Perspective on Knowledge: The challenge of emergence
Some birds flock in remarkably coherent and resilient ways. Hundreds of birds will maintain a fluidly morphing yet oddly consistent shape even as the wind and the landscape changes. The shape of these murmuring creatures may bend or twist, but it looks as if a single shape is being twisted or pulled, with all the pieces seemingly coordinated in their unpredictable movement.
This impresses us because we know the birds are not communicating over headsets to work out where each one should go. These flocks are emergent phenomena: They arise from simple behaviors that on their own have nothing to do with creating a particular multi-bird shape. The birds are trying to maintain some set distance from each other. When you put them all together, you get a phenomenon—a shaped flock that shares no properties with the birds themselves or even with the rule they’re following that produces the shape. Flocks are emergent.
A simple start
That flocking behavior strikes us as cool. In fact, I secretly hope that our autonomous vehicles will exhibit flocking behavior without having been programmed to, simply because that too would be cool. We’re awed by this sort of emergent phenomena in part because we expect complex effects—such as a resilient geometry of birds—to have complex, or at least complicated, causes.
That termites can build towers for themselves to a height of 8 meters is also an example of emergence. If you put together the brains of all the termites in a termite colony, you still wouldn’t have enough mental power to figure out who goes first at a stop sign. Yet, with the scraps of brain doled out to these tiny insects, they together can build structures 2.5 human stories high. And then they have enough synapses left to farm fungus.
The termite towers, similar to bird flocks, result from rules far simpler than the formations they result in. For example, a termite mound begins because termites follow two instructions: If you find a wood chip on the ground, you (the termite) should pick it up if you aren’t already carrying one. But, if you are already carrying a wood chip and come across a new one, drop the one you’re carrying and move on. These two rules are enough to generate random sets of small piles. When one pile gets a little larger than the others, it’s more likely to be encountered by chip-carrying termites, so it garners yet more chips, making it even more likely to be found by chip-carrying termites. A mound thus begins to emerge without a single termite ever thinking, “Hey, let’s build a pile!” If you look only at the behavior of individual termites, you don’t see a tower coming.