In 1986, a dive instructor named Kihachiro Aratake dropped into the water off the southern tip of Yonaguni Island — the westernmost inhabited speck of Japan, closer to Taiwan than to Okinawa — looking for hammerhead sharks. What he found instead was a massive terraced structure sitting on the seabed, its flat planes and right angles so regular they looked quarried. The formation rises 27 meters from the ocean floor, stretches roughly 150 meters long and 40 meters wide, and its uppermost surface hovers just five meters beneath the waves. Within a few years, it had acquired a name — the Yonaguni Monument — and a question that, four decades later, remains genuinely unresolved: did people build this, or did the Earth?

The Case for Human Hands

Masaaki Kimura, a marine geologist and professor emeritus at the University of the Ryukyus in Okinawa, has spent more time studying this structure than anyone alive. He first visited the site shortly after Aratake’s discovery and has since catalogued what he describes as ten distinct structures off Yonaguni and five more near the main island of Okinawa. His inventory reads like an archaeological survey: a stepped pyramid, castle ruins, a triumphal arch, temples, a stadium, roads, and carved monuments.

Kimura’s arguments are specific. He points to what he identifies as quarry marks — a straight row of small holes resembling the wedge-and-feather technique used in ancient stone splitting. He highlights two round holes, each about 61 centimeters wide, that he interprets as post supports. He claims to have found etched characters, including a plus sign and a V shape, and animal-shaped carvings. Most critically, he notes the relative absence of loose rubble on the flat terrace surfaces. If natural erosion produced these steps, where did the broken rock go?

His dating has shifted over the years. Initial estimates placed the structure at over 10,000 years old — a timeframe that aligns with when the site would have been above sea level, before postglacial flooding submerged it. A 2003 beryllium-10 survey led him to revise this to 2,000–3,000 years. He has also cited stalactite formations in nearby underwater caves as evidence of a 5,000-year-plus timeline. The inconsistency is a weakness in his case, and critics have noted it.

Toru Ouchi, an associate professor of seismology at Kobe University, has backed Kimura’s position. “I have never seen tectonic activity having such an effect on a landscape either above or below the water,” Ouchi stated, suggesting that natural geological processes alone cannot account for the Monument’s geometry.

The Case for Geology

Robert Schoch holds a Ph.D. in geology and geophysics from Yale and teaches natural sciences at Boston University. He has made dozens of dives at Yonaguni, starting with a 1997 expedition. His conclusion is blunt: the Monument is primarily a natural formation.

His reasoning is rooted in stratigraphy. The structure is composed of medium to very fine sandstones and mudstones of the Lower Miocene Yaeyama Group, rock deposited roughly 20 million years ago. These sandstones contain well-defined, parallel bedding planes — horizontal layers along which the stone naturally separates. The rock is also criss-crossed by sets of parallel, vertically oriented joints. When you combine horizontal bedding planes with vertical fracture systems and then subject them to wave action, tidal erosion, and tectonic stress in one of the most earthquake-prone regions on the planet, you get rectangular terraces and right angles without anyone picking up a chisel.

Schoch emphasizes one point above all others: the entire structure is solid, living bedrock. No part of it is constructed from separate placed blocks. Every terrace, every wall, every “step” is carved from continuous rock. In conventional archaeology, placed blocks — stones quarried, transported, and assembled — are the signature of construction. Their absence here is significant.

The alleged carvings? Natural scratches. The “roads”? Channels eroded along joints. The holes Kimura attributes to post supports? Created by underwater eddies exploiting weak points in the rock. The “walls”? Horizontal platforms that toppled into vertical positions as underlying rock eroded away.

And there’s a comparison that’s hard to dismiss. Similar formations exist above water on Yonaguni’s own south coast. The Sanninudai geosite, on dry land, displays the same terraced, angular morphology. If you saw these features on a hillside, you wouldn’t call them ruins. You’d call them geology.

The 2019 Study and Scientific Consensus

In 2019, Takayuki Ogata of the University of the Ryukyus and colleagues published what remains the most rigorous recent analysis of the site. Using digital elevation models and geological field investigations at three geosite locations — Tindabana, Kube Ryofurishi, and Sanninudai — they concluded that the Monument is a natural feature formed by weathering and erosional processes acting on bedding planes and linear joints in sandstone. The study noted that formations essentially identical to the underwater Monument are “commonly observed on the south coast of Yonaguni Island.”

