For half a century, a windswept site on the banks of a creek in southern Chile has sat at the centre of one of archaeology's most important questions: when did humans first reach the Americas? Monte Verde was the place that broke the old orthodoxy, pushing the first settlement thousands of years earlier than anyone had believed. Now, in 2026, the site is controversial again — and the row has become one of the most punishingly technical, and most consequential, arguments in modern archaeology.
In March this year, a team led by Todd Surovell of the University of Wyoming published a paper in the journal Science arguing that Monte Verde is thousands of years younger than claimed — perhaps as young as 8,200 to 4,200 years old, rather than the widely accepted 14,500. On 19 August, Tom Dillehay of Vanderbilt University, the archaeologist most associated with the site, fired back with an entire rebuttal paper in the journal PaleoAmerica. What follows is the rough outline of a dispute whose details could fill a book — and whose stakes concern the deepest history of two continents.
Why Monte Verde matters
The Americas were the last continents to be settled by humans. For most of the 20th century, the dominant story was the "Clovis First" model: that the first Americans were a people known as the Clovis, named after a site in New Mexico, who made distinctive pointed stone tools. The oldest Clovis sites are around 13,000 years old, and the consensus was so strong that any older claim was treated with intense scepticism.
Monte Verde was the first site to survive that scrutiny and become established. Dillehay and his colleagues have excavated there, on and off, since 1977, working two nearby locations dubbed Monte Verde I and Monte Verde II. At Monte Verde II they found artefacts of stone, wood and bone, plus what appear to be the foundations of a structure — perhaps a tent. Directly above the artefact-bearing layers lay a layer of peat, dated to 14,500 years ago, implying the settlement was at least that old. In 1997, a team of visiting archaeologists gave Dillehay's analyses an official stamp of approval, and subsequent evidence — including marine algae directly dated to between 14,220 and 13,980 years ago — appeared to support the dating.
Monte Verde I is murkier: its sparse traces were dated to between 14,500 and 18,500 years ago in a 2015 study, though even Dillehay does not put much store in the older end of that range. Today he says the evidence for the site's age rests on more than 1,000 radiocarbon and luminescence dates, intact chrono-stratigraphy, and diagnostic stone tools dated at other sites. Monte Verde effectively detonated the Clovis First model — and for two decades, the story it told was wildly controversial before becoming fairly settled. As of this year, it is controversial again.
The new challenge
Surovell first encountered Monte Verde as an undergraduate in 1994, when Dillehay came to give a guest lecture. "I was absolutely blown away," he recalls. "It really inspired my career." The irony is that the inspired student has become the establishment's most formidable critic.
Growing increasingly sceptical, Surovell and his colleagues visited the area in 2022 and 2023. Crucially, they did not excavate at Monte Verde I or II themselves, but along the banks of Chinchihuapi Creek — the small waterway the site sits on, a tributary of the larger Maullín river. From the sediment layers they found, they reconstructed a history of the whole area.
About 20,000 years ago, they argue, the region — deep in a glacial period — was the outwash of a glacier extending down from the Andes: a plain of braided streams, sand and gravel. As the climate warmed, a forest moved in, gave way to a wetland, and was buried, forming a black peat. Then, around 11,000 years ago, a volcano called Michinmahuida erupted, blanketing the area in a layer of ash known as the Lepué Tephra. Afterwards, the creek cut down through all these layers, forming an exposed surface where people lived.
The implication, Surovell's team argues, is that Monte Verde II must be younger than the ash layer. Dating sands and gravels at their excavation sites, they place the settlement at between 8,200 and 4,200 years old — roughly half the age claimed. The Ice Age wood found at the site, they contend, was eroded from older deposits and washed by streams onto these younger surfaces.
Dillehay's rebuttal
Dillehay's response has been blunt: Surovell's team did not excavate at the site, never saw the stratigraphy, and examined only the surface landscape from 50 to 4,000 metres away. On a flood plain, where rivers shift course over time and sediments can be mixed, extrapolating from one spot to another is hazardous.
His counter-arguments are specific. The peat layer Surovell's team identified, which they say corresponds to the peat overlying Monte Verde II's occupational layer, is not the same peat, Dillehay's team argues. The large quantity of wood at the site shows no sign of having been redeposited — none of the telltale damage from being bashed around in floodwater. And perhaps most decisively, there is no sign of the crucial Lepué Tephra ash layer at Monte Verde II at all, either above or below the artefact layer — so, Dillehay argues, it cannot tell us anything about the site's age.
The Science paper has now attracted seven "eLetters" responding to its various claims, and on 19 August Dillehay and colleagues published their full rebuttal in the journal PaleoAmerica. Dillehay also says an independent group of geologists and archaeologists has since visited the site, seen the stratigraphy, confirmed his team's work and identified major errors in Surovell's analysis — with their results to be published soon.
Why the argument is so hard to settle
Part of what makes this dispute so intractable is a problem philosophers of science call the "experimenter's regress". When a claim is disputed and the true answer is unknown, it becomes extremely difficult to judge whether an experiment — or an excavation — was performed correctly. If a result contradicts your favoured theory, you can simply argue the work was flawed; and your opponents can say exactly the same about yours.
That cycle of claim and counterclaim is now playing out in thousands of words of technical argument over peat layers, ash deposits, sediment transport and radiocarbon dates. As the New Scientist account of the dispute notes, every detail of both teams' excavations is being picked over, and an outsider is not well placed to adjudicate whether two peat layers 50 metres apart are or are not the same thing.
