Guido Borghi, a linguist at the University of Genoa, thinks Plato’s Atlantis is a memory of a real country, one that drowned under what is now the North Sea around eight thousand years ago. He means it literally: an actual lost land, remembered in a story and carried south to Greece. It sounds like exactly the kind of claim this site exists to take apart.
Except that the lost land is real. In September 1931 the trawler Colinda was fishing the Leman and Ower Banks, about twenty-five miles off the Norfolk coast, when its net dragged up a black lump of peat from a hundred and twenty feet down. The skipper broke it open and found a barbed point of antler inside, a harpoon or a fish spear, eight and a half inches long. It is still held by the Norfolk Museums Service. That one object proved something that sounds impossible: the floor of the North Sea was once dry land, and people hunted on it.
We call that land Doggerland, after the Dogger Bank. It was a whole country, with lagoons and saltmarsh and river valleys and hunting grounds, and the sea swallowed all of it. So half of Borghi’s claim is not in dispute. There really was a drowned world, in the right part of the world, at roughly the right time. Everything turns on the other half. Could anyone have remembered it, across eight thousand years and the length of a continent, with no writing to hold the memory still? That is the whole problem, and the rest of this piece is the test the claim has to pass.
The country under the water
The science of the drowned land is not in doubt. As the last Ice Age ended, sea level rose by roughly a hundred and twenty metres worldwide. The meltwater drowned the low plain of the southern North Sea over thousands of years. The Thames, the Rhine, the Meuse, and the Scheldt once joined into a single broad river that ran west through what is now the English Channel. The Dogger Bank, a ridge of glacial debris, stood up as high ground and became an island before it too went under.
The archaeologist Bryony Coles gave the land its name in a 1998 paper, and she argued for treating it as a place where people lived rather than a corridor they crossed. Since then the Europe’s Lost Frontiers project at the University of Bradford, led by Vincent Gaffney, has mapped tens of thousands of square kilometres of the seabed with oil-industry sonar. The maps recovered the shapes of drowned rivers and hills. Cores drilled into the seabed have yielded sedimentary ancient DNA, the genetic trace of the forests and animals that once grew there. The trawlers keep bringing up the rest: mammoth and lion bone, worked flint, and once a fragment of a Neanderthal skull more than forty thousand years old.
The harpoon from the Colinda was not recognised for what it was until a geologist named Clement Reid examined the peat around it. Reid had already argued, against ridicule, that a drowned forest lay under the North Sea. The fishermen who worked those waters had a name for the dredged-up peat long before the scientists arrived. They called it “moorlog.”
There was probably a final blow, too, though its force is now disputed. Around 6200 BCE a vast slab of the Norwegian seabed gave way in a submarine landslide known as the Storegga Slide, “the great edge.” It sent a tsunami across the North Sea. Geologists have found its deposits along the whole Norwegian coast, in the Faroes, in Shetland above twenty metres, and in eastern Scotland. For a while the Storegga wave was cast as the event that finished Doggerland in an afternoon.
The most recent work has cooled that story down. James Walker and colleagues, publishing in Antiquity in 2020, found the first traces of the tsunami in the southern North Sea and concluded the opposite of catastrophe. Much of Doggerland was already underwater before the wave came, and significant areas seem to have survived the wave and lasted into the Neolithic. The land was lost to “inexorable sea-level rise,” in their phrase, not to a single drowning night. Popular accounts get this wrong. There is no good evidence that the wave killed a quarter of Britain’s population, and the researchers who found the deposits argue against that reading.
So the drowning was mostly slow. Slow disasters make bad candidates for vivid memory, and that is the first obstacle in the way of any story carrying Doggerland forward.
Can memory last that long?
So could anyone have remembered it? The firm ground ends with that question. Doggerland’s last lowlands went under more than eight thousand years ago, five thousand years before anyone in the region could write. No text records it. The only way a memory of it could reach us is through oral tradition, and the durability of oral tradition is exactly what scholars fight about.
The strongest case on record is the Cascadia earthquake. On the evening of 26 January 1700, a great quake struck the subduction zone off the Pacific Northwest of North America. We know the date to the day, and not from any Western record on that coast. The quake sent a tsunami across the Pacific that struck Japan, where officials wrote down a wave that arrived with no earthquake felt beforehand, an “orphan tsunami.” Kenji Satake and Brian Atwater matched those Japanese records to the dead “ghost forests” drowned by the quake along the American coast, whose tree rings stop in the winter of 1699 to 1700.
The Indigenous peoples of that coast remembered it too. The Huu-ay-aht of Pachena Bay on Vancouver Island carried an account, recorded in 1964, of a winter night when every village on the bay was destroyed by a wave, except for the people of one settlement high on a mountainside. The oral memory and the geophysical event line up on the same night, fixed by an independent written source on the far side of an ocean. The memory there had to last only about two hundred and sixty years to reach the recorder. That is comfortably within what any folklorist will accept, which is exactly why it is the strongest case there is.
