The Bronze Age Collapse: The Fifty Years the Old World Came Apart

The Late Bronze Age Collapse is dated c. 1200 to 1150 BCE, and within approximately fifty years it destroyed or severely diminished virtually every major palace-based civilization in the eastern Mediterranean and the Near East. That fifty-year window is our file's own headline timing, and it is the only collapse date range this article will lead with.

Cline himself has said plainly that the year on his cover is a device. In a published interview he explains that '1177 BCE is a good reference point, for it is in that year that the Sea Peoples invaded Egypt for the second time', and that 'I use 1177 BCE as shorthand for the entire Late Bronze Age Collapse, just as we commonly use 476 CE as shorthand for the fall of the Roman Empire; we know that neither took place entirely in exactly that year.' On causation he is equally direct: the urge to find one explanation 'seems to come from a modern human need for an easy explanation', and 'it took much more than a single cause to bring down the Bronze Age civilizations'. Nobody here is being straw-manned. The tidy year was never claimed as a date.

Our file's counter-arguments section goes further, and it is arguing against the framework the same document otherwise endorses. A truly interconnected system should show synchronous collapse. The destructions actually span about fifty years with considerable local variation. Any sentence about this event that implies a single year, a single season or a simultaneous fall is contradicted by our own primary source before any outside critic gets involved.

A sister file puts a name to the same objection. Jan Driessen argued in 2009 that the destruction horizon across Mycenaean sites may not be synchronous at all: improved radiocarbon dating suggests the destructions occurred over several decades rather than simultaneously, which weakens the cascade narrative that the single-year framing depends on.

So what is the nearest real thing to 1177? Ramesses III, who reigned c. 1186 to 1155 BCE, repelled the Sea Peoples at the Battle of the Delta at c. 1178 BCE and in a simultaneous land battle in Djahi, in southern Canaan, and recorded both victories at Medinet Habu. That is the second Sea Peoples attack on Egypt, and it is the event Cline's title points at. It is the date of one Egyptian battle. It is not the date of a collapse.

In the Aegean: the citadel of Mycenae and its surrounding settlement were destroyed c. 1190 to 1180 BCE. The Palace of Nestor at Pylos was burned c. 1180 BCE, and the fire that ended it is the reason its Linear B archive exists, since the flames baked the clay tablets that would otherwise have dissolved. Among those tablets are records of coastal watchtower deployments, which our file reads as suggesting the rulers anticipated an attack. Anticipated, not identified: the tablets are evidence that someone was watching the sea, not evidence of who came over it or why. Tiryns suffered major destruction and was then partly reoccupied on a reduced scale, and our file gives it no absolute date.

In Anatolia and the Levant: Hattusa, the Hittite capital at modern Bogazkale, was burned and abandoned c. 1180 BCE, and the last Hittite king, Suppiluliuma II, simply vanishes from the record. An empire that had rivalled Egypt for two centuries collapsed entirely. Troy level VIIa was destroyed by fire c. 1180 BCE. Ugarit, a wealthy trading city on the Syrian coast, was destroyed c. 1185 BCE and never reoccupied. Megiddo carries a destruction layer between its Late Bronze and Iron I strata, with no absolute date given in our file.

Ugarit's last letters were preserved in the same clay the city died in, and they include desperate pleas for military aid. The most quoted line of them is a single sentence: 'the enemy ships are already here, they have set fire to my towns.' Our own file attaches that line to a specific tablet number and a popular nickname, and independent checking places that nickname on a different group of texts entirely, so this article quotes the line as what it certainly is, the famous surviving plea from Ugarit's final days, and does not attach it to a catalogue number.

Further out: Cyprus was devastated, with Enkomi and other cities destroyed, and the Kassite dynasty in Babylon fell. Our file gives dates for neither, and a sister file reads Cyprus the other way entirely. That disagreement is ours, and it is handled in section 10 rather than smoothed over here.

Egyptian records, principally the Medinet Habu inscriptions of Ramesses III at c. 1178 BCE, name a coalition of peoples who attacked Egypt by land and by sea: Peleset, Tjeker, Shekelesh, Denyen and Weshesh. That the inscriptions exist and say this is Tier 1. It is also worth noticing what kind of document it is. Medinet Habu is a pharaoh's own account of his own victory, carved on his own mortuary temple.

Who these people actually were remains partly unknown, and our file marks its own uncertainty with punctuation. Its identifications read Tjeker (Sikels?), Shekelesh (Sicilians?), Denyen (Danaans?), Sherden (Sardinians?) and Ekwesh (Achaeans?). Only Peleset (Philistines) and Lukka (Lycians) carry no question mark. Those question marks are the file's own and they survive into this article intact.

