Civilisational Collapse

There's a standard story about the Maya: unstoppable climate change withered their crops, killed thousands in overpopulated cities, and the jungle swallowed their pyramids until intrepid 19th-century explorers rediscovered them. There's a matching story about Easter Island: the inhabitants chopped down all their trees, wrecked their ecology, and devolved into a starving remnant eking out life among the ruins of former greatness. Both stories are wrong in ways that matter — not just for archaeology, but for how we think about our own civilisation's fragility.¹

States Collapse, Civilisations Transform

The archaeologist Guy Middleton makes a distinction that most popular accounts blur: states collapse; civilisations don't. States are tangible units — political structures with hierarchies, centralized functions, ideologies. When they collapse, what you see is fragmentation of political entities, abandonment of urban centers, breakdown of economic systems, failure of legitimating ideologies. But the broader culture — the language, the pottery styles, the religious practices, the agricultural knowledge — usually survives the death of the state that hosted it.¹

Take the Mycenaean collapse around 1200 BCE. The palaces burned. Linear B writing fell out of use. The wanax title for king was replaced by basileus, which had previously been a minor term. The elaborate megaron throne rooms with their central hearths and four-column layout were deliberately abandoned — not because the knowledge was lost, but because whoever came to power afterward chose to reject the old ideology. Mycenaean pottery traditions continued for another century and a half. The gods — Zeus, Poseidon — kept being worshipped into historical times. For small-scale farmers, the main practical change was that nobody demanded resources for palace feasts anymore.¹

The Maya case is even more instructive. What actually happened was a series of individual state collapses across roughly three centuries (750–1050 CE), not a single apocalyptic event. At Cancuén, around 800 CE, the archaeological evidence is vivid — unfinished palisades, spearheads scattered across the site, the king and his wife buried in shallow graves, buildings systematically defaced. A single, specific political collapse, violent and traumatic. But this was one of sixty to seventy independent Maya states, each with its own waxing and waning fortunes. After the "collapse," new cities were founded, trade continued, and millions of Maya people were still living in a complex society when the Spanish arrived centuries later. The real destruction of Maya culture came not from climate or overpopulation but from the Spanish, who burned thousands of Maya books (only four survive) and forcibly broke up families for re-education.¹

Easter Island follows the same pattern of misnarration. The eco-disaster story claims the population crashed after deforestation, leaving a degraded remnant. But the archaeologist Terry Hunt finds no evidence the population ever exceeded about 4,000 — far from unsustainable. The islanders had effective agricultural techniques: walled gardens, lithic mulching, adapted boat-building. Eating rats was a rational protein strategy, not a sign of desperation (the Romans ate dormice without anyone calling them primitive). The moai were still being actively venerated well into the period when "collapse" supposedly occurred — the evidence is that enemies kept toppling them, which you only bother doing if the statues still carry power. The real population crash came from European slave raiding in the 1860s, which reduced the islanders to perhaps 750 people.¹

The Appeal of Apocalypse

Why do the wrong stories persist? Middleton argues we're primed for them by deep cultural inheritance. From Sodom and Gomorrah to Atlantis to Pompeii, dramatic civilisational destruction is one of our oldest narrative forms. These stories let us feel superior to past peoples (we know why they failed), make ourselves feel important (we're living at a crucial turning point), and extract neat moral lessons (treat the environment well or perish). Jared Diamond's Collapse — the book that cemented the Easter Island narrative for millions of readers — is frankly a work of environmental parable dressed up as archaeology.¹

The problem isn't just that these stories are inaccurate. It's that they "appropriate the histories of past and indigenous peoples, and construct them as modern Western-focused parables." They erase agency — the Maya didn't just passively succumb to drought; some states adapted, others were conquered in war, and the population made deliberate choices about which cultural elements to keep and which to reject. The "collapse" framing dwells on supposed failures of pre-modern societies while ignoring the actual destruction inflicted by Western colonizers who arrived later.¹

The Very Long Run

Pull back further and civilisational collapse starts to look less like a catastrophe and more like the normal rhythm of human history. Matthew Yglesias frames this with alarming clarity: our species is about 300,000 years old. Agriculture started about 12,000 years ago. Sustained improvement in living standards has only been happening for 250 years, maybe less. The vast majority of human existence was the Paleolithic — wandering in small bands, changing very little over millennia that dwarf all of recorded history.²

The Acheulean hand axe was used for longer than Homo sapiens has existed. The Oldowan stone tool technology that preceded it lasted about a million years. "In an age where the phones in our pockets can become obsolete in a year, the idea of a single technology lasting that long is almost inconceivable."²

And when agriculture finally arrived, it was — in one of the more disquieting reappraisals of recent decades — probably a step backward for individual welfare. Farming made the land enormously more productive per acre, but that productivity was achieved through grueling labor and quickly eaten up by population growth. Skeletons of early farmers show worse nutrition, more disease, and shorter stature than their hunter-gatherer predecessors. The "benefits" of agriculture accrued to two groups: the aggregate population (many more people existed) and the extractive elites who could confiscate surplus from farmers too settled to run away. This is James Scott's point about grain specifically — because it's harvested at a specific time and stored, it's uniquely stealable, which makes it uniquely suitable for taxation, which makes it the foundation of the state.²

The dark portrait Marx and Engels painted of the Industrial Revolution as immiserating the working class was, in a sense, perfectly reasonable extrapolation from all prior history. Every previous technological improvement had led to brief betterment, then population growth, then a return to subsistence — with the gains captured by elites. The fact that the Industrial Revolution finally broke this pattern, with sustained per-capita improvement, is perhaps the most anomalous event in the entire human story. On the truly long timescale, it's an open question whether it will last.²

John Michael Greer pushes this perspective to its limit, imagining human history across ten billion years — hundreds of civilisations rising and falling, none achieving the escape velocity of interstellar expansion, the last human populations yielding to new intelligent species descended from raccoons, then crows, then (in a billion years) freshwater clams. It's a thought experiment, but the underlying logic is defensible: if you're plotting a trendline from the available data of 300,000 years, the last 250 years of sustained growth look more like a spike than a new normal.³

What's Actually Fragile

The real lesson from both the archaeology and the deep time perspective isn't "civilisations collapse, therefore despair." It's that what collapses is almost always political structure, not knowledge or culture. The Mycenaeans lost their palaces but kept their gods. The Maya lost their divine kings but kept their agricultural techniques and their language. What did destroy cultures — comprehensively, permanently — was usually conquest by outsiders who deliberately targeted cultural practices for destruction.

This has implications for the discourse around Scaling Laws and systemic risk. Complex systems don't typically fail all at once; they shed their most complex, most expensive-to-maintain structures first, while the underlying substrate persists. And the thing most worth worrying about isn't whether our systems are fragile — all states are fragile — but whether we're preserving the cultural knowledge and social connections that would allow recovery. The Polar Inuit lost technology when they were cut off from the broader Inuit network, not because they were conquered, but because their collective brain shrank below the threshold needed to maintain their knowledge base.