Introduction

In the winter of 1986, a diving instructor named Kihachiro Aratake dropped over the side of a boat off the southern tip of Yonaguni, the westernmost island in Japan, hunting for a reliable place to show tourists hammerhead sharks. Aratake, a director of the Yonaguni-Cho Tourism Association, knew these waters. Below Arakawabana cliff he found a wall of rock with flat faces, sharp corners, and a run of ledges that looked, to a swimmer suspended above them, like a staircase cut into the seabed. He surfaced with a story that would travel far past Okinawa.

Almost four decades later that story is still moving. In April 2024 the formation Aratake found became the flashpoint of a marathon podcast argument, and the phrase “Japan’s Atlantis” started trending again. The question people keep asking is simple. Did human hands build the Yonaguni Monument, or did the Earth?

No geologist who has published on the site apart from Kimura has argued that it is artificial. Every published assessment I can find reaches the same conclusion: the Yonaguni Monument is natural. What takes longer is why ordinary sandstone fools the eye so completely, why the man-made claim refuses to die, and what the sea itself records about when this rock last stood in open air.

What is the Yonaguni Monument?

The Yonaguni Monument (Japanese: 与那国島海底地形, Yonaguni-jima Kaitei Chikei, “Yonaguni Island Seafloor Topography”) is a submerged block of layered rock lying just off Arakawabana, the southern cape of Yonaguni Island. Yonaguni is the last scrap of Japan before Taiwan; about 108 kilometres separate the island from Taiwan at their closest points, close enough that Taiwan’s mountains are visible from Cape Irizaki on a clear day. The formation sits close to shore in shallow water, which is part of why it draws divers and part of why it became famous.

Reported dimensions vary from source to source, and that variation matters. Popular accounts describe the monument as roughly 100 meters long, about 60 meters wide, and around 25 meters high. Robert Schoch, a Yale-trained geologist who taught natural sciences at Boston University’s College of General Studies and who visited the site on two trips, in September 1997 and the summer of 1998, making several dives each time, described the main structure more conservatively as over 50 meters long east to west and over 20 meters wide north to south, with the top lying about 5 meters below the surface and the base at approximately 25 meters. Masaaki Kimura, the University of the Ryukyus marine geologist who champions the man-made view, gave larger figures still. His 2004 conference paper for Oceans ’04 MTS/IEEE Techno-Ocean, which is freely readable on his own site, puts the main monument at about 270 m long, 120 m wide and 26 m high, with the base at about 25 m, the upper terrace at 5 m and a metre of rock breaking the surface. The widely repeated figure of 290 meters does not appear in that paper; it is a downstream error that has been copied from blog to blog for two decades. The same paper puts the whole surveyed area at roughly one kilometre square with depths from 0 to 35 m, and it is not the same measurement as the 300 by 150 metres he gave National Geographic three years later, which covers the extent of the ruins he claims to identify rather than the area surveyed. Kimura told National Geographic that in total the ruins he identifies, ten structures off Yonaguni and five more off the main island of Okinawa, cover some 300 by 150 meters. No peer-reviewed metrology study of the underwater formation exists. These figures come from different instruments and different definitions of what is being measured, Schoch’s from hand-held reconnaissance dives, Kimura’s from SEABAT multibeam sonar, an underwater robot and SCUBA surveys, and they disagree with each other by a factor of five. Treat any single figure as approximate.

What everyone agrees on is the rock. The monument is carved from the local bedrock rather than assembled from separate blocks. The stone is sandstone and mudstone. The layers are almost horizontal. The block is cut by straight fractures. Those three facts explain nearly everything a diver sees.

Who discovered the Yonaguni Monument, and when?

Kihachiro Aratake is credited with the discovery in 1986. He was scouting dive sites for hammerhead viewing when he came across the terraced rock below Arakawabana cliff and recognized that it would draw visitors. Some sources give the date as 1987, and a few describe the discovery as a group of divers rather than a single person, so you will see both versions in circulation. Schoch’s 1999 acknowledgments settle the second question if not the first: he names Aratake as the man who discovered the monument, and as the guide for his own 1997 dives. The year itself stays a minor open point.