The Egyptian record also starts earlier than the destructions do. Merneptah, at c. 1208 BCE, defeated a coalition including Lukka, Sherden and Ekwesh. Whatever was happening in the eastern Mediterranean was already producing armed movements toward Egypt roughly thirty years before Mycenae, Hattusa and Ugarit fell.

One group can be followed to where it settled. Archaeological evidence confirms the Peleset settled in Canaan, in the Gaza Strip region, becoming the biblical Philistines. Excavations at Ashkelon, Ekron and Ashdod reveal Aegean-style pottery and architecture in early Iron Age levels.

Ancient DNA has entered this argument, and the careful statement of the result is narrower than the popular one. Feldman and colleagues, publishing in Science Advances in 2019, reported European or Aegean genetic input in Philistine burials at Ashkelon. One of our sister files calls that a confirmation of the identification; another states it in the weaker form, and the weaker form is the right one. Genetic input into a population is not the same thing as a confirmed migration narrative, and the study's own framing is of a transient gene flow.

Then there is the reading that turns the whole story around. Israel Finkelstein and others have argued that the Sea Peoples were as much refugees from the collapse as they were causes of it: displaced populations from Aegean and Anatolian societies that were already failing, moving east and south because home had stopped working. On that reading the ships in the Egyptian reliefs are a symptom of the collapse arriving in Egypt, not the mechanism that produced it elsewhere.

One ancient correspondent had the same suspicion at the time. Independent sources describe a group of three letters from the Rap'anu archive at Ugarit, exchanged between Ammurapi, the city's last king, and Eshuwara, a senior functionary of Alashiya on Cyprus. Ammurapi reports an enemy fleet of twenty ships sighted at sea. The Alashiyan reply asks where Ammurapi's own troops and the enemy fleet actually are, and includes the suggestion that the attackers may have been 'the people from your country and your own ships'. That is a complication of the pure external-invasion reading and nothing more. One functionary's suspicion, reported at a distance and known to us through secondary summaries rather than a primary edition, is not evidence that the raiders were local.

Popular claims linking the Sea Peoples to refugees from Atlantis have no archaeological, textual or geographic support. That is the whole of what this article has to say about it.

David Kaniewski and colleagues, publishing in PLOS ONE in 2013, documented a drought lasting about 300 years using pollen cores from the Larnaca Salt Lake on Cyprus, which record a dramatic shift from Mediterranean woodland to dry steppe. A note on the number, because our own file gives two different answers to the same question: its quick summary dates the drought c. 1200 to 850 BCE, while the section that carries the attribution says c. 1200 to 900 BCE. The published paper's own abstract gives 1200 to 850 cal yr BC. The summary matches the source and the attributed section does not, so this article uses c. 1200 to 850 BCE and states why.

Dafna Langgut and colleagues, publishing in Tel Aviv in 2013, found the parallel signal in pollen records from the Sea of Galilee: severe aridification beginning c. 1250 BCE and persisting for centuries. That timing matters more than it usually gets credit for. The drought opens about fifty years before the destruction window does, and it is still running long after the last palace has burned. It is not synchronous with the collapse. It brackets it.

Brandon Drake, in 2012, confirmed widespread drought across the eastern Mediterranean using speleothem records, that is, cave stalagmites, from Soreq Cave in Israel. A sister file in our library broadens the proxy base further: Dead Sea levels dropped dramatically during the drought window, confirmed by shore-terrace stratigraphy, alongside the Soreq oxygen-isotope shift to arid conditions and the pollen cores from the Sea of Galilee and Cyprus indicating agricultural decline. Four independent lines of physical evidence, from three different kinds of archive, agreeing.

The sharpest result of all is very recent and appears in none of our four files on this subject. Manning, Kocik, Lorentzen and Sparks, publishing in Nature in 2023, took juniper timbers from the Midas Mound Tumulus at Gordion in central Turkey, combined tree-ring widths with carbon-13 to carbon-12 isotope ratios, and identified three consecutive years of severe drought at approximately 1198 to 1196 BCE, give or take three years, inside an already dry period, coincident with the abandonment of the Hittite capital. Read the title's verb carefully: coincident with, not caused. Manning's own stated position is that severe climate events may not have been the sole reason for the Hittite Empire's collapse, and that this drought may have been a tipping point, at least for the Hittites. This is the finest resolution yet achieved for one stressor at one polity. It is not the answer to the question.

Amos Nur, of Stanford, documented in 2008 that clustered earthquake sequences, which he calls earthquake storms, can devastate multiple cities within decades. It is a real seismological phenomenon rather than a convenience, and it does something no other single suspect manages: it produces destruction across a wide area over exactly the kind of multi-decade span our timing evidence shows.

