Mill of Humanity – Swarms, emergence and us

Have you ever wondered how schools of fish move in mesmerizing unison? How termites build their colossal and complex homes? Where consciousness comes from?

At the heart of these phenomena lies one of the most fascinating concepts in science. It’s called ’emergence’, and we can learn an important lesson from it as a species.

The things ants do

Ants seem to roam this earth for the sole purpose of demonstrating what emergence is all about. Ant colonies show a remarkable and robust ability to solve problems.

Some species are able to build bridges out of themselves to cross gaps in the road:


Or they form rafts – to survive floods, or travel over water:


They build stunning cities:
Concrete cast of an ant city

In witnessing a colony perform stunts like these, one might wonder which one of these critters is giving the orders.

Who’s in charge?

The answer is: none of them.

A single ant always seems to know exactly what to do. It never runs off to the “Executive Officer Ant” for instructions. The key lies in the fact that each ant is following a tiny set of rules. Trivial rules, dealing only with its local environment. An ant is just minding its own damn business, really.

When a legion of these simple creatures live together, though, magic happens. The group starts displaying astonishing feats of organisation. It acquires problem solving capabilities a NASA rocket scientist would be proud of.

Systems consisting of simple components sometimes exhibit properties which are absent in any of the components themselves. This is called emergence, and it’s the underlying phenomenon at work in ant colonies. The resulting collective behaviour emerging from such a system has a more dramatic name: Swarm Intelligence.

How do ants gather food?

To illustrate the principle, let’s look at the way ants form those characteristic single file lines when gathering food:

Leafcutter ants gathering food

One might assume that an ant found the food source earlier, went to tell the others to come help collect it. And then promoted one of them to executive officer in charge of coordinating the enterprise.

In reality, a few simple facts explain the entire event:
  • Ants deposit pheromones when looking for food.
  • Ants tend to follow the strongest pheromone trail they can find when they don’t carry food
  • Ants explore randomly if there is no food or trail to follow
  • Pheromones decay over time
  • Short paths take less time to travel along

Let’s break the process down in steps.

 Step one: random exploration

Random exploration
Ants go out to find food, randomly exploring their surroundings. They leave pheromones along the way, like Hansel & Gretel left bread crumbs.

Step 2: Gary finds food

Gary finds food
Gary finds food and picks up a portion of it.

Step 3: Gary goes home

Gary goes home
Gary follows own pheromone trail back to the nest. He releases more pheromones on the way. This reinforces his trail.

Step 4: Cindy finds Gary’s trail

Cindy finds Gary's trail
Cindy runs into Gary’s strong trail.

Step 5: Cindy finds the food as well

Cindy finds food
Cindy follows Gary’s trail and finds the food source as well. She picks up a portion. Gary arrived at the nest and leaves his food there.

Step 6: Cindy reinforces Gary’s trail

Cindy reinforces Gary's trail
Gary sets back out along the strongest trail he can find, the one he started himself. Cindy picks the strongest available trail and follows it which will lead her to the nest. Kelly found the food by following her own random path, but will abandon it for the stronger path laid out by Gary and Cindy.

That’s the basic principle. More ants will stumble upon the successful trails of others. Short paths get reinforced while long paths decay.

Path optimisation

Whenever an ant follows a trail, it might deviate from it. It might need to avoid obstacles like other ants. Trails are fuzzy because they’re based on ‘smell’. Some of these deviations happen to be shortcuts.

This leaves pheromones on a shorter path, and shorter paths take less time to travel along. So, the pheromones on those shorter trails have less time to decay. Therefore shorter paths are more likely to be followed by other ants and get reinforced in the process.

Effects like these cause the ants to work towards a shorter path between nest and food over time.

And so, a bunch of ants start walking between their nest and your picnic basket in a single file line, stealing food in a coordinated manner. They seem quite smart collectively by acting quite stupidly on their own. (One would expect major swarm intelligence at frat parties.)

The same sort of simple rules give rise to the examples mentioned at the beginning of this article. (Even the ant cities!)

Emergent swarm stupidity

There are loopholes in the ways ants behave. Take army ants, for example. These ants are blind. They start following nearby ants when they lose a foraging pheromone track. This can culminate in a fascinating disaster:

This is called an ant mill, a clear indicator that there is no CEO ant coordinating the whole. None of the individuals caught in the mill have a clue they’re not making progress. Unless they get some external stimulus they will keep walking till they starve. Any CEO would be giving his employees stern talkings-to when noticing the problem.

These mills can be huge. There have been cases where it took a single ant over 2.5 hours to walk the full circle – a circle with a diameter of hundreds of meters. Some people make a game of nudging ants in such a configuration (please don’t).

I’ll come back to this phenomenon near the end of this article.

Other swarms in the animal kingdom

Interesting as ants are, there are many other examples of emergent behaviour to be found in the natural world.

Bees building honeycombs

A beautiful example of the power of emergence is this:


A honeycomb’s intricate structure might tempt one to start looking for a genius architect bee.

In truth, bees sit next to each other and start out creating their own circular cell. They then heat up the wax, making adjoining cell walls fall flat, like soap bubbles meeting each other. More details here.

The point being that each bee does its own local thing, and a hexagonal structure happens to emerge.

Earlier in evolutionary history, honeycombs must have been a lot messier. Natural selection has favoured species that made the most of available space and resources. Hexagons are the perfect solution to the problem.

Bird flocks

Another example would be the flocking of birds. When the birds doing the flocking are starlings, the swarm is called a murmuration. Usually caused by a predator, this is a sight to behold in many parts of the world:

Each individual bird merely reacts to what its seven nearest neighbours are doing. The amazing fluidity and cohesion displayed by the flock is the high level product.

