headings:

  • "I am my position"

  • "The enemy is out there"

  • "The illusion of taking charge"

  • "The fixation on events"

  • "The parable of the boiled frog"

  • "The delusion of learning from experience" ( 1990, pp. 17 - 26 )

Although each of these contains a distinct message, illustrated how traditional thinking can undermine real learning by following up on one example: "the fixation on events."

According to Senge, fragmentation has forced people to focus on snapshots to distinguish patterns of behavior in order to explain past phenomena or to predict future behavior. This is essentially the treatment used in statistical analysis and econometrics, when trying to decipher patterns of relationship and behavior. However, this is not how the world really works: events do not dictate behavior; instead, they are the product of behavior. What really causes behavior are the interactions between the elements of the system. In diagrammatic form:

systems ( patterns of relationships ) ---> patterns of behavior ---> events ( snapshots )

It is commonly recognized that the power of statistical models is limited to explaining past behavior, or to predict future trends ( as long as there is no significant change in the pattern of behavior observed in the past ). These models have little to say about changes made in a system until new data can be collected and a new model is constructed. Thus, basing problem-solving upon past events is, at best, a reactive effort.

On the other hand, systems modeling is fundamentally different. Once the behavior of a system is understood to be a function of the structure and of the relationships between the elements of the system, the system can be artificially modified and, through simulation, we can observe whether the changes made result in the desired behaviors. Therefore, systems thinking, coupled with modeling, constitutes a generative --rather than adaptive-- learning instrument.5

Thus, according to Senge:

"Generative Learning cannot be sustained in an organization if people's thinking is dominated by short-term events. If we focus on events, the best we can ever do is predict an event before it happens so that we can react optimally. But we cannot learn to create. "( 1990, p. 22 ) [emphasis added]

LEARNING IN ORGANIZATIONS

Once we embrace the idea that systems thinking can improve individual learning by inducing people to focus on the whole system, and by providing individuals with skills and tools to enable them to derive observable patterns of behavior from the systems they see at work, the next step is to justify why systems thinking is even more important to organizations of people. Here, the discipline of systems thinking is most clearly interrelated with the other disciplines, especially with mental models, shared vision, and team learning.

Patterns of relationships ( or systems ) are derived from people's mental models --their perceptions about how the relevant parts of a system interact with one another. Naturally, different people have different perceptions about what the relevant parts of any one system are, and how they interact with one another. In order for organizational learning to occur, individuals in the organization must be willing and prepared to reveal their individual mental models, contrast them to one another, discuss the differences, and come to a unified perception of what that system really is.

This alignment of mental models can be referred to as developing a shared vision, as is discussed in the first part of this paper. It is possible that mere discussion among individuals may lead them to a shared vision but, because problems are often too complex, usually this exercise requires the aid of some skills and tools developed by systems thinkers. Whether simple or complex frameworks are used ( such as word-and-arrow diagrams or computer simulation ), they are essential instruments to developing a shared vision.

When groups of individuals who share a system also share a vision about how the components of that system interact with one another, then team learning ( or organizational learning ) is possible. First, they learn from one another in the process of sharing their different perspectives. There are many organizational problems that can be solved simply by creating alignment. For example, cooperation is a lesson that is often learned by people who recognize that they belong to different interdependent parts of the same system.

Second, people learn together by submitting their shared vision to testing. When complex dynamics exist, a robust shared vision allows organizational members to examine assumptions, search for leverage points, and test different policy alternatives. This level of learning often requires simulation, which is a much more specialized systems technique. However, if the problems faced by the organization are among commonly observed patterns which have been previously studied, archetypal solutions may be available to deal with them. Later in this paper, we will discuss an example using an archetype commonly referred to as "growth and under-investment."

THE FIFTH DISCIPLINE, A METANOIA

Systems thinking represents a major leap in the way people are used to thinking. It requires the adoption of a new paradigm. Although there is no such a thing as a learning organization, we can articulate a view of what it would stand for. In this sense, a learning organization would be an entity which individuals "would truly like to work within and which can thrive in a world of increasing interdependency and change." ( Kofman and Senge, 1993, p. 32 )

And according to Senge, systems thinking is critical to the learning organization, because it represents a new perception of the individual and his/her world:

"At the heart of a learning organization is a shift of mind --from seeing ourselves as separate from the world to connected to the world, from seeing problems as caused by someone or something 'out there' to seeing how our own actions create the problems we experience. A learning organization is a place where people are continually discovering how they create their reality. And how they can change it." ( 1990, pp. 12-13 )

But, as we shall see next, systems thinking requires skills and tools which can only be developed through lifelong commitment. Plus, it requires that not just one, but many organizational members acquire them. Thus, some of the authors refer to learning organizations as ‘communities of commitment.’