Some of that damage is dated. A sister file records earthquake damage at Late Bronze Age destruction layers at Troy VIh at c. 1250 BCE, Mycenae at c. 1250 BCE, Tiryns at c. 1200 BCE and Enkomi on Cyprus at c. 1200 BCE. Two of those are half a century before the destruction window opens, and the Troy horizon is a different event from the Troy VIIa fire at c. 1180 BCE. Merging them would manufacture a synchrony the evidence does not contain.

More broadly, archaeological evidence of earthquake damage is documented at Mycenae, Tiryns, Midea, Knossos and various Levantine sites, though dating individual earthquake events precisely remains difficult. That difficulty is not a rhetorical hedge. It is the reason this suspect sits at Tier 2 while the drought sits at Tier 1: the events are real, and pinning any one of them to any one destruction is another matter.

Our file then writes the sharpest sentence in the entire document, and it is a sentence against its own suspect. The earthquake storm hypothesis explains physical destruction but not the failure to rebuild. Cities in this region had been knocked down and put back up for two thousand years. What is different about this episode is not that things fell. It is that so many of them stayed down.

Robert Drews argued in 1993 that the palaces were beaten by a change in tactics. On his account a military revolution, the shift from chariot-based warfare to infantry using javelins and long swords, made palatial armies obsolete and left them vulnerable to mobile raiders. It is a genuinely testable proposal about equipment and battlefield practice, and it carries an internal complication our library should own: our primary file and one sister file carry Drews at Tier 2 as a contributing factor, while another sister file files the same thesis at Tier 4 and marks it debunked, on the ground that he proposed it as the sole cause. Read as a contributing factor it is credible. Read as the whole answer it falls under the same Tier 4 refusal every other single cause does.

The texts also hint that some of the trouble was domestic. The Pylos tablets reference deployments against unspecified threats, and Hittite texts mention provincial rebellions in the empire's final years. Neither says who the threat was, which is exactly why the material is suggestive rather than decisive.

From that our file draws the possibility that some destructions may represent internal revolts against palace elites rather than external invasions. A burned palace does not record who lit it. That single fact is why this suspect cannot be ruled out and cannot be convicted, and it sits underneath the whole invasion narrative like a question the ash cannot answer.

Bronze is the weak link, and specifically the tin in it. Tin makes up typically about 10 percent of the alloy, and none of it is local: our library sources it to Afghanistan, naming Badakhshan, to Cornwall in Britain, and possibly to central Asia. A Mycenaean or Hittite armoury depended on a supply chain running thousands of kilometres through polities it did not control. The collapse severed those networks, depriving dependent economies of critical resources.

That the tin supply failing was a trigger of the collapse, rather than a consequence of it, is a much weaker claim than the dependency itself, and one of our sister files files it that way explicitly: direct evidence is limited. The route and the dependency are Tier 1. Tin disruption as a cause is Tier 3, and the distinction is not pedantry, because a severed network is exactly what a collapse produces as well as what it might produce a collapse from.

The eruption of Hekla 3 in Iceland, dated variously to about 1159 or 1021 BCE, has been proposed as a contributing factor through climate cooling, a volcanic winter. The dates remain too imprecise to establish a clear link with a collapse onset at c. 1200 BCE, and the honest way to say that is blunter: both dates offered are later than the onset, so this hypothesis fails on chronology before anybody tests it on physics. One of our sister files files the volcanic trigger a full tier lower than this, at Tier 4, on the grounds that it confuses chronologies.

Ancient texts mention plagues and epidemics, including Hittite plague texts describing outbreaks during the empire's decline, and scholars have speculated that epidemic disease contributed to population decline. No specific pathogen has been identified and the evidence is fragmentary. A sister file puts it slightly more sharply: the proposal lacks direct bioarchaeological evidence for the period around 1200 BCE. Neither statement is a refutation. Both are a description of what is missing.

One more thing this event is not. It is not Thera. Our primary file never mentions the eruption at all; of the sister files, one raises it only inside its own dubious tier, to say that linking the collapse to Thera confuses chronologies, and the other states at its speculative tier that the eruption, dated to c. 1628 or 1530 BCE, predates the collapse by several centuries and is not directly relevant. The corpus's own answer is that there is no usable connection, and the eruption has its own file and its own article in this wing, where the dating is the whole argument.

No, it was not one thing. Attributing the collapse to any single cause, whether the Sea Peoples alone, drought alone or earthquakes alone, is rejected by current scholarship. The evidence demonstrates multiple simultaneous stressors interacting within an interconnected system. Note what that does to the rivals: it does not retire them. It refuses each of them as a complete explanation while accepting several of them as real, well-evidenced stressors.