Sardine run

In the marine world, many species of fish form similar groups. Even though each fish is just minding its own immediate vicinity, these schools as a whole might well be mistaken for conscious entities.

This is on especially spectacular display during the ‘Sardine Run’, an annual natural event off the coast of South Africa. This event is a feast, though not for the sardines: they’re fed on by every predator in the vicinity. Many of which are in the vicinity at the time because they came from far and wide for the occasion.

The result – of which the word ‘epic’ is an inadequate description – is this:


Ant and termite colonies show many similarities. Termites build especially high-tech nests, as mister Brian Cox can explain to you:

Again, each termite just performs simple, local, even boring actions. The end product drives human architects to careers in music, or bicycle repairs.

Evolutionary Perspective

There doesn’t seem to be a consensus on the evolutionary mechanics explaining many cases of swarm behaviour. In this context, the often heated debate about group and kin selection rears its head.

In any case, no intelligent creatures had to evolve. That’s an important part of the puzzle. It’s easy to imagine small ‘mentality’ changes in the simple, local, behaviour of an ant. Such a small change might give rise to the emergence of complex swarm behaviour quite suddenly. One generation of ants roam on their own, the next is curious about pheromone trails. Small tweaks have the potential to generate a massive advantage.

I recommend The Selfish Gene by Richard Dawkins should this topic fascinate you.

Artificial Swarms

Science doesn’t mind stealing ideas from nature. Evolution has had ample time to come up with brilliant solutions to all sorts of problems. We’d be silly not to take advantage of its work.

In line with this tradition, swarm intelligence found its way to technology and engineering.

The key attractive features of swarms in engineering are:

  • Fault-tolerance. Swatting away at a couple of bees will not stop the colony as a whole to chase you off into the bushes. If a few members of a swarm fail to perform their tasks, it won’t matter much. There are plenty of others to maintain the behaviour of the group as a whole. This robustness is a much sought-after quality in technological applications.
  • Emergence. Creating many simple units to solve a problem is much easier than creating a complex single unit to do the same. Hence, emergence is very attractive to engineers without having to put on make-up and bat its eyelashes.

Artificial Intelligence

Swarms have been a research topic in computer science for a good while.

  • Ant Colony Optimisation. The way ants find a near-optimal path to a food source has been simulated inside computers. The approach proved applicable to a wide range of problems, network routing problems for example.
  • Particle Swarm Optimisation. A method of solving problems by having a swarm of candidate solutions ‘travel’ through the space of possible solutions. This method leverages the schooling/flocking mechanics we see in fish and birds to solve difficult problems.

Self-healing and self-assembling architecture

Fire ants are capable of forming surprisingly resilient structures out of themselves – like the rafts and bridges mentioned earlier. This has inspired some scientists to look into self-healing structures, made up of simple parts. Think bridges that repair themselves when they develop cracks or other weaknesses. Or even structures that build themselves from scratch:


It’s doubtful Apple will announce iSwarm tomorrow. However, imagine a personal swarm of nano-bots keeping you healthy. A swarm that pervades your body, doing maintenance. Doing repairs before you even know anything needs repairing.

Scientists are working on exactly that: swarms that can deliver medicine inside the body, perform surgery, hunt tumours, …

This might sound like science fiction, but robot swarms are already becoming a reality:

Swarms in fiction

Some works of fiction have featured swarms in prominent roles.

(Should you have come across other examples, I’d be interested to learn about them!)

As these examples indicate, humanity is starting to pay attention to swarms and emergence. We’ve started taking some lessons from nature on the subject. Great, but we must not overlook what might be the most important lesson of all. To see what I’m getting at, let’s do a little introspection first.

Consciousness as an emergent property

One particularly fascinating domain in which emergence plays a key role is one extremely close to our hearts: the human body.

Our bodies consist of a humongous collection of cells. None of these cells know where they are – which role they need to play within the whole. Each cell follows its own local rules – and out come the livers, eyeballs and appendices. Like schools and flocks of cells.

There is one area of the human body which deserves some special attention. It’s probably not the one you have in mind.

It’s the mind.

The cells that make up the brain are called neurons. A single neuron is no smarter than any other cell in your body. You will have a hard time pointing out a neuron that has your personality. When they get together in large numbers, though, ant-colony magic takes over. Our brains acquire the most spectacular emergent property of all: consciousness.

In other words, our intelligence is nothing more than the swarm intelligence generated by a colony of neurons.

This begs a multitude of questions. Are all swarms conscious at some level? Where is the threshold between appearing conscious and actually being conscious? Is there a difference?

These matters remain, as far as my swarm can tell, unresolved.

Mill of humanity

Zooming out, human beings are only slightly overcomplicated ants, named ‘Gary’ or ‘Cindy’. Human society is yet another swarm. Unfortunately though, it does not seem to be a very intelligent one.

We seem to be collectively determined to destroy our home planet. We like to think of ourselves as free, sophisticated and ‘more evolved’ than the rest of the animal kingdom. In truth, we all follow a limited and primitive set of rules, with little regard for indirect consequences. There’s every sign we’re all stuck in the human equivalent of an ant mill.

Complex systems are often very sensitive to small tweaks. This is the domain of chaos theory. Since our complex collective behaviour emerges from the relatively simple rules we follow as individuals, small changes in our mentalities can have dramatic high-level effects on our species.

This is good news: if enough of us follow the right set of principles, humanity can steer away from collapse. In contrast, if we keep following our neighbours without thinking, the collective outcome emerging from our individual behaviour might as well be the end of our species.

Are we any wiser than army ants? Time will tell.

Feel free to leave your comments, suggestions or (preferably constructive) criticism below!

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s