Here’s a question with no experiment to settle it: if you could rewind Earth’s history to the beginning and let it play again, would intelligent life emerge a second time? Stephen Jay Gould and Simon Conway Morris — two of the most influential evolutionary thinkers of the late 20th century — spent years arguing exactly this, and they reached opposite conclusions. The disagreement is still unresolved, and it sits directly beneath the entire search for intelligent life elsewhere.

Gould’s Answer: Rewind and It’s Gone

In his 1989 book Wonderful Life, Stephen Jay Gould argued that evolution is dominated by contingency — by accident, luck, and the specific sequence of events that actually happened. Replay the tape of life, he wrote, and you’d get a wildly different result. The particular path that led to humans depended on countless chance events: which species survived mass extinctions, which mutations happened to occur, which asteroid hit when.

Gould’s prime example was the Cambrian explosion, the burst of animal diversity around 540 million years ago. Many strange body plans appeared; most went extinct. Which ones survived, in his reading, was largely a matter of luck rather than superiority. Remove the lineage that led to vertebrates at that early stage — an entirely plausible accident — and there are no fish, no amphibians, no mammals, no us. On this view, human-level intelligence is a fluke, and expecting it to recur elsewhere is a mistake.

Conway Morris’s Answer: It Comes Back

Simon Conway Morris, who ironically did much of the foundational work on the same Cambrian fossils Gould wrote about, drew the opposite conclusion. In Life’s Solutions (2003), he argued that evolution repeatedly converges on the same solutions regardless of starting point — and that intelligence is one of those solutions.

His evidence is convergent evolution: the camera eye evolved independently in vertebrates and octopuses; flight arose separately in insects, birds, bats, and pterosaurs; complex problem-solving appears in lineages as distant as primates, crows, and octopuses. If the same complex traits keep reappearing along different evolutionary paths, Conway Morris reasoned, then they’re not flukes — they’re attractors that evolution finds again and again. Replay the tape, and something intelligent and technological would likely emerge once more, perhaps not human but recognizably mind-like.

The Problem Neither Could Solve

The frustrating thing is that both men were looking at the same evidence and weighting it differently. Convergence is real — eyes and wings and tool use genuinely did evolve multiple times. But human-level technological intelligence evolved, as far as we know, exactly once in nearly four billion years. Convergence shows that useful traits recur. It doesn’t obviously show that our specific kind of intelligence is an attractor, because we have only the single instance.

The evolutionary biologist Ernst Mayr made this point sharply in a famous exchange with Carl Sagan. Of the billions of species that have existed on Earth, Mayr noted, only one developed the kind of intelligence capable of building a radio telescope. If it were a favored outcome, why so spectacularly rare? One data point can be read as “it happened, so it can happen” or as “it happened only once in four billion years, so it’s nearly impossible.” The same fact supports both stories.

What the Timing Suggests

There’s a quantitative clue buried in the chronology, first pressed by the physicist Brandon Carter. Complex, technological intelligence took almost the entire available lifetime of Earth’s biosphere to appear — roughly four billion years, a substantial fraction of the time the Sun will remain stable. If intelligence were easy and favored, you might expect it to arise relatively early. That it took nearly the whole window hints that one or more steps along the way were extraordinarily difficult — rare enough that, on most planets, the star might burn out before they’re cleared.

This doesn’t prove intelligence is rare. But the timing leans, gently, toward the pessimistic reading. We arrived close to the last possible moment, which is the opposite of what you’d expect from an inevitable process.

Why the Search Settles It

The honest position is that we cannot currently answer the question from Earth alone. A sample size of one is consistent with both “intelligence is common” and “intelligence is a one-in-a-galaxy accident.” No amount of cleverness about the fossil record breaks that tie, because the entire fossil record is a single replay of the tape.

This is precisely why the search for extraterrestrial intelligence matters beyond its obvious appeal. Finding even one other technological civilization would prove that Earth’s path wasn’t unique — that Conway Morris was closer to right, that intelligence is a destination evolution can reach more than once. A long, careful search that keeps finding silence would, slowly, support Gould and Mayr — the possibility that what happened here was a fluke not repeated anywhere nearby. The debate that two brilliant biologists couldn’t resolve with fossils might finally be settled by a telescope. Until then, it stays genuinely, frustratingly open.

