BU Today

Science & Tech

What Computer Modeling Can Tell Us about Religion

STH researchers introduce high tech to a low-tech discipline

“I don’t even like computers,” Connor Wood confesses. So why is this postdoctoral fellow in religion studies so enthusiastic about the computer simulation displayed on his laptop, resembling the result of an explosion at a spaghetti factory and an adjacent paint store—a tangle of curved, intersecting, multicolored lines?

Because those lines represent the new frontier of computer-simulated religion studies being explored by Wood (GRS’16) and other researchers under Wesley Wildman, a School of Theology professor of philosophy, theology, and ethics. The answers they seek might help explain things from religion’s benefits (potentially better mental health) to its evils (violence in the name of God).

In the field of religion studies, Wildman says, he became an early adopter of computer simulation after a working lifetime had taught him that “social dynamics like religion are complex…and that makes them difficult to understand.” Beholding breakthroughs in other fields made possible by computers, it struck him that “virtual social worlds” could help religion scholars tackle previously unanswerable questions.

Case in point: the lines on Wood’s laptop, a computer simulation predicting answers to the question: Why do regular practitioners of religion seem to do better, mental health–wise, than the nonobservant? (They have slightly lower suicide rates, for instance, Wood says.) One theory holds that religions requiring disciplines like fasting or regular prayer build behavioral self-regulation in their members. Using a computer at Virginia’s Old Dominion University, where the BU researchers have collaborators, Wood made the simulation, which predicts how many people would stay in a religion based on its strictness.

He compared that to real-world data about defection rates from 18 Christian denominations, ranging from those with strict obligations like the Mormons (a fast-growing faith) to the more permissive United Church of Christ, which has shrunk from two million members in the 1950s to less than a million today. The computer model “corroborated and predicted that people’s subjective well-being and self-control would rise in a growing community” with rigorous rituals, Wood says.

They can do such simulations now in the comfort of Wildman’s STH office, where they’ve parked “the Beast,” as they’ve dubbed the $55,000 computer they had custom-built for their research. Resembling a black extra-large suitcase, the Beast’s capabilities match the scholars’ needs, Wildman says: “The particular models we’re using involve computer agents, the religious subjects being studied, with virtual minds.…To get religious cognition expressed in the computer model, you’re going to need a lot of memory.”

Money for the Beast came out of a $2.4 million grant, most of it from the John Templeton Foundation, to use computer modeling and simulation to test heretofore untestable hypotheses about religion. Now halfway through that three-year grant, Wildman is also a coresearcher in the early stages of another grant, of $2 million, from the Norwegian government’s Research Council. Norway is concerned about immigrants, mainly Muslim, who don’t assimilate, he says. “The grant’s challenge to us is to figure out…the process of integration and refugee flow and the risks of religious extremist violence.”

“This is a potential explanatory tool for understanding why people get radicalized, why religious violence is increasing, why we’re seeing culture wars about religion in our political discourse.”—Connor Wood

Wood is using the Beast to research a separate question that nevertheless may have bearing on Norway’s problem. Religion confers structure and hierarchical roles on life, he says, from tribal initiation rites to bar mitzvahs to weddings. He wondered whether such structure builds resilience in ways that unstructured societies—including secularized ones—don’t. The Beast simulated a group of individuals that it depicted by wave patterns onscreen; when Wood tells the computer to eliminate structures and hierarchy, the waves jump apart chaotically, a computerized depiction of nonresilience.

“This is a potential explanatory tool for understanding why people get radicalized, why religious violence is increasing, why we’re seeing culture wars about religion in our political discourse,” he says. While an un-hierarchical society might feel more equal, many people seek structure, he says, and violent religious groups are nothing if not structured according to a rigid insiders-vs.-outsiders outlook. To test the accuracy of the computer’s prediction, Wood will use a variety of real-life data, including how adolescent social networks work and how emotions can spread like contagious diseases among online communities.

Wildman’s research team has used the Beast to analyze people’s reactions to terrifying events such as natural disasters or disease outbreaks. The frightened often flood churches to “cope with death-salient events,” Wildman notes. A philosopher like himself, operating without simulation tools, “could sit down and say, it makes sense to me that people would react to this terrifying event by seeking supernatural agent help; they’d probably start going to church or temple more.…I could write a book about that, and it’d be the theoretical story behind my hypothesis.

“But it’s just a hypothesis. It really needs to be tested,” he says. Using the Beast, the team created a computer-simulated virtual human mind capable of simulating the impacts of terror on behavior. The model, in turn, suggests that religious ritual observance would spike after terror-inspiring events drove people beyond a certain threshold of fear. The simulated output will be tested against what sociologists and demographers have learned about spikes in religious observance following a frightening event, such as the 2011 New Zealand earthquake that killed 185.

