A stepped stone formation roughly 200 meters long and 20 to 25 meters high sits about 25 meters below the surface off Yonaguni Island, the westernmost point of Japan. For more than two decades, researchers have traded competing explanations for its flat terraces and straight edges, with one camp calling it a submerged prehistoric monument and another attributing it entirely to natural geology. Neither side has produced the kind of definitive evidence that would close the argument, and the absence of direct radiometric dating from the structure itself keeps the question alive.
Why the Yonaguni formation still divides geologists and archaeologists
The debate matters because its outcome would reshape what scholars know about human activity in the western Pacific during the last ice age. If the terraces are artificial, they would represent organized construction thousands of years before the earliest known monumental architecture in the region. If they are natural, they offer a case study in how tectonic jointing and wave erosion can mimic human design closely enough to fool trained observers.
Masaaki Kimura, a marine geologist formerly at the University of the Ryukyus, has been the most prominent advocate for the human-made interpretation. Drawing on sonar, underwater robot, and SCUBA surveys, Kimura has argued that artificial topographies spread across a claimed area of approximately 1 km by 1 km at depths from 0 to 35 m off Yonaguni. He describes features he interprets as stepped pyramids, roads, and canals, and he estimates the structures date to around 10,000 years ago, a period when post-glacial sea-level rise would have drowned coastal land across the Ryukyu Arc.
A separate line of research approaches the same island from the opposite direction. A peer-reviewed geomorphology study published in E-journal GEO by the Association of Japanese Geographers documents Yonaguni’s bedrock joints, faults, and erosion patterns using DEM analysis and field observations at multiple geosites. The paper catalogs natural mechanisms, including faulting, jointing, and weathering in Miocene-era sandstone and mudstone, that can produce flat surfaces and right-angle breaks without any human intervention. If terrace edges align more closely with regional fault trends than with known prehistoric quarrying angles, the formation is predominantly natural, and any human use would have been secondary at most.
Surveys and dimensions behind the Yonaguni claims
The primary technical description of the site comes from a 1998 paper published by the Nagoya University Dating and Materials Research Center. That study, which labels the main feature “No. 1 monument,” reports measured dimensions of roughly 200 m in length, 150 m in width, and 20 to 25 m in height at a depth of about 25 m. The authors noted flat terraces, straight walls, and what they described as wedge-like scars and stone-fence-like fragments, features they interpreted as signs of human construction rather than the product of random erosion.
Kimura expanded on those findings in a peer-reviewed paper in the Journal of Fossil Research, published by the Fossil Research Society of Japan. In that study, he placed the Yonaguni features within a broader narrative of Ryukyu Arc subsidence and post-glacial flooding, arguing that artificial monuments resembling stepped pyramids exist off the island. His estimated age of around 10,000 years ties the supposed construction to the transition from the Pleistocene to the Holocene, when sea levels rose rapidly enough to submerge large stretches of coastal terrain across East Asia.
The geological counter-argument does not require rejecting any of Kimura’s dimensional measurements. Yonaguni’s bedrock consists of well-bedded sedimentary layers prone to fracturing along predictable joint sets. When waves and currents erode rock that already contains regular fracture planes, the result can look strikingly architectural. Horizontal bedding produces flat platforms. Vertical joints produce straight walls. The combination can generate step-like profiles without a single chisel stroke.
Proponents of a natural origin also stress that similar-looking formations appear above water on Yonaguni and neighboring islands, where there is no suggestion of prehistoric quarrying. There, cliffs break into rectilinear blocks, and wave-cut notches create ledges that can resemble terraces when viewed from the sea. If the same rock types and structural controls extend offshore, they argue, the underwater “monument” may simply be a submerged continuation of the same geomorphological pattern.
Reading architecture into rock
For advocates of human construction, visual impressions still carry weight. Divers describe sharp-edged steps wide enough to walk on, apparent platforms that look like plazas, and isolated blocks that resemble carved megaliths. Some observers report what they interpret as tool marks on vertical faces and straight channels separating blocks that suggest deliberate cutting. The overall effect, especially in clear water and oblique light, can evoke a ruined ziggurat or ceremonial complex.
Geologists caution that human brains are primed to recognize patterns and structures, even where none exist. They point to examples from other coastlines where natural jointing produces right angles, including formations that have been mistaken for walls or paved roads. From this perspective, the Yonaguni terraces are an underwater Rorschach test: those expecting architecture see it, while those trained in structural geology see textbook fracture-controlled erosion.
Both sides agree that photographs and dive impressions are not enough. Camera angles, lens distortion, and selective framing can exaggerate straightness or hide irregularities. Accurate three-dimensional mapping, ideally combining high-resolution sonar with photogrammetry, would allow more rigorous tests: Are terrace risers consistent in height? Do edges follow a small number of orientations, as joints typically do, or do they form orthogonal grids suggestive of planning? At present, published data only partly answer these questions, leaving room for interpretation.
Missing data and what would settle the Yonaguni dispute
The most conspicuous gap in the published record is the absence of direct radiometric or stratigraphic dating from the monument itself. The 1998 Nagoya University paper discusses age estimates, but no publicly available study presents carbon-14 or other absolute dates obtained from material taken directly from the terraced surfaces. Without such dates, any age estimate for the structure relies on inference from regional sea-level curves rather than from the rock or any artifacts embedded in it.
A second gap involves independent geological fieldwork on the underwater formation. The published arguments for natural origins draw heavily on surface geology and DEM data from the island above water. Direct measurements of joint spacing, fracture orientation, and possible tool marks on the submerged terraces by geologists outside Kimura’s research group have not appeared in the indexed literature available for review. That kind of independent field comparison, matching the angles and spacing of underwater terrace edges against mapped fault and joint sets on land, would help determine how closely the “monument” follows natural structural controls.
Archaeological evidence is equally thin. Claims of carved symbols or arranged stone features at the site have not been backed by peer-reviewed documentation, detailed drawings, or recoverable artifacts. No ceramics, worked tools, or datable organic remains firmly associated with the terraces have been reported in scholarly venues. Without such finds, the case for a prehistoric culture capable of large-scale stone construction in this location remains speculative.
Researchers on both sides outline similar wish lists for future work. Systematic coring around the base of the formation could search for buried soils, plant remains, or cultural layers that might yield radiocarbon dates. Microscopic analysis of surface textures could distinguish between mechanical abrasion and intentional carving. Expanded sonar and photographic surveys, released in full rather than as selected images, would allow more independent scrutiny. Until those data exist, the Yonaguni formation will likely continue to occupy a gray zone between spectacular geology and possible archaeology.
For now, the site functions as a reminder of how much of the late Pleistocene landscape lies underwater and how difficult it is to reconstruct human activity along drowned coasts. Whether ultimately judged natural, modified, or fully artificial, Yonaguni forces geologists and archaeologists to confront the limits of inference from shape alone, and to acknowledge how the sea can both preserve and obscure the deep history written into stone.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
