Why water alone does not explain later cities
All six of the first civilizations sat on rivers, which makes water look like the whole answer. It is not. A large river is not required, and on its own it is not enough. Madrid grew large beside a small stream. Singapore has almost no land and no river, and it became one of the busiest cities on Earth because it sits on the strait that every ship in the region must pass through. A city does not need one advantage from geography. It needs several advantages in the same place at the same time. The biggest cities are where the largest number of them come together.
The journalist Tim Marshall, in his book Prisoners of Geography, shows how the physical map sets these limits for whole countries. His clearest point corrects the assumption about rivers directly. A small river that boats can travel along is worth more than a huge river that they cannot. The United States grew rich on the Mississippi system, because its rivers connect to each other, boats can travel them, and they all drain to a single outlet at the sea. Several of the largest rivers in Africa carry far less trade despite their size, because rapids and waterfalls block boats from passing. Marshall's other repeated point is the value of a port, particularly one that does not freeze in winter. That is why reaching open sea has decided the fate of countries for centuries. In his account, geography sets limits that a country cannot easily escape.
The things that make a city grow divide into two groups. Geography provides the first group. It provides water a city can use, which is not the same as a large river. It provides a position on a trade route, such as a river mouth, a strait, or a natural harbour. It provides a coast or a port for sea trade. It provides ground that is easy to defend, such as hills, islands, or a bend in a river. And it provides flat fertile land, with room for the city to expand and food to supply the people who arrive. Human decisions provide the second group. People build roads running in many directions, so the city can trade with everyone instead of only along one river. And rulers make political decisions, because a ruler can declare a place the capital and make it grow where geography alone would not have. Madrid, Brasília, and Washington all grew that way.
Geography decides where a city starts. What decides whether it works, and whether people want to live in it, is how the city is laid out. The planner Kevin Lynch spent his career on that question. In his book Good City Form, Lynch does not ask where cities appear. He asks what makes an existing city a good one: whether people can understand its layout and move through it, whether it suits the way they actually live, and whether it serves the people inside it. His work warns against the idea that geography decides everything. Geography makes a place possible. Human design decides whether what gets built there works or fails.
So the honest rule is this. No single cause builds a large city. Cities grow where several causes come together in the same place. In the ancient world water was the one cause that nothing could replace, so every early civilization sat on a river. Technology then changed that. Pipes carried water to places without it, engines carried goods across land, and ships carried them across oceans. The other causes became more important, and the combination could now form in places a river had never reached. That is why the first cities are on rivers and the largest cities today are on coasts and trade routes. Geography sets the limits. Human decisions and human design work inside those limits. The largest cities are where geography and human decisions agree.
The four sizes of human society
Before cities existed, human groups grew through four stages. Each stage held more people than the one before it, and each depended less on family ties than the one before it. Anthropologists use these four stages to describe how societies grow larger.
- The band, 10 to 50 people. Everyone is related to everyone else. They move from place to place, they hunt and gather their food, and they have no leaders and no ranks. Humans lived this way for almost the whole of our history.
- The tribe, hundreds to a few thousand people. Several bands connected by family ties, a shared language, and shared ancestors. There are still no ranks, and family still organises everything. At most there is a respected elder, who can advise but cannot give orders.
- The chiefdom, thousands to tens of thousands of people. The first permanent inequality appears. A chief and his family hold a position above everyone else, and a son inherits that position from his father. Family still organises the group, but now some families rank above other families.
- The state, tens of thousands of people and above. Family ties are no longer what holds the group together. People who are not related live together under a government, written laws, taxes, and a permanent armed force. Cities are what make this possible.
The size at which a group needs a government
There is a rough limit to how many people one person can know well enough to keep track of them. That limit is estimated at around 150, and it is called Dunbar's number. Below that size a group manages itself without formal rules, because everyone knows everyone, and a person's reputation is enough to keep order. Above that size the group contains people who do not know each other. Knowing everyone personally no longer works, so written rules, defined roles, and someone to enforce them have to take its place. Passing this limit is the point at which a society has to invent a government. The number itself is disputed. It comes from a line drawn through the relationship between brain size and group size in other primates, then extended to humans. A 2021 study repeated that calculation with updated data and modern statistics, produced averages far below 150, and found the range of uncertainty so wide that no single figure can be stated. Dunbar answered that the study used the wrong statistical method and ignored the evidence gathered since. The general point holds either way. A group small enough for everyone to know everyone needs no formal rules, and a group larger than that does.
