[dzdt]

Alternate hypothesis: locations 2000-year-old cities predicts modern day prosperity.

And also locations of 2000-year-old cities predicts locations of 2000-year old roads -- uncontroversial I think.

The proposed mechanisms are (1) cities are stable on millenial timescales, continuing to develop and attract wealth (2) geography is stable on millenial timescales, so the places where wealth-attracting cities tend to be located are the same then as now.

Contra (2), there has been a shift in importance of different transportation networks and energy sources. The roads discussed in the article are transportation network, but waterways historically have been even more important. And more recently rail networks as well. The shift from water to coal to grid-distributed electricity has loosened the connection of energy to geography.

[mcphage]

> And also locations of 2000-year-old cities predicts locations of 2000-year old roads

The article claims the correlation is the opposite of what you stated—roads were built, and then cities built up around them. From the article:

> “Roman roads were often constructed in newly conquered areas without any extensive, or at least not comparable, existing network of cities and infrastructure,” Dalgaard and his colleagues write. In many instances, the roads came first. Settlements and cities came later.

[dredmorbius]

The Romans were not unaware of principles on the siting and prospects of new settlements. See Vitruvius's De Architectura, or The Making of the Ancient Greek Economy by Alain Bresson:

As early as the fourth century BCE, ancient Greek city planning was very far advanced. When possible—that is, when a city was to be founded or refounded—the site was carefully chosen, taking into account first of all climatic conditions. Drawing on a tradition that goes back to the physician and hygienist Hippocrates and to Aristotle, and whose “intermediaries could only be the architects who built the Hellenistic cities,” the Roman architect Vitruvius recommended choosing a site where the temperature remained moderate and that was far from swamps, in order to avoid miasmas and fogs.

https://press.princeton.edu/titles/10376.html

[jakamau]

Then how about "Only 2,000 year old cities that were prosperous maintained their roads to survive until today"

> the Roman roads in the Middle East and North Africa (MENA) weren’t maintained the same way they were in Europe...

> The correlation between ancient roadways and modern-day development so prevalent in Europe is much smaller and less significant for the Middle East and North Africa.

[jacobolus]

I wonder how much of that can be attributed to using camels to get around vs. horses & carts, with roads less necessary/helpful for commerce/transportation.

[bobthepanda]

Most of North Africa and MENA's population centers are located near the coast or major navigable rivers (namely, the Nile). River and sea transport is much cheaper than land transport, even today. Why maintain a bunch of expensive roads when you can just have well developed waterborne trade?

[evgen]

The location for most major cities around the world is where there was a transition point between modes of transport. Sea <-> river, river <-> road, then later adding canals and rail into the mix. Where such transit points exist you are going to develop infrastructure for storing and re-packaging goods from one mode of transportation to another, which will lead to trade of the goods at these locations, which leads to more goods coming in to be traded, and round and round it goes.

[baud147258]

Even with camel, roads are still helpfull to move around.

[singularity2001]

Yet roads were probably build through towns, which were created in geographically advantageous locations.

[toasterlovin]

Most major historical cities are located near water of some kind (either a seaport or on a navigable river), so I find it hard to believe that roads were the driving force in determining where cities ended up.

[patall]

But often it was rather some way of crossing the water that was that driving force. Oxford and Cambridge come to mind, or in Germany: Frankfurt (furt = ford), Erfurt, Saarbrücken (brücke = bridge).

[toasterlovin]

If it was just about crossing water, you should also have random large cities in the middle of nowhere, but if you start looking at all the major cities of the world on a map, they’re almost all on a coast or a major river.

[baud147258]

I thought that settling on a river is more for ease of access to a water supply than access to water as a mean of transportation. I may (surely) be wrong too.

[toasterlovin]

I’m sure that’s a factor. But most of the largest cities in the world are specifically on the coast or on a large river very close to the coast. You can’t drink the water in the ocean, so that leaves trade as the main driver (by my reckoning, at least).

[dzdt]

Yes they note that in some instances the roads preceded the cities. But when they do analysis there is no effort to seperate cases where the cities preceded the roads.

[baud147258]

> (1) cities are stable on millenial timescales

Unless they are razed to the ground. But I agree that cities tend to stay on the same place.

[falsedan]

Yes I love travelling to Babylon, or visiting the sites of Troy. And the trip to Machu Picchu is definitely worth it!

I know London and New York look like that will be around forever, but no city in the US is older than 600 years. We don’t have enough evidence to support that cities are stable on millennial timescales, since a few hundred years is enough for drastic change.

[Zardoz84]

Cartago, Cartago Nova(Cartagena), Roma, some Egyptian cities, etc are far older. we are talking of two/three millenniums. But is true that some older cities disappeared, displaced by more modern cities, climate changes, geopolitical changes or recurrent races by frequent wars.

[nafey]

Indian cities like Patna have been around before Roman empire

[baud147258]

> We don’t have enough evidence to support that cities are stable on millennial timescales

Some cities have existed for millenia, like Rome, but with ups and down. And drastic change can be even quicker than hundreds of years, for example a plague or a conflict will cause change quickly.

[chimprich]

London is not a great counter-example because it has been a major settlement for at least a couple of millennia.

[falsedan]

No it hasn't, Londinium was abandoned when the Romans vacated Great Britian.

[TheCoelacanth]

The Roman settlement was, but London as a whole was only abandoned briefly if at all. There was a major Anglo-Saxon settlement outside the Roman walls in what is now Covent Gardens and the Strand definitely in the 6th century and possibly in the 5th. The Roman settlement started declining in population towards the beginning of the 5th century.