Let's take a look at the major cities of the world:


  1. London
  2. New York


  1. Tokyo
  2. Hong Kong
  3. Singapore
  4. Shanghai
  5. Sydney
  6. Dubai
  7. Beijing
  8. Paris

I'll point out features for a few of these

London: Located on a major river with easy access to the North Atlantic

New York: Located on an enormous natural harbor with easy access to the North Atlantic and the Eastern Seaboard.

Hong Kong and Singapore: Deep natural harbors protected by their consitutent island chains, both on the Pacific, Singapore located between the Indian Ocean.

Deterministic most simply means given the same input, the same output always results.

In other words, would the course of human history always end up making cities in these locations major world cities? Or at least some near variant thereof?

For example, if the Nazis had won WWII, Singapore would still have its natural advantages, and if the French had successfully colonized North America entirely, it's hard to imagine New York, given its harbor and placement for trans-Atlantic trade, would not grow as large as it has today.

Even if the United States had nuked Tokyo, it would have sprung right back, look at Hiroshima and Nagasaki.

A couple people have pointed out that the major cities of the world have changed over the course of history. This is, of course, true. I'm looking at a sampling of modern major cities, so the question is really about whether, in many alternate 2016s, major cities would always be in these locations.

  • 2
    $\begingroup$ One can argue that "if the Nazis had won WWII" means a pretty large deviation from "same input", in which case "deterministic" according to your definition becomes a moot point. (Also, at least unless I recall things completely wrong, I don't think the outcome of World War II had any major impact on what city was the capital of Germany.) $\endgroup$
    – user
    Commented Jan 20, 2016 at 12:21
  • 3
    $\begingroup$ Climate is also a significant factor. There is a wonderful natural harbor near Barrow, Alaska- population 4212. $\endgroup$
    – cobaltduck
    Commented Jan 20, 2016 at 15:02
  • $\begingroup$ Related, over on History.SE $\endgroup$ Commented Jan 20, 2016 at 16:11
  • 4
    $\begingroup$ Good city sites are good city sites regardless. So that much is deterministic. Whether cities get built there and whether those cities become "major" cities is less so. A great city site is probably more likely to spawn a great city, but there are many contingencies of history that can intervene. $\endgroup$
    – wjousts
    Commented Jan 20, 2016 at 16:45
  • 1
    $\begingroup$ The fact that the position of the biggest city change during the history, is a serious problem against the hypothesis of "Location of biggest city is deterministic". $\endgroup$
    – Kolaru
    Commented Jan 20, 2016 at 17:23

6 Answers 6


Not at all.

There are many places with these characteristics that do not have major city status and the elements that you have selected were often secondary.

London is the city I know best so we'll use that as an example. Access to the North Atlantic: Completely irrelevant for the first 1500 years of the city's existence. Even when it became relevant, there were better ports to use, Bristol, Southampton, Liverpool. London was built for easy access to France and is the merger of the cities of London and Westminster (aka Ankh and Morpork).

What you need to build a major city:

Let's start with what you need to build a city at all:

Water London sits on a massive underground fresh water reserves, it also has a lot of (mostly now underground) fresh water rivers running through.

To make it a major city:

Communication, trade, transport. For much of history these were the same thing. The only way to get anywhere was by sea, even until the industrial revolution, it was cheaper and easier to ship from Lisbon to Liverpool than to go overland from Liverpool to York. Trade and transport are still key, hence why the British Government is desperate to expand one of the London airports.

Make it your capital. How many of the cities on your list are not the (state or) national capital? (Hong Kong is a special case, Shanghai was a different special case). People are attracted to the centres of power for whatever reason, once this has started it tends to continue. People go to where the jobs are, the jobs go to where the people are. Outside Dubai City, the rest of Dubai is a wasteland.

The main reason London is a key global city is Empire. It was the centre of the world for a century. Capital of the last global empire. The wealth of the world passed through London, as did the people of the world. The city is built on that. It's not the best port, not the best harbour, it's not the best building land, most of the area close to the river is below high tide and marshland. The tide in the river flows too fast etc.

