Why Are There No Empires in Age of Empires?

The November 2019 post that opens Bret Devereaux's acoup.blog Collections series, prompted by the announcement of Age of Empires IV. The complaint in the title is literal: the games contain peoples who had empires, but no empires, and the distinction turns out to carry most of what's interesting about how large states actually worked.

What an empire is

An empire is a state where one set of people (the core) extract resources (typically by force) from another set of people (the periphery).

The periphery is linguistically, culturally, ethnically or religiously distinct from the core. That's the whole definition, and it comes straight from the Latin: imperium is a sphere of command, from imperio, to order or command. The Romans applied it to a consul's authority over an army and to the imperium populi Romani over non-Roman peoples — and, as Devereaux points out, one cannot have imperium over oneself. An imperium of the Roman people must be over someone else.

Two corrections follow. Empire does not require a monarchy: Rome held an empire as a Republic, France kept one after ceasing to be a monarchy, and Athens converted the Delian League into an ἀρχή (arche) while remaining internally democratic. And empire is not an exotic form — it has been the dominant mode of large-scale human organization since at least the Bronze Age. Federations and nation-states are recent, and many modern nation-states are consolidated empires whose state power manufactured the shared culture that later looked like a nation. empire-core-periphery carries the definition and its consequences as a standalone concept.

He notes the joke in the title of any game called Age of Empires: there has never been an age without them, ours included.

What the games model instead

In a typical game you start with settlers, harvest resources, build, expand, and meet rival players who are hostile by default. You don't conquer their settlements or reduce them to tributaries — you annihilate them and repopulate the ground with your own villagers. Villagers cannot normally be captured, only killed.

His verdict is blunt: this is not how empires work, because the entire point of having an empire is access to the resources and labor of a subordinate population, and exterminating that population defeats the purpose. What the games actually depict is "fanatically murderous nation-states, projected backwards in history hundreds — if not thousands — of years before any such idea of a state existed." Few real states have pursued this, and the clearest example that did is Nazi Germany.

He handles three objections directly:

• It's an abstraction, the empire-building happens off-screen. Fair for Civilization, but not for AoE, where the non-combatants are visibly on screen and visibly not capturable by military units.
• Priests can convert instead of kill. Three problems: real empires mostly didn't care about local customs so long as taxes arrived, so forced cultural conversion isn't the imperial behavior being modeled; swapping genocide for ethnic cleansing is not much of an improvement; and mechanically it's a one-unit-at-a-time tool with a long cooldown, a way to turn battles rather than a real alternative to annihilation.
• It's just a game, stop overthinking it. He calls this, paraphrasing C.S. Lewis, an unserious "boy's philosophy" — games are cultural products, and denying that cultural products shape political ideologies and social movements argues against the historically obvious.

Rome as the counterexample

The AoE Rome is Roman through and through and never stronger than its count of ethnic Romans. Devereaux notes he's not building a straw man — people have confidently told him that Roman military strength came from cultural and ethnic cohesion.

The actual Roman Empire was diverse in a way that was load-bearing at every stage. Peninsular Italy in 218 B.C. held at least three religious systems (Roman, Etruscan, Greek), half a dozen languages some unrelated to each other, and many distinct groups with strong local identities and fierce local rivalries. The army was split between citizens and the socii — the polite word for subjugated Italian communities — and the socii always made up more than half of it. Military service was the resource being extracted.

It only gets more diverse outward: Sicily held Greeks, Punic settlers and non-Greek natives; Cisalpine Gaul was Gallic; Spain added Iberians, Celt-Iberians and more Punic settlers. His summary line: Roman armies were essentially never majority "Roman" in any period, save possibly the third century. Rome was a fusion society from its foundation, sitting on the boundary between Latin and Etruscan zones.

Where military force wasn't extracted directly, revenue was, which is the point of the whole arrangement — making other people do the farming and mining that supports your society. Keeping that revenue flowing requires being good at managing culturally varied subject populations. "Empires that are bad at handling a wide array of different cultures/religions/languages do not long remain empires."

The two distortions

Diversity as strength, not malus. Well-run empires drew on the particular capabilities of many cultures. Mongol subsistence patterns produced excellent cavalry and poor infantry and siege skill — so a Mongol army arrived with cutting-edge Chinese siege technology and local infantry alongside its horsemen. A Roman army at its height might field Italic infantry, Sarmatian or Gallic heavy cavalry, Numidian or Arab light cavalry, eastern composite bowmen and light infantry from the rough country. The same held in engineering and administration. Games that do acknowledge subject populations — he names Europa Universalis, Stellaris, Imperator — typically render them as a flat penalty ("non-primary culture, -x% production/stability"), which inverts the historical reality.

Hierarchy erased. The games let expansion mean growth of the core. Real imperial expansion meant subjugating a much larger periphery to a core that stayed small. Until the Constitutio Antoniniana in 212 A.D., most people in the Roman Empire were not Romans and held no citizenship — they were subjects of a populus Romanus that was itself a minority within its own empire.

The closing point is the one the rest supports: the dominant human experience of empire is not conquest but subjection — having resources extracted from you by a state that likely inflicts violence on you and may care little for your safety.

He notes the piece covers pre-modern tributary empires and not modern mercantilist ones, which were no less diverse and on the whole more brutal and less tolerant.

Cross-references

• empire-core-periphery — the definition and its mechanics as a concept page
• acoup-blog — Devereaux's blog
• soviet-control-rooms — the wiki's other page on how design artifacts encode a political worldview