Overlooked No More: James Sakoda, Whose Wartime Internment Inspired a Social Science Tool

This article is part of Overlooked, a series of obituaries about remarkable people whose deaths, beginning in 1851, went unreported in The Times.

Unlike most of the 120,000 Japanese Americans detained in internment camps in the United States during World War II, James Sakoda had a mission: to document the experience of incarceration. He took about 1,800 pages of notes, largely in private, lest he be accused of being a traitor or a spy.

Those notes would form the basis of his 1949 dissertation on the dynamics of individuals and groups at one of these camps, the Minidoka War Relocation Center in Idaho. Tucked into Appendix B of the paper was possibly the first example of what is known as an “agent-based model” — a simulation of how individual actions can add up to large-scale patterns.

The tool is essential in a wide variety of fields, and has helped social scientists, epidemiologists, financial regulators, city planners and wildlife experts do their work. During the coronavirus pandemic, for instance, agent-based models were essential for forecasting the spread of the virus and prioritizing vaccines for certain groups of people.

To develop the model, Sakoda used the home computing technology of the time: a checkerboard. Each checker was given a simple rule for movement, based on its immediate surroundings. By changing the rules even only slightly, Sakoda showed that the pieces could mingle freely, or they could quickly segregate by color.

Ecologists and environmentalists have used agent-based models to investigate the interactions between shipping boats and beluga whales in Canada’s St. Lawrence River estuary; between humans and elephants in Tanzania; and between scuba diving tourism and coral reefs in Thailand. Transportation agencies use the models to predict how even minor changes, like expanding a bus stop, could affect the flow of traffic.

“James Sakoda was perhaps the first social scientist ever to apply computational modeling for unraveling the complexity of social processes,” Andreas Flache, a sociologist at the University of Groningen in the Netherlands, said in an email.

Despite the widespread use of his model, Sakoda did not get much credit for his innovation.

James Minoru Sakoda, who was known as Jimmy, was born on April 21, 1916, on an alfalfa ranch in Lancaster, Calif., in northern Los Angeles County. His conservative Buddhist parents, Kenichi and Tazu (Kihara) Sakoda, were both from Japan.

After moving around the Los Angeles area, his parents took their four children to Japan, where James attended high school for three years and Tokyo University for another three.

With $100 in his pocket, Sakoda returned to California and enrolled at the University of California, Berkeley, where he studied psychology. It was during his second year there that the secretary of war established detention camps on the West Coast for Americans of Japanese heritage.

Sakoda was still at Berkeley when he began documenting Japanese-Americans’ reactions to the crisis. Through a classmate, he met Dorothy Swaine Thomas, a sociologist who was recruiting soon-to-be-incarcerated fieldworkers for a project called the Japanese American Evacuation and Resettlement Study.

All four Sakoda siblings were back in the United States by the time they were ordered to one of these camps; their parents remained in Japan during the war. Sakoda, his brother, George, and his sisters, Ruby and May, were initially incarcerated in 1942, at the Tulare Assembly Center in the San Joaquin Valley in central California.

“Soldiers stood watching with rifles and Tommy guns,” Sakoda wrote in his journal, noting that tall grass poked through the asphalt floors of his barracks, and that the condition of the latrines was “open to criticism.”

He went on to chronicle daily camp life for Thomas’s project, always in a detached, analytical way. “I never talked about this happening to us,” he told the historian Art Hansen in 1988. Instead, he said, he looked at it as, “It happened to them.”

The study “gave him a sense of purpose,” Hansen said in a phone interview. “He played a salvation sort of role for not only his community, but generally for American history.”

The Sakoda siblings were later moved to the Tule Lake Relocation Center, near California’s northern border, where James taught psychology to detainees and met his future wife, Hatsuye Kurose, who was known as Hattie — the “smartest girl in my class,” as he called her in a letter to Thomas.

James and Hattie then spent two years at the Minidoka camp in Idaho, where they married before returning to Berkeley shortly before the camp was closed in 1945.

Sakoda was working toward a Ph.D. in psychology at Berkeley when a fellowship took him to Harvard. It was there that he developed his checkerboard model, examining the interactions among various groups at the internment camps: the “clannish” Nisei; children of Japanese immigrants; more reclusive detainees; and camp administrators.

After earning his doctorate from Berkeley in 1949, he briefly taught at Brooklyn College, then joined the psychology faculty at the University of Connecticut. There he developed an interest in the potential of computing in studying human behavior.

In the summer of 1956, Sakoda learned to program on early IBM punch-card computers at M.I.T. Then, with his wife and their son, Bill, he moved to Providence, R.I., hired by Brown University, where he became the director of a social science computer laboratory.

At a time when the study of human behavior was largely isolated from computing, Sakoda pushed for better tools with which to merge the two; the checkerboard model, which he taught to students over the next three decades, was just one of them.

In 1963, he was invited to a summer institute at the RAND Corporation in Santa Monica, Calif., to trade ideas about modeling cognitive processes using computers. While there, he began developing his own computing toolbox for social scientists, calling it DYSTAL. A 1971 paper, “The Checkerboard Model of Social Interaction,” modernized his 1949 model through computer-run simulations.

After retiring from Brown in 1981, Sakoda told Hansen, “I think the best thing I’ve done is the social interaction model, which solved the problem in social psychology of going from the individual level to the group level.”

But in the 1990s and 2000s, as agent-based modeling became fundamental to studying infectious diseases and the movements of humans on a large scale, a different origin story emerged.

Thomas Schelling, a well-connected Harvard economist and White House adviser, was on a plane bound for Boston when he started noodling with Xs and Os moving along a line. It would eventually become a checkerboard model strikingly similar to Sakoda’s. Schelling mentioned it in a 1969 RAND research report and expanded it into an article in 1971, shortly after Sakoda had published his, in the same journal.

Decades later, it was Schelling’s article that became widely credited as the first in which the checkerboard model appeared.

It is possible that Schelling encountered the seed of the idea at RAND — he completed a residency there a year after Sakoda visited. But when asked in a 2001 interview if the checkerboard model devised by Sakoda had influenced him, Schelling replied, “I have never heard of him.”

In 2005, Schelling was awarded the Nobel Prize in economics, with Robert J. Aumann, for “having enhanced our understanding of conflict and cooperation through game-theory analysis.” In a biographical statement accompanying the prize, Schelling wrote of the checkerboard model, “Without knowing it I was pioneering a field of study that later became known as ‘agent-based computational modeling.’”

In his later years, in Barrington, R.I., Sakoda focused on gardening, his family and a longstanding mathematical side interest: origami. His book “Modern Origami,” published in 1969 and still in print, showcases his own designs and made him notable among enthusiasts. (He decorated his computer laboratory at Brown with his origami.)

His nephew Jim Kurose said in an interview that at family gatherings Sakoda “would usually go sit by himself quietly in the living room and take out his paper, and he’d start folding, and he would just keep kids absolutely entranced.”

He died on June 12, 2005. He was 89.

Sakoda’s agent-based modeling innovations are being rediscovered thanks to the research of Rainer Hegselmann, a philosopher and social scientist at the Frankfurt School of Finance and Management in Germany. In a 2017 article in the Journal of Artificial Societies and Social Simulation, Hegselmann speculated that the timing of Sakoda’s retirement, in 1981, before the personal computer became ubiquitous, may have led to the erasure of his achievement.

“Maybe that life punishes those that are late,” he wrote. “But sometimes it punishes those that are early as well.”

Sakoda, however, was “not much concerned with getting explicit credit for what he did,” his son, Bill, a computer scientist, said in an interview.

Instead, he added in an email, “He worked magic for a lot of people very quietly.”