Political Will. This is the difference between a merely big city and a major one. 4 of the cities in the top 10, Hong Kong, Singapore, Shanghai and Dubai, were built by a government basically saying, "we want a major city, put it here". In these cases a small settlement or port already existed, but was expanded by political will.

In conclusion. The factors you provide are required to make a great city, but they will not alone cause a great city to exist.

  • 2
    $\begingroup$ Where a large city is a state capital, that's usually because it's a big city, and not vice versa. And most of the cities on your list are not national capitals: 5 out of 12 are capitals of a nation that's larger than just the city (can't count Singapore). And Tokyo was not the capital of Japan for most of its history. $\endgroup$
    – Mike Scott
    Commented Jan 20, 2016 at 12:04
  • 4
    $\begingroup$ Capitals move, but even artificially created capitals become big cities. London also wasn't the capital a lot of the time. $\endgroup$
    – Separatrix
    Commented Jan 20, 2016 at 12:19
  • 1
    $\begingroup$ No artificially created capital is on the top 12 list in the original post. And London has been the capital of England continuously since the 11th century. $\endgroup$
    – Mike Scott
    Commented Jan 20, 2016 at 12:32
  • $\begingroup$ It's only a list of 10 not 12. 9 of the next 14 are current or former capitals. 3 (maybe 4) of the top 10 are artificial imperial cities, I'm still counting Singapore even though you aren't, it's also one of the imperial cities. (Singapore, Hong Kong, Shanghai (Dubai)) $\endgroup$
    – Separatrix
    Commented Jan 20, 2016 at 12:50
  • 3
    $\begingroup$ In engineering we call these "necessary" and "sufficient" conditions. The cited conditions may be "necessary" for a major city but there are locations that have these necessary conditions but no major city - meaning the conditions are not, of themselves, "sufficient". $\endgroup$
    – Jim2B
    Commented Jan 20, 2016 at 14:38

There is no definite answer, because we've never had an alternate history, only the one we have, but I'd risk a "no, it's not deterministic."

There are certainly statistical commonalities between major cities. In particular, you'll see common patterns like the major cities forming in locations with good harbors, because good harbors are good for city making. However, it would be very surprising if human history had to always end up making all major cities in the same places. First off, that implies that "major city" has some fundamental underlying meaning, which it doesn't. Second, we can find easy reasons why this wouldn't be the case. While there are "good" places for cities to be founded, there are clearly more pressing concerns in the minds of a founder when they're planting their city. A city might move up or down the coast depending on factors. For example, if a neighboring country's borders pushed closer to a capitol city, it is highly likely that a different city would become the capitol of your nation. If a capitol moves, all sorts of commerce and government lines get redrawn, and different cities may be accentuated.

I think a corollary to your claim would be to argue that a human being develops from a fetus to an adult in a deterministic fashion, which we know is patently false, and yet nearly every person you met has two eyes, a nose, walks on their legs rather than their arms, etc. Deterministic is a powerful word, and I just don't think it works out here.

  • $\begingroup$ As to "planting a city," it would make no difference if London was founded 20 miles farther downstream, it would still be an enormous megalopolis. I don't follow your second point. What do you mean by patently false? As far as I've learned, fetal development happens in a very linear and consistent fashion. $\endgroup$ Commented Jan 20, 2016 at 6:07
  • $\begingroup$ @Azor-Ahai I think the issue may be the definition of "no difference." Yes, if you are willing to ignore all the things that change, then the remaining things stay the same. If you ignore all of personality and the effects of personality on the body, such as musculature strength, a baby grows into an adult in a deterministic fashion. The vocal box descends at a certain time, puberty happens at a certain time, etc. However, many people consider the stuff inbetween to have at least some importance. $\endgroup$
    – Cort Ammon
    Commented Jan 20, 2016 at 7:14
  • 1
    $\begingroup$ For example, it would be easy for you to identify histories where London got founded hundreds of years later than it did in our world. It might not become a bussling metropolis until the year 2100. Does that count as "the same" for purposes of determinism? If a city's population peaks at 500,000 instead of 1,000,000 is that "the same?" $\endgroup$
    – Cort Ammon
    Commented Jan 20, 2016 at 7:16
  • $\begingroup$ Likewise, if there was a future "big city," but instead a nuclear event happened, it is highly unlikely that a big city will appear there. $\endgroup$
    – Cort Ammon
    Commented Jan 20, 2016 at 7:20
  • 7
    $\begingroup$ @Azor-Ahai London would never have been founded 20 miles downstream, because the Thames wasn't crossable there until the 19th century. It's not a coincidence that London is at the exact lowest point on the Thames that can be forded or bridged using mediaeval technology. $\endgroup$
    – Mike Scott
    Commented Jan 20, 2016 at 12:07


