The power of self-organizing systems

Before I went to university to get my engineering degree, I taught myself to program in C on my old 386 computer. One program I especially enjoyed working on was based on a Martin Gardner column from an old issue of Scientific American. The program was based on a large, two-dimensional array of numbers (between 0 and 15, say), and displayed visually as a field of pixels each number being represented by a different colour. The program would begin with the numbers (and thus the colours) assigned randomly, but would step through the array and apply a rule to each member. This rule would determine whether the pixel would change colour in the next refresh of the grid. The rule was generally simple, such as if a member directly adjacent to you is larger than you by one, increment yourself by one. (The numbers wrap around, so that for the sake of this program 15 + 1 = 0.)

It was fascinating to run the program and watch the initial digital Jackson Pollack painting evolve into something different. For example, using the sample rule above, order formed with a twinkle of change starting here and there, growing more frequent until small loops were created where the colour was changing each cycle. These loops grew wider and wider, drawing more pixels into their vortex until the loops collided with each other and every pixel in the grid had fallen into some sort of step with its neighbor. Order had grown out of chaos. Different initial rules resulted in different forms of order.

I didn't know it at the time, but I had created a self-organizing system (SOS). This is a (relatively) new area of mathematically based scientific investigation with applications in all fields of study. The basic premise is that small elements or components in a system, working with each other on a local level, create a higher-level order that is not easily predicted. There can be larger, global forces working on elements in an SOS, but the complexity and unpredictability of the comes from the local interactions. In the physical world there are plenty of examples of SOSs: the formation of a snowflake, the patterns made by the wind on a sand dune, the shapes of the galaxies, or the beads of condensation on a bottle of beer.

But you can also see SOSs comprised of elements with more complex behavior than molecules, stars or grains of sand. An ecosystem is an SOS. Each living thing reacts to and is affected by others in its immediate surroundings. The brain is composed of cells which selectively receive and transmit signals to other cells through the synapses. An economy is comprised of individuals and corporations making innumerable exchanges with each other, each believing to have gained value in the trade. Each of these SOSs have larger macro effects that can be seen and quantified, but the power and dynamism of these systems is in the low-level exchanges and interactions.

Humans have a hard time understanding self-organizing systems. They're chaotic and unpredictable, and for governments, uncontrollable. Most of the 20th century's great man-made catastrophes were the result of various vicious leaders attempting to impose 'order' on the 'inefficiencies' of a unregulated society. Though the brutality of the methods used to impose this order have (mostly) changed in the 21st century, the essential goal has not. Modern governments don't trust people to create their own order. The natural feedback loops that encourage some behaviors and discourage others are warped by interventions that regularly harm far more than they help. On the other hand of course, the most natural form of social organization is the gang, tribe, or clan. It's a form of order, and it's quite resistant to change. We clearly need some form of regulation -- where the right balance is is the issue.

My mind has been wandering around these ideas for the past few weeks. They're certainly not original, but have given me a different perspective on the libertarian issues I'm interested in right now. I'll be coming back to them regularly.

Posted by Bruce Gottfred at February 4, 2004 10:03 AM | TrackBack