What Travels Far
Before cheap transport, only certain things could cross a continent, and the filter was ruthless: value per unit weight. If a cargo cannot pay for a thousand kilometres of camel, it does not go a thousand kilometres.
| Cargo | Value density | Range |
|---|---|---|
| Silk, spices, gems, gold | Enormous | Anywhere |
| Slaves | High, and self-transporting | Anywhere |
| Manufactured metalwork | Moderate | Regional |
| Wine, oil | Modest — but amphorae go by sea | Coastal, far; inland, short |
| Grain | Low | Sea only; overland it eats itself |
Grain is the instructive one. Move grain far enough overland and the oxen consume the cargo — the transport eats the freight. Which is why every pre-modern city large enough to need imported bread had to be on water, without exception, and why Rome fed itself from Egypt and North Africa rather than from Italy fifty kilometres inland.
So pre-modern long-distance trade is systematically biased toward luxuries. Not because merchants were frivolous, but because luxuries were the only things the arithmetic permitted to travel. A trade route is a value-density filter, and the goods you find on it are the ones that passed.
The Silk Road Was Not a Road
Two corrections, and the second is the interesting one.
First, it was not a road but a shifting web of routes across oases, and the name was coined in 1877 by Ferdinand von Richthofen — nobody who ever travelled it called it that.
Second, and more importantly: almost nobody travelled its length. Goods did; people mostly did not. Silk moved from hand to hand through dozens of intermediaries, each moving it a few hundred kilometres and taking a margin. A bolt of silk reaching Rome had been traded twenty times, and no single merchant knew both ends. This is why Rome's knowledge of China was so spectacularly wrong: the goods arrived, the information did not, because information does not survive twenty relays even when the silk does.
The relay structure explains the whole system. It explains the markups (each hand takes a cut), the ignorance at both ends, the resilience (no single point of failure, so a war closes one segment and the web reroutes), and the fact that the network's most consequential cargo was never intended: ideas, religions, and disease. Buddhism travelled the routes to China. So did the plague. A trade network is a disease network with better public relations — the same relays that pass a bolt of silk pass a flea, and the flea does not need a merchant who knows both ends either.
The Container Broke the Filter
For most of history, transport costs shaped the map. Then, in 1956, Malcom McLean put cargo in standardized steel boxes, and the cost of moving goods collapsed toward irrelevance.
The effects were structural, not incremental:
- Loading time collapsed — the port, not the ocean, had always been the bottleneck. Break-bulk cargo was handled piece by piece by armies of dockers; a container is one crane movement.
- Value density stopped mattering. When shipping is nearly free per unit, you can ship anything anywhere. Grain crosses oceans. So does bottled water, which is a genuinely absurd thing for the arithmetic to permit.
- Location de-anchored. Manufacturing could go anywhere with labour and a port, because proximity to the market stopped being a cost worth paying for.
- Old ports died and new ones appeared wherever the water was deep enough for the ships and the land flat enough for the stacks. The great labour-intensive dock districts emptied — a container terminal is a car park with cranes.
The historical point is what this reveals: almost everything that looked like a permanent law of economic geography — luxuries travel, bulk does not; industry clusters near markets; cities cling to natural harbours — was a statement about the price of moving things. Change the price and the "laws" evaporate within a generation. Geography did not change in 1956. The price list did, and it turned out the price list had been doing all the work.