Why Independent Paths Meet

ai blog
July 18, 2026

Joel Kowalewski, PhD

Two letters, one idea, June 1858

In the third week of June 1858, a parcel reached Down House in Kent, carried by the ordinary post from an island in the Malay Archipelago. Charles Darwin opened it and read, in the hand of a younger naturalist named Alfred Russel Wallace, a short essay written in a fever on the island of Ternate: “On the Tendency of Varieties to Depart Indefinitely from the Original Type.” It described, in a few thousand words and without a single one of Darwin’s twenty years of barnacles and pigeons and notebooks behind it, the mechanism Darwin had told almost no one he had found — the differential survival of small variations, working over time, as the engine that makes and unmakes species. Darwin had been sitting on the idea for two decades, afraid of it. Wallace, half a world away and gravely ill, had arrived at the same idea in a night.

Darwin’s first response was not triumph but a kind of vertigo. He wrote to his friend the geologist Charles Lyell on the eighteenth of June, and the letter has the sound of a man watching his life’s work slip sideways out of his hands.

I never saw a more striking coincidence; if Wallace had my MS sketch written out in 1842, he could not have made a better short abstract! … So all my originality, whatever it may amount to, will be smashed.

Charles Darwin to Charles Lyell, 18 June 1858

What happened next is famous. Lyell and the botanist Joseph Hooker arranged a compromise that satisfied the etiquette of priority without robbing either man: on the first of July, 1858, extracts from Darwin’s unpublished work and Wallace’s essay were read together to the Linnean Society of London, credited to both. Neither author was in the room. Wallace was still in the Malay Archipelago and did not yet know it was happening. Darwin was at home burying his infant son, who had died of scarlet fever three days before. And so the theory that would reorganize all of biology entered the world not as one man’s revelation but as a coincidence — two people, unknown to each other’s work, separated by an ocean and by every difference of temperament and training, reaching for the same idea at nearly the same hour.

We have three habitual ways of explaining a moment like this, and I think all three miss what it is telling us. The first is coincidence: a fluke, a story for the anniversary dinners. The second is the spirit of the age: the idea was “in the air,” the culture had ripened, and any two competent naturalists were bound to catch it. The third, uglier and never far away, is theft: someone must have taken from someone. My argument is that the striking coincidence is neither an accident nor a mood nor a crime. Independent paths meet because the world is a definite set of relationships, and any path that honestly tracks those relationships must arrive where the others did. Convergence — two minds, or forty separate lineages, landing on the same solution — is not a curiosity at the edge of discovery. It is the clearest evidence we ever get that what we found was real, and not merely ours.

The rule, not the exception

The first thing to establish is that Darwin and Wallace are not an anomaly to be marveled at but an instance of the ordinary. In 1922 the sociologist William Ogburn and his colleague Dorothy Thomas published a paper with a title that is already half the argument: “Are Inventions Inevitable? A Note on Social Evolution.” Appended to it was a list of roughly a hundred and fifty discoveries and inventions that had been made independently, by two or more people who did not know of one another, at very nearly the same time. The calculus, arrived at separately by Newton and by Leibniz. The theory of natural selection, our case in point. The telephone, whose two inventors, Alexander Graham Bell and Elisha Gray, filed at the United States Patent Office on the same day — the fourteenth of February, 1876. Oxygen, sunspots, the conservation of energy, the pendulum clock discovered three times over, the thermometer seven. The list goes on, and it has kept growing since.

The sociologist Robert Merton spent much of his career on this pattern, and drew from it a conclusion that inverts our usual picture of discovery. In his 1961 essay “Singletons and Multiples in Science,” he argued that independent multiple discovery is not the exception in science but the rule — that the singleton, the idea genuinely found only once and by only one mind, is the case that actually needs explaining, and probably owes its solitude to the others having gone unrecorded or unpublished. We tell the history of science as a procession of lonely geniuses because the biography sells better that way. The ledger of the discoveries themselves tells a different story, one in which the same door is very often opened by several hands at once.