Word spread, though not immediately. Kimura, then a professor in the Department of Physics and Earth Sciences at the University of the Ryukyus, began his own study of the site in 1992, six years after Aratake surfaced with the story, though a team from his university had visited the formations not long after the discovery. In 1997 the Japanese businessman Yasuo Watanabe funded an informal expedition that put Schoch, the author Graham Hancock, the photographer Santha Faiia and the author John Anthony West in the water together, with Aratake himself as their dive guide. Schoch returned in the summer of 1998 with the Team Atlantis expedition and its documentary crew. That expedition matters, because it is where the two poles of the modern argument first met face to face over the same rock.

Is the Yonaguni Monument man-made or natural?

Almost everyone’s first reaction to the photographs is that someone built this. That reaction deserves a fair hearing before it gets taken apart.

Supporters of an artificial origin point to the concentration of unusual features in a small area: flat platforms, parallel faces, near right-angle corners, what look like steps rising from one terrace to another, straight channels resembling roads or drainage, two paired holes and a line of smaller holes that resemble the wedge sockets ancient quarrymen used to split stone, and a rock that from certain angles reads as a carved face. Kimura has argued over the years that he can identify a pyramid, castles, roads, monuments, and a stadium within the site. He has also pointed to markings he reads as resembling Kaida script, a set of pictographs once used in the Yaeyama Islands. Hancock, arguing the same corner, has stressed that while single features like these appear in natural sandstone worldwide, finding so many together in one place seems unlikely, and that the flat surfaces show a relative absence of the loose rubble you would expect from pure erosion and collapse.

The counter-case is geological, and it is the one that has carried the scientific community. The rock at Yonaguni is a sedimentary sandstone with strong internal structure: flat bedding planes stacked like the pages of a book, cut by two or more sets of steep, roughly perpendicular joints. Sandstone with that architecture, in a region shaken by frequent earthquakes, breaks into blocky, angular shapes on its own. Waves and strong tidal currents then scour away the loosened material along the fractures, leaving behind clean faces and flat treads. The result can look startlingly built, without anyone having built it. Crucially, the same step-like and terraced forms appear on the dry coast of Yonaguni Island and in the shallows nearby, where no lost civilization is invoked to explain them.

The man-made interpretation rests on pattern recognition and on the argument from improbability. It does offer physical evidence: quarry and wedge marks, characters said to be etched on carved faces, rocks read as animal figures, and the objects Kimura reports recovering, two chipped stone tools, an inscribed shale palette, a carved cobble. None of it has been published in a peer-reviewed archaeological venue or re-examined in print by independent specialists. And the one thing that would settle the question, cut blocks transported and set into place, has never been found by anyone.

What rock is the Yonaguni Monument made of?

The monument is built from the Lower Miocene Yaeyama Group, a sequence of medium to very fine sandstones and mudstones deposited roughly 20 million years ago. This is not a contested point; it comes from the geologist most often quoted on the natural-formation side, and it comes from rock samples, not from a distance.

Schoch collected samples during his dives and scraped away the coating of algae, coral, and sponge that makes the surfaces look smoother and more regular than they are. His description, published in 1999, remains the most-cited passage on the rock: the Yonaguni Monument “is composed predominantly of medium to very fine sandstones and mudstones of the Lower Miocene Yaeyama Group (the rocks themselves were deposited about 20 million years ago). These rocks contain numerous well-defined, parallel bedding planes along which the layers easily separate. The rocks of this group are also criss-crossed by numerous sets of parallel and vertical (relative to the horizontal bedding planes of the rocks) joints and fractures. Yonaguni lies in an earthquake-prone region; such earthquakes tend to fracture the rocks in a regular manner.”

Schoch is not a neutral outside referee here, and it is worth saying so plainly. He is best known for the Sphinx water-erosion hypothesis, an argument for a far older Great Sphinx that Egyptology rejects, and his popular books sit well outside the mainstream. He also reached Yonaguni in 1997 on a trip Graham Hancock helped arrange, and dived it with Hancock. None of that makes his rock description wrong: sandstone with parallel bedding planes and crossing joint sets is a routine field observation, and it is corroborated by everyone who has looked. But the argument here does not rest on Schoch, and should not be read as resting on him. It rests on the rock, on Ogata and colleagues’ onshore mapping, on Nunn, and on the sea-level curve.

