Book review – The Ecology of Collective Behavior

8-minute read
keywords: ecology, ethology

This is the third of a trio of reviews in which I take a brief detour into ants and collective behaviour more generally. I previously reviewed The Ant Collective, a graphical introduction to ant behaviour, and entomologist Deborah M. Gordon’s Ant Encounters, a primer on how collective behaviour in ants comes about. The Ecology of Collective Behavior is the second book by Gordon that I will examine. It proposes a research programme to figure out both how collective behaviour responds to changing environmental conditions, and how it evolves. Though squarely aimed at professional biologists, this brief and interesting book is nevertheless accessible to a wider interested audience and makes its case with nary an equation in sight.

The Ecology of Collective Behavior, written by Deborah M. Gordon, published by Princeton University Press in October 2023 (paperback, 184 pages)

One of the most familiar examples of collective behaviour is swarming, where large numbers of animals move in a coordinated fashion, whether flocks of birds or schools of fish. However, collective behaviour can take many different forms and it has proven incredibly tricky to develop a general theory that successfully captures the diversity of processes generating it. As far as Gordon is concerned, one habit of thought that is holding us back is reductionism: “it is not possible to learn how natural systems function collectively by considering the components separately and independently of the world they inhabit [as it] severs the relations that matter” (p. 2).

A brief tangent here, as the limitations of reductionism have been a recurrent theme in recent reviews. Gordon’s book thematically sits smack in the middle of several others I have covered lately. Safina made exactly this observation (that reductionism overlooks relational dynamics) in Alfie & Me. Also, this is the second book in short order that mentions panta rhei (“everything flows”) and refers to Nicholson & Dupré’s book. But where I gave Slijepčević‘s Biocivilisations a mixed review for, amongst others, dunking on reductionism, Gordon is intellectually honest enough to admit that without centuries of reductionist research: “we could not have arrived at the recognition that all of this knowledge is not enough” (p. 34). Her book is far more grounded and presents an actionable framework.

Gordon’s goal here is to outline how she researches collective behaviour. She is particularly interested in how it changes in response to changing environments because, again, panta rhei: “response to changing conditions is fundamental in living systems” (p. 5). You start to see why studying this can be quite challenging as you are effectively considering second-order dynamics (the dynamics of dynamic behaviour). The book’s nine chapters break down into roughly three parts that I will discuss further below.

“There is an interesting discussion here as to why attempts at formulating an overarching theory [of collective behaviour] have failed.”

Chapters 2 to 4 form an introduction to the topic. Gordon describes the two model species that she has worked on for several decades: the desert-dwelling red harvester ant Pogonomyrmex barbatus and the rainforest-dwelling turtle ant Cephalotes goniodontus. Their collective behaviour differs and, as later chapters explore, this can be linked to differences in their environment. She then zooms out to consider collective behaviour more generally, whether at the level of molecules, cells, or organisms, and discusses two well-studied examples. Lastly, she considers how collective behaviour results from individual interactions and reflects on that buzzword “emergence”. A younger Gordon was once a fan; today she is ready “to replace its mystical glow with something more substantial” (p. 33). There is an interesting discussion here as to why attempts at formulating an overarching theory have failed: similar outcomes do not necessarily result from similar processes; model output resembling what we see in nature does not necessarily mean that you have identified the processes at work; and models that fail to predict an outcome can be incomplete rather than incorrect, since the same process can have different outcomes under different conditions.

The next two chapters present her core framework that explains how collective behaviour depends on, and responds to, environmental dynamics. To reiterate: collective behaviour consists of individuals interacting with each other. She considers three parameters of collective behaviour: rate (fast/slow: how quickly interactions respond to environmental change), feedback (stimulation/inhibition: how interactions change the collective behaviour), and modularity (uniform/modular: a description of network structure in terms of node connectivity, i.e. are all individuals in contact with each other or do they form distinct, internally connected clusters). She then links these three behavioural parameters to three environmental parameters: stability (stable/unstable: how prone the environment is to large or frequent changes), resource distribution (patchy/uniform: how resources are distributed in time and space), and energy flow (low/high: the ratio of energy or resources obtained versus spent; economists also speak of EROI: energy return on investment).

