We tend to imagine deep time as a kind of blank slate — aeons of bacteria, then fish, then dinosaurs, then us. The notion that we are the first technological civilisation is so embedded in how we frame human history that it rarely gets examined. But it is an assumption, not a finding. And the Silurian Hypothesis, a scientific paper published in 2018 by NASA climate scientist Gavin Schmidt and astrophysicist Adam Frank, asks us to examine that assumption with considerably more rigour than we usually bother to.

The stakes are not merely historical. How we answer this question has direct bearing on the Drake equation — the famous framework for estimating how likely intelligent civilisations are to arise in the universe. If we want to know how rare or common technological civilisations might be across the cosmos, we have to start by honestly auditing our knowledge of Earth. And our knowledge of Earth, it turns out, is far patchier than most people realise.

There is also a present urgency. We are living through what geologists now call the Anthropocene — the age in which human industrial activity has become a primary force shaping the planet's geology, chemistry, and climate. The study of what signatures our civilisation will leave in the rock record is not idle speculation. It is active science. And that science opens a door nobody expected: if we can describe our own geological fingerprint, we can ask whether any previous fingerprint of comparable character might have been missed, eroded, or misinterpreted.

The future dimension matters too. Any honest reckoning with the Silurian Hypothesis forces us to sit with the possibility — not the certainty, but the genuine possibility — that intelligence arising from biological life may be a more recurring feature of Earth's history than we assume. Whether that possibility is comforting or disquieting depends entirely on why previous civilisations, if they existed, are no longer here.

First, a clarification the popular press often skips: the Silurian Hypothesis is not a claim that a prior civilisation existed. It is a question about detectability. Schmidt and Frank are not arguing that lizard-people built cities during the Cretaceous. They are asking something more methodologically interesting: if they had, what would that look like in the geological record today — and would we recognise it?

The name is a gentle nod to the Silurians, a fictional reptilian civilisation from the British science fiction series Doctor Who, who supposedly ruled Earth before the rise of mammals. Schmidt and Frank borrowed the name with obvious affection for the conceit while making clear that their paper is serious science, published in the peer-reviewed International Journal of Astrobiology. The playfulness of the name should not obscure the sharpness of the question.

Their approach is rigorous and somewhat humbling. They begin by asking: what geological signatures is our industrial civilisation likely to leave behind? They then examine whether those signatures would be distinguishable, millions of years from now, from naturally occurring geological events. And then they flip the question: are there anomalies in the existing geological record that could, in principle, match those signatures?

This is not fringe science. Gavin Schmidt directs NASA's Goddard Institute for Space Studies. Adam Frank is a professor of astrophysics at the University of Rochester. The paper is dry, careful, and measured — which makes its implications more remarkable, not less.

To understand what we might be looking for, we first need to understand what we are leaving behind. The Anthropocene — however one dates its beginning — is producing a suite of geological signals that future geologists or, hypothetically, alien archaeologists would potentially identify as anomalous.

Consider the chemical record. Industrial combustion of fossil fuels releases carbon that has been sequestered underground for hundreds of millions of years. This produces measurable shifts in the ratio of carbon isotopes, specifically a decrease in the proportion of carbon-13 relative to carbon-12, a signal geologists call a carbon isotope excursion. Industrial agriculture and fertiliser use are dramatically altering the global nitrogen cycle, leaving signatures in sediment layers. Plastics, synthetic molecules, and rare industrial metals — including technomarkers like plutonium isotopes from nuclear testing — are being deposited in sedimentary layers right now.

There is also the biological record. The Anthropocene is associated with a significant mass extinction event, as human activity eliminates species at a rate estimated to be hundreds to thousands of times the natural background extinction rate. This would show up as an abrupt faunal turnover in the fossil record — a sharp line between one assemblage of species and another, with many of the previous species simply gone.

Temperature and sea level changes driven by greenhouse gas emissions would also be recorded in isotope ratios preserved in ice cores, deep-sea sediments, and certain rock formations — though ice cores themselves would not survive over geological timescales of millions of years.

Schmidt and Frank conclude that our civilisation would leave a detectable, if not always immediately interpretable, signature. But here is the critical catch: they also note that this signature would not necessarily look unique. It would look, in many respects, like other known anomalies in the geological record.