Schoch’s essay, “An Enigmatic Ancient Underwater Structure off the Coast of Yonaguni Island, Japan,” is the original English draft of a piece that ran in Spanish in the popular-mystery magazine Más Allá de la Ciencia in May 1999. The English text has lived since then on the Circular Times site, and Schoch developed the same material in Voices of the Rocks, written with Robert Aquinas McNally and published by Harmony that year. It has never been peer reviewed, and yet it is the origin of the passage that thousands of blogs, videos and encyclopaedia entries have since repeated. So it deserves to be quoted accurately, but its status should be stated: this is a trained geologist’s field reconnaissance, self-published, not a study that passed through journal review.

The Yaeyama Group itself is well documented in the geological literature. Its sandstones crop out across Yonaguni and neighboring Iriomote Island, forming the low hills of the islands. Yonaguni sits at the far western end of the Ryukyu arc, along the subduction zone between the Eurasian and Philippine Sea plates, an active tectonic setting that keeps fracturing the rock.

How the sandstone fractures into “steps”

A geologist looking at this outcrop sees three things stacked on top of each other, and once you can see them too, the steps stop being mysterious.

Sandstone deposited in flat layers inherits a set of horizontal weaknesses, the bedding planes, where one depositional event met the next. When such rock is later stressed, buried, uplifted, squeezed by plate motion and jolted by earthquakes, it also develops joints, which are fractures where the rock has pulled apart without much sliding. Joints often form in sets that are roughly parallel to each other, and different sets tend to meet at high angles. Combine one set of horizontal bedding planes with two sets of steep, near-vertical joints and you have, in effect, a natural grid built into the stone.

Now remove material along that grid. Wave action, tidal currents, and the slow prying of weathering exploit the planes of weakness first. Blocks bounded by a bedding plane below and joints on the sides come loose and are carried off. What remains is a landscape of flat tops and abrupt steps, with straight walls between them, all oriented along the directions the fractures happened to run. The regularity that looks designed is simply the regularity of the fracture pattern expressed in the shape of the outcrop.

Yonaguni sits in one of the more seismically active parts of the planet, which sharpens the effect. Earthquakes fracture rock in a repeated, regular manner, refreshing and extending the joint sets over geological time. Strong currents then do the removal. The same combination produces angular, stepped coastal platforms in many parts of the world; the Ryukyu setting just supplies unusually favorable ingredients and then hides the result under clear water, where a diver’s sense of scale and a camera’s flattening of depth make the geometry look even more deliberate.

This is also why the “absence of rubble” argument does not carry as much weight as it first appears. In a high-energy underwater setting swept by strong currents, loosened blocks are not guaranteed to pile up conveniently at the base of a face. They can be transported away. Clean terraces are what a swept, high-energy site looks like.

What Masaaki Kimura claims, and what the evidence shows

Masaaki Kimura, born in 1940, is a professor emeritus of the University of the Ryukyus and a marine geologist and seismologist with a long publication record on the Okinawa Trough and on earthquake processes. He is the central figure on the man-made side, and he has studied the Yonaguni formation more intensively than anyone, over a mapping effort that National Geographic News described in 2007 as spanning more than 15 years of diving, measuring and mapping.

Kimura’s own paper is more candid than his public statements. It records that the monument’s beds dip 10 to 12 degrees to the southeast, that layered sandstone and mudstone of this kind separates readily along its bedding, and that the separated, fractured and slipped layers produce exactly the flat terraces and straight walls a diver sees. He then argues that specific features cannot be explained that way. But the baseline concession is his: the structure, in his own words, looks natural at first glance.

