Joshua Epstein is a professor of epidemiology in the School of Global Public Health at NYU and director of NYU’s Agent-Based Modeling Laboratory, with affiliated appointments at NYU’s Courant Institute of Mathematical Sciences and Department of Politics. He is the author of Agent_Zero: Toward Neurocognitive Foundations for Generative Social Science. Recently he spoke by telephone about the COVID crisis with Commonweal contributing editor Rand Richards Cooper. The interview has been edited and condensed.
Rand Richards Cooper: In a Politico article, you wrote about “the contagion of fear” and its role in pandemics. Can you describe your epidemiological modeling work and what it reveals about fear and contagion?
Joshua Epstein: Most classical mathematical epidemiology ignores behavioral adaptation. People in the models just keep mixing together and spreading disease, for example, and don’t self-isolate on their own. I introduced methods to study how fear of the pathogen can produce self-isolation. I call these coupled contagion models, where a disease is spreading, but fear of the disease is also spreading, and the two contagions interact to produce multiple waves. The core narrative is this: if the disease is high, people get scared, and they spread their fear and put themselves in isolation. This takes fuel out of the epidemic, suppressing the spread. When the spread gets low, people's fear starts to diminish, and they come out of isolation. The problem is that there can still be infection in the community. Prematurely relaxing self-isolation and distancing pours fuel—in the form of susceptible healthy people—onto these infective embers, and the thing can explode in a new wave. That's what we saw in the Spanish flu pandemic of 1918. And it's what we're worrying about with COVID-19—cycles of vigilance and complacency that can produce oscillations of just this sort.
RRC: By mid-March, my family had self-isolated completely, even though at the time we didn't know a single person who’d even been sick. I’d always assumed that people need fear to be immediate and personal in order to be sufficiently motivated to make such drastic changes. Is that wrong?
JE: Fear of disease can spread without direct exposure to the disease itself, and can spread faster and wider than the disease. To get the disease you have to bump into a sick person, but there are many more channels for transmitting fear. And things can spread like wildfire on social media. We see that a lot—not just in epidemics but in financial panics and other collective dynamics. There's a lot of fascinating work using social media to study this spread. At my lab we build agent-based models. These are artificial societies of software people that we build on computers. We let them interact, and they can spread disease, they can spread rumors, they can spread fear. I developed a next-generation software person named Agent Zero who has a fear module, a deliberative module, and a social module. Agent Zero’s overt behavior is the result of all these forces.
The system of fear is very ancient. The amygdala complex in the brain is as old as vertebrate history. We're hardwired to learn fears—it’s a big survival component. And we can spread the fear to one another. There are experiments in neuroscience where one person is conditioned to fear a blue light through having a shock cuff activated when the light goes on; eventually the person clenches up in fear from the light alone. The real core of the experiment is where someone is not wearing the shock cuff, but just observing the person who’s being conditioned. And then you shine the light at the observer, and she’s afraid, too. Watching people experience fear can make you afraid.
RRC: In the context of an epidemic, fear sounds almost entirely salutary. What’s the downside?
JE: Yes, fear helps suppress the disease, as we discussed before. But it is also double-edged. That same fear also closes down a lot of social interaction and economic life. And we can acquire and spread all kinds of baseless fears and biases by the same mechanism. An important step, I think, is to understand that we’re subject to these unconscious evolved forces. If we had a better appreciation of our fears and how they happen, and how baseless they can actually be, it would help us avoid being manipulated politically. And that brings in another fear that worries me a great deal— fear of vaccines. The WHO is now listing vaccine refusal as one of the top ten threats to global health. We're seeing the resurgence of measles and even polio out of these really baseless fears of vaccination.
RRC: I’ve been struck by how the daily lifestyle precautions we’ve adopted create what we'd normally think of as a kind of paranoid state. It’s like living your whole life as a TSA airport-security procedure. How can people sort out prudence from panic? Are we too afraid or are we not afraid enough?
