Empires and exchange

Foundations

One machine, many badges

Four thousand years of empires, on every inhabited continent, with different languages, gods, art and law. Underneath, they were all running the same machine.

An agrarian state is a device for moving grain from the people who grow it to the people who rule. Before industrialisation, almost all production was agricultural, so any state that wanted soldiers, officials, temples or roads had to take it from farmers. Everything else is variation in method.

Some took a share of the harvest. Some took labour instead: so many days a year building walls or clearing canals. Some took through intermediaries who kept a cut. Some legitimated it through divine kingship, some through law, some through simple force.

But the constraint is the same everywhere, and it explains far more about these societies than the personalities of their rulers.

The limits nobody could cross

Two hard limits shaped every pre-industrial state, and neither could be argued with.

Transport. Move grain overland by pack animal and the animals eat it. The load shrinks as it travels, and within a few hundred kilometres the cargo is gone. Water transport is roughly an order of magnitude cheaper.

The consequence is visible on any map of ancient empires. They cling to rivers, coasts and a few corridors. Mesopotamia on the Tigris and Euphrates. Egypt as a ribbon along the Nile. Rome as a ring around a sea. The Indus along its rivers. The shape of an agrarian empire is the shape of its cheap transport.

Communication. A message travels at the speed of a horse or a ship. Rome’s relay system managed perhaps eighty kilometres a day under good conditions; ordinary messages went far slower. The Persian royal road and the Mongol relay system were both attempts to buy speed at enormous cost.

That means a provincial governor a month’s travel from the capital is effectively autonomous. Every large agrarian empire faced the same dilemma: delegate enough authority for the province to function, and the governor may rebel; delegate too little, and nothing happens. Nobody solved it. The failures are much of what we call political history.

What made bigger possible

Three developments raised the ceiling, and none was a form of government.

Secondary products. Early herders kept animals for meat, which means killing them. Then came the realisation that a living animal produces milk repeatedly, grows wool repeatedly, and can pull things. Traction is the big one: an ox pulling a plough does work no household could do by hand, and it is what turns heavy soils into farmland. It also, as lesson 7.5 noted, made holdings expandable and inequality compoundable.

Metals. Bronze, then iron. Iron matters more than bronze precisely because iron ore is common, so tools stop being elite goods. An iron ploughshare and an iron axe change what ordinary farmers can do to land.

Writing and bureaucracy. Lesson 7.3 showed writing beginning as accounting. Scale it up and you get a state that can record who owes what, across a territory too large for anyone to hold in memory. Administration is a memory technology, and it is what allows a polity to outgrow the reach of a single person’s knowledge.

The people who did not farm

The most persistent military problem for agrarian states came from people who had refused the whole arrangement.

The Eurasian steppe runs from Hungary to Manchuria, and it is grassland: poor for cultivation, excellent for grazing. The societies that formed there were mobile pastoralists, and the horse gave them something no farming state could match. They could move faster than armies could react, they carried their food supply with them, and their entire adult population was, in effect, cavalry.

The result is one of the great recurring patterns of Eurasian history. Small steppe populations conquering enormous settled ones, over and over, culminating in the Mongol empire, which by the late thirteenth century covered around twenty-four million square kilometres, roughly a sixth of the world’s land.

There is a quantitative version of this. Analysis of the historical distribution of empire sizes found that the largest polities arose disproportionately close to the steppe frontier, consistent with intense and sustained military competition driving political scale. Empires, on this reading, grow where they must or die.

The connections

While states fought, goods moved, and the network mattered more than any of the states in it.

The Silk Roads are misnamed twice over. There was no single road, and silk was one commodity among many. The reality was a web of overland and maritime routes, and almost nobody travelled the whole way. Goods moved in relays: a merchant carried them a few hundred kilometres and sold on. A bolt of silk might change hands dozens of times between Chang’an and the Mediterranean, and no participant needed to know where it started or ended.