Kimura’s interpretation has shifted over time, and the shift is itself instructive. He first argued that the monument must be at least 10,000 years old, reasoning that it had to have been carved when the area stood above water, and he linked it to the mythical lost continent of Mu, a Pacific counterpart to Atlantis. Speaking to National Geographic News in 2007, he described the site in his own words: “The largest structure looks like a complicated, monolithic, stepped pyramid that rises from a depth of 25 meters,” and said he was convinced he was looking at the remains of a 5,000-year-old city. The conference he was speaking about was the 21st Pacific Science Congress, held in Okinawa that June, and the two surviving accounts of that one presentation do not agree with each other. Nat Geo reports a five-thousand-year-old city drowned by an earthquake around two thousand years ago. Secondary accounts have him dating the construction itself to two or three thousand years ago, on the grounds that sea level then was close to today’s, though the most-cited version of that claim traces not to a congress document but to a note on Kimura’s own website dated May 2006, a year before he spoke.Either way the mechanism he invokes for the drowning is tectonic subsidence rather than sea-level rise. That two readings of a single talk differ by three thousand years is a fair measure of how stable his chronology has been. He has variously connected the site to Mu and to the ancient Yamatai kingdom. Kimura also came to describe the monument as “terraformed”, a natural rock mass that, in his reading, was modified by human hands rather than built from scratch.

Kimura’s specific observations deserve an answer. He notes pairs of steps he finds hard to explain by erosion alone, the scarcity of fallen debris in certain channels, symmetrical holes that resemble post sockets, and markings he reads as carvings. The trouble is that each of these has a competing natural explanation that fits the local geology, and none of them has been backed by the kind of evidence archaeology treats as decisive. Schoch, working the same rock, reported finding depressions and cavities forming naturally at the surface on the island that look exactly like the supposed post holes, and he could not find tool marks or transported cut blocks that would substantiate artificiality. Kimura’s work on Yonaguni has been described in the scholarly and popular literature as pseudoarchaeological, and his related claims, his location for Yamatai, his earthquake-prediction method, are also regarded as controversial within his field.

It is fair to Kimura to note where he and his leading critic actually converge. Both accept that the underlying rock is natural. The disagreement is about how much, if any, of what a diver sees is the result of human work. On that question, the physical evidence for human work has not materialized in nearly forty years of attention.

Did humans modify the Yonaguni Monument? What Robert Schoch actually concluded

Robert Schoch is often filed as the man who proved Yonaguni is natural. His actual position is more careful.

Schoch concluded that the monument is “primarily of natural origin” and that “it should be considered a primarily natural structure until more evidence is found to the contrary.” He was explicit that this was a working hypothesis, not a closed case. And he left a door open that pop accounts often ignore. “We should also consider the possibility that the Yonaguni Monument is fundamentally a natural structure that was utilized, enhanced, and modified by humans in ancient times,” he wrote, floating the idea that the site might even have served as a quarry, worked along its natural joints, or as a foundation or dock for structures long gone.

That middle position, natural rock, possibly touched by people, is a reasonable one to hold, and it is a long way from “an ancient civilization built a pyramid city.” Nothing about a fisherman scratching a mark or an islander cutting a block along an existing fracture would make the monument a designed monument. The claim that actually matters, the one that generates the Atlantis headlines, is that the overall structure is artificial. That claim is the one the evidence does not support.

What the 2020 geological study of Yonaguni Island found

The most rigorous recent scientific work touching on the formation is a 2020 paper by Takayuki Ogata, Makoto Otsubo, and Hideyuki Itoh, “Geosites of Yonaguni Island Located at the Westernmost Point of Japan,” published in E-journal GEO, the journal of the Association of Japanese Geographers (volume 15, issue 1, pages 44–54; DOI 10.4157/ejgeo.15.44).

The study is not primarily about the underwater monument. It is a landform analysis of Yonaguni Island, built on a 10-meter digital elevation model and field observation at three onshore “geosites”: Tindabana, a cliff exposing the unconformity between the Yaeyama Group sandstone and the younger Ryukyu Group limestone; Kubura-furishi, a rocky coast of shore platform and cliffs in Yaeyama sandstone with pronounced tafoni (honeycomb weathering hollows); and Sanninu-dai, which displays normal faults and many joints. The English abstract lists weathering and erosion among its keywords, reports remarkable tafoni (honeycomb weathering hollows) in the Yaeyama sandstone at Kubura-furishi, and records that Sanninu-dai shows normal faults and many joints derived from horizontal stresses, closing on landform development in relation to back-arc tectonics.