Eric Cline, of George Washington University, synthesised the multi-causal model in 2014, arguing that no single factor, not drought, earthquakes, Sea Peoples, internal revolt or trade disruption, was sufficient to cause the collapse alone, but that their simultaneous convergence overwhelmed interconnected systems. The model draws on complexity theory and on Joseph Tainter's 1988 work on societal collapse, which argues that complex societies experience diminishing marginal returns on investment in complexity, and so become vulnerable to stressors they could once have absorbed.

Our file marks one implication as a key finding while keeping the hedge in its own wording: the Bronze Age collapse may be the earliest well-documented example of systemic cascading failure in interconnected civilizations, a concept with relevance to modern globalized systems. Relevance is the word the file uses, and this article will not upgrade it into a lesson. A cascade at the far end of the Bronze Age is an object worth studying, not a verdict on the present one.

The first attack on the framework is that it may not be falsifiable. Our file states the critique and leaves the critics anonymous: scholars argue the systems collapse model is too vague and potentially unfalsifiable, explaining everything and therefore explaining nothing. The named source for exactly that critique is sitting unused in the same document's own bibliography. Knapp and Manning, in the 2016 American Journal of Archaeology review already quoted above, write that explanations for the crisis are numerous, that there has never been any overarching explanation to account for all the changes, and that there is no final solution, concluding instead that the collapse presents multiple material, social and cultural realities demanding continuing collaborative attention. That is a real position held by named people, and this article names them rather than writing 'some scholars'.

Assyriology specialists note that Assyria actually grew in power during this period, under the Middle Assyrian Empire of Tiglath-Pileser I, who ruled about 1114 to 1076 BCE. The file's own reading of why is the most useful sentence it has on the subject: the collapse primarily affected palace economies dependent on international trade, while less-complex societies adapted. If the system was so interconnected that its failure cascaded everywhere, a major power inside it should not have been expanding on the far side of the cascade.

A sister file makes the same point across the map, and in doing so contradicts our primary file. It records that the collapse was not uniform: Egypt survived though weakened, Cyprus experienced transformation rather than destruction, and some Levantine coastal cities, Byblos and Sidon among them, continued operating, which suggests differential resilience based on geographic and economic factors. Our primary file's summary says instead that Cyprus was devastated. Those are two of our own documents reading the same island the opposite way, and this article carries the disagreement rather than choosing one silently.

Guy Middleton went after the word itself. Our sister file attributes to him, in 2012, a criticism of the collapse framework as privileging elite palatial perspectives: for the rural farming population, roughly 80 percent of people, the end of the palaces may have represented liberation from exploitative palatial redistribution systems rather than catastrophe. Whether or not one accepts that, it exposes something structural about the evidence base. What we can see collapsing is what left archives, and archives were kept by palaces. A world in which nobody was writing anything down looks emptier than it necessarily was.

The collapse ushered in a prolonged Dark Age, c. 1150 to 900 BCE, characterised by population decline, loss of literacy, reduced long-distance trade and simpler material culture. The literacy loss is stark and specific: Linear B disappeared, and Hittite cuneiform ceased. Two writing systems in use for centuries stopped being written.

The one solid number for the human scale of it comes from a sister file. Population estimates for the Argolid, the Mycenaean heartland, show a decline of about 75 percent in the centuries following the collapse, a figure our file attributes to Dickinson in 2006. That is one region, measured across centuries rather than across a season, and it should be read as exactly that rather than as a Mediterranean-wide death toll.

What eventually came out of the dark is the world most readers think of as antiquity. The collapse created the conditions for the emergence of Iron Age societies: the Greek poleis, the Phoenician maritime network, and the kingdoms of Israel and Judah.

Recent archaeology has also nuanced the word 'dark'. There was genuine decline in material complexity and literacy, and at the same time some regions, particularly Cyprus and parts of the Levant, show continuity and even innovation during the early Iron Age. The period saw the invention of the alphabet, out of Phoenician consonantal script, and the spread of iron technology. A world losing two syllabic scripts was simultaneously acquiring a writing system that almost anyone could learn.

That the collapse caused those innovations is a further and weaker claim, and a sister file says so directly: the framing that the collapse enabled them is a historiographic interpretation rather than an empirically demonstrable cause-and-effect chain. This article therefore describes what followed and does not assert that the collapse produced it.

Iron is the one suspect whose arrow our library points firmly downstream. The development and adoption of iron technology, which does not require tin, accelerated after the collapse and eventually displaced bronze, which suggests that tin scarcity was a factor in the technological transition. Iron is what the collapse produced, not what produced the collapse.