The Indian Ocean carried more, and it was in some ways more sophisticated. Monsoon winds reverse seasonally, blowing reliably northeast for part of the year and southwest for the rest. Sailors who learned that pattern could cross open ocean and return, on a schedule. Regular long-distance sea trade linking South Asia, the Persian Gulf, the Red Sea and East Africa was operating well over a thousand years before anyone crossed the Atlantic. Indian ports were central to it, and pepper, cotton textiles and gemstones went west while horses, silver and glass came east.

What travelled was more than cargo.

Technologies. Papermaking moved west from China over centuries. Gunpowder and the compass followed. Cotton and sugar cultivation spread across the network.

Mathematics. The decimal place-value system with a symbol for zero was developed in India and transmitted westward through the Islamic world into Europe. Every calculation performed by every computer descends from that transmission, and it happened along these routes.

Religions. Buddhism moved from India into Central and East Asia along the same corridors. Islam later spread along the Indian Ocean routes as much as by conquest.

Diseases. The same connections that carried paper carried plague. A network that lets goods travel lets pathogens travel, and epidemics repeatedly followed trade routes across Eurasia. Lesson 7.5 built the crowd diseases; this is the delivery system.

By around 1000 CE, most of Afro-Eurasia formed one connected zone. The Americas formed another, with no sustained contact. Those two zones met in 1492, and chapter eight begins with what happened.

Collapse, and why the word misleads

Empires end. The pattern is regular enough that it has generated an entire literature, and most of that literature spends its first chapter correcting the word.

Collapse usually does not mean people died. It means administration failed. The tax records stop. Monuments stop being built. Standardised goods stop being made. Long-distance trade contracts. Writing sometimes disappears entirely.

Meanwhile, in most cases, people continue. They disperse into smaller settlements, farm at lower intensity, and often live somewhat better than they did while supporting a state. The Indus de-urbanisation in lesson 7.4 is a clean example: the cities emptied and the population moved east.

Joseph Tainter’s explanation, which remains the most influential, is about diminishing returns. A society facing a problem responds by adding complexity: more officials, more soldiers, more infrastructure. Early on, this pays handsomely. Over time, each additional increment yields less while costing the same or more. Eventually the complexity costs more than it returns, and when a shock arrives, the society cannot afford the response. Simplification is then not a catastrophe but an economising.

In practice, documented collapses show several stresses arriving together: climate variability, epidemic disease, fiscal strain, elite factionalism, external pressure. Isolating one cause is usually impossible, and monocausal explanations should be treated with suspicion, whichever cause they favour.

Four thousand years of not breaking through

Here is the summary that sets up the next chapter.

Between the first cities and about 1500 CE, agrarian societies achieved a great deal: mathematics, astronomy, medicine, engineering, law, philosophy, literature and art that has never been surpassed. They built the pyramids, the Parthenon, Angkor and the Great Wall.

And they did not escape the constraint. Living standards for ordinary people fluctuated around a level set by land and technology. Productivity gains were absorbed by population, as lesson 7.5 described. Energy came from muscle, wind, water and burning wood, all of which are limited by how much sunlight falls on how much land. Empires grew until transport and communication failed them, then fragmented, then regrew.

Four thousand years of a cycle that did not break. Nothing in this period suggests it was going to.

Then, in one corner of Eurasia, somebody started burning something that was not wood, and the entire structure of the constraint changed. That is chapter eight.

Deeper

The arithmetic of overland transport

A pack animal consumes roughly two to three percent of its body weight in fodder daily and carries perhaps a fifth of its body weight. Over a long journey the fodder requirement compounds, so the delivered fraction of the original load falls sharply with distance; classic estimates put the economic limit for bulk grain at a few hundred kilometres. Water transport avoids this because the vessel does not eat, which is why pre-industrial cities of any size sit on navigable water and why inland populations remain dispersed until railways.

What the empire-size analysis did

Turchin compiled the territorial extent of the largest polities through history and tested where they arose relative to the steppe frontier, finding a strong spatial association. The proposed mechanism is that intense and sustained military threat selects for larger, more cohesive polities, because small ones are absorbed or destroyed. The approach treats history as a source of testable spatial patterns rather than as a sequence of narratives, and it inherits the usual difficulties of that approach: the sample is small, boundaries are uncertain, and territorial extent is a crude measure of power.

