To understand why early cities thrived, look not to the temples of kings but to their subjects’ bustling neighbourhoods

All cities have neighbourhoods. This may not sound like much of an observation, but it is in fact a powerful claim for archaeologists of early cities. We now know that neighbourhoods are the only true urban universal – a feature found in every city that has ever existed, past and present. Other seemingly ‘urban’ traits, from streets and big buildings to markets and specialists, are absent from many cities and urban traditions. But neighbourhoods are playing a crucial role in the transformation of archaeological research on the earliest cities. Older views of cities focused on pyramids, tombs, greedy kings and oppressive states. In recent decades, archaeologists and others who study cities have turned to focusing on life and society in ancient cities. But now archaeologists have started to analyse these sites as cities, as urban settlements, comparable with cities in the modern world.

Neighbourhoods provide a good introduction to the different forces that shape cities. In them, we can see the ‘top-down’ forces, or the actions of governments, economies and other institutions that set the scene and circumscribe life. When planners – today and in the past – design cities from the ground up, they usually start by establishing neighbourhood units. Most neighbourhoods throughout history, on the other hand, were not planned by authorities; rather, they grew gradually or organically: what social scientists call from the ‘bottom up’ – through the daily activities of residents. It is the generative quality of these bottom-up processes that concerns most of the important scientific research on both ancient and contemporary cities today.

In their bestseller The Dawn of Everything (2021), David Wengrow and the late David Graeber claim that, by focusing on bottom-up, generative forces, they succeed in ‘overturning conventional wisdom on the origin of cities’. Early cities, they argue, could be organised by councils, using democratic methods based on egalitarian principles. Apart from the fact that such models have become rather common in archaeology over the past decade, this claim is dubious because the authors exhibit a very incomplete understanding of generative processes. They emphasise grass-roots actions, where people gather to make decisions together in small groups. But there is another type of generative process – spontaneous order – in which social change arises as the byproduct of interactions among individuals.

Walking to work, shopping for basic needs, making friends, attending religious ceremonies – these activities all lead to the concentration of social interactions in neighbourhoods on a scale much smaller than the city. Daily social activities within households and blocks, largely independent of governments and controlling institutions, are the building blocks of the spontaneous-order type of generative force. In his book The City Shaped (1991), the architectural historian Spiro Kostof described the totality of these social interactions and relations as ‘energised crowding’. This is now recognised as one of the most fundamental processes creating growth and change in cities. It is the centrepiece of my latest book, Urban Life in the Distant Past: The Prehistory of Energized Crowding, but it plays no part at all in The Dawn of Everything (forthcoming, 2023).

The early archaeologists – and travellers – devoted themselves to the monumental remains of ancient urban sites: tall pyramids, luxurious palaces and royal tombs with spectacular funerary offerings. They saw such monumental works as proof of the actions of powerful rulers who controlled large amounts of labour and resources. They were also a source of loot to carry off for museums. It would take two types of conceptual and methodological advances to allow these spectacular ruins to be analysed as urban settlements. One development was to transform the obsession with monuments and elites into more realistic models of the top-down processes that operated in early cities; the other was to supplement the focus on pyramids and kings with a new emphasis on generative processes.

The early archaeological fixation on palaces, tombs and pyramids fed into a story that all early rulers were cruel tyrants who ruled people with an iron fist until the Greeks invented democracy. This view has been overturned by the political scientist David Stasavage, and the archaeologists Richard Blanton and Lane Fargher. Their books show that early governments exhibited a wide range of forms, some more autocratic and some more collective or democratic. Autocratic kings ruled the Maya cities of Central America (200-900 CE), for example. These kings carved their portraits on stone billboards in public locations. When they died, their subjects buried them in luxurious tombs. These are signs of ancient autocratic government. Meanwhile, in central Mexico, archaeologists can’t even identify the presence of royalty at the huge metropolis of Teotihuacan. There are no royal tombs, and no public images of powerful individuals. Most archaeologists believe that Teotihuacan was ruled by a council or a coalition, not an autocratic king. This political distinction may help explain one of the notable differences between the Maya cities and Teotihuacan: the level of wealth inequality was much lower in Teotihuacan. Most commoners in that city were more prosperous, and ruling elites were less powerful and wealthy, than in the Maya cities.

Archaeologists turned away from temples and tombs and started excavating houses

Another advance is in the study of ancient economies. For decades, economic anthropology was locked into a sterile debate between two polarised positions. The formalists claimed that economic models of capitalist economies could be applied without modification to non-Western and ancient societies, whereas the substantivists countered that these systems were so radically different that they required their own economic models. This same debate raged in the field of Classics, between the ‘modernists’ and the ‘primitivists’. Of course, both positions were wrong, and methodological advances in archaeology paved the way for a more productive approach. I was part of a group of archaeologists who developed methods to identify markets and commercial economies in early societies. We found that most cities had commercial economies that were not capitalist. That is, they had money, markets and merchants, but wage labour and property markets were minimal.

