The Swarm Within at the Edge of Chaos

“By operating without a leader, the scout bees of a swarm neatly avoid one of the greatest threats to good decision making by groups: a domineering leader. Such an individual reduces a group’s collective power to uncover a diverse set of possible solutions to a problem, to critically appraise these possibilities, and to winnow out all but the best one.”

— Thomas D. Seeley, Honeybee Democracy (2021)

During late spring and early summer, a honeybee hive becomes a hectic, overcrowded place as multitudes of new bees emerge fully grown from their cells, ready to join their elders who have overwintered. Meanwhile, a new queen, still furled tight in the confines of a sealed cell, signals – through distinct noises, termed “quacks”– that she is ready to emerge. The time has arrived for the colony to split – an act of fission wholly necessary for survival in which the old queen, along with up to one half of the colony departs to find a new home.

The swarm takes flight, forming a living cloud that soon condenses into a loud and tremulous mass with an ever-shifting surface. Thousands of honeybees crawl over one another while enveloped in a haze of hundreds more coming and going. This phenomenon – which lasts a few hours to a few days – is at once unsettling and beautiful, offering a rare glimpse into the elusive inner workings of a decision.

While the swarm clusters at an interim site, hundreds of scouts fly in all directions searching for a suitable home. Fortunately, honeybees are excellent navigators and exacting judges of real estate. Each scout flies up to several kilometres, seeking a well-protected cavity in a high position with room to build a comb. Tree hollows often serve this purpose well, as might a gap between the walls of a building; a chimney can do quite nicely too, until it doesn’t. In the northern hemisphere, bees prefer a south facing entrance to make the most of the winter sun. The size of that entrance is important, too, as is the volume of the entire cavity. Too big and there will be little structural support for the comb, too small and crowding quickly becomes a problem. There are multiple considerations, each critical. Honeybees can neither store honey nor produce a new generation without a good site to build its comb, so it is not a decision to be made too hastily. Yet lengthy deliberation is dangerous – so long as the swarm remains at the interim site, it is exposed and vulnerable.

On the hunt for a new home, a swarm will often find itself presented with more than a dozen scouts dancing the praises of as many potential sites, and despite such hierarchal monikers as queen and workers, there is no top-down decision-making structure in play. Thomas Seeley is a biologist with decades’ experience studying honeybee behaviour. As he and his colleague Susannah Buhrman noted in the journal   Behavioural Ecology and Sociobiology (1999), “there is no omniscient supervisory bee that compiles all the evaluations and selects the best site.” It is the group that decides. What happens next relies on – and reveals – the intelligence of a swarm.

A scout, having identified a potential spot, returns to the swarm to perform a “waggle” dance rich with information. By its orientation and duration, the dance relays the direction of the site and its distance. The location’s quality, meanwhile, is conveyed by the dancer’s enthusiasm – the more ebullient the waggle and the more return trips the scout makes to the site, the better the accommodation. This enthusiasm is an excitatory signal, one that recruits other scouts to follow the dancer to the promising site. If convinced of its high quality, they return to perform the same dance. This piques the interest of yet more scouts who also visit the place. On the pattern goes, following the inviolable mathematics of contagion. Meanwhile, a scout that discovers a less compelling location will not dance as fervently, so provokes less interest. Fewer trips are made to that spot by fewer followers. If, amongst all the options, there is one particularly outstanding site, it will be advertised longer and by more bees, and a decision is reached quickly.

The availability of multiple good quality sites complicates matters. It’s not just the degree of infectious enthusiasm that matters now. Inhibitory signals also occur – down votes, if you will. An advocate of site “A” can stop a neighbour dancing for site “B” with a little head-butt or a brief vibration. In this way, small-scale feedback loops play out – support for competing sites rises and falls. As hours pass and then days, small differences are amplified, and even the slightest advantage of one place over another becomes critical, leading, inevitably, to the moment of quorum.

As honeybee scouts investigate a high-quality site, they can sense how many of their compatriots have arrived to do the same. They perceive – via vision, tactile interaction or olfactory cues – that this has become a well-recommended location. The evolutionary calculus is such that only a highly suitable site could lead to this. Having sensed that a quorum has been reached, the scouts return to the swarm to announce via a “piping” signal that a decision has been made.

It is a decision by quorum, not unanimity. The overwhelming majority of the swarm – around 95% – have been standing-by during all this, concerned largely with the matter of keeping the population together in an uncertain environment. They trust the hundreds of scouts to deliberate and then tell them where to go. Following the announcement, these thousands of honeybees warm up their flight muscles before they will journey to their new home.

The swarm intelligence of honeybees is employed for many decisions – to identify a new home, to reveal a good foraging site, to avoid a predator, or to provide collective strength against the sheer forces of a fierce storm. In so doing, a swarm becomes a whole that far outsizes the sum of its parts – something biologists often refer to as a “complex adaptive system” or, more evocatively, a superorganism.

