The Silurian Hypothesis and Why a Techno-Sophisticated Predecessor to Humans is Unlikely
In 2018, astrophysicist Adam Frank and climatologist Gavin Schmidt published a paper in the International Journal of Astrobiology that formalised a provocative thought experiment: if an industrial civilisation had dominated Earth tens or hundreds of millions of years ago, would any trace of it remain? They called it the Silurian Hypothesis, a nod to the intelligent reptiles of Doctor Who. The authors were clear from the start that they strongly doubt any such civilisation existed. Their goal was not to argue for lost Silurians, but to define what evidence for one would look like, and, by extension, to sharpen our understanding of the Anthropocene and the search for technological life elsewhere.
The premise begins with Earth's restless geology. Plate tectonics, erosion, weathering, and the subduction of ocean crust constantly recycle the planet's surface. The oldest large-scale surviving land surface is only about 1.8 million years old (in the Negev). Ocean sediments do not preserve a continuous record older than the Jurassic, roughly 170 million years ago, because the seafloor is recycled. Exposed pre-Quaternary land is sparse, limited to outcrops, mines, and drill cores. In this environment, physical ruins: cities, roads, machinery, are vanishingly unlikely to survive intact or even as recognizable debris after tens of millions of years. Urbanisation today covers less than 1% of Earth's surface; the odds of any of it being preserved, re-exposed, and discovered after geological deep time are tiny.
Fossilisation is equally unforgiving. Dinosaurs roamed for roughly 180 million years, yet only a few thousand near-complete specimens have been found, roughly one good fossil every 10,000 years of their existence. A species as short-lived as Homo sapiens so far (a few hundred thousand years) might leave no fossil record at all if it vanished tomorrow. Schmidt and Frank estimate that for any civilisation older than about 4 million years, the chance of finding direct evidence in the form of artifacts or fossils of the builders themselves is small.
That leaves indirect "technosignatures": chemical and isotopic fingerprints left in sediments. An industrial society burning fossil carbon would produce a negative spike in carbon isotopes (δ¹³C), because fossil carbon is depleted in the heavier isotope. Synthetic fertilizers would perturb the nitrogen cycle. Plastics, persistent organic pollutants, unusual metal concentrations, or long-lived radioactive isotopes from nuclear activity could appear in the rock record. The Anthropocene layer forming now will be abrupt and multi-proxy: simultaneous spikes in several independent markers, plus faunal changes, appearing almost instantaneous on geological timescales (a few centimetres of sediment). Some markers, such as certain synthetic compounds or specific nuclear isotopes, would be unique; others would not.
When Schmidt and Frank compared this expected signature to the deep past, they found familiar patterns. The clearest analogues are hyperthermal events, most famously the Paleocene–Eocene Thermal Maximum (PETM) around 56 million years ago. Temperatures rose 5–7 °C, carbon isotopes dropped sharply, ocean chemistry shifted, and there were extinctions and faunal turnovers. Similar signals appear in smaller Eocene hyperthermals and in Cretaceous Oceanic Anoxic Events. In every case, however, the evidence points to natural drivers: large igneous provinces (such as the North Atlantic Igneous Province), thermogenic methane from magma interacting with organic-rich sediments, or other geological carbon releases. The PETM's carbon injection, while geologically rapid (perhaps a few thousand years for the main pulse), was slower than the modern industrial rate and is consistent with volcanic and sedimentary sources rather than industry. No unique synthetic compounds, anomalous nuclear ratios, or other unambiguous industrial markers have been reported in these intervals.
There is also a sustainability paradox. A civilisation that lasts long enough to leave a thick geological layer must become relatively sustainable; the more sustainable it is, the smaller and less distinctive its footprint becomes. Brief industrial bursts lasting only centuries would be nearly invisible in the compressed rock record. The continuous fossil record of life, from the rise of complex terrestrial ecosystems in the Devonian onward, shows no abrupt, unexplained interruptions that would require an industrial episode. Evolutionary trends proceed without the kind of ecological resets or novel chemical anomalies that a global industrial society would be expected to produce.
Geologist Jan Zalasiewicz has argued that Schmidt and Frank may underestimate the durability of modern industrial materials and the resolving power of the fossil record; plastics, fly ash, and certain metals could leave clearer traces than anticipated. Even so, the absence of such signatures in carefully studied sections remains telling. The authors themselves emphasize that the hypothesis is not supported merely by the lack of alternative explanations for past events. Occam's razor favors the natural mechanisms that already account for the isotopic, climatic, and biotic changes observed.
In short, the Silurian Hypothesis is a valuable framing device. It forces a precise inventory of what our own civilisation is writing into the rocks and clarifies what we should look for, or not expect to find, on other worlds. But the combination of geological erasure, the rarity of fossilisation, the natural explanations available for past hyperthermals, and the lack of unique industrial markers, makes a techno-sophisticated predecessor on Earth extraordinarily unlikely. The planet's memory is incomplete, yet what it does preserve has so far told a coherent story of life without industrial interruption.