The question has a (somewhat arbitrary) list of cities which are important and prosperous right now. We have no reason to assume their prosperity will continue forever.

We can see this by considering cities which have stopped being rich and important:

  • Venice was once the richest trading city in Europe, but now is of interest only as a tourist attraction.
  • Istanbul has an excellent natural harbour and strategic location. For centuries, it was the capital of the Byzantine (and later Ottoman) Empire and legendary for its wealth and culture. It is still the most important city in Turkey, but no longer one of the top handful in the world.
  • Cairo with its predecessor Memphis was one of the greatest cities in the world for centuries. Today it is large but poor and beset with problems.
  • Baghdad again was famous for its wealth, but it never really recovered from the Mongol sack of 1258. Needless to say, subsequent wars have not been kind to it.
  • Buenos Aires has a similar position to New York, at the mouth of a great river, but is not nearly as rich as it once was.

All of these cities have natural advantages, particularly access to water in the form of a river, the sea, or both. This helped them rise to become rich and powerful, and remain at least somewhat important today, but it did not prevent them from falling. The reasons why these cities declined go far beyond the local geography; they include complex military, political, economic, and technological factors.

There has been a world in which Istanbul/Constantinople was an "Alpha++" city; we know that from the history books. There is also a world where it is not; we know that because we live in it. There is no reason to suppose the cities listed by the question are any different. A thousand years ago, London was an unimportant backwater and Manhattan Island was a wilderness; a thousand years from now, they may be so again.

  • $\begingroup$ Well, I don't know about "somewhat arbitrary," since these cities were chosen based on their economic importance to the world. I presume there's an algorithm behind it. Of course, if you look at history in a sliding window you can point out major cities that once were, but for this question I meant to imply we're looking at one time frame, that is the present. I'll edit my question to clarify. $\endgroup$ Commented Jan 21, 2016 at 21:55
  • $\begingroup$ The Wikipedia article describes one particular method of ranking. It acknowledges that the authors made a choice of economic metrics, and intentionally excluded political and cultural factors. In that sense, it's arbitrary. It's far from obvious that Sydney is more important to the world than San Francisco, or Paris more important than Berlin. $\endgroup$ Commented Jan 22, 2016 at 16:23
  • $\begingroup$ You missed Detroit from the list of cities that waxed and waned. Should be noted for the speed of 20th century change if nothing else. $\endgroup$
    – nigel222
    Commented Jan 25, 2016 at 16:17

Not really, but it kind of will always look like it might.

Cities are basically centers of logistics, in the broad sense that includes the dissemination and distribution of information and control as well as simply goods. As such they will invariably grow in places where logistics is practical and cities where the logistics is bad tend to not grow very large.

Historically efficient logistics has meant water transport and cities have grown along coasts at places with natural harbours and along navigable rivers. Places where the two meet and transshipment from ships to barges happens have been prime locations for large cities. So places where you find large cities are fairly predictable.

But if you step up from good location for cities and which cities have chances to grow "major" you'll find lots of variables.

The obvious one is that trade routes are connections between supply and demand connected with the technology available within the political limitations of the time.

Supply depends on natural resources and their development. It depends on locations of large centres of manufacture. What is manufactured and how and how well it can be transported depends on technologies available.