SYSTEMS THINKING SKILLS AND TOOLS

At the foundation of systems thinking is the identification of circles of causality or feedback loops. These can be reinforcing or balancing, and they may contain delays. But before we "close" the loops to distinguish among these terms, let’s examine two examples of flawed ( or incomplete ) thinking which take into account only partial relationships between elements of systems.

The first example is an unilateral perception of the arms race. The word-and-arrow diagram below illustrates, from the point of view of an American, the logic behind building U.S. armaments:

Foreign arms ---> Threat to the U.S. ---> Need to build U.S. arms 6

The diagram can be read as follows: The more foreign arms, the greater the threat to the United States and, thus, the greater the need to build U.S. arms to defend the country from these potential aggressors. This non-systemic view suggests that U.S. arms are a defensive response to the threat posed by other nations: "If only the other nations would reduce their armaments, then so would the United States."

The second example illustrates a simple view of the mechanism involved with adjusting the temperature in a room during a hot summer:

Current temp. too hot ---> Turning on the air-conditioner ---> Results in lower temperature

For all of us who know about the developments of the cold war, or who have experienced first-hand the extremely cold temperatures inside movie theaters in mid-July, it is no surprise that these two diagrams tell only part of the story. Yet, if asked to tell the whole story, many of us would draw alternative diagrams, instead of complementing these. Over time, systems thinkers developed conventions to illustrate relationships, and to capture the whole story in just one diagram. Moreover, they found it useful to distinguish between stories such as the ones told above.

REINFORCING FEEDBACK

The arms race is an example of reinforcing ( or positive or amplifying ) feedback. Not only do more foreign arms increase U.S. arms, but more U.S. arms also tend to provoke increases in foreign arms. One reinforces the other:

Image2.gif (3839 bytes)

Although reinforcing feedback is commonly labeled as "positive" or "amplifying," this does not carry any value judgment. It simply means that a change in one part of the system causes a change in another part of the system which, in turn, amplifies the change in the first. Things do not always have to grow either. For example, a reduction in foreign arms will reduce the threat to Americans, which will probably cause a reduction in U.S. arms, which is likely to lead to further reductions in foreign arms ( since U.S. threat to foreign nations is reduced. )  

Image3.gif (2978 bytes)

By itself, reinforcing feedback leads to either exponential growth or decay.      

Balancing Feedback

Controlling room temperature is an example of balancing ( or negative or controlling ) feedback. In this case, a change in one part of the system causes a change in another part of the system which, in turn, counteracts the change in the first:  

Image4.gif (4217 bytes)

If the Perceived Gap is positive, i.e., Current Room Temperature is greater than Desired Room Temperature, the A/C is adjusted upwards increasing the flow of colder air, thus reducing the gap. This is a balancing system because more adjustment means less gap, not more ( unless, obviously, the adjustment is made in the wrong direction! ). The leverage point in this system is desired room temperature. If it is set too low, as seems to be the case in shopping malls and movie theaters, the resulting room temperature may be too low for the casual wear people tend to use during the summer.

By itself, balancing feedback leads to goal-seeking behavior.

Image5.gif (3520 bytes)

 

Delays

The time dimension is another factor which tricks people who fail to think systemically. For example, because it takes time to build up foreign arms, an American may not perceive that action as resulting from a response to increases in U.S. arms, but rather as an independent aggressive initiative. Thus a more accurate representation of the arms race would be:    

Image6.gif (4520 bytes)

Sound systems thinking requires the utilization of a combination of reinforcing and balancing feedback loops, and the accurate identification of delays. Complex systems are composed of multiple feedback loops laid upon one another. Often, the behavior of
the variables in these systems can only be understood through simulation. But, before we discuss simulation, let’s recognize the existence of certain archetypal structures which are commonly found, and for which behaviors are already well understood. 

