Where the Flood Is Not Remembered

Four hundred and fifty-two, and where they are

This Atlas maps flood traditions that can plausibly be about the end of the Ice Age: a sea that rose and kept the land, a glacial lake that drained, a strait that opened, a plain that went under. Stories about a river in flood, or a wave a thousand years ago, do not count. The water has to have changed the map and kept it changed.

Applied region by region, it produces a distribution that is nothing like even:

Australia and New Guinea hold nearly a hundred. Africa holds eight, and seven of them stand on its Mediterranean and Red Sea rim, where Egypt faces the flood world of the Near East — one of them a written text insisting that the floods were many and happened to other people. South of the Sahara the whole continent holds one, and Siberia three. That is not a difference of degree.

The obvious explanation would be that nothing happened in the empty places. It is the wrong explanation: more of the world’s map was rewritten in Africa and northern Asia than almost anywhere else.

The plain that nobody remembers

The Palaeo-Agulhas Plain lay off what is now the southern Cape of South Africa. Through the glacial maximum it was open grassland with rivers running across it, carrying grazing herds on a scale the region has not seen since, and people hunted it from the caves along what was then an inland escarpment — Pinnacle Point, Blombos, Klasies River, some of the most important archaeology of modern human origins anywhere. Then the sea came up and took the whole of it, and the escarpment became the coast.

That is San and Khoe country. And the ǀXam archive assembled by Wilhelm Bleek and Lucy Lloyd in the 1870s is the single richest record of a hunter-gatherer cosmology ever taken down in Africa: thousands of notebook pages of narrative dictated by ǀa!kunta, ǁkabbo, ǀhanǂkass’o and Diä!kwain, transcribed phonetically and translated line by line.

Go through it for water and the sea and the flood, and here is what is there. The rain is a bull or an eland, and can be captured or offended; a girl of the Early Race kills the Water’s children and a whirlwind carries the camp into the spring, and the family become frogs. A hunting simile says the springbok resemble the water of the sea, meaning a herd so vast that the Boers ran out of shot. And ǀhanǂkass’o describes, in November 1878, a flood at Victoria West that carried houses and stock down a Karoo riverbed — his wife was there, it had rained in the night, and he calls the sound of it sorcery’s bell.

That last one is a real flood, remembered exactly, by a named man who could tell you who was present. It is also a nineteenth-century cloudburst. The ǀXam were interior Karoo people; they had a word for the sea, and ǁkabbo had seen Cape Town. They did not tell of a grassy world that the water took.

They did have a bullroarer. The !goin !goin is in the same collection, swung to bring the bees. So the instrument this Atlas is built around is present in the ǀXam record, in the hands of the people who lost the largest drowned plain in Africa, and the flood is not there with it.

This is where the obvious objection dies. Africa’s flood record was largely collected by missionaries, from informants who had already met Noah — so a thin record could just be a contaminated one. Bleek and Lloyd were not missionaries. Their informants were convicts on loan from the Breakwater prison, dictating in their own language to a philologist who wanted the phonetics. It is the least evangelised body of narrative on the continent, it is the largest, it belongs to the people who lost the plain, and it is empty. No archaeologist of the Palaeo-Agulhas programme has set a living tradition beside that drowning, because there is not one to set beside it.

The quietest coast on earth

The Agulhas Plain was Africa’s largest single loss, and it invites a suspicion: perhaps Africa forgot because there was less to remember. That can be measured. Run the same water over every region in this series — the same terrain, the same sea-level curve the maps draw — and ask how far each coast actually moved: take the land that stood above the Ice Age lowstand and is under the sea now, and divide it by the length of the coastline it drowned along. The result is the average distance the shore walked inland while people watched.

Africa is last, and by a distance. Everywhere the traditions run thick, the shore moved between roughly forty and sixty-five kilometres. Africa’s moved twenty-three — the least on earth. The continent has almost no shelf: its floor drops away close to the beach, so a hundred and sixteen metres of water bought the sea less African ground than anywhere else it was spent. Where the Sunda coast came inland a village at a time and kept coming for ten thousand years, most of the African coast stepped back once and stopped. The drowning was real. It was also the quietest on the planet. And the water that did remake African country in those millennia was inland — Victoria refilling from a grass plain, Mega-Chad spreading toward an inland sea, Kivu rising four hundred metres — which is where the continent’s one tradition south of the Sahara sits.

The top of the table is what keeps this from becoming a just-so story. If retreat made memory, northern Asia would out-remember the world: the Siberian shelf is the widest on the planet, its shore walked a hundred kilometres, and it holds three traditions — every one of them already wearing Noah’s clothes. Geometry sets how much there was to see. It does not decide whether the telling survived, and the rest of this essay is about what decides that. Africa’s blank is overdetermined: the least visible drowning on earth, with scripture arriving early on top of it. Siberia’s blank needed the second mechanism alone — and it is the one place the overwriting was caught in the act.