This is where the paper's most unsettling implications emerge. The geological record is not an archive. It is a ruin. Of everything that has ever existed on Earth's surface, only a tiny fraction survives in the rock record — and of that fraction, only a further tiny fraction has been found, examined, and correctly interpreted.

Diagenesis — the physical and chemical transformation of sediments as they are buried, pressurised, and heated — destroys enormous quantities of chemical and biological information. Tectonic activity recycles the ocean floor every 200 million years or so through subduction, feeding old seafloor rock back into the mantle. Erosion strips continental surfaces. Even the most chemically robust signatures become diluted, altered, or simply lost across timescales of tens of millions of years.

Consider what this means practically. If an industrial civilisation existed 100 million years ago — during the Cretaceous period, while non-avian dinosaurs still roamed — the sedimentary record from that precise time has been subjected to a hundred million years of compression, heating, chemical alteration, and tectonic movement. Surface structures would be gone entirely. Plastic, if used, would have broken down. Most organic molecules would have degraded. What might remain is a thin chemical anomaly in a layer of rock: a subtle shift in isotope ratios, perhaps an unusual spike in certain metals, perhaps an abrupt biological turnover that looks like one of Earth's periodic extinction pulses.

In other words, the signal would look a great deal like noise. And here is the honest admission at the heart of the Silurian Hypothesis: we might not be able to tell the difference.

Schmidt and Frank do not simply describe the problem in the abstract. They point to real events in Earth's geological history that exhibit some of the characteristics we would expect from an industrial civilisation's footprint — while being careful to emphasise that natural explanations exist and are, in most cases, the scientific consensus.

The most striking example is the Palaeocene-Eocene Thermal Maximum, or PETM, which occurred approximately 56 million years ago. Over a geologically brief period — estimates range from a few thousand to around twenty thousand years — global temperatures rose by 5 to 8 degrees Celsius. Ocean chemistry changed dramatically. There was a significant carbon isotope excursion, indicating a massive, rapid injection of isotopically light carbon into the atmosphere. Many deep-sea organisms went extinct. Then, over the following roughly 200,000 years, the climate recovered.

The natural explanations for the PETM include massive volcanic activity, the destabilisation of frozen methane deposits on the seafloor (known as methane clathrates), or some combination of these. These explanations are plausible and most earth scientists accept them. But Schmidt and Frank note, carefully, that the source and mechanism of the PETM's carbon injection remains genuinely debated in the scientific literature. The event's speed, in geological terms, is remarkable. And it does happen to fall in a window — 56 million years ago — where, if any prior intelligent species existed among mammals' early ancestors, some geological evidence might still be preserved.

They are not claiming the PETM was caused by an industrial civilisation. They are asking: how would we know if it were? And they conclude that, given the limitations of the geological record at that age, we probably could not determine this with certainty. That is an honest and important admission.

Other anomalies they and subsequent researchers have gestured toward include several abrupt carbon isotope excursions during the Mesozoic and early Cenozoic, as well as certain rapid faunal turnovers that do not map cleanly onto known impact events or the largest volcanic episodes. None of these are strong evidence of prior civilisations. All of them are reminders that Earth's history contains genuine mysteries.

Here the hypothesis requires a different kind of thinking — biological rather than geological. For a prior industrial civilisation to have existed, there must have been a lineage capable of developing sufficient intelligence and dexterity to build one. And this, for most of human history, is where speculation would have ended: humans are obviously the first, because what else could it have been?

But Earth's evolutionary history is longer and stranger than intuition suggests. Non-avian dinosaurs dominated the planet for roughly 165 million years — a span of time so vast that our entire evolutionary history from the first Homo species to today would fit into it fifty times over, with room to spare. Among dinosaurs, the troodontids — a family that includes the famous Troodon — are notable for their relatively large brain-to-body ratios and, in some species, evidence of stereoscopic vision and grasping forelimbs. Palaeontologist Dale Russell famously proposed a thought experiment in the 1980s imagining how a troodontid lineage might have evolved given another few million years — arriving at what he called the dinosauroid, a hypothetical bipedal, large-brained descendant.