 

System Archetypes

A number of system structures or patterns of relationships are commonly found in a variety of settings. Some of these have been carefully studied, and their patterns of behavior and leverage points have been identified. Senge discusses them in The Fifth Discipline, Appendix 2 ( pp. 378-390 ):

  • "Balancing process with delay"

  • "Limits to growth"

  • "Shifting the burden"

  • "Eroding goals"

  • "Escalation"

  • "Success to the successful"

  • "Tragedy of the commons"

  • "Fixes that fail"

  • "Growth and under-investment"

The arms race discussed previously could be used as an example of the "Escalation" archetype if we told the story using two balancing feedback loops, instead of just one large reinforcing feedback loop:

Image7.gif (5045 bytes)

The management principle derived from it is to look for a way for both sides to win, since their continued competition will lead to great costs and inefficiencies. Cooperation or mutual understanding is called for.
A practical application of a combination of the "Growth and under-investment" and "Eroding goals" archetypes was recently applied in a strategic planning effort for the Office of Disabled Student Services ( DSS ) of the University at Albany, State University of New York. Appendix A contains a copy of the analysis that was done for DSS. In this study, the authors suggested that the only way to respond effectively to the increased demand for services for disabled students at the University at Albany would be by increasing work capacity. Although this insight was not particularly dazzling by itself, when coupled with an evaluation that in the absence of resources to increase capacity, there would be slow but unequivocal tendencies to allow for the erosion of the quality of services traditionally offered by the Office, DSS’ leadership recognized that this process was already in place, but no one had really noticed it. This is because there are delays in the system.

When system archetypes apply, it becomes easy to focus on high leverage points, and to identify and avoid symptomatic solutions to real problems. This is because the analysis which serves as the foundation for the archetypes has already been done. On the other hand, when the systems under study are more complex because they are composed of a combination of structures, it becomes important to build models and to simulate to confirm assumptions about behavior.

 

Modeling and Simulation

Model building involves the conceptual formalization of mental models about the interrelationships between important elements in a complex system, for the purpose of examining the behavior of the variables of interest. Unfortunately, a great deal of modeling training, and experience is required to build good models, even simple ones. For this reason, so far, the literature in systems thinking for learning organizations has only traced a few steps in this arena. Usually, when modeling work is required, professional modelers are involved in the analysis to serve as the interface between those who know the system ( the clients ), and the mathematical formalization of the model.

The distinction between qualitative and quantitative systems thinking is commonly made by referring to the former as soft and the latter as hard system dynamics. At present, the contribution made by Senge to the field of organizational learning has relied primarily upon soft system dynamics. However, it is important to emphasize that the knowledge available today in the form of general principles, archetypes, etc., is the product of 30 years of hard system dynamics research and development. Thus, in general, the development of knowledge in systems thinking is highly dependent upon the latter, while its application has been very successful in the former.

Yet, system dynamics technology has progressed tremendously in the last few years. The availability of low-priced, user-friendly software for PCs ( such as Stella II, produced by High Performance Systems ) is extending the realm of quantitative analysis to amateur modelers. Moreover, the skills and tools needed are becoming available in a variety of settings, including K-12 education. Still, only a handful of people qualify as professional modelers, a fact which should serve as an alert with respect to the quality of the modeling being done in the field.

 

Micro-Worlds and Games

Where formal models do exist, they serve the function of a learning laboratory for managers. Some of the commonly used micro-worlds are:

  • The People Express simulator

  • The Boom & Bust game 

  • The Beer Distribution game 

  • Fish Banks

  • Stratagem

Each of these captures the dynamics of different systems, with different behaviors, leverage points, principles, etc. For example, the Boom & Bust and Beer Distribution games deal with different dynamics of the business cycle. Fish Banks, on the other hand, is modeled after the tragedy of the commons problem. In Stratagem, players make decisions about investment and consumption practices which carry short- versus long-term tradeoffs.

In each of these games, the objective is to understand the nature of the system at hand, and to extract some lessons about how to improve the conditions of the system or how to avoid problems inherently associated with the systems because of the nature of their structures. The underlying message is that structure determines behavior, and people can generally learn to identify what has to be done to deal with problematic behavior by "playing" with the system until they "understand" how it behaves.

blog comments powered by Disqus