Frazer got there first

James Frazer went through the deluge stories of the world for the first volume of Folk-lore in the Old Testament, and he was not a man inclined to report an absence when he could report a parallel. On Africa he had to.

“It is curious, that while legends of a universal flood are widely spread over many parts of the world, they are hardly to be found at all in Africa. Indeed, it may be doubted whether throughout that vast continent a single genuinely native tradition of a great flood has been recorded.” — J. G. Frazer, Folk-lore in the Old Testament (1918), vol. 1, p. 329

What he could print in place of the missing traditions ran to three items and a shrug: a Lower Congo story in which the sun plasters mud on the moon and the ancient people put their porridge sticks to their backs and turn into monkeys; a one-line notice that the Bapedi are said to have a flood that destroyed nearly all mankind; and Livingstone at Lake Dilolo in Angola, recording a village that sank with its fowls and its dogs and adding that it was the only such thing he met with in the whole country. Nothing in ancient Egyptian literature. And the Rev. John Roscoe, after twenty-five years among the peoples of the Uganda Protectorate, telling Frazer he had found no native flood legend at all. The East African ark stories Frazer set aside as Bible, and they plainly are.

Robert Moffat had already run the experiment in the south. His inquiries came back empty, and the one man who claimed a deluge of the forefathers turned out to have learned it from a missionary named Schmelen. Moffat drew the moral himself: stories picked up at a mission station or from a godly traveller get so mixed and metamorphosed by heathen ideas, he wrote, that they come to look exceedingly like native traditions.

That is the whole problem in one anecdote. The tradition Moffat was hunting had already been replaced by the thing he was hunting it with — and he could tell, because he knew Schmelen.

The largest floods on earth, unremembered

The other blank runs across northern Asia, and it comes with the most extravagant geology in the world attached.

In the Altai, on the Russian–Mongolian border, glaciers repeatedly dammed the Chuja and Kuray basins into deep lakes, and those lakes repeatedly burst. The floods that came out of them are the largest freshwater floods known to have happened anywhere on the planet — discharges beyond anything in the modern world, which scoured the Chuja and Katun valleys and left giant current ripples across the Kuray basin, gravel dunes tens of metres high that you can stand on.

People live there. The Telengit of the Kosh-Agach district live in that exact terrain, and Kuzelesh Yadanova has collected their oral prose village by village — Kurai, Chagan-Uzun, Ortolyk, Mukhor-Tarkhata, Bel’tir, Kokorya. The flood is in there, ten texts of it, and it is attached to the sacred mountains: the survivors climb the ыйык, Kökörü and Irbistü, and on the tops, people say, you can still find the broken timbers of the raft and a bronze cauldron.

Not one of them describes a lake going out. The valleys those people graze were cut by the largest floods in the history of fresh water, and what the stories describe is high water and a landing place — which is the shape of Noah, with Kökörü and Irbistü as the local Ararats.

The same holds along the whole belt. The Khakas raft lands on Mount Borus and gives it its name. The Shor have the ark, the devil, the mouse that gnaws the hull and the cat that stops it — the standard Eurasian legend complex, intact. The Kazakh and Kyrgyz have the ark on Qazygurt. And among the Ob-Ugrians and the Samoyeds the water is doing something different again: in the Mansi, Khanty, Nenets, Nganasan, Selkup and Ket traditions the world begins as water and land is fished up out of it by a diving bird. That is a cosmogony in which land is made. It has no slot for land being lost.

One more thing about the north, and it settles the geology. The great ice-dammed lake of the West Siberian Plain — the Ob and the Yenisei blocked by the Kara ice sheet and ponded into an inland sea — is real, and it is far too old. Its well-attested phases run to ninety and sixty thousand years ago, tens of millennia before anything oral tradition could carry. It cannot be the memory behind the Samoyed flood, and it should stop being offered as one.

What did survive: one lake, and one that only looks like it

Africa is not quite empty, and the single thing that came through is worth having.

Imana gives a farmer a bull on the condition that he tell nobody. His wife finds out and the secret gets out, and Imana floods the man’s whole country. The water stays: it is Lake Kivu, on the border of Rwanda and the Democratic Republic of the Congo, and the few hilltops that stood above the flood are its islands.

Lake Kivu did that. Douglas Wood and Christopher Scholz reconstruct the basin as very nearly dry around twenty thousand years ago, and then, at about twelve thousand two hundred, a transgression of more than four hundred metres that established the southern overflow the lake still uses. Jean-Pierre Descy, François Darchambeau and Martin Schmid, writing the standard limnology of the lake, open their history of it with the Imana story and set it beside that refilling. Four hundred metres is a lake climbing the hills, and the islands in it are the hilltops the story says they are. Albert Pagès had the narrative down in Anthropos in 1919, from the Banyarwanda, before any of the geology existed to be fitted to it.