How the city divided society into levels
Farming produced more food than the farmers needed, and that extra food could be stored. Stored food can be owned, counted, and taken, and whoever controls it controls the people who need it. That control became a permanent position, and the position became something a son inherited from his father. The flat society of the band became a pyramid with levels, built roughly in this order from the top down.
- The priests. The temple stored the grain and decided who received it, so religious authority and control of the food supply combined into one thing first. In Sumer the temple was also the first place that held wealth on behalf of others and lent it out.
- The king and his government. The role of the chief became permanent, and a son inherited it from his father. Officials supported him, and they collected taxes, kept records, and gave orders.
- The soldiers. A permanent group of fighters, paid to defend the stored food and to take more of it from others. They protected the levels above them and enforced their orders.
- The merchants. People who lived by trading and moving the stored food and other goods, instead of producing food themselves.
- The farmers. The majority of the population. They produced the food that supported every level above them, and they kept very little of it.
- The slaves. The bottom level, created by capture in war and by unpaid debt. The law of the state treated them as property rather than as people.
Where the biggest cities are today
If water and trade routes built the first cities, the same pattern should appear in the biggest cities now. It does, with one difference. The largest cities on Earth today are not on big rivers far inland. They are on coasts, on deltas, and at river mouths, where a river meets the sea and river trade connects to sea trade in the same place. The type of water changed from river to coast as trade became worldwide, but water is still there underneath every one of them. The exact order by population is disputed, because different counting methods put Tokyo, Jakarta, or Guangzhou first. The list below is therefore grouped by the water advantages each city has, not by a single ranking.
- Tokyo, Japan. On Tokyo Bay, on the Pacific coast. It has a natural harbour, and the flat Kanto plain surrounds it. The water gave it trade and the plain gave it room to expand.
- Jakarta, Indonesia. On Jakarta Bay, on the north coast of Java. Thirteen rivers run through the city and into the sea. River and coast meet in the same place.
- Shanghai, China. At the mouth of the Yangtze, on the East China Sea, built along the Huangpu River. Having both a river and a coast made it one of the busiest ports in the world.
- Delhi, India. On the Yamuna River, which flows into the Ganges. It is one of the few very large cities that is inland. Its river supplies the water, and its position as the national capital is a political decision added on top of the geography.
- Mumbai, India. On the coast of the Arabian Sea, built on a group of islands that were later joined into one piece of land. A deep natural harbour made it the main entry point for trade into India.
- Cairo, Egypt. On the Nile, at the point where its delta begins. It is the only city on this list supplied by one great river alone, the same river that supplied ancient Egypt.
- São Paulo, Brazil. Inland, on high ground near the coast. It is the exception on this list. It grew on wealth from coffee and then from industry, not on water. It shows that when enough other advantages come together, they can outweigh a weak position for water.
The cities in this list
Each card below opens that city's page on this site.
The pattern is clear. Almost every one of these cities sits where a river meets the sea. That position combines river water, sea trade, flat land, and in many cases the status of national capital, all in one place. The few very large inland cities, such as Delhi, are supported by other advantages, mainly political decisions. This is the same rule as before. No single cause explains any of them, and water is present in nearly every case, now as it was at the start.
The same structure today
Modern countries are the same structure at a much larger size. Law, money, and public administration have replaced family ties and personal reputation, and those three are what allow millions of people who do not know each other to live together. The older behaviour still continues underneath. Neighbouring groups that are almost identical still draw firm boundaries between themselves, because in the past their families and their social connections stayed inside the group and rarely mixed. Groups that are very similar tend to make their small differences seem larger, so that they remain separate. Freud named this pattern the narcissism of small differences. Someone from outside can gain authority in a divided community for exactly this reason, because he takes no side in the local disputes, and each side accepts him more easily than it accepts the other. Population growth and people moving between places slowly remove these boundaries, but a shared sense of identity continues for a generation or more after the conditions that created it have gone.