A caveat about how this paper gets cited. The English abstract stops at the three onshore geosites and the tectonic setting; it makes no claim about the underwater monument, and it does not spell out the bedding-plus-joints mechanism that produces stepped forms. English-language accounts of the paper, Wikipedia’s among them, report that the Japanese body text does make both points: that the monument, however artificial it looks, is a natural feature produced by weathering and erosion acting on bedding and linear joints in the sandstone, and that comparable features occur at the Sanninu-dai geosite and along the island’s south coast. The full PDF is free to download from J-STAGE, and readers who want to check that for themselves should; I am reporting a secondhand summary of a Japanese-language text, and saying so. Readers who follow the DOI and find only the abstract should know why the paper’s conclusions are not visible there. What is directly checkable in English is that Yonaguni’s sandstone coast is dense with joints and actively weathering, which is the substrate the mechanism needs. Either way, the study’s contribution is to show that the same processes producing tidy steps on the dry island are ordinary, documented, and local.

One correction has to be made here, because the error has spread widely. Some blogs cite Toki and colleagues’ 2022 paper ‘Spring discharge mechanism along the southeast coast of Yonaguni Island in the southern Ryukyu forearc’ (Journal of Hydrology: Regional Studies 40, article 101051) as evidence about the jointing or origin of the monument. That paper is a study of groundwater and spring discharge on Yonaguni. It is not evidence about how the formation formed, and it should not be cited that way.

How old is the Yonaguni Monument? What sea level tells us

The strongest check on the whole debate is the water itself. If any part of the monument was shaped by human hands, that had to happen when the rock stood above the sea, because you do not quarry or carve a terrace 25 meters underwater with strong currents running. So the question becomes: when was the base of this formation last dry land?

The history of global sea level since the last ice age is one of the better-constrained records in Earth science, so the question is answerable. The benchmark study is Kurt Lambeck and colleagues, “Sea level and global ice volumes from the Last Glacial Maximum to the Holocene,” published in the Proceedings of the National Academy of Sciences in 2014 (volume 111, issue 43, pages 15296–15303; DOI 10.1073/pnas.1411762111). Drawing on about 1,000 sea-level observations from sites far from the former ice margins, chosen precisely to minimise local glacial-isostatic distortion, the authors reconstruct the rise and fall of the oceans over the past 35,000 years.

Their key numbers are these. At the Last Glacial Maximum, roughly 21,000 years ago, global mean sea level stood at its lowest. The paper’s abstract records “a slow fall to −134 m from 29 to 21 ka BP with a maximum grounded ice volume of ∼52 × 10⁶ km³ greater than today”, a lowstand of about 134 metres on their preferred solution, with a companion solution putting it nearer 140. Most of the great post-glacial rise then happened between about 16,500 and 8,200 years ago, when meltwater from collapsing ice sheets poured into the oceans. After 8,200 years ago the rate of rise decreased progressively until about 2,500 years ago, at which point ocean volumes became nearly constant and stayed that way until the renewed rise of the last century or so. Between 6,000 and 150 years ago, Lambeck’s team finds no evidence of oscillation exceeding 15 to 20 centimeters over any interval of 200 years or more.

Apply that curve to Yonaguni. If the base of the monument sits at about 25 meters depth, then reading off Lambeck’s global eustatic curve, the sea last stood 25 meters lower somewhere around 9,000 years ago. That reading is mine, not the paper’s: Lambeck and colleagues make no claim about Yonaguni. The number cuts both ways. Ten thousand years ago the ocean was closer to 40 meters below its present level, which means the entire formation, base, terraces and all, was standing in open air. The top, about 5 meters down today, went under far later, somewhere near 7,000 years ago. So the window in which a builder could have worked the base closes around 9,000 years BP, and the window in which anyone could have walked on the upper terraces closes roughly two millennia after that. Anything carved at the base would have to predate the flooding of the base, which pushes a construction hypothesis back to the early Holocene at the latest, into a period from which the southern Ryukyus have almost no archaeological record at all. Richard Pearson’s synthesis of Ryukyuan archaeology, Ancient Ryūkyū: An Archaeological Study of Island Communities (University of Hawai’i Press, 2013), describes late Pleistocene human remains in the Sakishima Islands followed by a long hiatus, with the first ceramic culture, the Shimotabaru phase, appearing only around 2900 BC, and even that start date is contested; Summerhayes and Anderson’s calibrated radiocarbon series, below, puts the phase later, at 4500 to 3900 cal BP. Glenn Summerhayes and Atholl Anderson’s survey of early occupation in the Yaeyamas (Bulletin of the Indo-Pacific Prehistory Association 29, 2009, pp. 76–91) adds that on Yonaguni itself the site of Toguru-hama is contemporaneous with Shimotabaru sites elsewhere in the group, though no pottery has been associated with it.

