Post

The authors gained access to a collection of almost 12,000 deciphered and edited texts that were excavated primarily at the archaeological site of Kültepe, ancient Kaneš. The ruins (pictured) are located in central Turkey, in the province of Kayseri.

201

27.2K

Crémieux

@cremieuxrecueil

Aug 26

The texts look like this. They were inscribed on clay tablets in the Old Assyrian dialect of Akkadian in cuneiform by ancient Assyrian merchants, business partners, and their family members. This tablet is dated to between 1930 and 1775 B.C.

173

The tablets were all from between 1930 and 1775 B.C., and 90% of the sample came from just one generation of traders, between 1895 and 1865 B.C. The reason is that Kaneš experienced a major fire in 1840 B.C. and the commercial archives in the city were sealed off.

174

Tablets were largely business letters, shipment documents, accounting records, seals, and contracts. A typical shipment document or expense account in which a merchant would inform partners about their cargo and expenses would read like this:

204

Some business letters would contain information about market and transport conditions, like this:

172

The tablets are spread across the world in museums and institutions, but many have been transcribed. The transcribed ones mentioned 79 cities distributed across modern-day Iraq, Syria, and Turkey and 2,806 mentioned at least two Anatolian city names simultaneously, like so:

161

That tablet identified three shipments: Durhumit to Kaneš, Kaneš to Wahšušana, and Durhumit to Wahšušana. So the itinerary is A→B→C, and there were 227 of these, with 391 examples of travel between city pairs. Specifically, 25 city pairs: 15 known (gray), 10 lost (black).

168

Using trade among known cities, they estimated the distance elasticity of trade (how sensitive trade btwn cities is to the distance btwn them), so they could estimate the prbblity of shipments from city i to city j given their distance Thus, probable locations for 10 lost cities

236

These estimates largely concurred with those of historians, and since the historians' conjectures weren't used in the model, this suggests people should start pursuing those estimations.

190

In fact, this modeling exercise might help to decide among the different proposals made by historians.

174

But the authors weren't done. They supplemented their analysis with data from merchant itineraries. For example, consider this letter:

164

That letter was submitted to the Assyrian port authorities at Kaneš from emissaries in Wahšušana, and it described how missives would travel through two different routes: Wahšušana→Ulama→Purušhaddum W→Šalatuwar→P But only Wahšušana, Ulama, and Šalatuwar are known cities.

134

Using every multistop itinerary, a model with just two constraints offers a lot of info. The constraints are simple: 1. When deciding itineraries, merchants like direct routes. 2. Caravans have to make stops to rest, replenish supplies, feed pack animals, and make side trades.

152

With estimates constrained to regions that are admissible given those constraints (dashed lines), the locations of the newly-identified lost cities are now more certain! With the exception of Purušhaddum.

143

But how do we know this method works? Easy! Just lose known cities and see if the method rediscovers them. As the picture shows, the average distance between estimated and known city locations wasn't huge. In fact, estimates were a median of 33km away (mean = 40km).

166

This method also helps to identify the names of sites that people have continued living in, like Kırşehir Kalehöyük, which might have been located under where the Alaaddin Mosque and a high school were later built.

134

There are other interesting findings here, too. Consider this: geography has deep and persistent impacts on the economy of the area, and cities tend to show up where there are "natural roads".

199

Ancient cities were estimated to be larger when the natural roads were better!

159

And, modern cities are larger when nearby ancient cities were estimated to be larger as well. The deep geographic reasons for cities to crop up in certain locations are still powerful forces today!

162

And for the real nerds, Zipf's law looks to basically hold for ancient city populations.

166

There you have it: economists might have discovered the locations of ancient lost cities from the Bronze Age, and supported a number of other fun facts while they were at it. Only time will tell if these discoveries end up being true

203

Link: academic.oup.com/qje/article/13 The model the authors used was the gravity model: the workhorse model of trade. Just as all roads lead to Rome, many trade theorists have derived the gravity equation for trade: cepr.org/voxeu/columns/

264

16.6K

Show more replies

To view keyboard shortcuts, press question markView keyboard shortcuts

Post

Conversation

To view keyboard shortcuts, press question mark
View keyboard shortcuts