Your city may have resources suitable for making excellent glass, but if the techniques required are unknown or nobody has figured the weird sand is valuable or somebody else already has the market, the growth potential will remain limited.

Similarly unless there is population or industry that needs a resource being well positioned to supply it is not very useful. So the population and infrastructure at nearby areas is vital. This is really a major issue since population and infrastructure concentrate around cities, so interactions between cities are very important in which will grow really major. If a city nearby is already much large it will take unusual events to displace it as the major city of the area even if your location is now otherwise better. Similarly if there are no other large cities nearby, having good logistics is not that useful. Trade and industry will go to cities with larger markets.

And the "nearby" is dependent on technology and politics. A person building a factory in South Korea because just to the North on the other side of the border there is population that has need for the product will probably be denied funding. Being on the Atlantic or Pacific was much less valuable before crossing oceans became reasonably safe and economically viable.

It also naturally varies over time. When Rome was able to supply itself with food produced nearby building it away from the sea made sense. When it later grew large enough to require importation of food by ship, the city was not moved to the coast, nor did trade and population move to Ostia.

And then there is the inconvenient fact that humans change geography to allow easier logistics. They make rivers more navigable and dig canals where no connection existed. They make artificial harbours of concrete where no natural harbour existed. They build roads and railroads to supply effective logistics where nothing natural existed. Many river valleys were converted from marshes or jungles to rich agricultural land supporting large populations along the rivers. Large water projects can turn deserts into something capable of supporting cities.

So while you can say that major cities be themselves are "deterministic" the geography of transport routes, population, resources, and industry that determines which cities grow and which do not is dynamic, complex, and variable product of the past history and technological progress.

This turned out to be fairly incoherent and rambling even by my standards. Hope it is still possible to determine what I was trying to say. Suggestions on fixes welcome.


I'd do this as a comment but I can't at this point, so I'll try to bring something more in the discussion.

Human settlement are always built around three concepts :

  • Ease of defense
  • Easy of reach (includes waterway)
  • Close to an exploitable resource

The first and third made for hard to develop township because the locations were often hard to reach, the best case being a combination of the second and third or of all the three. Then you need space around the settlement to develop.

Later settlement like New York were built with different precondition than their earlier counterparts : the defense part was more manpower than strategical location, the ease of reach meant funnelling immigrants from oversea and with more modern transportation means, the exploitation sites got closer.

You can say that the location of cities used to be deterministic, in that the location of the settlements were chosen on similar criteria for similar results.

Now, with modern technologies, you can settle on a plain in the middle of nowhere and build the needed defense and road (train, plane,...) network to get close to your resources then start a flourishing town out of nothing, for no other reason than "you can". You'd need some good PR people and huge assets, but it could work.
Of course, one might wonder why you'd do that...


Yes, there is a deterministic factor.


No, they would not always end up becoming major world cities.

The deterministic factor of becoming major world cities is economic importance, given enough economic importance as input, every city could become major city.

But economic importance changes overtime due to many complicated reasons.

For example, the city with highest economic importance in West Asia was once Baghdad. However, this later change to Istanbul, and now it is Dubai.

However, Istanbul is trying to change its position in recent years as being the bridge between Europe and Asia.

The only reason that many major city is near coastline or river is simply because waterway promotes trading, so it could easily gain economic importance, but this is not always the case.

For example, inland Mecca is on land trading route, so it gains economic importance. Inland Tehran is the hub of Iran, so it gains economic importance.

However those cities usually have great transport network.

Countries tend to send their economic hub as capital, but again this is not always the case.

Sydney, New York, Istanbul are examples, the capital is Canberra, Washington Dc, Ankara

All cities on the list (except Dubai and Beijing) have well developed financial market and exchange to gain economic importance

London has London Stock Exchange and London Metal Exchange

New York has NYSE and NASDAQ

Tokyo has Tokyo Stock Exchange

Hong Kong has Hong Kong Stock Exchange

Singapore has Singapore Exchange Limited

Shanghai has Shanghai Stock Exchange

Syndey has Australian Securities Exchange

Paris has Euronext Paris


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .