American Scientist

What do financial markets, pandemics, and ecosystems management have in common? They are all complex systems that can be described by mathematical principles, as demonstrated in the research of Simon A. Levin, the James S. McDonnell Distinguished University Professor in Ecology and Evolutionary Biology and Director of the Center for BioComplexity at Princeton University. Levin is known for the development of the foundations of spatial ecology and his work on pattern and scale. More recently, his research has focused on the interface between ecology and economics, especially regarding problems of public goods and communally pooled resources. His book, Fragile Dominion: Complexity and the Commons (Perseus, 1996), along with his research, weaves these themes together, invoking ecological and evolutionary theory to inform principles for management practice. Levin was a recipient of the Gold Key Award at Sigma Xi’s 2025 International Forum on Research Excellence (IFoRE) and spoke with editor-in-chief Fenella Saunders about his work. (This interview has been edited for length and clarity.)

How did you start out in mathematics and end up in ecology?

I always liked mathematics when I was growing up, and when I went off to college, I really didn’t have much idea what I was going to do, but my brother encouraged me to major in mathematics because he knew that I enjoyed doing it. My approach to doing things is always what in optimization theory would be called hill climbing: When you get to the top of the hill, you can worry about what to do from there.

I was really interested in applying mathematics even then, and so I took a course in operations research. This field searches for the best way to do a certain task, and there are parallels in evolutionary biology. For example, optimal foraging would be an operations research problem: If you’re feeding in some patch, when do you leave it and go to some other patch? But imagine that you have a landscape full of hundreds of patches. How do you search them? A classical problem in operations research is the Traveling Salesman Problem. How does the salesman visit all these cities in order to minimize travel?

So I knew I wanted to do applied things, and I knew pretty early on I wanted to do biology. I was interested in the outdoors. This was a time when Rachel Carson had published her book Silent Spring. And so I thought, there’s someplace where I could make a contribution. Maybe there’ll be mathematical research problems even there.

I wanted to be able to attack the problems that ecologists were worrying about. So I transformed myself into an ecologist. But I still like the mathematical approach of looking at the simplest model possible, because I like to isolate the factors.

What work have you done on modeling epidemics as larger systems?

Epidemiology has a great historical tradition that goes back to the first mathematical models. I’ve been working on influenza since the late 1980s. I worked on myxomatosis, which is a disease used to control rabbit populations that were rampant in Australia. I’ve worked a lot on antibiotic resistance. In 2019, with my graduate student Chadi Saad-Roy, we started working on predicting when viruses could have an asymptomatic stage. COVID must have read our paper. Then we worked a lot on COVID, and the modeling there has expanded out to look at things like individual behavior during epidemics. We’ve found that human adaptive behaviors can mitigate the epidemics if people avoid contact. We also explored things like vaccination strategy. What nobody I know would have predicted is that this would become such a politically hot issue. This led me to be interested in political polarization and how it arises. During flu season, everybody in Asia walks around in masks. If you went to the Nordic countries, nobody ever did that, and they still didn’t do that during COVID. The United States and much of western Europe was a third set of countries in which we had a transformation, and everybody started walking around with masks. So what was it about those three different cultures that led to these different strategies? A lot of what we’ve learned was about how the social dynamics can develop to create a switch. This actually resonates with a lot of the work we’ve done on attitudes toward decarbonization and how that becomes politicized. So there’s got to be a coevolution between popular sentiments and decision-makers. Understanding the social dynamics and how it’s coupled into that has been a big part of what we’ve been doing on disease models.

Are there unified methods of modeling systems across many disciplines?

That’s what makes it possible to make nonlinear advances with the cross-fertilization across fields. The fact that the systems are similar and people in different fields have thought about similar questions, but in different ways, means that there may be low-hanging fruit. Certainly ecological systems and socioeconomic systems are very similar in their structure. Ecology and economics both start with eco, and that’s not an accident, because the root word means “house” in Greek. There’s a lot of economics in natural systems. But it’s the same for physical systems. When you boil water, you’ve got a lot of individual agents. They’re not people, they’re molecules, but boiling is an emergent outcome. As a mathematician, I’m drawn to finding the core organizing principles. All of these systems are what we call complex adaptive systems. That is, they’re systems that have to be thought about on at least two levels. One is, you’ve got these individual agents who are moving around. They might be birds in a flock, they might be humans in a society, they might be molecules in your tea kettle. There’s some self-organization process, or you’re applying an external change, maybe delivering heat to the tea kettle, or delivering heat through climate change to our society, which changes the interactions. And then you’d like to understand how the microscopic dynamics give rise to the macroscopic dynamics. Individuals are organized into groups, those groups into populations, and those groups may be overlapping, too. You’d like to understand the relationship. You’d like to understand when there’s going to be a phase transition, whether it’s water boiling or a sudden change in the structure of your society. You’d like to understand how robust your systems are, and how that depends on the individual agents. And for all the systems I’m interested in, you’d like to understand the conflict between the interests of the individuals and the interests of the collectives. How can we learn from how evolution has solved similar problems to deal with international cooperation on the problems that underlie sustainability?

How do you handle that there is no way to account for all of the variables in complex systems?

I think that’s where mathematics in particular has a lot of power. When you see the same thing happening over and over again, as a mathematician, I think there must be a fundamental reason that that’s happening that can’t depend on all the details of all of the interactions that are going on. How do we extract the signal from all that noise? What are the essential interactions? Einstein said that your models should be as simple as possible, but not simpler. There are two approaches to dealing with that. One is, you start with simple models and say, What if this was all that was going on? What would we see? But we don’t see that, so why not? And you start adding complications to see what they would do. The other approach—these are not mutually exclusive—is you start from a very detailed model with lots of variables. You start looking at sensitivity.

In physics, when we average over the detail in a model, this is called coarse graining. There’s an art to it. If we were in the same room and I threw a ball to you, whether you catch it or not depends on your system of muscles and nerves and so on. But nobody sensible would attempt to build a model that took into account every neuron and every muscle. Something simpler is going on at the macroscopic level. You need that, because otherwise the system would be very sensitive to small errors. With the detailed model, it’s impossible to fit the parameters properly, because the model itself generates enough noise that you don’t know what’s going on.

What problems are there in the world that still bother you?

That’s very simple for me. I mean, the problems are hard, but identifying them is very simple. We’re all facing major threats that affect us all around the globe that will affect our children and grandchildren. We all should have a collective interest in solving those problems. Why can’t we? We know about threats having to do with climate change, biodiversity loss, the spread of infectious diseases, new pandemics, nuclear disarmament. How do we realize that although our strategies have developed within groups, often to aid us in conflicts with other groups, how do we make the next step to realize that the threats are not from other groups, but they’re from the global deterioration of our resources and the sustainability of the planet, and to work together? The things that I’ve directed the most attention to in the last few years are around political polarization. How does it arise? What can we do about it? Can we reduce the effects? How do we get across the message where we can all cooperate in addressing the problems of the commons? This is not a new problem, but those dwarf all the others. The one problem we haven’t solved is how do we cooperate with others? How do we value natural capital and the interest of future generations?

Can systems approaches help people feel like they can influence society?

I’m working more with social scientists to find ways to make people feel motivated to do things. You have to close the loops so that people at least see the local benefits. We have shown that local prosociality could lead to global cooperation. But how does prosociality evolve? Avinash Dixit and I developed a model that shows that you’d be better off in a prosocial world; you care about your children and want them to live in prosocial worlds. Therefore you’re willing to make a contribution, say, to education systems that will encourage prosocial behavior, because the payoff goes to your children. That’s the power of simple mathematical models, to show it doesn’t take much to make this happen.