Two caveats. First, the Lambeck curve is a global eustatic reconstruction. Local relative sea level at any specific coast is modified by the region’s own vertical movement. Yonaguni sits on the tectonically active Ryukyu arc, where uplift and subsidence are ongoing, and by glacial isostatic adjustment, the slow flexing of the Earth’s crust and mantle as ice and water loads shift. Both can shift the exact local timing by centuries or more. Kimura’s own 2007 revision leans on exactly this kind of local reasoning to argue for later submergence by tectonic drop. Second, the depth figures themselves are approximate, as noted above. So treat “about 9,000 to 10,000 years ago” as an order-of-magnitude anchor from the global curve, not a precise date for Yonaguni specifically. What it rules out cleanly is the idea, sometimes floated, that a city was carved here and drowned within the last couple of thousand years by ordinary sea-level rise. The ocean had already finished most of its climb by then.

Is it connected to Atlantis or Mu?

“Japan’s Atlantis” originated as a headline, and it has been doing headline work ever since. The link to Atlantis and to its Pacific cousin, the supposed lost continent of Mu, comes from the writers and researchers who wanted the monument to be the ruin of a vanished high civilization, not from the geology.

Atlantis is a literary invention from Plato’s dialogues. Mu and Lemuria are nineteenth- and twentieth-century inventions layered on top of misread science and colonial-era speculation. There was never a sunken Pacific continent; plate tectonics leaves no room for one, and the ocean floor records no such lost landmass. Patrick Nunn, then professor of oceanic geoscience at the University of the South Pacific and author of the 2009 University of Hawai’i Press book Vanished Islands and Hidden Continents of the Pacific, treats these myths seriously as cultural history while dismantling them geologically, and classes the Yonaguni formations as natural. Writing later about the book on his own site, he described visiting the island to examine the cliffs beside the underwater monument, and reported that erosion along bedding planes in these rocks produces flat smooth surfaces and sharp risers, natural, he concluded, as is the monument itself.

The recurring pull toward Atlantis and Mu tells us more about the appeal of lost-civilization stories than about Yonaguni. It is the same instinct that turns a fossil elephant skull into a Cyclops, a theme explored in this look at how an elephant skull may have inspired the Cyclops of the Odyssey. A striking natural object, no framework to explain it, and a story rushes in.

The Joe Rogan debate: Graham Hancock vs Flint Dibble on Yonaguni

On 16 April 2024, the Joe Rogan Experience released episode #2136, a debate that ran for hours between Graham Hancock, the journalist and author behind the Netflix series Ancient Apocalypse, and Flint Dibble, an archaeologist at Cardiff University. Published on the PowerfulJRE YouTube channel, where it has drawn millions of views, the episode reached an audience larger by a factor of thousands than any journal paper on the subject. Hancock has spent decades arguing that an advanced Ice Age civilization existed and was erased by a cataclysm at the end of the last Ice Age, roughly 12,000 to 13,000 years ago, and Yonaguni is one of his showcase sites.

During the episode Hancock walked through his underwater footage, a tunnel his fins disappear into; terraces he reads as cut steps, pressing his case that human hands worked them. He told Rogan he was certain he was looking at a rock-hewn structure, a natural rock face, in his words, that people had cut and shaped. He told Rogan he and Santha Faiia had done more than two hundred dives at Yonaguni, and that they were not dilettantes, the conviction, in his framing, was hard-won. Dibble’s reply captured the mainstream response. Looking at the same images, he said, “even the photos you were showing at Yonaguni showed a lot of natural fractures along straight lines,” and added that it is easy to confuse what happens naturally and geologically with something that looks anthropogenic. He was equally unmoved by the other underwater sites Hancock screened alongside it.