These two sets of three parameters relate to each other as follows, as exemplified by the ant species she studies. Rate depends on stability and resource distribution. Harvester ants live in stable desert environments with uniformly distributed resources and adjust their foraging activity slowly. Turtle ants live in the exuberant chaos of a rainforest where they form rapidly changing foraging trails to exploit ephemeral resources. Feedback depends on stability and energy flow. For harvester ants, foraging is energetically expensive and they need stimulation (positive feedback) from patrollers before they leave the nest. For turtle ants, foraging is energetically cheaper and they will continue to forage until inhibited (negative feedback) by e.g. a shortage of mouths to receive the liquid food they carry home. Modularity (like rate) depends on stability and resource distribution. In harvester ants, modularity is low and individuals are highly connected with information transfer happening in a centralized fashion in the nest. In turtle ants, modularity is high, with foraging decisions resolved locally through interactions of small groups of ants. If the above all seems rather impenetrable, that will be my attempt to condense her framework into just two paragraphs. Fortunately, her treatment of it is more accessible: she presents it as a verbal argument and pads it out with lists of examples from across the biological spectrum. Having quickly perused the copy I have here, those interested in a mathematical treatment of collective behaviour should instead turn to Sumpter’s book Collective Animal Behavior.

“The emphasis on an actionable framework illustrated by her decades-long research on ants makes for a particularly grounded, can-do book.”

The third part, then, wraps up the discussion. One chapter provides a very practical research programme for investigating how collective behaviour works: observation followed by experimental disruption. There are some excellent tips here on how to disrupt a system (stay within the ordinary range of fluctuations) which she contrasts with other standard approaches such as knockout experiments and factorial design. She illustrates her approach by taking the reader through decades of her research on harvester ants.

The last two chapters consider the evolution of collective behaviour. In some ways, this is little provocative: participation in collective behaviour is an aspect of an organism’s phenotype and thus a trait subject to natural selection, just like any other trait. Where it gets more provocative (depending on your viewpoint) is that, as a pupil of Richard Lewontin, she embraces the Extended Synthesis, including environments (in addition to genes) influencing phenotypic expression, organisms altering their environment (niche construction), and the always hotly debated notion of group selection. I am receptive to these ideas and hope to explore them more in future reading. Gordon offers interesting critiques here of divisions such as nature/nurture and (in ethology) instinct/learning and how they are holding back our understanding of collective behaviour. Particularly insightful is the last chapter that contrasts her framework with the prevailing idea that collective behaviour is marked by a conflict of interest between the individual and the group, i.e. that strategic individuals always seek to maximize fitness and prioritize individual benefits over the collective good. If I understand her correctly, she argues that, because collective behaviour has consequences for individuals, selection can act on how individuals produce collective behaviour. In other words, selection can act on multiple levels simultaneously, without group selection necessarily conflicting with individual selection.

Are there limitations to Gordon’s ideas? She indeed expects that her hypotheses will not cover all possible ways that collective behaviour can interact with a changing environment. Chapter 5 explicitly opens by writing that “these hypotheses are meant as a framework for research on collective behavior, not as a deductive theory” (p. 46). Though this book is aimed at biologists studying collective behaviour, the subject matter here is easily accessible to ecologists and evolutionary biologists more generally. The verbal argumentation and numerous examples also put this book within reach of interested general readers. The emphasis on an actionable framework illustrated by her decades-long research on ants makes for a particularly grounded, can-do book. Overall, a very interesting perspective and a satisfying conclusion (for now) to my brief detour into collective behaviour in ants.

Disclosure: The publisher provided a review copy of this book. The opinion expressed here is my own, however.

The Ecology of Collective Behavior