Some ancient societies – most notably the Inka of Peru and Dynastic Egypt – had command economies, where the state dominated the economy. Command economies leave their traces in a proliferation of central storage facilities, and a lack of shops or marketplaces. These developments in political and economic analysis have led to more realistic models of the institutions and governments – the top-down forces – that structured life in early cities.

Beginning in the 1970s, archaeologists who wanted to study life and society in the past turned away from temples and tombs and started excavating houses. To interpret the new domestic remains, we drew on the flourishing field of household analyses in anthropology and history. The field of ‘household archaeology’ remains a vibrant approach to past social life in all parts of the world. My colleagues and I now have a much better grasp of the activities and social conditions of people in the past, both commoners and elites. In my excavations of households in a provincial area of the Aztec empire, I was surprised to find that these farmers and crafters had been quite prosperous. The trash-heaps behind each house yielded many fragments of imported goods (ceramics, obsidian, bronze tools), as well as many non-utilitarian goods, from incense burners and ritual figurines to musical instruments. I had expected to find these people had been poor, isolated and downtrodden; instead, they were prosperous and well connected. My colleagues and I have recently turned to studying wealth inequality, using the size of houses as a measure of household wealth. This quantification allows us to use measures from economics, such as the Gini index, which in turn permits rigorous comparisons of sites and regions.

While the household turn has been very productive, the lack of a theory about cities and city life – beyond simplistic views that rulers tried to control everything – has held back the recognition of ancient cities as urban places. While field archaeologists were busy excavating houses, Ian Hodder was building a postmodern approach known as ‘post-processual’ archaeology. Processual archaeology was the name of a movement in the 1960s, headed by Lewis Binford, to make archaeology more scientific. Hodder and the post-processualists drew on literary theory to reject scientific theory and hypothesis testing in favour of speculative interpretations of the ‘meaning’ of artefacts at individual sites. Post-processualism dominated archaeological theory – particularly in Great Britain – for decades, and its antagonism toward science and comparative analysis held back advances in understanding life and society in early cities.

In his pathbreaking book The Limits of Settlement Growth (1995), the archaeologist Roland Fletcher developed a model that relates the size of cities to their growth and characteristics. Fletcher noted that human settlements (cities, towns and villages, today and in the past) tend to grow, but if they grow too large – in terms of either absolute population numbers or population density – stress and social problems will set in. Fletcher identified several thresholds of size and density that he called ‘limits to settlement growth’. Once these limits are reached, if the people and society cannot make advances in technology and communication, social stress will lead to chaos and unrest, preventing additional city growth. For early urban centres to cross a threshold from villages to towns or cities (of approximately 10,000-20,000 people, living in around 1 sq km), they needed to develop communication methods such as writing, and architectural features that allowed privacy and the segregation of activities in different rooms or buildings. Excavations have vindicated the theory.

People created neighbourhoods to make urban life legible and social life more congenial

Fletcher applied his model of settlement growth quite widely, including the earliest sedentary villages (prior to the early cities), modern cities after the Industrial Revolution, and even cities going into the post-industrial future. However, because he limited his publications to an archaeological audience, his work remains poorly known in other disciplines. For our understanding of life in early cities, Fletcher’s central insight is that population size and density are crucial qualities of a city with vital implications in many social realms. Sociologists, economists, geographers and urban historians have all contributed to the case for a dynamic relation between threshold points and urban population size and density. There is wide support for the fact that size matters for cities and urban life, in ways that can hardly be exaggerated. For whatever reason, Graeber and Wengrow, in The Dawn of Everything, ridicule this finding and assert – without evidence – that cities with large populations can be governed effectively without rulers, managers or bureaucrats.

Population size structures urban life and society, in every city in history, by making neighbourhoods. People in cities do not want to interact and engage with tens of thousands of people on a daily basis. They instead created neighbourhoods to make urban life legible and social life more congenial. For cities, neighbourhoods create social order, they also provision services and facilitate the communication of information to residents. Of course, even though in some ways people live in a neighbourhood more than in a city, any given individual’s face-to-face social interactions may take place over the entire city. Social exchanges at one’s place of employment, walks in the park, purchases at a downtown store, conversations while commuting all offer transient contacts with strangers in public, in shops and other places.

Fletcher’s work rests on the fact that, as the population and density of a settlement increase, the number of potential interactions for each person increases exponentially. Energised crowding occurs when large numbers of social encounters take place within a settlement. In large settlements – cities – the joint effects create a ‘buzz’ whose consequences go far beyond the immediate results of individual interactions. These interactions among individuals have social and economic consequences. Perhaps Graeber and Wengrow’s refusal to acknowledge the role of spontaneous order and energised crowding is responsible for their claims that population size is not important in cities. Another contributing factor may be their distrust of both the natural sciences and the social sciences. Yet it is the methods and concepts of quantitative scientific research that are making the greatest advances in urban understanding today.

