Book Review-Thinking in Systems: A Primer By Donella Meadows

I had worked in companies that supplied the transportation industries, both aviation and automotive industries, which was where I became acquainted with the idea of systems engineering. There was also a school of Industrial and Systems Engineering at my alma mater, Georgia Tech, but I was too distracted by my gradual studies to pay any attention to what they were doing at the time. I was curious but never curious enough to dig into the granularities of the idea of systems thinking until later in my engineering career, when I organized a technical session in a conference, the topic was an introductory session on systems engineering. That experience piqued my interest, which is how I came to possess this book by Donella Meadows. It was recommended to me by people in the systems engineering area as a starting point to trying to understand what the idea of systems thinking.

The book sat atop of my To Be Read list until I finally got around to it.  While the book was indeed an excellent primer for someone who is a complete neophyte in the area, the book also left much to be desired. At this point I am not sure whether it is the explanation in the book, my own lack of understanding of the foundations of the area, or whether I was just assuming too much about the subject which left me grasping.

The author split the seven chapters into three parts. Part 1 defines the underlying structures, foundational beliefs, and behavior of systems. Part 2 Digs into the insights from the author’s work in the area, giving us behind the curtains views on what the author believes to be the key insights that she had internalized in her years of practice as a systems thinker, as well as the common pitfalls and false steps that system thinkers tend to fall into in their practice. The author also appeals to our curiosity when she tries to sell the readers on how the pitfalls and false steps can be turned into advantages. Finally, Part 3 is where the author lists the tactical points in a system view where intervention into the system can be made, where the systems thinker can actuate some change by applying leverage to change the trajectory of the system. The last chapter is the sales pitch, this is the chapter where the author makes the argument that if we were perspicacious about the world around us, we can become prescient about how the world will behave if we analogized all that we know to a system, and thinking in systems will clarify how we understand and perceive the world around us.

As an electrical engineer, one of my great downfalls occurred when I was preparing for my doctoral qualifying exams. I failed because my theses is based on the theory and practice automatic control systems. My committee allowed me to bone up on the subject and retake the exam. I made it my mission to understand the subject and I passed the exam the second time around.  Thus, automatic controls concepts, theories, and practice has been at the center of my applied engineering career; indeed, my world view is informed by the automatic controls structure and framework that I had worked so hard to understand. It is my habit, it is actually more of a reflex, to draw connections between whatever I am working on, technical or not, in automatic controls terms. It is therefore natural that I drew the analogy between amorphous and deliberately ambiguous systems thinking concepts that I had read about in this book and the technical and mathematical automatic controls ideas that have become part of my procedural thinking.

After having read Part 1 of this book, it became obvious to me that the idea of system thinking is based on the feedback control paradigm from automatic controls, whether it was obvious to anyone else or not. The ideas presented in the book of stocks, flows, and block diagrams loaned themselves readily to the concepts that are in my mind, which made me jump into analogy mode. The ideas of stabilizing feedback and reinforcing feedback became negative and positive feedback loops. The ideas presented in the book about the importance of time delays became time constants that are inherent to automatic control systems. The descriptions of the intrinsic nonlinearities in systems parallels nicely with how I understand the coupled system plants are modelled in automatic controls. The idea of self-organizing systems is indeed what automatic controls designers seek to do on a local scale with adaptive controls, amongst numerous other  techniques.

After I discovered that the recognized originator of the precursor to Systems Engineering and  Systems Dynamics, was  Jay W. Forrester (Forrester, Industrial Dynamics, 1961), (Forrester, Principles of Systems, 1968), an electrical engineer who had worked on the magnetic core memories, I was more convinced than ever that Forrester had taken the strict and mathematical practice of designing automatic controllers and eased the technical constraints to the ideas used in automatic controls so that can be analogized and applied to creating a perspective that is used to make sense of reality through the lenses of engineering controls theory. Indeed, that is the framework and perspective that I have taken towards learning about system thinking, taking advantage of my existing knowledge and using its heuristics to gain understanding of systems engineering.

My idea of using my prior knowledge in automatic controls fell apart somewhat as I read further into this book. There were major gaps in the analogy that I was drawing between system thinking and automatic controls. In consulting with some people, they pointed out that the basis of automatic control systems is assuming linear behavior, which is not true, nonlinear systems are a very large part of automatic controls writ large. We use techniques to treat nonlinear systems as point wise linearized to simplify the mathematics or we use piecewise linear models, once again to simplify the mathematics, or we go through complicated nonlinear solution methods to solve nonlinear systems.

The author also staked a claim in the book that while engineering controls places prediction and control at the center of its focus, system engineering does not. I can accept that paradigm, but I am confused as to how the system engineers define and identify what they mean as systems? More importantly, since they place such a premium of the value of feedback paths — as they rightly should — how do they identify the feedback loops in the amorphous and ambiguous “plant” or system? One of the lessons I garnered from automatic controls education is that the best laid plans with regard to the controls are often obliterated by the unmodeled dynamics that are hidden and they are always the nasty surprises when they are inadvertently excited by an input. Many test pilots have given their lives because they were flying airplanes that did not respond to calculated controls that failed to take into account the unknown dynamics.

In reading the chapter on traps and pitfalls for system thinking, and the chapter on points of intervention for systems, the idea struck me that both chapters, while valuable, seem to consist of heuristics and ad hoc solutions and observations. They are tools that can be used to resolve issues in designing or resolving system behavior, they are not systematically consistent solutions, which means that there is a possibility that their use might also introduce undesirable bahviors. This is the point where I realized that the author is serious about abandoning the prediction and control purpose of automatic controls. The idea here is to design and identify the system in a piecewise and progressive fashion so that the designers do not fall into the trap of creating system designs based on ever changing system parameters in a model. The system can never be modelled for all time, it is always evolving, which makes sense but quite disconcerting for the controls engineer in me.

Finally, I am curious to understand how Forrester and the systems engineering colleagues went from point A to point B, how the ideas of automatic controls evolved into the systems thinking that Meadows is writing about. I am quite curious, and I would appreciate any advice or resources that I can consult with in order to find a path to those answers. In addition, I wonder if the later concepts that have become integrated into automatic controls have been similarly evolved and broadened into systems engineering; ideas about stability, adaptation, controllability, and observability since Forrester published his work in the 1960’s.

This was a very readable, concise, and well written book. It serves as a good introduction to the idea of system thinking. Unfortunately, it raised more questions in my mind than it answered, although it could be interpreted as being a fortunate event, since it will allow me to dig further into the granularities.

References

Forrester, J. W. (1961). Industrial Dynamics. Waltham MA: Pegasus Communications.

Forrester, J. W. (1968). Principles of Systems. Waltham MA: Pegassus Communications.