Emergence: The Connected Lives of Ants, Brains, Cities, and Software, by Steven Johnson.
Emergence is one of those slippery ideas that is hard to define, other than to say “I know it when I see it”. Loosely it occurs when a system containing many interacting “atomic” components exhibits patterns at a higher level of abstraction, especially when these patterns are hard to explain in terms of atomic interactions. So the behaviour of the air in a balloon is fairly easy to describe as the average of the movement of all its individual molecules, but the behaviour of an ant colony is more than just the sum of the behaviours of the individual ants. We say the latter is “emergent” while the former is not.
The popular science press has got a lot of mileage out of emergence. There are certainly a lot of “gee-whiz” stories you can tell about it. Ant colonies, human societies and the human brain are all remarkable things and there is something tantalisingly mysterious about how they work. Steven Johnson writes about these things, and while his prose is good and his topics engaging, he ultimately doesn’t provide much insight into how it occurs. The closest he gets to an explanation is to offer five “fundamental principles”:
- More is different.
- Ignorance is useful.
- Encourage random encounters.
- Look for patterns in the signs.
- Pay attention to your neighbours.
Complex behaviours arise in systems of many small atomic parts. Macro-behaviour is different to micro-behaviour.
Interactions at the atomic level should be simple.
The interactions between individual atoms are ‘random’ or ‘arbitrary’.
Atoms react to certain gross properties of the greater system.
Atoms act based on local information from neighbouring atoms.
At least, this is my interpretation of his principles. He never really expresses them in general terms. While these do describe some of the common features we see in complex systems, they do not work very well as design principles because they still fail to distinguish systems that exhibit real ‘complexity’ (like an anthill) from those that do not (like a balloon full of air). Studies of systems like cellular automata show that complexity lies on a narrow edge between vast plains of ordinariness.
Johnson does talk about efforts to engineer emergence in software systems (such as Slashdot’s model for group filtering) but it doesn’t really identify why one such system works but another runs amuck. This example is especially relevant to us as designers because it is a system that incorporates human choice and self-interest. Indeed, it is a system which many people are deliberately playing to “win” — or to grief. Designing such a social system to optimise the participants’ experience is as much about game design as anything else. It requires both an understanding of game theoretic and psychological models of multiplayer interaction. A simple theory of emergence is unlikely to help here.
However there are less complex forms of emergence in our games that can be understood in terms of Johnson’s rules. He cites the example of Eric Zimmerman’s game Gearheads (which really deserves a revival). In that game players do battle with one another with armies of wind-up toys. There are twelve different kinds of toys you can deploy. Individually they each have a simple behaviour, but they react to each other so combinations of toys have emergent behaviour. Zimmerman describes his delight at “[seeing] a player take what you’ve designed and use it in completely unexpected ways.”
I share his delight in such situations, both as a player and as a designer. Emergence is indeed fascinating, but it seems that a real understanding of it remains outside our grasp. Maybe it is for the best. Perhaps if we understood it, it wouldn’t seem so remarkable anymore.