Ants, Genes, and Robots

Have you ever watched an ant trail and wondered how the apparent order in these insects come about? Ants are determinedly running back and forth, carrying food and building materials – somehow, you may have thought, this order must have been created. You might have imagined an ant queen ruling over her kingdom, or ants that are genetically programmed to perform their tasks. Indeed, your imagination might tell a lot about yourself1 – as it does about the French revolutionary Latreille, who thought that the colony has “a single will, a single law” based on the love the ants feel for each other.

As Deborah M. Gordon’s recent book “Ant Encounters” shows, the reality might be even more fascinating than Latreille’s altruistic phantasy. Ant behaviour, she writes, is determined by “interaction networks”: “An ant colony’s behavior is guided by a pulsing, shifting web of interactions, in which the pattern of interactions is more important than the content.”

“Understanding how ant colonies actually function”, Gordon writes in an article for the wonderful Boston Review, “means that we have to abandon explanations based on central control”. Each ant responds only to its immediate surroundings and to its interactions with other ants nearby, yet from this interaction network, coordinate behaviour emerges.

One of the most fascinating parts of “Ant Encounters” is devoted to the question how ants communicate. “An ant uses its recent experience to decide what to do. The pattern of interaction itself, rather than any signal transferred, acts as the message”, writes Gordon. It’s not important what ants tell each other when they meet, but simply that they meet.

The author herself reminds us of the stunning parallel between ant behaviour and the self-organization that forms the human body: “Ant colonies, like genes, work without blueprints or programming”, she writes. Just as in ants, the messages of neurons are not transmitted by one neuron, but a multiple. A single neuron can only send an excitatory or an inhibitory signal, or not fire at all. Yet one excitatory signal is not enough, just as one ant can’t tell another what to do – a whole pattern of interactions is necessary to trigger an effect.

On the other hand, Gordon points out where the scientific strife to create cognitive systems still falls short of its aspirations. Engineers have started to model robots after insects, and ants in particular. But even as robots communicate amongst each other to coordinate behaviour, they are far from living beings, writes Gordon: “[T]he complexity of complex biological systems is not what makes living systems unique. One way that living systems are unique […] is that they cause their own development and activity.” A robot is still programmed to achieve a certain goal – an ant can change its task by simply encountering enough nest mates.

Deborah M. Gordon’s “Ant Encounters” gives a fascinating insight into the organization of an ant colony. Most of all, however, it is a great read because it inspires to question common place understandings of communication and organization, far beyond the world of insects.

Deborah M. Gordon: Ant Encounters. Interaction Networks and Colony Behavior. Princeton University Press. 2010.

Crossposted from the BeTA Lab website. BeTA Lab is led by Dr. Sennay Ghebreab, who teaches my course Information, Communication, Cognition. The lab operates “at the crossroad of the brain sciences and information technology”.

  1. If this intrigues you, you might be interested in Diane M. Rogder’s “Debugging the Link Between Social Theory and Social Insects”, which explores the link between political fashion and the interpretation of insect behaviour in depth. []

You may also like