A thousand kilometres north there is a second candidate, and following it to the bottom is more instructive than adding it to the table.

The story goes that a woman named Sepenya lived on a dry plain in what is now the Turkana basin, drank at a spring, forgot to put the lid back on it, and the spring came up and made the lake. She and her son are the ancestors of the El Molo, who call themselves Gurapau, the people of the lake, and who fish it still. The geology cooperates handsomely: Turkana rose after about twelve thousand years ago, flooded wide low plains and overflowed north toward the Nile.

The story has no published record behind it. It surfaces in 2018 on a Kenyan blog, credited to a tour guide, and every later appearance descends from that page. The El Molo are among the most-studied small peoples in East Africa — Dyson and Fuchs took the first ethnography in 1937, when there were eighty-four of them left, and Neal Sobania, Bernd Heine and the National Museums of Kenya have all worked on them since. Not one of those records carries Sepenya. The autonym is real. The origin tradition the scholarship does carry is a migration out of Arbore country in Ethiopia, which the language corroborates and which is not a flood.

So Africa south of the Sahara yields one. And the near-miss is the more useful half of the result, because it shows the mechanism working in the present tense: a people who live on an island in a lake that really did arrive, telling a story about the lake arriving, that nobody wrote down in ninety years of visiting them.

The survivor is an inland Rift lake, and it is not a coast. That is the second half of the pattern.

What got there first

Set the blanks against the geology and the explanation cannot be that nothing happened.

Lake Mega-Chad, at its African Humid Period highstand, covered on the order of three hundred and fifty thousand square kilometres of what is now the Sahara, and then collapsed to the remnant we call Lake Chad; the Bodélé Depression is its floor and the Bahr el Ghazal its drain. The Green Sahara itself — rivers, lakes, fish, hippo, cattle and the people who painted them at Tassili and Acacus — ran wet from roughly fifteen thousand years ago and then dried to hyper-arid desert at about five and a half thousand, in front of witnesses who had been painting the animals on the rock for millennia. Lake Victoria was dry, or nearly, through the glacial maximum, and refilled from about fourteen and a half thousand years ago: an inland sea the size of Ireland appearing where there had been grassland. Palaeo-lake Makgadikgadi filled the Kalahari. The Gulf of Guinea, Namibian and Mozambican shelves all drowned on schedule.

Africa has as much end-of-Ice-Age landscape change as any continent on earth, and south of the Sahara it narrates none of it — the traditions on its northern rim are literate observers reading the ground, not the ground’s own peoples telling what it did. Northern Asia has the largest floods in the world and narrates high water on a mountain.

What both regions do have is a book. Christianity reached Ethiopia in the fourth century and the Nile valley before that; Islam crossed the Sahara from the eighth; missionary work covered most of the rest of the continent in the nineteenth. Orthodoxy moved through Siberia with the Russian state, and Islam through the Turkic south. Everywhere they went they brought a flood — a very good one, with a boat, a mountain and a bird, portable, memorable, and endorsed.

A flood story is exactly the thing that gets overwritten by a better flood story, because the two occupy the same slot. You cannot hold both, and one of them arrives with literacy, authority and a reason to prefer it. Nadezhda Oinotkinova, working through the southern Siberian corpus, identifies it outright as biblical legend adapted into local mythology — the Altai narrators are not telling a garbled Ice Age memory, they are telling Genesis with the names changed to the mountains outside.

The obvious counter-example is in the table above, and it is worth meeting head on. The Mediterranean Near East — the Levant, Anatolia, Mesopotamia, the Black Sea rim — is the deepest scriptural ground on earth and holds forty-six traditions. North America was evangelised coast to coast and holds sixty-seven. Neither was emptied, and the reason is that displacement needs the incoming story to want the same piece of ground. Around the Black Sea and the Gulf, Noah is the local tradition: Utnapishtim, Ziusudra, Atrahasis and Deucalion are the same flood in older clothes, attached to the same rivers and the same shore, and the book that arrived was a cousin rather than a rival. In North America the missions came late, in force, and mostly in the nineteenth century — by which time Boas, Swanton, Teit and Dorsey were already in the villages with notebooks, taking down the flood the missionaries were displacing. Africa and the Altai got neither reprieve. Scripture arrived early, arrived alone, and arrived before anyone was writing the local version down.

The earth-diver does something subtler in the one belt where it rules alone. Where the cosmogony says the world began as water and land was brought up out of it by a diving bird, there is no conceptual room for land that used to be there and went; losing ground is not a thing that can happen in that account of how ground came to exist. The earth-diver is told across half the northern hemisphere and mostly it is one story among many — in North America it sits beside deluges, escapes and remade coasts without troubling any of them. Across Samoyedic and Ob-Ugrian Siberia and into Finnic and Karelian northern Europe it is not one story among many. It is the account of origins, with no competitor, and the water in it only ever recedes.