Replaying the Tape for Real

Gould’s “replay the tape” was a thought experiment — but biologists have since run a version of it in the lab, and the results complicate both sides. The longest-running example is Richard Lenski’s experiment with E. coli, begun in 1988, in which twelve identical populations of bacteria have been evolving in parallel for tens of thousands of generations. For decades they followed broadly similar trajectories, suggesting evolution is somewhat predictable. Then, around generation 31,000, a single population suddenly evolved the ability to metabolize citrate — something E. coli normally can’t do. The others never did.

That single divergence is the whole debate in miniature. Most of the time, the parallel lines converged on similar adaptations, supporting Conway Morris’s view that evolution finds the same solutions repeatedly. But the citrate breakthrough depended on a specific chain of earlier, contingent mutations — remove them and it never happens — which is exactly Gould’s contingency at work. The experiment suggests both men captured something real: evolution is statistically predictable in its common adaptations and radically contingent in its rare breakthroughs. The unsettling possibility is that technological intelligence is a citrate-style event — dependent on a precise, unrepeatable sequence — rather than a reliable convergent outcome.

The Wildcard of Mass Extinction

There’s a factor neither contingency nor convergence fully captures: catastrophe. Earth’s history is punctuated by mass extinctions, and at least one of them was decisive for our existence. For more than 150 million years, dinosaurs dominated the planet’s large-animal niches, and mammals remained small, nocturnal, and marginal. Then, 66 million years ago, an asteroid roughly ten kilometers across struck near the Yucatán, and the resulting catastrophe wiped out the non-avian dinosaurs.

Only in the vacated ecological space that followed did mammals diversify, grow large, and eventually produce primates and us. Without that impact — a random event from space, nothing to do with biology — the dinosaurs might still dominate, and there’s no obvious reason a technological species would have emerged from them. This injects pure chance into the story of intelligence at the largest scale. The path to us ran through a cosmic accident that cleared the board. On another world, the equivalent board might never be cleared, or might be cleared in a way that favors something else entirely. Mass extinctions are wildcards that no model of “natural progression toward intelligence” can predict.

Is There a Direction to Evolution at All?

Underlying the whole dispute is a question biologists still argue about: does evolution have any inherent direction, any tendency toward greater complexity or intelligence over time? It’s tempting to see a trend — life did, after all, go from simple cells to brains. But the appearance of progress may be an illusion of where we choose to look. Life started simple because it had nowhere to start but simple; some lineages grew more complex while most stayed microbial, and bacteria remain, by biomass and diversity, the planet’s dominant life form.

Stephen Jay Gould argued there is no drive toward complexity — only a random walk away from the minimum possible complexity, which inevitably produces some complex outliers without any directional force behind them. If he’s right, intelligence isn’t a goal evolution moves toward; it’s a rare tail of a distribution that mostly stays simple. That reading aligns with the data: four billion years of life, overwhelmingly microbial, with technological intelligence appearing once, at the very end. Whether that single appearance reflects a faint directional pull or pure statistical luck is precisely what we can’t determine from one planet — and precisely what finding life elsewhere could finally illuminate.

SETIworld follows the science and the argument over whether intelligence is inevitable or accidental — join the portal to track where the evidence points next.

References

• Gould, S.J., Wonderful Life: The Burgess Shale and the Nature of History, Norton 1989
• Conway Morris, S., Life’s Solutions: Inevitable Humans in a Lonely Universe, Cambridge 2003
• Carter, B., The anthropic principle and its implications for biological evolution, Phil. Trans. R. Soc. 1983
• Lineweaver, Paleontological Tests: Human-like Intelligence is not a Convergent Feature, 2008
• Mayr, E., The probability of extraterrestrial intelligent life, 1985
• Lenski et al., Long-term experimental evolution in E. coli, various 1991-2020