Thinking in Systems: A Primer by Donella Meadows December 15, 2009Posted by Sean Simplicio in Systems Thinking.
I feel somewhat ripped off after reading this book: its content is so valuable that it’s almost a crime to read it and not spend the next several months unlocking its mysteries. Presented at a fairly high level, Thinking in Systems helps the reader map out systems (defined as interconnected sets of elements coherently organized an ways that achieve something) in an effort to truly diagnose problems, be they societal, business, familial, etc. The author describes the essential elements of any system (stocks, flows, and feedback loops), discusses common system frameworks (renewable vs. non-renewable stocks, competing feedback loops, etc.), and also enumerates numerous system traps that recur again and again (the tragedy of the commons, the drift to low performance, escalation, etc.). The book is so jammed-packed with such frameworks and mnemonics, I felt dizzy when I finished it.
The book is called Thinking in Systems: A Primer for a reason. Meadows does a fine job of helping the reader conceptualize systems and their different parts. But practice is definitely necessary here: it can be harrowing to even start to map out a system in the way she describes: where does one draw the boundaries? How deep should one get? The goal of systems thinking is to accurately depict how systems function, in an effort to rightly describe why certain outcomes happen. The problem is that there are a lot of assumptions that one need make about the amount of detail to map out, the causality and connection between certain elements, the relative importance of others, etc. In some ways there’s a chicken-and-egg problem: to map effectively, you need to know how elements interact, but in many cases system mapping helps you determine how the elements interact. This is where the practice comes in. Meadows does warn us that models are never perfect, and reminds us that one has to start somewhere. One rule of systems thinking is that systems can change over time; we must then be prepared to adjust our own models during the process of their creation.