Push the clock back and the ground softens. At Crater Lake in Oregon, the Klamath tell of a battle between Llao, a spirit of the underworld, and Skell, a spirit of the sky, that destroyed a mountain. A mountain was indeed destroyed there: Mount Mazama collapsed into the caldera that now holds the lake about 7,700 years ago. Archaeologists have found human sandals buried under the Mazama ash, so people were certainly in the region when it blew. That proves presence. It does not prove that the Llao and Skell story is an unbroken memory of the eruption rather than a later explanation for a strange and sacred landscape. The National Park Service is careful to say only that the traditions “indicate that people witnessed” the event. The line between witnessed memory and inherited explanation is thin, and it is the line this whole subject turns on.
In New Zealand, the first Polynesian settlers left their footprints in the fresh ash of Rangitoto, the island volcano in Auckland’s harbour that erupted around 1400 CE, roughly a century after they reached the country. Any memory of that eruption had to cross only about six centuries, which is no challenge to oral transmission at all.
The real frontier is Australia. In 2016 the geographer Patrick Nunn and the linguist Nicholas Reid published a study of Aboriginal stories from twenty-one places around the Australian coast, each describing a shoreline that now lies underwater. For every location they worked out how much sea-level rise the story required, then checked it against the known post-glacial curve. Their conclusion was that most of these stories most plausibly refer to coastlines drowned more than seven thousand years ago, which would mean transmission across hundreds of generations. They proposed a mechanism for it: a culture that treated knowledge as law, and a kinship system in which several lineages held the same story and corrected each other. That cross-checking, they argue, could hold a memory steady across millennia.
Read their own words and they are careful. The conclusions hold, they write, only “if these stories are accepted as authentic and based on observations of coastal flooding.” They concede that Australia may be “genuinely unique.” They know that folklorists usually reject any claim of oral survival beyond about a thousand years. The popular press was not careful. Headlines stretched the figure to ten thousand years, then thirteen thousand, which is a small lesson in how a number grows in the retelling, the very effect that makes the underlying claim so hard to test.
The case for doubt
Against all of this stands a disciplined skepticism, and it deserves a fair hearing rather than a quick dismissal. The center of gravity in folklore studies is that traditions about specific events last on the order of a thousand years, not ten thousand. The sharpest statement of the problem came from the historian David Henige, whose 2009 paper carries a title that is also its argument: “Impossible to Disprove Yet Impossible to Believe.”
Henige’s point is precise. A modern story cannot be cleanly separated from the accretions of the centuries it passed through. When a researcher matches a present-day story to an ancient event, there is no way to know whether the matching detail is genuine memory or a later addition read backward into the past. A deep-time oral claim can rarely be confirmed and rarely be refuted, and that, Henige argues, is a reason to keep it out of the evidence column, not a license to put it in.
There are three more reasons for caution, and Patrick Nunn himself supplies one of them. In a separate study of volcanoes, he found that of more than two thousand known eruptions, only about twenty have surviving stories, and about as many that should have been remembered were forgotten. Oral memory is selective, not total. The stories that survived and matched something get noticed; the thousands that vanished or never formed go uncounted. That is survivorship bias, and it drives the worst habit in geomythology: take any flood myth, find a flood somewhere near it, and announce a connection.
The honest position sits between the believers and the debunkers. The thousand-year ceiling is the safe default. A handful of cases, Cascadia firmest among them, reach further when an independent record pins them down. The Australian claim is a real and unresolved frontier, argued by serious scientists and not yet proven. None of that licenses the leap to an eight-thousand-year-old memory of a drowned North Sea. It only keeps the door from being slammed.
Geomythology, and its temptation
The field that studies all of this has a name. The geologist Dorothy Vitaliano coined “geomythology” in 1968 and developed it in her 1973 book Legends of the Earth: the study of traditions that record real geological events. It has a respectable ancestor in euhemerism, named for the Greek writer Euhemerus of the late fourth century BCE, who argued that the gods were once real kings, remembered and exaggerated into divinity. And it has an exact opposite in the theologian Rudolf Bultmann’s “demythologization,” coined in 1941, which strips the mythic husk away to reach the meaning underneath. Geomythology does the reverse. It digs into the husk looking for a real event at the core.
Greek myth offers a clean test case and a warning at the same time. The Thessalian version of the Deucalion flood has Poseidon split the mountains at the Vale of Tempe with his trident to drain the floodwaters off the plain. Thessaly really was an enclosed basin that once held a lake, drained by the Pineios river cutting the Tempe gorge between Mount Olympus and Mount Ossa. The fit is tidy. But the basin drained in the deep geological past, long before any human stood there, so the myth cannot be a memory of the draining. It is an explanation invented later for a dramatic landform, a story reverse-engineered from the gorge. Geomythologists call this etiological: myth as explanation, not myth as record.