A superorganism alludes to a population of individuals that cooperate to such an extent that they behave en masse like a single organism. In this, honeybees are not alone. Other social insects are famed for such behaviour, too, including ants, termites, wasps, aphids and more. A slime mould made up of single celled amoeba will function as one entity as well. Portuguese man-o-war jellyfish appear to behave similarly: each one is a colony of genetically identical organisms performing distinct functions. At times, superorganisms seem to emerge in fleeting states of collective behaviour – a flock of European starlings in the midst of a hypnotic murmuration, a silvery shoal of Atlantic mackerel forming a colossal whirling mass to deter predators. In some cases, they even extend beyond the bounds of a single species, such that entire coral reefs can function as complex adaptive systems. So too forests, with their staggering networks of roots and symbiotic fungi. Indeed systems resembling superorganisms can also exist within the confines of a single organism: in the immune system; the microbiome and the particularly interesting case of the human brain.

Precisely how we make decisions is not fully understood – quite a lot remains unknown. Nevertheless, as Seeley and his colleagues noted in the journal   Science there are certainly many parallels in how honeybee swarms and complex brains make decisions. This becomes especially apparent through certain “psychosocial” laws the human brain obeys during decision-making that seem to employ some remarkably similar mechanisms. The Hick-Hyman law, for instance, holds that the more options there are, the more difficult it becomes to make decisions. Whereas Weber’s law clarifies that as the options increase in quality, the difference between them must also become more pronounced in order for a decision to be made easily. Yet Piéron’s law states that a decision between high quality choices will be made quicker than a decision between low quality choices. A 2018   study published in the journal Scientific Reports by researchers at the University of Sheffield demonstrated that honeybee swarms follow these very same laws when selecting a new hive site.

As with honeybees, no single neuron in the human brain carries out complex decision-making on its own, nor is there any top-down hierarchy: it is a group endeavour. To highlight their similarities, Seeley in the aforementioned journal Science describes individual neurons and individual honeybees as simply “excitable units”. Each makes local interactions: neurons excite neighbouring neurons by releasing certain neurotransmitters, honeybee scouts excite nearby honeybees with a waggle dance. In the same publication, Seeley and his colleagues also explain that for a decision involving multiple options, a population of excitable units forms in favour of each alternative. Sometimes one population will be much larger than the others or will have stronger activity, but sometimes competing populations of excited neurons may be very similar and the most favourable choice is not clear.

Like honeybee swarms, our brains also use “stop” signals. Just as scouts dancing for one site will send a stop signal to scouts favouring a different one, a network of neurons favouring one outcome will exchange inhibitory signals with a group of neurons favouring an alternative. These chemical messages quell the opposing neurons, dampening their signal, which demonstrates that there is more than contagious agreement at play – there is also dissent. It’s an evolutionary strategy that, among other things, prevents deadlocks. Like honeybees sensing a quorum, when activity in a network of neurons finally reaches a certain threshold, a decision is made. It is hard to say which is the more compelling implication: that, in certain moments, the human brain behaves much like a swarm or that a swarm behaves much like a human brain.

When comparing bees and brains like this, we can see functional complexity emerging from myriad small-scale interactions. Yet, mathematically, there’s no guarantee that numerous local exchanges will lead to anything functional at all. As swarm-intelligence researcher Martina Szopek and her colleagues recently pointed out in the journal Frontiers in Physics (2021) it could easily lead to chaos. Yet, this doesn’t happen in bee swarms. Feedback loops between individuals – including all those stop signals – enable the swarm to reside “on the edge between chaos and order”. Likewise, according to certain theories in the field of neuroscience, such as the critical brain hypothesis, the human brain functions optimally at the critical boundary between order and disorder. In other words, at the edge of chaos. This is not uncontroversial but increasing evidence suggests there could be something to it.

Functioning at the edge of chaos may be a defining feature of complex adaptive systems, and for good reason. As science writer and former physicist Mitchell Waldrop put it in Complexity: The Emerging Science at the Edge of Order and Chaos (1992), the boundary between order and chaos is where “the components of a system never quite lock into place, yet never dissolve into turbulence, either.” And so, the edge of chaos appears to be a goldilocks zone, where a system is neither so disordered as to become non-functional nor so stable that it has no capacity for change. This brings us to why bees, ants, starlings, and our brains – separated by millions of years of evolution – would converge at the point of almost-chaos: it allows us to change.

The capacity to change has always been essential for survival, but change is messy by nature. To shift from what was to what will be always requires a degree of unravelling, however small. A complex system at the edge of chaos is in a state of almost falling apart, and it is there, in the loosening – yet not entire loss – of order that something novel can take shape, from the momentary form of a starling murmuration to the arrival of a quorum decision. Indeed, some argue that this “in-betweenness” is precisely where creativity emerges, providing, as ever, an avenue to resilience.

We are immersed in a world of superorganisms and complex adaptive systems from roiling shoals to financial markets; from the smallest of food webs to our planetary climate system; from the intricacies of a single decision to the widest of kinships. Yet, a lot remains unknown about how precisely they work, or indeed, what causes them to collapse. It seems that honeybees still have a lot to teach us about ourselves and each other. If we listen closely, we might learn how to exist at the edge of chaos without falling in.

Fiona McMillan-Webster is an Australian science writer with a Bachelor of Science in physics and a PhD in biophysics. She has written science stories for National Geographic, Forbes, COSMOS magazine, Australian Geographic, and other publications. Her writing has also appeared in the Best Australian Science Writing anthologies for 2015, 2016, 2018 and 2021. Her forthcoming book The Age of Seeds: How Plants Hacked Time and Why Our Future Depends on It will be published by Thames and Hudson Australia in 2022.