Russell's thought experiment is speculative and has been criticised on various grounds; most palaeontologists regard it as imaginative but not predictive. The broader question it raises, however, is legitimate: given enough time and ecological pressure, could another lineage have crossed the threshold we associate with technological civilisation? We do not know. Intelligence and tool use have evolved independently many times on Earth — in corvids, in cephalopods, in cetaceans, in primates. The specific suite of traits that produced Homo sapiens involved enormous contingency, but contingency is not impossibility.

More intriguingly, there are lineages far older than the dinosaurs. The Carboniferous period, around 300–350 million years ago, produced ecosystems of staggering complexity. The Permian period, which ended 252 million years ago in Earth's largest mass extinction, included the therapsids — the mammal-like reptiles that were, in some ways, sophisticated animals. And before that, there is the murky deep of the Palaeozoic, where soft-bodied creatures and early vertebrates proliferated in ways we are still reconstructing.

None of this constitutes evidence for prior civilisations. What it does is establish that Earth's history has contained, repeatedly, the raw material of complexity from which intelligence might emerge — given enough time, stability, and perhaps a little luck.

Schmidt and Frank, being scientists, do not leave the hypothesis at the level of philosophical wonder. They propose a framework for what testable evidence might look like — a kind of checklist for future geologists or astrobiologists looking for the fingerprint of lost technology.

The most promising signals would be geochemical anomalies in precisely dated rock strata. A narrow layer showing an unusual spike in industrial metals — iridium occurs naturally in impact events, but bismuth, antimony, or lead in unusual concentrations and isotopic profiles might be more suggestive — combined with a contemporaneous carbon isotope excursion and an abrupt biological turnover would, together, constitute a pattern worth investigating.

Artificial isotope ratios are another potential marker. Nuclear technology, for instance, produces isotopes that do not occur in natural processes — or occur only in vanishingly small quantities. Plutonium-239, for example, has a half-life of about 24,000 years, meaning it would be essentially undetectable after a few hundred thousand years. Most nuclear-era isotopes would be long gone across tens of millions of years. However, fission products leave subtle long-term isotopic anomalies in certain elements, and some researchers have proposed that sufficiently advanced analysis of isotope ratios in ancient rocks might, in principle, detect such residues. This is at the frontier of what is analytically possible and remains speculative.

Thermogenic hydrocarbons — the kind produced by burning fossil fuels — leave characteristic molecular signatures in sediments. These differ from biologically produced organic compounds in detectable ways. If a prior civilisation had burned significant quantities of fossil carbon over centuries, a sufficiently well-preserved sedimentary record from that period might show an unusual pulse of thermogenic organic molecules.

The honest summary is that we have not found any of these signatures in a form that compels a civilisation interpretation. But Schmidt and Frank's point is that we have also not been looking for them — not systematically, and not with this specific question in mind. The absence of evidence is not, in this case, strong evidence of absence.

The Silurian Hypothesis intersects, unexpectedly, with one of the oldest puzzles in the search for extraterrestrial intelligence: the Fermi Paradox. Physicist Enrico Fermi famously asked, roughly, if intelligent civilisations are common in the universe, why haven't we heard from any of them? Where is everybody?

One of the many proposed answers to the Fermi Paradox is the Great Filter — the idea that there is some step in the progression from simple life to starfaring civilisation that almost nothing survives. The Great Filter could lie behind us (life is rare, or intelligence is rare) or ahead of us (civilisations reliably destroy themselves before spreading to other stars). If the Silurian Hypothesis were confirmed — if prior civilisations had existed on Earth and perished — it would suggest the Great Filter, or at least one filter, is something that recurs within a planet's history. Not only might civilisations fail to escape their home solar system; they might fail to survive long enough to leave a legible record even on their own world.

This connects to what Frank has written about elsewhere under the framework of planetary intelligence — the question of whether industrial civilisations are a naturally unstable configuration, a kind of phase transition that planets go through, burning hot and fast before either collapsing or transforming into something more sustainable. If Earth has already been through this transition once, or more than once, that is information of extraordinary weight — both for understanding our own trajectory and for calibrating our expectations about intelligent life elsewhere.