The warning is Thera. The volcanic island of Santorini erupted around the sixteenth century BCE in one of the largest blasts of the Holocene. It buried the Minoan town of Akrotiri and laid down ash across the eastern Mediterranean. Here was a real, dated, enormous catastrophe sitting right inside the Greek cultural world. The archaeologist Spyridon Marinatos proposed that it was the original of Plato’s Atlantis, the island that sank in a day and a night. The idea is seductive and it is not accepted. Thera shows how a spectacular event lures the pattern-matching mind into a connection the evidence will not carry, which is worth remembering when Atlantis comes back into the story.
The road north, and the amber island
Suppose a memory could last. It would still have to travel. Was there any road from the North Sea down to Greece? There was, and it is not mystical. It is trade. Baltic amber moved south to the Mediterranean from at least the sixteenth century BCE. Amber beads turn up in the shaft graves of Mycenae, and a bead believed to be Baltic amber sits in a pectoral ornament of Tutankhamun, who died around 1323 BCE. Infrared analysis can fingerprint the resin, and most of the amber found around the Mediterranean carries the chemical signature of the southern Baltic coast near the Bay of Gdańsk. The route was so well established that under Nero, Pliny tells us, the official who ran the emperor’s games sent a Roman knight from Carnuntum on the Danube to the Baltic shore, and he came back with enough amber to decorate the nets of an arena.
Around 325 BCE a Greek from Massalia, the city we call Marseille, sailed north along that channel. His name was Pytheas, and his book On the Ocean is lost, surviving only in fragments quoted by later writers who often quoted him in order to call him a liar. Strabo named him an “arch-falsifier.” Modern scholars are kinder, because Pytheas got a great deal right: he was the first Greek to describe Britain, the frozen sea to its north, and the coasts where amber washed ashore.
In those same fragments sits an island. Pytheas described amber gathering on the shore of an island a day’s sail off the coast of a Germanic people. Pliny reports that the historian Timaeus of Tauromenium, a Sicilian Greek of the third century BCE and no relation to Plato’s dialogue of the same name, called that island Basilia, the “royal island.” Other writers gave it other names, Baltia and Abalus among them. Where it was, nobody can say for certain. The guesses run from Heligoland to Gotland to the amber coast of the Baltic. But the tradition is genuine: a real Greek source preserved the name of a northern island where treasure rose out of the sea.
Basilia equals Atlantis: one scholar’s wager
The amber road gives Borghi his channel, and the amber island gives him a name to work on. In his 2022 paper in the Italian onomastics journal Rivista Italiana di Onomastica, titled “Basilia (Dogger Island) = Atlantide (Doggerland),” he argues that the amber island Basilia and Plato’s Atlantis alike preserve, through Celtic etymology, the memory of the drowning. His training in Indo-European and Celtic onomastics sets the argument apart from the cranks who usually claim a northern Atlantis. It is a philological case, made on the evidence of words.
Start with the name itself. To a Greek, Basilia meant the royal island. Borghi argues it carries a Celtic reading underneath, Bāssalēdā, the plain of death by drowning. On that reading the very word the Greeks wrote down for the amber island already described a flat country lost beneath the sea, and he ties it to the waves that did the drowning, the Storegga slide and an earlier surge around 3500 BCE.
He has a second move, and it is the one most readers skip in Plato. In the Critias, Plato says the names in the Atlantis story are translations, Greek renderings of names first coined in another tongue. Plato then lists the ten sons of Poseidon who ruled Atlantis. Borghi reads those ten names as Celtic ethnonyms and place-names, with the firmest anchor being the pairing of the son Gadeiros with Gádeira, the ancient name of Cádiz. If the ten names decode as a roster of real northern peoples, he argues, the coincidence is too large to be chance.
The argument is a real one, and its standing has to be stated exactly. The wider field has not engaged with it. A search of the scholarly citation record turns up no responses and no reviews, and no other researcher has built on it or tried to knock it down. Unreviewed is a long way from vindicated, and a long way from refuted. One credentialed linguist published the idea in a real journal, and the discipline has so far walked past it. That is the true state of the claim, and anyone who tells you a scholar has proven Atlantis was Doggerland is overstating a paper that the experts have not yet read closely enough to argue with.
Borghi himself does not overplay it. He lays the find out as two possibilities and lets them stand. Either Plato, an Athenian working from the materials of his own day, drew a picture that happens to match a real drowned country, or a memory of that country reached him by some route we can no longer trace. That is the fork the whole question stands at.