This tracks with the broader scientific consensus. Neither the Japanese Agency for Cultural Affairs nor the Okinawa Prefectural government recognizes the Yonaguni Monument as a cultural artifact. Neither agency has conducted preservation work or designated it as a protected site. In the world of Japanese cultural heritage — a country that rigorously catalogues its archaeological record — this absence speaks volumes.

Patrick D. Nunn, a professor of oceanic geoscience at the University of the South Pacific, concurs with the natural-formation hypothesis. Wolf Wichmann, a German geologist who studied the site during expeditions in 1999 and 2001, was even more direct: “I didn’t find anything that was man-made.” It took him three dives to reach that conclusion.

The Uncomfortable Middle

Here’s where it gets complicated. Schoch himself doesn’t entirely close the door. While he maintains the structure is “primarily natural,” he acknowledges evidence of ancient human habitation on Yonaguni and allows that inhabitants may have “touched up” portions of the natural rock or drawn cultural inspiration from its geometry. This is a genuinely different position from saying it’s all just sandstone doing sandstone things.

And there’s a strange data point from the 1997 expedition. John Anthony West — the American alternative-history author best known for championing the Sphinx water erosion hypothesis — was on that trip alongside Schoch and Graham Hancock. West was, by disposition, inclined toward ancient mysteries. He had spent decades arguing that the Egyptian Sphinx was thousands of years older than mainstream Egyptology allowed. Yet after diving at Yonaguni, West agreed with Schoch that the formations were natural. He specifically stated that Kimura “had not looked carefully enough at the natural processes at work.” When even the alternative-history camp splits on your mystery, the picture is genuinely ambiguous.

Graham Hancock, for his part, has been more cautious than his reputation might suggest. After making over 100 dives at the site — documented in his 2002 book Underworld: The Mysterious Origins of Civilization and the accompanying Channel 4 series — he wrote that he “cannot totally discount the possibility that the Yonaguni Monument was at least partially worked and modified by the hands of humans.” Partially. That hedge matters. Hancock’s broader thesis — that rising seas at the end of the last ice age, between roughly 17,000 and 8,000 years ago, submerged coastal civilizations — doesn’t strictly require Yonaguni to be artificial. It just needs it to be possible.

What It Would Mean

Assume, for a moment, that some portion of the structure was deliberately shaped. What follows?

During the Last Glacial Maximum, sea levels were 110 to 130 meters lower than today. Yonaguni was connected to Taiwan by a land bridge. The site would have been dry land roughly 8,000 to 10,000 years ago — potentially earlier. If humans carved even part of this structure, they did so at a time when, according to conventional archaeology, no civilization in the region possessed monumental stone-working capability. The oldest known megalithic construction, Göbekli Tepe in southeastern Turkey, dates to approximately 9500 BCE. A worked-stone monument off Yonaguni from the same era would mean that the impulse to reshape landscapes at massive scale emerged independently, in widely separated populations, at roughly the same moment in human history.

That wouldn’t overturn archaeology. But it would force a serious rethinking of how and when complex social organization developed in East Asia and the Pacific. It would suggest that the hunter-gatherer-to-agriculture-to-monument pipeline isn’t the only path — that coastal peoples we know almost nothing about, whose settlements are now drowned under dozens of meters of ocean, may have been doing things we haven’t imagined.

The counterargument is straightforward: extraordinary claims require extraordinary evidence, and a natural geological explanation that accounts for every observed feature exists. The steps are too large to function as stairs. There are no placed blocks. Identical formations sit on dry land nearby. You don’t need humans to explain this.

What We Can’t See

The ocean is not kind to evidence. If coastal civilizations existed during the last ice age — and people certainly lived on coastlines; we know that much — their traces are now under water that is deep, dark, and largely unsurveyed. We have mapped more of the Moon’s surface than the shallow continental shelves where ice-age humans would have built their settlements. The absence of evidence is not evidence of absence, but it’s also not evidence of presence.

The Yonaguni Monument sits in this gap. Not proven natural, not proven artificial, not likely to be definitively resolved without technology or methodology that doesn’t yet exist. Kimura’s wedge marks could be tool traces or pressure fractures. Schoch’s joint-and-bedding explanation could account for 100% of the geometry, or 95% — and that remaining 5% might be everything.

So here’s the question that actually matters, the one that outlasts the Yonaguni debate itself: how much of the human past is sitting on drowned continental shelves, in places we haven’t looked, in forms we wouldn’t recognize — and what are we prepared to do about it?