Yesterday I called ants “mindless,” but that’s not really fair. In fact, they have the largest brain in proportion to their body size in the entire animal kingdom. The average ant brain has about the same processing power as a early 1990’s Macintosh. (When ants are thinking it looks like an hourglass turning over: .) They invented farming, herding, and slavery far before these were a glimmer in a bipedal monkey’s eye. But ants’ real genius comes not from their brains, but from the collective wisdom of emergence. The philosophical concept of emergence can be summed up thusly: the creation of a complex system or pattern by a multiplicity of simple individual actions. Crystals, those highly geometric forms, are forged from the random movement of individual molecules, and hurricanes do not take their classic galaxy shape by choice but instead are crafted by the force of individual winds in and around them. Emergence also explains swarming activity. A flock of birds or a school of fish is not made in a mold, but is rather made by the choices of many individuals to stay close to others of its kind. Ant colonies are particularly beautiful examples of emergence. The queen isn’t telling her millions of workers what to do. Rather, the workers are following pheromone trails and interpreting each other’s actions and chemical signals, and in acting on their own, create an incredibly complex and mathematical organization.
For the largest and most complex organization of individuals in the animal kingdom, we look to an old favorite: the leafcutter ant. Leafcutter societies can tell us a great deal about how emergent systems, like this here World Wide Web, work. They can also tell us a bit about that other great organizer of random information: the human brain.
Leafcutters are not one ant species but 41 species in two genera. The leafcutter system is such an effective one that it seems to have evolved independently twice. They are the first farmers of the world, cutting and carrying pieces of foliage below ground to mash up and feed to a symbiotic fungus, which in turn feeds their larvae. They are the primary consumers of foliage in the South and Central American forests, shearing off 15% of its vegetation; each colony of 8 million ants does the same damage to green matter as a cow. Different species of ant cultivate different fungi, but the fungi are entirely dependent on the ants, and vice versa. The two are so reliant on the other that ants can sense when a certain plant is toxic to the fungal stock, and cut off supply of those leaves. The fungus, in addition to supplying food, also breeds an antibacterial microbe which keeps the ant healthy. It is the same bacteria responsible for most of the world’s antibiotic medicine today.
Leafcutter workers are organized into at least four castes. The smallest workers, called the minima, tend the fungus gardens, chewing the foliage into a digestible pulp and clearing the fungus of harmful molds. They will also ride on the backs of the mediae as they forage for leaves, in order to chase off the parasitic wasps and flies that try to lay eggs in their heads. Between the minima and mediae are the minors, the scout caste which patrols the area looking for intruders to dispatch. The majors, largest of the bunch, are soldier ants built to defend the nest and to clear debris for the incoming mediae.
Between the four, they create the largest structures in the ant world, a network of tunnels, farming chambers, nurseries, entrances and ventilation ducts that forms a geometric structure often 20 feet deep. Waste deposit sites outside the nest are always the maximum distance from any entrance. It is a city without an architect, a building without blueprints, its construction based entirely on the choices of individuals cued by the chemical commands of other ants.
Scientists call this “swarm intelligence,” and have begun to use it to design algorithms which optimize the efficiency of computational systems, and which may eventually lead to a form of Artificial Intelligence. Here’s how it works in the ant world: A single ant leaves a pheremone trail as it goes. If it finds food, it beelines back to the nest, leaving a different and stronger pheremone trail for other ants to follow. These scents evaporate quickly, so only the trails traveled by many ants will survive long. The shorter, more efficient routes are better traveled and have a stronger scent, and so are more attractive to individuals. The less efficient trails disappear.
It is essentially a computer made of many tiny computers. In an article in The Economist, the researchers who discovered swarm intelligence and developed the algorithms based on it discuss the hundred of known uses for them, including optimizing the routes of information bytes within a computer or the routes of delivery trucks in a city, of augmenting the capability of a computer to find mechanical failures or to make medical diagnoses. The ultimate goal: a robot made of a swarm of tiny robots, an artificial intelligence based on on the human brain but on the emergent wisdom of ants. But while this [skin-crawlingly creepy] system might be the most efficient route to artificial intelligence, it may not be very different from the way our brains work after all. Some neuroscientists believe that our brains themselves function like a hive, a “swarm cognition” of neural signals that follow each other at the speed of electricity. Consciousness itself be a form of emergence as the cells behave like ants and bees, directing each other toward via chemical signals toward reasoning. Within a human, a swarm of tiny humans.