One of the true breakthroughs in understanding how population size affects other aspects of cities and settlements has come from the new field of settlement scaling research. In the 1990s, Geoffrey West, Brian Enquist and James Brown – scientists at the Santa Fe Institute (a think tank in the science of complexity) – had worked out a quantitative model that explains how the metabolic rates of mammalian species are predicted by their body size. After West – who trained as a physicist – had contributed to this advancement in biology, he turned his gaze to the social sciences. Could similar quantitative regularities be identified, and could they be predicted with a model? West gathered a group at the Santa Fe Institute that included the physicist Luís Bettencourt, the economist José Lobo and several others, and they decided to attack the problem of the size of cities.

West and his group quickly discovered two quantitative regularities: city population predicted both the area of a city, and its level of social and economic outputs. But how did this work? What were the underlying mechanisms that produced the observed patterns? In 2013, Bettencourt published a mathematical model that predicted these regularities based on social interactions in the urban built environment. Unlike the models of economists, which typically require wage labour, capitalist firms and the modern financial system, Bettencourt built his model on more fundamental assumptions: movement has a cost; social interactions have benefits; and interactions increase with the size and density of a city’s population. As a result, the size of a settlement is both cause and effect of the overall productivity of the settlement.

Bettencourt’s model predicts specific quantitative outcomes. Within a system of cities, quantities such as the area of the settlement, the amount of infrastructure and the economic outputs all increase exponentially with population, with very specific rates (as measured by an exponent). Area in structured settlements, for example, scales with population with an exponent of 5/6. In other words, with an exponent less than 1.0, area increases less rapidly than population, a pattern called sublinear scaling. A city with twice the population of another city will occupy less than twice the area. Another way of saying this is to note that – within a given urban system – larger cities are more dense than smaller cities. Because energised crowding is based on the number of persons and the population density, this means that the higher densities of larger cities will have greatly increased levels of energised crowding.

Population density and energised crowding lead to greater wealth creation in larger cities

A more surprising scaling relationship holds for economic outputs – GDP, the number of patents, income, as well as negative quantities such as the amount of crime or poverty. These quantities scale at a rate greater than one (7/6), called superlinear scaling. A city that is twice as large as another will have more than twice the GDP per capita, and more than twice the crime, compared with the first city. The scaling exponents predicted by Bettencourt’s model (5/6 and 7/6) were derived to explain the data from contemporary urban systems. Some of the measurable consequences of the population size of cities are quite regular and precise. Such results are examples of what economists call agglomeration effects – the advantages that come from the concentration of population and firms within settlements. These results led Bettencourt to coin the term ‘social reactors’ to characterise cities: social interactions generate outcomes that go beyond the sum of the individual encounters. In the words of the historian Fernand Braudel: ‘Towns are like electric transformers. They increase tension, accelerate the rhythm of exchange and constantly recharge human life.’

Scott Ortman – an archaeologist then at the Santa Fe Institute and now at the University of Colorado – realised that Bettencourt’s model is sufficiently general to apply to settlements in the ancient world. So, he took the next step by applying Bettencourt’s scaling formulae to an ancient settlement system. Archaeologists can easily measure the area of sites, but to estimate their population independently of the area is trickier. Ortman reworked published settlement pattern data from the Basin of Mexico archaeological survey project and found that the same scaling exponents reported by Bettencourt for modern urban systems held for pre-Spanish central Mexican sites as well. Was this a signal of fundamental similarities in urban processes through history? Or, was it a fluke, an accidental result?

To find out, Bettencourt, Ortman, Lobo and I established the Social Reactors Project in 2013. Our goal was to test the application of Bettencourt’s scaling model to settlements before the modern era. We have collaborated with groups of specialists, and investigated more than 15 early systems. My student Rudolf Cesaretti used historical records to show that that medieval towns in Europe exhibit sublinear scaling of area with population, as predicted. Using archaeological data, our group has identified superlinear scaling of wealth (house size) with population for many cases. That is, as the size of a settlement increases, the amount of wealth per capita (using house size as a measure of household wealth) goes up. Just as big cities today have higher GDP per capita than smaller cities, large settlements in the past had higher household wealth per capita than smaller settlements. Population density and energised crowding lead to greater wealth creation in larger cities, compared with smaller settlements, in the past and in the present.