Frank has been careful, publicly and in his writing, to note that the Silurian Hypothesis is a thought experiment as much as a research programme. But thought experiments with rigorous logical structures have historically been among science's most productive tools. The question it poses is genuine: we cannot rule out prior civilisations on Earth, and we have not tried very hard to do so.

It would be incomplete — and perhaps intellectually dishonest — to address the Silurian Hypothesis without acknowledging that many human traditions contain stories of prior ages of the world, antediluvian civilisations, and peoples or beings who preceded us and were destroyed. The Hindu cosmological concept of yugas — vast cycles of time in which civilisation rises and decays — the Greek mythological memory of lost Atlantis as recounted by Plato, the Mesoamerican concept of previous Suns or world-ages destroyed by catastrophe, the Sumerian antediluvian kings said to have ruled for impossibly long periods before a great flood: these stories are remarkably widespread.

The standard historical explanation is that these narratives are mythological, not historical — expressions of human psychological needs around origins, time, and catastrophe. That explanation is likely correct in most cases, and nothing here should be taken as an argument that ancient mythology constitutes evidence for pre-human or early-human industrial civilisations. It does not.

What these traditions do illustrate, perhaps, is the persistent human intuition that our civilisation is not the first chapter of the story, but a later one — that the world we inherited is older and stranger than it appears. Whether that intuition reflects some fragment of cultural memory, a general psychological tendency, or simply the fact that landscapes, ruins, and bones were always visible to our ancestors, prompting questions they answered with narrative, is itself an interesting question.

The Silurian Hypothesis does not vindicate these myths. But it does take seriously, in a rigorous way, a question those myths have always been circling: were we preceded?

The Silurian Hypothesis is, at its core, a collection of open questions. Some of them are answerable in principle, if we develop better tools and ask more precisely. Others may be permanently beyond our reach.

Could we develop a systematic programme to search for geochemical civilisation signatures in the rock record? Schmidt and Frank gesture toward this but acknowledge that no such programme exists. Given that we now understand the Anthropocene's fingerprint well enough to describe it, a targeted search for comparable anomalies in well-preserved sedimentary sequences from the Palaeocene, Cretaceous, or Permian seems scientifically feasible. Nobody has funded it yet.

Is there a lower bound on how long a civilisation would need to exist to leave a detectable trace? Our industrial civilisation is roughly 300 years old. Schmidt and Frank estimate that, even at current rates, the Anthropocene's signature might be reduced to a thin chemical layer by the time millions of years have passed. A civilisation that lasted only a few centuries — or even a few thousand years — might leave almost nothing. Does this mean we should expect the record to be empty even if prior civilisations existed?

What would the evolutionary pathway to a pre-human technological species actually look like, and do we see any candidates in the fossil record? The troodontid thought experiment is decades old, but palaeontology has advanced dramatically since then. Are there lineages in Earth's fossil record that exhibit the specific combination of traits — large brain, manipulative appendages, social complexity — that we associate with proto-technological intelligence, and which disappeared before reaching that threshold?

How does the Silurian Hypothesis change our interpretation of the PETM and other rapid carbon cycle events? Not much, in terms of the best current evidence — the scientific consensus still favours natural causes. But the question deserves formal methodological treatment: what additional analyses, at what level of resolution, would be needed to definitively rule out an anthropogenic cause for events like the PETM?

And perhaps most provocatively: if Earth has been through something like the Anthropocene before, and the civilisation responsible is gone, what does that imply about the endpoint of industrial development? Not necessarily catastrophe — a civilisation might transform rather than collapse, leaving fewer traces precisely because it became more efficient, more biological, less geological. Or it might have been destroyed by the same dynamics that researchers now study under the heading of existential risk. The silence of the deep past might be informative, but we have not yet learned how to read it.

The Silurian Hypothesis will not be resolved by this article, or likely by any single study. What it offers is something rarer than an answer: a genuinely new way of asking an old question. We have always wondered whether we are alone. Schmidt and Frank suggest we have not even finished asking whether we were first — and that asking that question rigorously, humbly, and with full awareness of what the geological record can and cannot tell us, might teach us something essential. About the planet. About intelligence. About what it means to leave a mark on a world that was old, and will be old again, long after we are gone.