Gunung Padang
Indonesia · written from the published paper and its retraction
The paper said 27,000 years. Five months later the journal withdrew it. What is actually under the hill is still a fair question.
- Latitude
- 6.9947° S
- Longitude
- 107.0561° E
- Site type
- punden berundak
- 2023 claim
- retracted 2024
- Open questions
- 4 logged
What is actually there
A hilltop in West Java, near the village of Karyamukti, covered in columnar basalt, the naturally hexagonal volcanic rock that fractures into long prisms. The stones are arranged into five terraces climbing the summit. Locals call this kind of structure a punden berundak, a stepped sacred terrace, and the form is known across Indonesia.
It is a real archaeological site, and nobody disputes that the surface terraces are human work. The disagreement is about how far down, and how far back, that work goes. Published estimates for the surface structure are in the first millennium CE: Lutfi Yondri put construction between the second and fifth centuries CE, Harry Truman Simanjuntak proposed the sixth to eighth, and pottery fragments from the site have been dated to about 45 BCE to 22 CE. All of that is within the last two thousand years.
The claim, and what happened to it
In October 2023 the journal Archaeological Prospection published a paper by Danny Hilman Natawidjaja and colleagues, based on ground-penetrating radar, electrical resistivity tomography, seismic tomography, core drilling and radiocarbon dating. It argued that buried structures beneath the visible terraces were a constructed pyramid, with the deepest phase dating to as much as 27,000 years ago.
If correct, that would be the oldest known monumental construction on earth by a very wide margin.
In March 2024 the paper was retracted.
The publisher and the editors retracted it after letters of concern from specialists. The notice says the radiocarbon dating was applied to soil samples that were not associated with any artefacts or features, and concludes that the pyramid interpretation is incorrect.
Why it was retracted, and the part worth understanding
The core problem was not the fieldwork. It was what the dates were attached to. Radiocarbon dating tells you the age of organic material, in this case soil from drill cores. It does not tell you that a person put that soil there. Old organic soil buried under a structure dates the soil, not the structure.
Other critics went further. Sutikno Bronto, a volcanologist, concluded that the hill is the neck of an ancient volcano and not a pyramid. Thirty-four Indonesian archaeologists and geologists signed a petition questioning the team’s motives and methods, and Víctor Pérez called the conclusions pseudoarchaeology. I have taken these from a secondary summary and have not read the originals.
What is still open
What is under the terraces?
The geophysics did detect subsurface structure. Retracting an interpretation does not delete the data. The options include natural columnar basalt, a volcanic neck as Bronto argues, earlier phases of ordinary age, or a mixture. What would settle it is excavation that establishes stratigraphic context, with dated artefact-bearing layers, not more remote sensing.
How old is the surface structure?
The published estimates run from about the first century BCE, from pottery, to the eighth century CE, from architecture and comparison, and none is close to 27,000. Even the uncontroversial part is not pinned down. What would help is dating from layers that contain worked material tied to the terrace construction.
How much did the damage outlast the correction?
The claim reached far more people than the retraction did, and articles are still circulating with the original figures. This is a structural problem with how extraordinary claims move, and it is not specific to this site.
Can dating ever settle a question like this?
The lesson generalises. Almost every “far older than we thought” claim turns on what exactly is being dated, and whether it can be tied to human activity. That question is worth asking every time, including when the answer is one you like.
The retraction is the right call on the evidence. But “retracted” has become a way of not thinking about the site at all. What if the interesting question was never the 27,000-year claim, and is simply this: why did people build a terraced sanctuary on a remote volcanic summit, out of stones that fracture into columns all by themselves?
Why it is worth asking: because the surface terraces are certainly human, and the choice of a hill that offered ready-made columns is a choice. What supports it: the published dates show people used the hill within the first millennium CE, and the form is a recognised Indonesian one. What speaks against it: if Bronto is right that the hill is a volcanic neck, then the columns are simply what the rock does, and the site would be less a puzzle than a use of local stone. Continuous use for thousands of years is not documented. What would tell the readings apart: stratigraphic excavation of the terraces. What stays unknown is what the terraces were for.
On Ancient Apocalypse
Gunung Padang opens the first episode of the Netflix series, presented as potential proof of a lost Ice Age civilisation. The series aired in 2022, before the paper it drew on was published and well before it was retracted. If you watched the show and came here looking for what happened next: this is what happened next.
Sources
- Natawidjaja, D.H. et al., “Geo-archaeological prospecting of Gunung Padang buried prehistoric pyramid in West Java, Indonesia”, Archaeological Prospection, 2023. The retracted paper.
- The formal retraction notice, Archaeological Prospection, March 2024.
- Secondary summary of the retraction, the critics (Bronto, Pérez, the 34-signature petition) and the conventional age estimates (Yondri, Simanjuntak, pottery dating), as collected in Wikipedia. I have not read the primary sources for those.
Written from the published paper, the retraction notice, and the peer commentary. Last reviewed 12 October 2026.