The timeline of prehistory

The timeline does not begin with the species; it begins a little before, so that the beginning has something to be measured against. This band is that margin: thirty-five thousand years of running room, ending at the oldest fossils a specialist would call ours.

Two kinds of uncertainty meet here, and they are not the same size. The measurement carries an error of 34,000 years, which is honest and bounded.

Three hundred thousand years in which stone was shaped only by breaking it, and in which nothing that a modern reader would call an institution existed. It is by far the longest band on this timeline, and the one that makes all the others legible: everything since farming fits in the thin strip at its right.

The word was invented in nineteenth-century Europe and it travels badly. African archaeologists divide the same span into Early, Middle and Later Stone Age.

People stop following their food and start growing it. Within six millennia the species becomes sedentary, far more numerous, and dependent on a handful of plants it has reshaped. Almost everyone alive still lives under that arrangement.

There is no single Neolithic, and treating it as one event is the commonest mistake made about it. Farming was invented independently in at least a dozen regions.

The band in which the past begins to speak in its own voice. Not because more was happening, but because someone started writing things down, and because some of what they wrote survived.

This is the most local boundary on the timeline, and the one that most rewards suspicion. Writing appears in Mesopotamia around 5,300 years ago and in Egypt almost.

For two hundred and sixty-five thousand years the way of making a tool barely changes, while almost everything else about our species does. This is the technical world in which Homo sapiens appears, spreads across Africa, and leaves it.

Neither edge is a date in the way a battle has one. The left one follows the oldest well-dated assemblage, which records the state of excavation rather than the state.

In the space of a few millennia, and across most of Eurasia, the things people leave behind change character: long regular blades, tools of bone and antler, beads worn on the body, animals painted on walls, and graves furnished for the dead.

The opening at 50,000 years is a rounded convention, and a generous one. Assemblages called Initial Upper Palaeolithic are dated to before 45,000 years ago.

People stop following their food and start growing it. Within six millennia the species becomes sedentary, far more numerous, and dependent on a handful of plants it has reshaped. Almost everyone alive still lives under that arrangement.

There is no single Neolithic, and treating it as one event is the commonest mistake made about it. Farming was invented independently in at least a dozen regions.

An alloy nobody can make alone. Copper comes from one region, tin from another, usually far away, and the metal exists only where the two are brought together. This is therefore the age in which distance becomes an economic fact.

The name describes a technique, and techniques travel at their own speed. Bronze working reaches southwestern China and mainland Southeast Asia only in the eleventh.

Iron is harder to work than bronze, and far easier to find. Where bronze had depended on trade routes that a rival could cut, iron could be dug up almost anywhere, and a decent blade came within reach of people who would never have afforded the older metal.

Its bounds are the round dates of European and Near Eastern textbooks. Elsewhere the sequence they imply breaks down: in sub-Saharan Africa iron working appears.

The stretch of the past where reading begins to rival digging. Between the eighth century before our era and the fifth after it, several states kept records dense enough that we can follow their harvests, their lawsuits and their quarrels.

A European frame, and an old one. Its closing date, the fall of the western Roman Empire, meant nothing in China, in Mesoamerica or across most of Africa.

A thousand years named by the people who came afterwards, and named as the gap between two ages they liked better. The name has stuck so well that it is easy to forget it was an insult.

Purely conventional, openly Eurocentric, and kept anyway. The alternative would be to invent a neutral name that no reader recognises.

Three centuries in which ships join the continents, and what happens on one becomes the fate of another. It is the period when the history of the species stops being several histories in parallel.

Both ends, around 1500 and around 1800, are classroom dates and nothing more. Historians dispute them, and a few deny the period any coherence beyond the convenience.

Two centuries in which the species goes from one billion to eight, and becomes the main agent of change on the surface of its own planet. Measured against everything to their left on the frieze, they are a rounding error.

The only band with no closing boundary. Its opening date is conventional and moves with the author and the country, from the 1780s to the 1850s.

The frieze opens in the middle of a glacial stage, and that is the first thing it shows: this history begins in a cold world, not a temperate one.

The band is truncated at the left edge of the window: the stage actually begins around 374,000 years ago, well before this timeline starts.

A warm stage, and the one in which the oldest fossils of our species are usually placed. If the species has a climatic setting at its origin, this is it.

Two datings meet here and neither is precise enough to constrain the other. The stage boundaries come from a smoothed global curve.

A cold stage of fifty-seven millennia, during which the species exists but leaves few traces anyone has yet found.

An empty band on a timeline is ambiguous by nature: it can mean that little happened, or that little has survived, or that little has been looked for.

Warm by convention, complicated in detail: not one interglacial but a set of mild intervals separated by cold ones, a useful reminder of how much these bands smooth over.

The band hides its own structure. Where a stage is drawn as one colour, the reader should assume oscillation underneath: the LR04 curve is continuous.

Sixty thousand years of cold, and the longest glacial stretch our species had faced by then. Ice sheets reach far into northern Europe, and the Sahara becomes a barrier rather than a corridor.

The name is regional, taken from northern European glacial geology, while the boundaries drawn here come from the global isotope curve.

The last interglacial before our own, warmer than the present, with seas that stood several metres higher. Nothing in the recent past comes closer to a rehearsal of a warm world.

The sea level figure is a probability, not a measurement: it comes from combining many local records, each corrected for how the ground beneath it has risen.

A hundred thousand years of cold, and the stage during which our species leaves Africa for good, reaches Australia, enters Europe, and crosses into the Americas. Almost everything on this timeline happens inside it.

A single colour for a stage that in fact oscillates violently. Three synchronised Greenland ice cores resolve a long series of abrupt climatic changes within it.

The warm stage we are in. Eleven thousand seven hundred years of unusual stability, and every farm, city and state that has ever existed fits inside it.

One of the very few boundaries on this timeline that were decided deliberately, written down, and voted on. That does not make it a natural threshold.

The coldest seven and a half millennia of the last hundred thousand years, when the ice sheets stood at their greatest extent and the ocean sat roughly 130 metres below its present level.

A maximum is easier to name than to date. Different ice sheets peaked at somewhat different times, and the interval given here is the window in which nearly all.

The first age of our epoch, running from its floor to a sharp drying eight thousand two hundred years ago. Farming spreads through it, and the first villages with it.

Ratified, and therefore official, but recent enough that some geologists question whether such fine divisions of a single epoch are useful at all.

The middle age of the Holocene, four thousand years in which cities, writing, metal and the first states all appear, in succession and largely without borrowing.

Defined by ice chemistry, not by human events. A reader who sees the first cities appear inside this age should resist reading the age as their cause.

The age we are living in, opened by a two-century drought that several of the world's first states did not outlast.

Ratified in 2018 and criticised ever since: partly because the drought's global synchrony is less certain than the definition implies.