The exchange did not resolve anything on air, and it was never going to. What it did was hand a decades-old geological curiosity to a vast new audience, most of whom had never heard of Aratake or the Yaeyama Group. The “Japan’s Atlantis” framing returned to circulation, and the same question landed in millions of feeds: is any of this real?

The debate establishes less than either side’s fans think. A long, watchable argument between a bestselling author and a working archaeologist settles nothing about the rock. The weight of the scientific case does not move because a podcast was compelling. It rests on the rock, the fracture patterns, the onshore analogues, and the sea-level record, none of which changed in April 2024.

Official recognition: what Japan says

A telling and often overlooked fact: no Japanese authority treats the Yonaguni Monument as an archaeological site. Neither Japan’s Agency for Cultural Affairs nor the government of Okinawa Prefecture recognizes the features as important cultural artifacts, and neither has carried out research or preservation work there. This was reported by Julian Ryall for National Geographic News on 19 September 2007, in an article that also laid out Kimura’s case in his own words. No subsequent designation or state-led survey has been reported in the two decades since.

Japan is meticulous about its cultural heritage, with one of the world’s most developed systems for designating and protecting sites. If the responsible agencies believed there were a genuine sunken city off Yonaguni, the site would be studied and protected as a matter of national importance. Its absence from those registers reflects the professional judgment that there is no archaeological site to protect, only a remarkable piece of natural geology.

Can you dive the Yonaguni Monument?

Yes, and people do, though it is not a beginner’s outing. The monument lies off the south coast of Yonaguni, a short boat ride from Kubura port, with its top only about 5 meters below the surface and its base near 25 meters. Because it faces the open ocean, the site is exposed to strong, fast currents, and the standard technique is drift diving with a negative entry: divers roll off the boat and descend head-first quickly before the current sweeps them off the site.

Operators typically expect advanced certification and real experience, with recommendations commonly around 100 logged dives, because conditions can be demanding. One dive blogger who visited recorded that she did not go deeper than about 16 meters because the currents near the steps were too strong to safely descend further. Visibility can be excellent, and summer months (roughly June through October) tend to offer the clearest water at the monument, while winter (broadly November through March) is the season for the region’s famous schooling hammerhead sharks at nearby sites. Reaching Yonaguni usually means a short flight from Ishigaki, or a longer, weather-dependent ferry.

How the Yonaguni case compares to other “too neat to be natural” features

Yonaguni belongs to a family of natural features that look manufactured and end up teaching the same lesson: geometry is not a signature of intent. Basalt cooling into hexagonal columns builds the Giant’s Causeway in Northern Ireland, a colonnade so regular that legend credited it to a giant. Sandstone and other layered rocks around the world weather into stairs, pavements, and box-like blocks along their joints. In each case the pattern is inherited from the rock’s internal structure rather than imposed on it.

The habit of reading design into geology runs deep, and reading it correctly is often what turns a mystery into science. When petroleum geoscientists Simon Stewart and Philip Allen spotted a set of concentric rings in North Sea survey data, the structure was strange enough to invite exotic explanations before careful work confirmed its origin, a story told in this piece on the Silverpit Crater and its confirmation as an asteroid impact. Even the interior of our planet holds features that sound impossible until the evidence is examined, as in this explainer on the ocean’s worth of water locked inside the mantle. Yonaguni is the same lesson, taught underwater and at a scale that makes it harder to see.

Weighing the evidence: the verdict

Set the two accounts side by side. One of them needs a great deal more to be true.

The natural account needs only what is already known to be present at Yonaguni: layered sandstone with strong bedding and joint sets, a highly active seismic setting that fractures rock in regular patterns, powerful currents that remove loosened material, and post-glacial flooding that put the whole thing underwater. Every headline feature, terraces, straight walls, steps, channels, the “post holes”, follows from those ingredients, and the toolkit’s output can be inspected on the dry parts of the same island.