“And that,” Wildman says, “makes for a less speculative and more interesting book.”

Besides predicting contemporary realities, he says he plans to turn his computer modeling toward the past, studying religion’s role in such epic developments as humans’ transition from hunting and gathering to agriculture.

The Beast requires the researchers to articulate as precisely as possible the hypotheses they hope to model. “Computers are dumb,” Wood says. “They’re not conscious, no matter what Silicon Valley says….They can’t do context, they can’t get a facial expression, they can’t make a leap to understand what you’re saying.” That may explain what Wildman calls the aversion in the humanities to the kind of abstraction that computers permit. But while fieldwork is priceless, he says, modeling stretches the realm of academic theorizing.

“You’ve got a big, complicated system in the real world; you try and approach it from the top, from sociology, you can only get so far,” he says. “You approach it from the bottom, from psychology and neuroscience; you can only get so far.…How do you get to the actual system dynamics? The thing to do is to simulate the complicated social system in a computer so that you can slowly study it.”

Rich Barlow

Rich Barlow can be reached at barlowr@bu.edu.

6 Comments on What Computer Modeling Can Tell Us about Religion

  • Andrew Wolfe on 06.01.2017 at 9:32 am

    The research seems to make a horrifyingly unscientific and irrational assumption, specifically that there is a single set of social dynamics around “religion” and that these dynamics are not affected by the behavioral codes and mindsets of the specific religions. Why do Wood and Wildman treat religious “radicalization” as if it were a universal process independent of the specific religion? Are they afraid to confront the possibility that different religions have different effects on the behaviors of their adherents?

  • Ray on 06.02.2017 at 2:36 am

    We are living in a world filled with exciting opportunities. One of these is utilization of computer technology to understand impact of religion on human behavior and psyche. I am very excited about what Connors, W. Wildman, and his team would eventually come up with. As a BU alum, I wish you guys best of luck in your unique adventure into the complex world of religion.

  • Connor Wood on 06.02.2017 at 2:42 pm

    In reply to Andrew’s comment, we’re very aware of the complexity of religion and we don’t claim that cultural dynamics don’t matter. My PhD is in religious studies, where the problem of defining religion – and avoiding simplistic definitions of religion – is a central concern for the field. What the tools of modeling allow us to do is to explore those aspects of “religion” that are generalizable across contexts. We’re not claiming that those aspects exhaustively define religion.

    And no, I think we’re perfectly aware that different religions and different doctrines can and do influence adherents differently. But these differences can be studied using social scientific and computational tools. For instance, the Jewish faith emphasizes ritual much more than Protestant Christianity does. One researcher, a psychologist, found that Protestants feel more guilty about their thoughts than Jews do, because in Judaism it’s actions that count, not thoughts or beliefs. A Jewish believer who has a passing thought about doing something sinful would probably not feel too bad about it, as long as she didn’t act on that thought. But a Protestant would be more likely to feel guilty about having simply thought about it.

    The key is that these differences in moral cognition are correlated with cultural attitudes toward ritual. Some religions heavily emphasize ritual and behavior – e.g., Judaism, Confucianism, Catholicism – while others emphasize inner states and belief. In turn, anthropologists have found that a culture’s general attitude toward ritual has a lot to do with its economic modes and its level of collectivism. Specifically, societies in which economic activity requires a lot of intensive collaboration tend to have heavily ritualized religions, while societies that encourage individualism (like capitalist North Atlantic societies) tend to have less-ritualized, more inner-focused religions. (For readers who know Max Weber, this is essentially Weber modernized, with some Mary Douglas and cultural psychology thrown in.)

    Upshot: there are generalizable differences between cultures that allow us to model the dynamics that give rise to those differences. Those are the things we can model. We can’t model particular historical events or other contingent circumstances, and we don’t deny that those things have a major influence on outcomes in the real world. This is why models are only models. They’re useful for understanding general principles and causal dynamics, but they’re not a substitute for close reading, historical scholarship, or ethnography. Modeling is a complement to those approaches, not a replacement for them.

    • Julie on 06.06.2017 at 3:18 pm

      This is an important topic that needs new language to understand our language-limited interactive discourse on feelings, euphoria, morality, spirituality as well as a thoughtful study of brain activity, genetics and our human experience related to these.

  • Deb Scott on 06.06.2017 at 12:37 am

    Super exciting stuff! As a current grad student, I’d be interested to learn the languages you’re modeling in?

    • Connor Wood on 09.07.2017 at 5:41 pm

      Deb, my apologies for the long delay in replying. We’re using primarily AnyLogic and NetLogo.

Post Your Comment

(never shown)