Secondary products, dated

Sherratt’s original formulation proposed a relatively sharp horizon in the fourth millennium BCE for the adoption of milking, wool, traction and wheeled transport. Subsequent work, especially lipid residue analysis of pottery detecting dairy fats, has pushed milking substantially earlier in some regions, so the model is now generally treated as a set of related innovations adopted at different times and places rather than a single revolution. The core observation, that using animals alive is worth far more than using them dead, is unaffected.

Reading collapse from the ground

The archaeological signature of collapse is the disappearance of the markers of complexity: administrative texts, monumental construction, mass-produced standardised goods, imported materials. Settlement survey typically shows dispersal rather than abandonment, and skeletal indicators sometimes show improvement in nutrition after collapse, since the surplus that fed the state stays with the producer. This is why “collapse” is a description from the perspective of the state rather than of the population.

The transmission of Indian mathematics

The decimal place-value system with zero as a numeral appears in Indian mathematical and astronomical texts in the first millennium CE, with Brahmagupta’s seventh-century work giving rules for arithmetic including operations with zero. Transmission westward proceeded through translation into Arabic and thence into Latin, with the numerals themselves still called Arabic in European usage despite their Indian origin. Plofker’s history is careful about distinguishing what is documented from what is often asserted, which matters because this transmission is frequently narrated loosely in both directions.

Frontier

Whether collapse has a general theory

Tainter’s model competes with explanations emphasising climate forcing, elite overproduction, disease and network fragility, and with the position that no general theory is possible because the cases are too dissimilar. Quantitative approaches using large historical databases aim to test these against one another, and the databases themselves are contested for how they code uncertain historical variables.

How much the networks actually carried

Estimating the volume and economic significance of long-distance pre-modern trade is difficult, because the surviving evidence over-represents luxury goods that were light, durable and valuable. Isotopic and residue analyses of bulk commodities are beginning to give a better picture, and current estimates of the share of production that moved long distances remain low relative to the attention the routes receive.

Transcript and shot list

Target runtime about 8 minutes 30 seconds. Built as one mechanism with many instances: establish the shared constraint, show what varied, then show what happened when the constraint broke. Shot types: ANIM is rendered from code with Manim, AI is a generated shot from Higgsfield with the shared style reference, REAL is agency imagery with credit, CAM is the narrator.

[00:00] S01 · ANIM · Four empires across four thousand years with different scripts, gods, art and law; then the surface stripped away to reveal the same underlying diagram: grain moving from farmers to rulers.

Four thousand years of empires, on every inhabited continent, with different languages, gods, art and law. Underneath, all running the same machine. An agrarian state is a device for moving grain from the people who grow it to the people who rule. Before industry, almost all production was agricultural, so any state that wanted soldiers, officials, temples or roads had to take it from farmers. Everything else is variation in method.

[00:35] S02 · ANIM · Four extraction methods: a share of harvest, days of compulsory labour, intermediaries taking a cut, and force; each legitimated differently by divine kingship, law or nothing at all.

Some took a share of the harvest. Some took labour instead: so many days a year on walls or canals. Some worked through intermediaries who kept a cut. Some legitimated it through divine kingship, some through law, some through simple force. Same machine, different badges. And the constraint explains far more about these societies than the personalities of their rulers.

[01:05] S03 · ANIM · A pack animal setting out with a full load; the load shrinking as the animals eat it; the cargo gone within a few hundred kilometres. Then a boat carrying the same load at roughly a tenth of the cost.

Two hard limits shaped every one of them. First, transport. Move grain overland by pack animal and the animals eat it. The load shrinks as it travels, and within a few hundred kilometres the cargo is gone. Water transport is roughly ten times cheaper.

[01:35] S04 · ANIM · Ancient empires drawn on a map, each hugging its water: Mesopotamia on two rivers, Egypt as a ribbon on the Nile, Rome as a ring around a sea, the Indus along its rivers.