The man-made account needs all of that geology to be present anyway, since even its advocates agree the rock is natural, and then it needs an additional actor: a society, present in the early Holocene, capable of monumental stone engineering, that left the monument but left no tools in context, no transported cut blocks, no unambiguous inscriptions, no settlement of matching scale, and no trace in the meticulous Japanese heritage record.

On the balance of evidence, the Yonaguni Monument is a natural sandstone formation, shaped by fracturing and erosion and then drowned by the rising sea. The most one can responsibly grant to the human-involvement idea is Schoch’s modest, unproven possibility that ancient people may at some point have used or lightly modified a natural feature. Even that remains unsupported by direct evidence. The lost-city, sunken-pyramid, Atlantis-of-the-Pacific reading is not supported at all.

Frequently asked questions

Is the Yonaguni Monument man-made?

The scientific consensus is that it is natural. The formation is carved from solid local bedrock, sandstone and mudstone of the Miocene Yaeyama Group, that fractures along flat bedding planes and steep joints and then erodes into stepped, angular shapes in a seismically active, current-swept setting. Marine geologist Masaaki Kimura argues it is an artificial or human-modified structure, but that view has been described as pseudoarchaeological and lacks the independently verified tool marks, transported cut blocks, and in-context artifacts that would confirm human construction.

How old is the Yonaguni Monument?

The rock itself is roughly 20 million years old, dating to the Early Miocene. That figure dates the deposition of the Yaeyama Group sandstone and says nothing about when anything was built. As for when the formation was last above water, the global sea-level curve of Lambeck et al. (2014) implies that a base at about 25 meters depth was last exposed somewhere around 9,000 to 10,000 years ago, though local Ryukyu-arc tectonics and glacial-isostatic adjustment can shift that timing. Kimura first proposed the site was at least 10,000 years old, then revised it downwards. Accounts of that revision disagree, some give a 5,000-year-old city, others construction 2,000 to 3,000 years ago, but all have tectonic subsidence rather than sea-level rise doing the drowning.

Who discovered the Yonaguni Monument and when?

Diving instructor Kihachiro Aratake, a director of the Yonaguni-Cho Tourism Association, is credited with the discovery in 1986 while looking for a good place to observe hammerhead sharks. A few sources give 1987 or describe a group of divers, so both the year and the sole-discoverer detail carry minor uncertainty.

Can you dive the Yonaguni Monument?

Yes. It lies off the south coast of Yonaguni Island, with its top about 5 meters down and its base near 25 meters. Strong currents make it an advanced drift dive, and operators generally expect experienced divers (often around 100 logged dives). Summer usually brings the best visibility at the monument; winter is hammerhead-shark season at nearby sites.

Is the Yonaguni Monument connected to Atlantis or Mu?

No. “Japan’s Atlantis” is a nickname, not a scientific claim. Atlantis is a literary invention of Plato’s, and Mu and Lemuria are debunked pseudo-historical continents. Plate tectonics and the ocean-floor record leave no room for a sunken Pacific continent, and oceanic geoscientist Patrick Nunn, who has examined the site, classifies the formation as natural.

What rock is the Yonaguni Monument made of?

Medium to very fine sandstones and mudstones of the Lower Miocene Yaeyama Group. The layers are nearly horizontal and are cut by parallel and vertical joints, which is exactly the structure that produces the formation’s step-like, geometric appearance.

Is the Yonaguni Monument real?

The formation is entirely real.

The bottom line

Kihachiro Aratake found something worth finding in 1986. The Yonaguni Monument is one of the more arresting rock formations on any coast, above or below water, and the pull people feel looking at those terraces is honest. The mistake is to read the geometry as a message. Layered sandstone in an earthquake country, cut by joints and scoured by currents, makes flat faces and clean steps on its own, and the same process is visible on the dry shore of the same island. The sea-level record explains when the rock went under, and Japan’s own cultural authorities decline to treat it as a ruin because there is no ruin to treat. The formation is natural, and you can walk up to the process that made it on the dry shore of the same island and put a hand on it. That is the more interesting fact, and it does not need a lost civilization to be worth the dive.