Now look at where the traditions did survive. Arnhem Land and the Sepik. The interior of Amazonia. The Northwest Coast. The eastern Himalaya. These are places a book reached late, reached partially, or reached after the collectors did — and Sahul is the extreme case at both ends, with the deepest continuous occupation on earth and the shallowest overlay of anybody else’s flood.

Sundaland and the Andes look at first like exceptions and are in fact the confirmation. Both were converted early and hard — Islam through the archipelago from the thirteenth century, Catholicism in the Philippines from 1565, the extirpation of idolatries in Peru from the 1530s — and both still hold nearly sixty traditions each. What survived there is not a religion’s failure to arrive. It is a flood already welded to a named place: this strait, this ridge, this lake, this stretch of drowned shelf. A story pinned to ground is much harder to lift than a story about the world in general, which is why what the imported flood emptied out first, everywhere it went, was the unlocated kind.

Which yields a prediction. If displacement is the mechanism, then whatever survives in Africa should be found where scripture reached last: inland, small, out of the way, ideally speaking a language nobody else speaks. The survivor is Kivu: an inland Rift lake in the hills of Rwanda, a thousand kilometres from any coast and from any early mission, written down by a priest working in the local language in 1919. One survivor south of the Sahara, and it is exactly where the mechanism says to look.

The argument from the blanks

The universal flood myth is one of the durable ideas in comparative mythology, and it is usually offered in one of two forms: either every people has a flood because every people saw the water rise, or every people has a flood because the human mind produces one. The first is diffusionist, the second is psychological, and both need the distribution to be even.

It is not even. On this Atlas’s maps it is ninety-five to one.

The flood tradition that this Atlas tracks — a real body of water that took real ground and left the world arranged the way it is now — is a regional inheritance. It is dense on the Sunda and Sahul shelves, along the Pacific coast of the Americas, around the Mediterranean and Black Sea rim, through the eastern Himalaya and down the Atlantic seaboard of Europe. It is absent from Africa and from northern Asia. And the boundary between those two zones is not a boundary of geology, or of antiquity, or of how much land the sea took. Africa has the deepest human past on earth and lost a grassland the size of Ireland. Siberia has floods no other continent can match. Neither remembers.

There is a way to test this that does not depend on the filter at all. If the Ice Age filter is simply screening out Noah’s shape — high water, a mountain, a landing — then the empty regions are empty by construction and the argument is a circle. So look for a tradition inside the Noah belt that is not flood-shaped and see whether it survived. There is one, at the bottom end of Lake Baikal. Buryat and Evenk narrators recorded in 1936, 1937 and 1938 tell of Siberia under ice, of the first people living on the patches that stand out of it, of the meltwater gathering in the lake until the lake cannot hold it, and of the rock at its lower end giving way with a sound they were standing there to hear. They call the breach angara, the gaping mouth, and the river that came out of it still carries the name. Nobody is saved in a boat. No mountain is landed on. Nothing about it is shaped like Genesis, and it came through three centuries of Orthodoxy intact — because the incoming story had nowhere to put it. That is what displacement looks like from the other side.

The boundary is a boundary of what arrived afterwards. Where the water rose and nothing came later to displace the account of it, the account is still there — on the Sepik, in Arnhem Land, on the Sunda shelf, in the Kathmandu Valley, at the Point of Ayre. Where a book came, the book is there instead, and the mountain the ark landed on is the one you can see from the village.

An old priest at Saïs said all of this to Solon twenty-four centuries ago. You Greeks are always children, he told him: the world is wrecked over and over, most often by fire and by water, and the deluge the Greeks remember was only one of many. When the flood comes, the herdsmen high in the hills live and everyone in the cities drowns — so the survivors are the ones who cannot write, and each time the towns, the records and the names go under together, and a country begins again from shepherds who know nothing of what came before. Egypt is the exception, because the Nile spared it: no sea came in, no storm rain fell, the river rose gently from below, and the priests kept writing. The proof he offered Solon was a war the Greeks had forgotten entirely — Atlantis, nine thousand years gone. Memory of floods, the priest is saying, is not a function of floods; it is a function of who survives them and who keeps the records. That is this essay’s argument, stated on the African shore, by the one tradition on this map whose whole business was writing things down.

The empty half of the map is a measurement. It says what the full half cannot say alone: that these are transmissible, replaceable objects with a history, that they can be lost, and that we can see the shape of the thing that replaced them. Frazer looked at Africa in 1918 and reported that the flood was not there. He had no way to know that the silence was itself the evidence.

Mapped traditions