JE: Well, it’s pretty hard to tell because we’re not doing enough testing. It’s hard to know what the actual prevalence of the disease is. We saw the spectacle of the administration issuing guidelines that require fourteen days of demonstrable reduction in transmissions before you begin opening, and at the same time encouraging people to “liberate” Michigan—and this, when we didn’t have enough testing to even conform to those guidelines. What's the appropriate level of fear? We still don’t have a real, accurate appraisal of the extent of the disease or how deadly it is. The true case-fatality rate is total deaths over total cases, but we don’t know what the total cases are. And of course, getting a reliable antibody test would be very useful. If it turns out that we have conferred immunity, then people who were positive for antibodies could go back to work safely, and we might be able to navigate this trade-off between public health and economic life. We don't have enough testing to determine that. Not even close.
RRC: Tell me more about Agent Zero. How does it work, and what does it tell you about where we are now?
JE: The whole idea of Agent Zero began with the fact that social sciences are dominated by the notion of the rational actor—homo economicus, the agent who has good information and clear objectives. Homo economicus is an optimizer. He’s well-informed; he evaluates his options and rationally chooses the one that maximizes utility or economic gain. But in fact, humans are not entirely rational. My book, Agent_Zero, begins with a quote from Hume—“Reason is a slave to the passions”—and in fact we have overwhelming psychological and neuroscientific evidence that humans depart from these canonical, rational behaviors.
Yet even though there’s a lot of evidence against the rational-actor picture, there haven’t been many formal alternatives to it. The Nobel economist Paul Samuelson once said, “You can only beat a model with another model.” So I spent five years working on a formal agent that would be an alternative to the rational actor. That’s Agent Zero. The idea was that, yes, people do have some deliberative capacity, but they’re lousy at statistics and they usually have crummy information. So I gave them some reasoning ability, via a limited deliberative module, but also a distinct passion module based on one of the passions we know a good deal about: fear. We know a lot about the neuroscience of fear, and we even have simple equations for the acquisition of fear. I put those equations into Agent Zero. And then I put a lot of these agents together, for them to transmit their dispositions, their emotions, their affective states to others. When you put a bunch of these Agent Zeros together, they do a lot of disturbingly familiar things. They engage in genocide, financial panic, contagious dysfunctional health behaviors. They also can do very constructive things together. They can rebel against autocratic regimes. They can form their own networks and dissolve them. They are building blocks for a social science that takes seriously these evolved, not entirely deliberative, aspects of the human condition. I like to think of it as an alternative foundation for a more unified and realistic social science. We are building models populated with Agent Zeroes to study all sorts of things, including contagious diseases. We want to make it one of the pillars of epidemiology broadly construed.
RRC: Has the pandemic played out more or less the way you had expected?
JE: Well, it has played out roughly as the models suggest. What I think hasn’t been helpful is the public discussion of it. For example, the way debate has focused on the peak and on whether we’re past it. Lost in the fray is the fact that in most epidemics, half the spread happens after the peak. So the idea that “Hey, we’re over the peak, let’s go back to work”—that’s just reckless. I think states have opened up in a perilous and disorganized way. A lot of them never came close to the administration’s own expert guidelines about opening, partly because Trump encouraged them not to, to “liberate” their states. They didn’t demonstrate sustained decline before reopening. I mean, Georgia was on the rise when they opened. We’ll see what happens here and worldwide as people lift these distancing measures. As far as I’m concerned, we’re still deep in the woods.
RRC: On a personal note, what has it been like living in New York City during the pandemic?
JE: I live in a two-bedroom NYU apartment in the heart of Greenwich Village. I’ve been here with my wife, our two children (a college student and a grad student), our dog, and two cats. We all hunkered down here. It’s actually been wonderful for us to have everybody together. And we’ve luckily dodged the disease so far. To a New Yorker the city these past months has been just utterly surreal. New York is such a distinctively physically close city. The essence of it is that everybody’s on top of each other. To have it be dead like this was just unbelievable. It’s the city that never sleeps, and it was in a coma. What a bizarre sensation, to see Times Square empty! It’s just a very otherworldly feeling.
RRC: You’re a student of fear. Has there been any particular moment when you felt it?
JE: I haven’t felt afraid personally. I know a lot about epidemics, and we have our face coverings and we’ve stayed isolated here in our little cave. I do feel afraid for others. I’ve been very afraid for the first responders. But to be honest, our current political situation and the fate of our democracy—that’s scarier to me than the disease.
