The inner simplicity of complexity using Cynefin and TOC

This article is a translation of its german original post by Paul Bayer.

Can complexity be defined? Some have tried to do this. And the result is only paraphrases. But "the meaning of a word is its use in language". There are two ways of using complexity, which we better separate:

1. We call something complex when we cannot understand or explain it.
2. We call systems with certain characteristics complex.

There is a connection between the two uses. Systems having, many components, with many, partly non-linear dependencies, and feedback show behavior that is difficult to understand, complex in appearance, which we cannot predict or influence. As soon as we can, the system loses its horror and it is no longer necessary to emphasize the complexity of the system.

We want to find a way to deal with complexity! We don't need to understand it in all its details, we just want to find a practical way to deal with it or to influence it.

Dealing with complexity

A pragmatic understanding of complexity therefore builds on the first way of use: a complex system (a situation) appears to us to be difficult to understand, uncertain, hardly predictable. But we want to find a way to deal with it! We don't need to understand it in all its details, we just want to find a practical way to deal with it or to influence it.

As human beings we often deal with complex situations, living beings, organizations.... That is part of our life. Dealing with complexity follows a certain pattern:  

1. Aspects of a system or situation are closed and complex to our understanding. If we want to cope with it, we have to experiment with it, find starting points and patterns etc..
2. We find out parameters, e.g. by experiments, how we can influence the system, e.g. an optimum, an operating range, a model, the dependencies. We may be able to find out rules. This is complicated, often requires experts, etc.
3. Some things we know and learn about the system. When we find a leverage point or bottleneck through which we can influence the system, its behavior, its functioning, it becomes easy.
4. Through more insight into a system, we find easier and more effective ways to explain or influence it. However, we must not simplify or exaggerate our explanations and actions too much. Otherwise, we overstrain a system and provoke chaotic conditions. Our simplified explanations and approaches repeatedly reach certain limits.

This process can be represented as dynamics within the Cynefin model:

Our cognitive process runs clockwise through the Cynefin domains.

The inner simplicity of complexity

The postulate of science and the Theory of Constraints is that complex systems have inner simplicity. They are ultimately based on a few simple laws. They can therefore be influenced in a simple way. But these laws and starting points are not obvious. "Simple" does not mean "easy" and requires hard work. The tangle of dependencies is difficult to resolve.

Theory of Constraints suggests that the most effective way is to focus only on the constraint that prevents the system from reaching its goal, the bottleneck.

In reality, complex systems are characterized by their strong dependencies. The Theory of Constraints suggests that the most effective way is to focus only on those of them that prevent the system from reaching its goal, the system constraint, the bottleneck. The less we have to intervene to achieve a goal, the more effective we become.

The more complex a system or problem is, the simpler the solution must be if we want to influence it effectively.

Uncertainty and stability

Some contemporaries try to control complex systems like (simple or complicated) machines. But trying to control all interactions in a complex system is nonsense. Even if we can influence it in a simple way, its peculiarities: dependencies, feedback, non-linearities ... remain and become noticeable in fluctuations and unforeseen reactions. We cannot fully control it (move into the complicated or simple domain). Any such attempt to override or "reduce" it will destroy it (move it into the chaotic domain). So we can control only a few crucial points and must not do more than necessary.

We must give space to complexity, self-organization, development, learning and adaptation processes, and provide buffers and margins.

For this we need suitable economic methods to determine the necessary size of these leeways, and mechanisms to keep the system in this area of operation. As long as we succeed, the system is stable, its behavior predictable within certain limits and can be "managed". In this sense, Deming said, "Management is prediction!“.