Chaos Theory for Sustainable Design

John Body, assistant commissioner for information management for the Australian Taxation Office and CMU School of Design research fellow, gave a really interesting talk on understanding design through chaos theory yesterday. I'll summarize his thoughts.

The discipline of design is changing its face because design thinking is being applied to more complex challenges, challenges with multiple intents and many stakeholders that have to deal with experiences, products and services, processes, technologies, and people. The bigger the system you are dealing with, the more design needs different tools and techniques.

There are currently several ways of working with this complexity: strategic conversation, systems thinking, systems mapping and modeling, management theory, complexity theory (systems of many agents), and what John discussed: chaos theory. Chaos theory really began in 1961 with Edward Lorenz and really took off thanks to computing. Computers could run the many simulations and iterations that chaos theory requires. It could be shown over time that small changes have large consequences to systems (the famous butterfly effect).

There are four principles of chaos theory:

• Order and Chaos. Both things are in all systems, but too much of either thing is a bad thing. You need the proper balance. Too much chaos leads to being out of control. Too much order is equally unsustainable and leads to rigidity, a lack of variety, and ultimately death. Things get more interesting as they reach the edge of chaos, because there are more states the system or object can move into. Chaos has more variety. A glass on the edge of the table is more interesting than a glass in the center of the table.
• Attractors. Attractors are things within a system around which other things, people, or activities cluster. Attractors give order and form to systems. It is much easier to work with attractors than against them.
• Fractals. Fractals are about zooming in and out, showing a macro view, then a micro view. The two views are similar, but not identical. You don't find simplicity by zooming it; there is always more depth. Too high a view is bad. So is too low.
• Non-Linear - Bifurcation. Systems are non-linear and seemingly direct paths often diverge, going in different places than what you expect. Expect the unexpected.

So what does this mean for design? Here's some lessons that came out of the discussion.

• Working at extremes, too high or too low, isn't very effective. You can either get locked into categories or else get out of control. You need to dip in and out of chaos and order.
• Sometimes a design team needs to stir things up, sometimes it needs to provide order. You need to know which one is required and use different strategies for each. If a product is too stable and at the end of its useful life, it might need to be disordered.
• In any project, define the attractors. Seek out what attractors have been missed and which have operative force. Find the relevant ones and use them. Look for attractors that offer opportunities because they are neglected. There are often unspoken attractors like values that can affect a project and put blinkers on evaluation. Formal structure in organization can be a key factor of attractors, but isn't often the primary one.
• If something is becoming more stable, that usually means it's working.
• At what level do you begin looking? You need to zoom in and out during the design process so that you can see multiple levels of the project. Zooming can break down the scope of the project. If you are too high, you get scope creep. If you are too low, you get lost in the weeds of detail.
• Encourage people to push towards instability. Non-linear shifts can lead to new innovations and inventions.
• Design for the rare 1% of the time, not just the 99% normal times. You have to design for sub-optimal environments. Not efficiency, but redundancy.

By understanding chaos theory and its implications, you can design so that the system continues to be successful, not just one product of the system. Chaos theory helps us understand how you can sustain a system over a long period of time: by getting the right balance of order and chaos, by working with attractors, by looking at multiple levels of the project, and by expecting the unexpected.

We can use chaos theory to support what we already know. But we can also use it to add something extra. In all systems, there are a whole lot of elements working randomly, but somehow all working together. Everything is interconnected, therefore unpredictable things happen.

Originally posted on Friday, October 29, 2004