And the consequence is visible on any map. They cling to rivers, coasts and a few corridors. Mesopotamia on the Tigris and Euphrates. Egypt as a ribbon along the Nile. Rome as a ring around a sea. The Indus along its rivers. The shape of an agrarian empire is the shape of its cheap transport.

[02:00] S05 · ANIM · A message leaving a capital at the speed of a horse; a relay system managing perhaps 80 km a day; a provincial governor a month away shown as effectively autonomous; the delegation dilemma stated both ways.

Second, communication. A message moves at the speed of a horse or a ship. Rome’s relay system managed maybe eighty kilometres a day in good conditions, and ordinary messages went far slower. Which means a governor a month’s travel from the capital is effectively independent. Every large agrarian empire faced the same dilemma. Delegate enough authority for the province to work, and the governor may rebel. Delegate too little, and nothing happens. Nobody solved it, and the failures are much of what we call political history.

[02:40] S06 · ANIM · Secondary products: an animal killed once for meat, against a living animal giving milk repeatedly, wool repeatedly, and traction; then an ox pulling a plough through heavy soil no household could work by hand.

Three things raised the ceiling, and none of them was a form of government. Secondary products first. Early herders kept animals for meat, which means killing them. Then came the realisation that a living animal gives milk repeatedly, grows wool repeatedly, and can pull things. Traction is the big one. An ox on a plough does work no household could do by hand, and it turns heavy soils into farmland. It also, as the last lesson said, made holdings expandable and inequality compoundable.

[03:15] S07 · ANIM · Bronze as an elite good; then iron, with common ore, becoming ordinary tools; an iron ploughshare and axe in a farmer’s hands. Then accounting scaling into bureaucracy: a state recording who owes what across a territory nobody can hold in memory.

Then metals. Bronze, then iron, and iron matters more precisely because its ore is common, so tools stop being elite goods. An iron ploughshare and an iron axe change what an ordinary farmer can do to land. And then writing. Lesson 7.3 showed it starting as accounting. Scale that up and you get a state that can record who owes what across a territory far too large for any person to hold in their head. Administration is a memory technology, and it is what lets a polity outgrow one person’s knowledge.

[03:55] S08 · ANIM · The Eurasian steppe from Hungary to Manchuria as continuous grassland; poor for cultivation, excellent for grazing; mobile pastoralists with horses, food supply travelling with them, the whole adult population effectively cavalry.

And the most persistent military problem came from people who had refused the whole arrangement. The Eurasian steppe runs from Hungary to Manchuria, and it is grassland: bad for crops, excellent for grazing. The societies there were mobile pastoralists, and the horse gave them what no farming state could match. They moved faster than armies could react, carried their food supply with them, and their entire adult population was in effect cavalry.

[04:30] S09 · ANIM · Repeated conquests of large settled states by small steppe populations; then the Mongol empire at ~24 million km², about a sixth of the world’s land; then empire sizes plotted against distance from the steppe frontier, with the largest clustering near it.

Which produces one of the great recurring patterns of Eurasian history: small steppe populations conquering enormous settled ones, over and over, ending with the Mongol empire at around twenty-four million square kilometres, roughly a sixth of the world’s land. And there is a quantitative version. Analysis of historical empire sizes found the largest polities arose disproportionately near the steppe frontier, consistent with sustained military competition driving political scale. On that reading, empires grow where they must, or die.

[05:10] S10 · ANIM · The single-road image of the Silk Road crossed out; replaced by a web of relays, a bolt of silk changing hands dozens of times between Chang’an and the Mediterranean, no participant knowing both ends.

While states fought, goods moved, and the network mattered more than any state in it. The Silk Roads are misnamed twice: there was no single road, and silk was one commodity among many. It was a web of overland and sea routes, and almost nobody travelled the whole way. Goods moved in relays. A merchant carried them a few hundred kilometres and sold on. A bolt of silk might change hands dozens of times between Chang’an and the Mediterranean, and nobody in the chain needed to know where it started or ended.

