Sundaland: The Continent That Really Did Drown

Harold Voris published reconstructions in the Journal of Biogeography in 2000 showing Sundaland's coastlines at 5,000-year intervals through the post-glacial rise. Our file adds two details worth keeping: the paleo-Sunda river system drained north into the South China Sea with a catchment rivalling the modern Amazon, and the flooding was not gradual. Meltwater pulses at about 14,600 and about 11,300 years ago drowned vast lowlands within centuries, possibly within decades.

Modern human remains from Niah Cave in Sarawak are dated to about 45,000 years ago, among the oldest Homo sapiens fossils in Southeast Asia, published by Graeme Barker and colleagues in the Journal of Human Evolution in 2007. Homo luzonensis remains from Callao Cave in the Philippines are dated to about 67,000 years, published in Nature in 2019. And Sundaland was the staging ground for the colonisation of Sahul at about 65,000 years ago, which required open-water crossings of at least 60 to 90 kilometres. Both of those identifiers, in a file rated one out of five for its sources, resolve exactly.

Vincent Macaulay, Catherine Hill, Alessandro Achilli and colleagues published in Science in 2005. Verbatim from the abstract: "We show that mitochondrial DNA variation in isolated 'relict' populations in southeast Asia supports the view that there was only a single dispersal from Africa, most likely via a southern coastal route, through India and onward into southeast Asia and Australasia. There was an early offshoot, leading ultimately to the settlement of the Near East and Europe, but the main dispersal from India to Australia ~65,000 years ago was rapid, most likely taking only a few thousand years." OUR FILE CITES THIS PAPER CORRECTLY AND GIVES NO IDENTIFIER FOR IT. One was found and is linked below.

Our file's section on deep antiquity gives the Tam Pa Ling remains in Laos as about 63,000 years old, citing Demeter and colleagues, 2012. The bibliography entry for that exact paper, five screens further down the same document, is titled "Anatomically Modern Human in Southeast Asia (Laos) by 46 ka". The registry confirms that title word for word. THE BODY SAYS 63,000 AND THE ROW BELOW IT SAYS 46,000. No outside source is needed to see it.

The Austronesian language family is the largest in the world by number of languages, about 1,257 of them, running from Madagascar to Easter Island and Taiwan to New Zealand, with the main expansion between roughly 5,500 and 3,000 years ago. The Out of Taiwan model has the languages originating in Taiwan about 5,500 years ago and spreading south. Gray and colleagues applied Bayesian phylogenetic methods to the languages in 2009 and got a Taiwan origin at about 5,200 years.

Its suggestion is that Austronesian languages may indeed have spread from Taiwan while being carried by populations with deep genetic roots in Sundaland and Southeast Asia: a linguistic expansion over a pre-existing genetic substrate. That is a more useful sentence than either camp's slogan, and it is our file's own rather than borrowed.

The bibliography entry for the Gunung Padang paper in our own research file carries the note: RETRACTED 2024, peer-review irregularities, methodological concerns. That checks out. The registry now lists the paper's own title beginning with the word RETRACTED, and a separate retraction notice exists in Archaeological Prospection, volume 31, page 199, 2024, recorded with update type retraction, sourced to Retraction Watch, dated 18 March 2024. NOT ONE OF THE FOURTEEN HIGHER-RATED RESEARCH FILES AUDITED IN THIS SERIES HAS FLAGGED A RETRACTION IN ITS OWN BIBLIOGRAPHY. This one, rated lowest of all of them, did.

Eileen Ernenwein published a reflection on the episode in ArchéoSciences in 2025. Its abstract is three lines long and is quoted here entire, because the shape of it is the point: "False claim of World's oldest pyramid published due to peer review oversights, Evolving scholarly publishing practices are negatively impacting academic integrity, Archaeological Prospection methods are increasingly exploited for fringe narratives." Our file does not cite it.

Our file refuses a list, and the refusals are good ones. That Sundaland was Atlantis or Mu, which confuses a real geological question with fiction. That it hosted a technologically advanced Pleistocene civilisation, when every known site shows hunter-gatherer or early horticultural economies. That satellite imagery shows pyramids or cities on the Sunda shelf, which are natural features or sonar artefacts. That the Yonaguni Monument is a Sundaland-era construction, when the geology shows natural sandstone jointing. AND IT NAMES THE ACTUAL HARM: alternative-history authors cite Sundaland to borrow scientific credibility, which distorts both the real evidence and Oppenheimer's own carefully argued case.

Voris on the sea-level maps, Barker on Niah Cave, Demeter on Tam Pa Ling and Détroit on Homo luzonensis all resolve exactly, matching author, title, journal, volume, issue, pages and year in every case. The fifth is attached to Oppenheimer's 1998 book and resolves to John Edward Terrell's two-page review of it in Current Anthropology in 1999. The tell was visible before anything was resolved: a University of Chicago Press identifier sitting on a Weidenfeld and Nicolson trade book.

The header of this file says Source Count: 0 and Source Confidence: 1 out of 5. THE BIBLIOGRAPHY HAS TWENTY ENTRIES. The footer of the same document says, in its own words, consolidated from 20 sources. The confidence rating is derived from a count that the document refutes two screens further down. This is not a weak signal being over-read. It is an instrument reading a broken input and producing a number that is not merely imprecise but backwards.

The last entry in this file's bibliography reads: Pelejero and colleagues, 1999, "High-Resolution U-Y_1_06 Temperature Reconstructions in the South China Sea", Paleoceanography 14(2):224-231. The real title is "High-resolution UK37 temperature reconstructions in the South China Sea over the past 220 kyr". UK37 is the alkenone unsaturation index, ordinary palaeothermometry notation with a superscript and a subscript. Y_1_06 IS A DOCUMENT IDENTIFIER FROM THIS LIBRARY. Something that walks the corpus rewriting cross-references has read a scientific notation as a link and written a filing number into the title of a citation. THE JOURNAL, VOLUME, PAGES AND YEAR ARE ALL UNTOUCHED AND CORRECT, which is precisely why the mechanism is visible instead of just the damage.

Fifteen entries in this bibliography carry no identifier at all, and five of them were found: Pelejero 1999 in Paleoceanography, Lambeck and colleagues 2014 on sea level and global ice volumes in PNAS, Clarkson and colleagues 2017 on human occupation of northern Australia by 65,000 years ago in Nature, Macaulay and colleagues 2005 in Science, and the Gunung Padang retraction notice itself. EVERY ONE OF THOSE CITATIONS WAS ALREADY CORRECT except for the missing pointer, and in Pelejero's case the corrupted title. The other ten were not searched individually and are recorded here as unchecked, which is not the same as clean and not the same as wrong.

Verbatim: "No significant counter-arguments exist in the scholarly literature for the core claims presented here... with no active scholarly dispute over the fundamental claims presented in this document." The same document calls Oppenheimer's thesis contested, describes a live dispute between two named research groups over Gunung Padang, and devotes a whole section to refusing four popular claims. THIS IS THE THIRD FILE IN THIS SERIES CARRYING THAT EXACT SENTENCE while the dispute sits a screen above it. It reads like boilerplate because it is.