Some of the results of the social reactors project have been surprising. For example, we found the same scaling results for a command economy as for commercial economies. Cities in these two types of economy differed in their layout and operation, so we wondered whether the scaling patterns identified for commercial economies would hold for command economies as well. While there are not enough settlements with data to analyse Egyptian settlements for scaling, we were able to use settlement pattern data for the Mantaro Valley of Peru, a pre-Inka settlement system likely organised as a command economy. The superlinear scaling of house size with population matched our cases with commercial economies rather precisely. Ortman then found the same results for two ancient village societies in the upper Great Plains of North America. This case shows that the measurable effects of energised crowding are evident not only in urban systems, but also in systems of smaller, village settlements.

That nearly identical scaling relationships exist for a wide variety of premodern settlement systems points to the importance of the underlying mechanism. Bottom-up social interactions, leading to energised crowding, have structured urban life and societies through the ages. This does not mean that cities or behaviour or social patterns were identical in all these cases. Medieval European cities, for example, differed from those of the pre-Inka Mantaro Valley of Peru. They were larger settlements, with many commercial institutions, and technology – in everything from wheeled transport to ironworking – was more advanced. Yet, underneath this great variation, and perhaps in spite of it, the basic processes by which social interactions and energised crowding led to patterns of population density and wealth creation must have been quite similar. This is a remarkable finding.

The results from the Social Reactors Project confirm the role of energised crowding in cities and settlements. Settlements with higher populations and higher densities have greater energised crowding, and this in turn changes urban life in three primary ways. First, energised crowding leads to neighbourhood formation. With too many people in a city, people create communities on a smaller scale – neighbourhoods – and energised crowding then continues to operate within those smaller units. Second, energised crowding leads to economic growth and development. This growth in per-capita outputs creates the superlinear scaling patterns described above. For archaeologists, economic growth is difficult to identify directly, and we therefore use the scaling results as a measure of growth. Finally, energised crowding also brings negative effects: stress and anxiety on the individual level, and crime and poverty on the city level.

Do the results of the settlement scaling research imply that cities and villages did not differ in ancient times? Not at all. Most archaeologists use one of two definitions to identify cities as settlements distinct from villages or towns: sociological, and functional. The sociological one borrows the sociologist Louis Wirth’s classic definition of a city as a permanent settlement with a large population, a high density, and evidence of social heterogeneity (social or economic variation). While this applies to the largest and most complex early settlements – eg, Rome, Chang-an, Teotihuacan – as cities, it leaves out most ancient centres that were not as big, dense or complex.

Works promote the idea that there was something wrong with the Maya. They flourished for centuries!

The functional one defines cities as settlements that are home to urban activities and institutions that shape a hinterland. This approach focuses on what cities do, not on what they look like. It has the advantage of placing cities into their specific geographies. Some cities defined this way will lack large populations, high density or social heterogeneity. The Classic-period cities of the Maya (200-900 CE) are a prime example. With small populations spread out at low urban densities, their influence extended beyond the capital. The kings whose palaces dominated Maya cities ruled petty kingdoms, and the towering temple-pyramids served hinterland populations. For archaeologists using the sociological approach to urban history, Maya cities like Tikal or Copán were simply too small to qualify. But if we look using the functional method of identifying cities, we see that these settlements indeed existed in a dynamic relation with a hinterland. Residents of nearby villages came into town to attend a market or witness a ceremony, and those organised activities constitute urban functions.

The spectacular collapse of these Maya cities has blinded scholars to their value for understanding cities and urbanism in scientific terms. Many archaeologists devote so much time looking for the reasons these cities collapsed that they have not bothered to analyse them as urban settlements. Cities collapsing naturally leads to many questions, including how long early cities lasted. Works such as Jared Diamond’s bestseller Collapse: How Societies Choose to Fail or Succeed (2004) promote the ideas that there was something wrong with the Maya, and that their cities were fragile and doomed to failure. Yet Tikal, Copán and the other Maya cities were in fact quite durable; they established themselves, and grew and persevered in a highly difficult jungle environment. They flourished for five to 10 centuries!

Why did some early cities persist for centuries while others lasted only decades? Questions of durability and sustainability bear on the pressing problem of how cities are adapting to the climate crisis today. Cities have responded to shocks and adversity over the entire range of urban history. Some survived for millennia (think of Rome or Beijing), and others only a few years (these tend to have names like Tx-LT-43). When archaeologists determine the reasons for the success and failure of individual cities in the past, that knowledge could be useful in plotting the course of urban change into the future.

Cities initially developed several thousand years ago and, since then, the level of urbanisation of the world (the percent of the population living in cities) has increased relentlessly. Our future will be an urban future. How can we best understand how cities work, what makes them grow and shrink, and what determines the quality of life of their residents? While past kings, priests and other leaders are part of the story, it is clear that in the past, as today, generative processes – the actions and interactions of people in cities – were the driving forces in the expansion and success of cities.