Updated at 9:53 a.m. ET on May 12, 2021.
Historians have always assumed that the medieval city of Angkor, today located in Cambodia, was huge, simply based on how much land its kings commanded. From the ninth to the 15th centuries, Angkor was the capital of the Khmer empire, which at its zenith stretched across modern Cambodia, Laos, Thailand, and Vietnam. The city was thronged with visitors from all over Southeast Asia—royalty and peasants alike—and was home to large numbers of farmers who kept the city fed, as well as workers who built its palaces, canals, and reservoirs. But precisely how many people lived in Angkor is one of the enduring mysteries in archaeology.
The problem is that, centuries after the city’s decline, only the great walled temples at Angkor Wat and Angkor Thom remain standing. The city’s residential neighborhoods were made entirely of perishable materials such as wood. Angkor’s city grid quickly disappeared beneath thick vegetation, and farmers ploughed over its far-flung neighborhoods. Though generations of experts have studied the city’s ruins, they’ve been unable to come up with a reliable estimate of its population that would help them make sense of how such a large city was run.
Now the Leiden University archaeology researcher Sarah Klassen and her colleagues believe that they’ve found the magic number. Based on a comprehensive analysis using nearly 150 years of data, they peg the peak population of the greater Angkor region at 700,000 to 900,000 people in the 13th century. The only European city approaching that size at the time was Constantinople.
“Population is one of those fundamental building blocks to understanding an archaeological site,” Klassen told me. “This number changes everything.” With a solid population number, Klassen can extrapolate how much rice Angkorians needed to grow to feed their neighbors, for example, or how much wood and water they consumed. The estimate restores human activity to the empty halls of the city, revealing all the work people were doing to maintain it. But the estimate also reveals more fundamental patterns in how cities develop over time, as their swelling populations change the urban landscape.
Klassen and her colleagues arrived at Angkor’s peak-population number by reading the landscape as a palimpsest left behind by the ancient Khmer peoples. Back in 2012, a team used helicopter-mounted lidar—3-D laser scanning—to measure minute differences in ground elevation beneath tree cover. It revealed an unmistakable grid of housing foundations, roads, farms, and canals sprawling for 1,158 square miles in what the team calls the greater Angkor region, similar to a modern-day metro area like the San Francisco Bay Area, with its multiple urban centers and low-density residential areas in between.
Meanwhile, archaeologists broke ground in a number of places across Angkor to verify that the mounds on the lidar maps were truly the remains of homes and not simply natural features. Around the same time, Klassen and her colleague Alison Carter spotted an unexpected pattern. Carter, a University of Oregon anthropologist, was looking at lidar data from what archaeologists call a “ceremonial center”—a city center with temples and other public buildings—which would ordinarily have attracted a higher population than outlying areas. But when the two researchers looked deeper, at the settlement that existed before the temples were built, the quantity of debris and other human-made items suggested the area had already had a very high density of residents.*The researchers theorize that local built homes in these proto-centers because they sat atop valuable farmland. Then the Khmer king ordered temples and water infrastructure to be built on top of the already valuable land, making the place even more valuable, in a virtuous feedback loop.*
Angkor’s density, then, seems to have grown from the ground up, starting with local farmers and ending with kings laying claim to the land with their temples. The process didn’t end there, though. In the most densely packed downtown ceremonial area around Angkor Thom, the researchers discovered that the 12th-century King Jayavarman VII had evicted an entire neighborhood to build the roads around his palace. “When you look at the lidar, you can see … roads and causeways which used to be houses,” Klassen said. “You can just imagine them saying, ‘Sorry, we’re building a road here.’”
For Klassen, the most intriguing question is whether the sequence of development she found at Angkor is part of a more universal pattern. Are all cities doomed to see displacement follow growth? In modern megacities, Klassen told me, it’s hard to know for sure what’s causing that pattern—we can attribute growth to a wide variety of technological advances, to transit, or to economic booms. But what if we could strip away all the gleaming machines and look at how cities develop without the thumb of industrial capitalism on the scale? We might get a metropolis like Angkor. “Angkor is like a lab for investigating [urban] relationships without those variables introduced by technology,” Klassen said.
If that’s true—and there are some reasons to be skeptical—Angkor offers one kind of origin story for the process that today we call gentrification. Stacey Sutton, an urban-planning and -policy professor at the University of Illinois at Chicago, describes gentrification in contemporary cities as the process by which wealthy, privileged populations displace poorer, marginalized ones. One might certainly characterize Angkor’s history in those terms too. In retrospect, it appears that the city’s growth involved wealthy elites pushing people off the land they had made valuable, so that the richest residents could live in city centers while the poorest were displaced to Angkor’s distant suburbs.
But scholars who study the early history of cities elsewhere in the world are hesitant to endorse the claim that this sequence is inevitable. Rebecca Boyd, an archaeologist affiliated with University College Cork who studies the origins of Dublin, told me via email that Dublin began as a meeting of two great rivers and five great roads across Ireland. There is no evidence of extensive farming or population density on the site until the Vikings arrived and set up their own town in the winter of 840–1. Local Irish people kicked out the invaders in 902, and continued to develop the city on their own. Great cities, in other words, don’t always begin organically. Sometimes they need Viking invaders to jump-start the process.
The archaeologist Sarah Parcak has identified hundreds of urban sites from ancient Egypt using satellite data. Parcak told me that Klassen’s work on population size is incredibly valuable, and provides an excellent model for anyone using remote sensing like lidar or satellite maps in combination with old-fashioned data on the ground, but Parcak is very cautious about applying the growth models Klassen found to ancient Egypt. “Every landscape is different,” Parcak said. “There is so much nuance to understanding specific cultures and places.” In Egypt, for example, some capitals did start as small, organic farming communities based around turtlebacks, or sand formations high enough to stay dry during the Nile floods. But later, once the Romans had colonized Egypt, new cities were built on the orders of the occupying government. “The way cities grow changes over time,” Parcak concluded.
Still, the ancient pattern of urban growth identified by Klassen and her colleagues resonates in the modern world: As population density increases, a city’s original inhabitants are typically pushed out, literally marginalized. But if the study of Angkor reveals that the roots of gentrification stretch deep into the urban past, the diversity of ancient urban forms suggests that the pattern represents a choice—and that alternatives exist, if we’re willing to deliberately embrace them.
*This article previously misstated how the older settlements below a ceremonial center in Angkor were discovered.