[05:50] S11 · ANIM · Monsoon winds reversing seasonally; ships crossing open ocean northeast in one season and southwest in the other; a network linking South Asia, the Persian Gulf, the Red Sea and East Africa; pepper, cotton and gemstones west, horses, silver and glass east.

And the Indian Ocean carried more, and in some ways was more sophisticated. Monsoon winds reverse seasonally, blowing reliably one way for part of the year and the other way for the rest. Sailors who learned that could cross open ocean and get back, on a schedule. Regular long-distance sea trade linking South Asia, the Persian Gulf, the Red Sea and East Africa was running well over a thousand years before anyone crossed the Atlantic. Indian ports were central to it: pepper, cotton textiles and gemstones going west, horses, silver and glass coming east.

[06:30] S12 · ANIM · What else travelled: papermaking moving west over centuries; gunpowder and the compass; Buddhism moving out of India; the decimal place-value system with zero passing through the Islamic world into Europe; then plague following the same lines.

And what travelled was more than cargo. Papermaking moved west from China over centuries. Gunpowder and the compass followed. Buddhism moved out of India along the same corridors. The decimal place-value system with a symbol for zero was developed in India and transmitted west through the Islamic world into Europe, and every calculation done by every computer descends from that transmission. And the same connections carried plague. A network that moves goods moves pathogens, and epidemics repeatedly followed trade routes across Eurasia. The last lesson built the crowd diseases. This is the delivery system.

[07:10] S13 · ANIM · Collapse shown as an administrative event: tax records stopping, monuments halting, standardised goods ceasing, trade contracting, writing sometimes vanishing. Then the population continuing, dispersing into smaller settlements, sometimes with better nutrition.

Empires end, and the word collapse misleads. It usually does not mean people died. It means administration failed. The tax records stop. Monuments stop going up. Standardised goods stop being made. Long-distance trade contracts. Writing sometimes disappears completely. Meanwhile, in most cases, people carry on. They disperse into smaller settlements, farm less intensively, and often live somewhat better than they did while supporting a state. The Indus de-urbanisation from the last lesson is a clean example.

[07:45] S14 · ANIM · Tainter’s curve: early complexity paying handsomely, each increment yielding less, the cost line crossing the benefit line; then a shock arriving that the society cannot afford to answer; simplification as economising rather than catastrophe. Then five stresses arriving together.

Joseph Tainter’s explanation, still the most influential, is diminishing returns. A society facing a problem adds complexity: more officials, more soldiers, more infrastructure. Early on that pays handsomely. Over time each increment yields less while costing the same or more, until the complexity costs more than it returns. Then a shock arrives that the society cannot afford to answer, and simplification is not a catastrophe but an economising. In practice, documented collapses show several stresses arriving together: climate variability, epidemic disease, fiscal strain, elite factionalism, outside pressure. Isolating one cause is usually impossible, and single-cause explanations deserve suspicion whichever cause they pick.

[08:20] S15 · AI into ANIM · Achievements of the agrarian era in sequence: the pyramids, the Parthenon, Angkor, the Great Wall; then a living-standards line fluctuating around a flat level for four thousand years; then a flame of something that is not wood, and the line beginning to lift. The calendar returns.

So, the summary. Between the first cities and about 1500, agrarian societies achieved an enormous amount. Mathematics, astronomy, medicine, engineering, law, philosophy, and art that has never been surpassed. The pyramids, the Parthenon, Angkor, the Great Wall. And they did not escape the constraint. Ordinary living standards fluctuated around a level set by land and technology. Productivity gains were eaten by population. Energy came from muscle, wind, water and burning wood, all of which are limited by how much sunlight falls on how much land. Empires grew until transport and communication failed them, fragmented, and regrew. Four thousand years of a cycle that did not break, and nothing in this period suggests it was going to. Then, in one corner of Eurasia, somebody started burning something that was not wood.

Shot list

One AI shot, fourteen code-rendered shots. Estimated Higgsfield use: 2 to 4 generations including retries.