Borghi follows the drowned land up into later legend. He reads Atlas, the first king of Atlantis, as the Celtic *Abalū, the root behind Avalon, the island where the dying King Arthur is carried away over the water. The Irish kept the Land under the Waves, Tír fo Thuinn, and the historian Procopius wrote of the dead being ferried across to the island of Brittia. The same lost country keeps surfacing in the stories of the people who lived around the sea that took it. Tolkien reached for the identical idea when he drowned Númenor; see Tolkien’s real sources.
Where the real cranks live
It is worth marking the boundary clearly, because the moment you put “Atlantis” and “North Sea” in the same sentence you attract a long history of nonsense. In the 1950s a German pastor named Jürgen Spanuth placed Atlantis off Heligoland and tied it to a glorified Nordic Bronze Age. Professional archaeologists rejected it immediately, and historians have since traced its roots to the völkisch, racially charged “Ario-Atlantist” fantasies of the early twentieth century. More recently a nuclear engineer named Felice Vinci has argued that Homer’s Iliad and Odyssey are really set in the Baltic. Mainstream classicists do not accept it, for the plain reason that there is no archaeological thread connecting Mycenaean Greece to the Baltic and the chronology does not work.
These are the company a careless northern-Atlantis claim keeps, and the distance between them and Borghi is the distance between a fringe ideology and a falsifiable linguistic argument the field has ignored. Naming the cranks draws the fence that lets the serious question stand on its own ground.
What stays, what drowns
The solid part of this is the country itself. Doggerland was real, it was inhabited, and it is gone, and the trawlers still bring its bones to the surface off Norfolk. The world is full of such places. Helike sank off the Greek coast in 373 BCE. Pavlopetri lies under the water off Laconia. The Australian shoreline drowned slowly enough that, just possibly, someone watched it go and told their grandchildren, who told theirs.
What drowns with the land is the certainty about memory. The Storegga wave can be dated to the year. Whether anyone carried it forward in a story, and whether that story ever reached a coastline where Greeks were listening, is past recovering. Our own Atlantis article treated Doggerland as a parallel to the story and ruled it out as a source. Borghi makes the opposite case, and the answer he proposes sits in a journal the rest of the field has not opened. Doggerland is the better mystery than Atlantis precisely because it is not a fable. It is a real drowned world, and the only thing we have lost completely is the knowledge of whether anyone remembered losing it.
Sources
Bibliography. The same list is held in the article’s frontmatter for the citation tools that read it programmatically.
- Coles, Bryony. ‘Doggerland: a Speculative Survey.’ Proceedings of the Prehistoric Society 64 (1998): 45-81
- Walker, James, et al. ‘A great wave: the Storegga tsunami and the end of Doggerland?’ Antiquity 94, no. 378 (2020): 1409-1425
- Bondevik, Stein, et al. ‘Record-breaking height for 8000-year-old tsunami in the North Atlantic.’ Eos 84, no. 31 (2003): 289-293
- Nyland, Astrid J., James Walker, and Graeme Warren. ‘Evidence of the Storegga Tsunami 8200 BP? An Archaeological Review of Impact After a Large-Scale Marine Event in Mesolithic Northern Europe.’ Frontiers in Earth Science 9 (2021): article 767460
- Vitaliano, Dorothy B. Legends of the Earth: Their Geologic Origins. Bloomington: Indiana University Press, 1973
- Nunn, Patrick D., and Nicholas J. Reid. ‘Aboriginal Memories of Inundation of the Australian Coast Dating from More than 7000 Years Ago.’ Australian Geographer 47, no. 1 (2016): 11-47
- Nunn, Patrick D., et al. ‘The Selective Geography of Volcanism in Oral Traditions.’ Quaternary Research (2024)
- Nunn, Patrick D. The Edge of Memory: Ancient Stories, Oral Tradition and the Post-Glacial World. London: Bloomsbury Sigma, 2018
- Henige, David. ‘Impossible to Disprove Yet Impossible to Believe: The Unforgiving Epistemology of Deep-Time Oral Tradition.’ History in Africa 36 (2009): 127-234
- Satake, Kenji, et al. ‘Time and Size of a Giant Earthquake in Cascadia Inferred from Japanese Tsunami Records of January 1700.’ Nature 379 (1996): 246-249
- Plinius Secundus, Gaius (Pliny the Elder). Naturalis Historia, Books IV and XXXVII (1st century CE)
- Borghi, Guido. ‘Basilia (Dogger Island) = Atlantide (Doggerland).’ Rivista Italiana di Onomastica XXVIII (2022)
- Borghi, Guido. ‘Onomastica (e Indoeuropeistica) per la Scuola.’ Rivista Italiana di Onomastica XXIX, no. 2 (2023)
