The Collapse Of The Bronze Age World

AThe story most often told about the end of the Bronze Age begins with ships. Egyptian temple reliefs carved a little after 1200 BC show a sea battle against strangers, and the accompanying inscriptions name a coalition of peoples whose homelands are never stated; from the moment those reliefs appeared in a scholarly edition in 1922, the figures on them have been asked to explain the fall of every kingdom between the Aegean and the Euphrates. The attraction of the account is narrative rather than evidential. A burnt layer records that a settlement was destroyed, but it does not record who destroyed it, and the same ash will be produced by a raid, by a civil war, by a lightning strike or by a deliberate act of abandonment carried out by the inhabitants themselves. Of the eleven major destruction horizons now dated to the period, weapons of foreign type have been recovered from three. Even those three are ambiguous, since weapons travelled as trade goods long before they travelled in the hands of enemies. What the invasion hypothesis explains best, in short, is the shape of the story rather than the distribution of the ruins.

BIlse Radek, an archaeometallurgist at the Tarnow Institute for Metals Studies, has proposed that the crisis should be read as the failure of a supply chain. Bronze requires tin, and the eastern Mediterranean has almost none of it; the metal had to be carried from deposits thousands of kilometres away, through intermediaries who kept the routes secret and the prices high. Between 1998 and 2004 Radek's team measured lead isotope ratios in 61 bronze blades from four coastal sites, and found that the signature characteristic of the older eastern source disappears from the assemblage in 23 of them, all deposited within the final two generations of palatial administration. In their place appears a signature matching a western deposit that had never before supplied the region in quantity. Radek argues that the substitution was not a preference but an emergency. A palace that could no longer guarantee tin could no longer equip its dependants, and a palace that could not equip its dependants had little to offer in exchange for the grain they delivered, so the obligation that held the system together dissolved from the inside before any outsider arrived. The ships, on this reading, found a structure already hollow.

CThe argument gains force from what did not collapse. Radek maintains that the disaster was confined to the institutions that depended on metal as an instrument of government, and that the countryside, which needed little bronze and produced most of its own food, passed through the same decades with its population broadly intact. Survey work in an upland valley behind the coast records 23 per cent more occupied farmsteads in the century after the destructions than in the century before them. Pottery from those farmsteads is coarser and locally made, which she reads as independence rather than poverty. The claim is bounded in two ways that she states plainly. It holds only for regions where settlement lay above the level the tax collectors reached, and it holds only where the local diet did not depend on grain shipped in from elsewhere; along the Levantine coast, where both conditions fail, the country sites empty as thoroughly as the citadels. The model therefore predicts a patchwork rather than a uniform catastrophe. Radek dismisses as sentimental the older picture of a whole world going dark at once.

DMeret Osei, a palaeoclimatologist at the Kalmar Centre for Environmental Records, objects that the sequence has been assembled in the wrong order. Her own evidence is a stalagmite from a cave in southern Anatolia whose growth bands, counted and dated in a study published in 2013, indicate rainfall roughly a third below the long-term average for some ninety years, with the driest interval beginning about a generation before the earliest destruction layer. If the drought came first, she insists, then the change of metal source that Radek dates so carefully is a symptom of a trade network already failing under crop failure, and not the cause of anything. Osei concedes two weaknesses. The stalagmite is a single column in a single cave, and a record of that kind describes the catchment above it rather than a region; extending it across the eastern Mediterranean requires an assumption about the coherence of ancient weather that nobody can test. The dating, moreover, carries an uncertainty of about thirty years on either side, which is wider than the interval her argument depends on. Radek is quick to point out that a margin of that width would allow the dry years to have followed the destructions rather than preceded them.

EMost specialists now hold some version of a two-stage account, though they differ over which stage did the greater damage. A run of dry years reduced the surplus; a system that converted surplus into metal, and metal into obligation, had no reserve to draw on when the surplus shrank; the raiding that followed struck communities already unable to pay for their own defence. Neither researcher rejects this outline, and the disagreement is now about proportion rather than about mechanism. Two obstacles keep it from being settled. The first is chronological resolution: a review of published radiocarbon determinations in 2021 found that 61 per cent of the destruction layers rest on a single sample, and no argument about what preceded what can be built on evidence of that grain. The second is that the two proxies are not commensurable, since a stalagmite records a climate but not a decision, while a blade records a decision but not a climate. The writer's own inclination is to treat the drought as the trigger and the metal system as the reason the trigger proved fatal, on the understanding that the chronology may yet be revised.