This is exactly why the history of science is so thick with priority wars, and the wars are worth a moment because of what they assume. When Newton and Leibniz fell out over the calculus, the Royal Society convened a committee in 1712 to settle the matter; its report, the Commercium Epistolicum, found for Newton — unsurprisingly, since Newton himself had secretly drafted it. We litigate these disputes as though an idea were a piece of property that one person must have taken from another, because we cannot easily hold the stranger possibility: that two people found the same thing because the thing was there to be found. The statistician Stephen Stigler once formalized the whole comedy as his “law of eponymy” — no scientific discovery is named after its original discoverer — and, with a straight face, credited the law itself to Robert Merton, so that the law obeys the law. The joke has a serious edge. If discovery were a private act of individual genius, simultaneity would be a rare embarrassment. Instead it is so common that we have built an entire apparatus of patents, priority claims, and eponyms to manage the traffic.

There is a respectable way to deflate all of this, and I want to grant it its full strength before turning it. The deflation says: of course discoveries cluster: an idea becomes findable only when its prerequisites are in place. Wallace could reach natural selection in 1858 because Malthus and Lyell and the fossil record had already laid the track; nobody reaches it in 1300. The biologist Stuart Kauffman gave this reachable frontier a name, the adjacent possible — at any moment, only certain next steps are within reach of what already exists, and discovery is the exploration of that shifting edge. On this view the paths were never really independent. They shared a culture, a literature, a moment, and the convergence is just what it looks like when many people stand at the same frontier and take the one step available. This is true, and it explains a great deal. But notice that it only moves the question back a step. It tells us that there is an adjacent possible, a structured space of reachable next moves that many minds find their way to. It does not tell us why that space has the shape it does, or why the shape is one that different explorers, coming from different directions, keep mapping the same way. To answer that, we need a case where the paths share no culture at all — no literature, no moment, no possibility of collusion. We need to leave the history of ideas and go to biology.

Nature solves the same problem twice

Convergent evolution is the phenomenon in which unrelated lineages, starting from different ancestors and separated by tens or hundreds of millions of years, independently arrive at the same biological solution. It is the decisive case for my argument precisely because it strips away everything the “spirit of the age” explanation relies on. Two lineages of animals share no journals and attend no conferences. They cannot read one another’s work, cannot be influenced, cannot steal. Whatever they have in common, they came to alone. And yet they converge, constantly, and on the hardest problems.

Take the eye — the organ Darwin himself singled out as the hardest test of his theory. By the much-cited estimate of the zoologists Luitfried von Salvini-Plawen and Ernst Mayr (1977), eyes have arisen independently in the animal kingdom somewhere between forty and sixty-five times. The most striking instance is the camera eye: a lens focusing an image onto a sheet of photoreceptors. Vertebrates have one; so does the octopus. Our last common ancestor with the octopus was a wormlike creature more than half a billion years ago that had nothing of the kind. The vertebrate eye and the cephalopod eye were built separately, by lineages that diverged before eyes of that grade existed anywhere, and they converged on nearly the same design — so nearly that the differences (the octopus, tellingly, has no blind spot) are the exception worth remarking. Two entirely independent paths, and they meet at a lens.

The pattern repeats at every scale, down to the molecule. Echolocation — navigating by the echoes of one’s own cries — evolved separately in bats and in toothed whales, lineages that parted ways in the age of the dinosaurs. For a long time this was a textbook example of convergence in form. Then the genomes were read, and the convergence turned out to run deeper than anyone expected. Beginning around 2010, researchers found that a hearing gene called Prestin, and then a broad swath of other genes involved in hearing, carried strikingly similar sequence changes in echolocating bats and dolphins — the same edits, written into the same genes, by two lineages that had no contact for sixty million years. A 2013 paper in Nature, “Genome-wide signatures of convergent evolution in echolocating mammals,” found these parallel changes scattered across hundreds of sites in the genome. Nature had not merely arrived at the same answer twice; it had, in places, spelled it the same way. And lest this seem confined to the exotic, consider that the crab body plan — the flattened shell, the tucked tail — has evolved independently at least five separate times, a habit so pronounced that in 1916 the zoologist Lancelot Borradaile coined a word for it: carcinisation, nature’s recurring tendency to make a crab.

Now ask what is forcing the repetition, because this is where the whole matter turns. It is not copying. I have argued at length elsewhere that nature does not clone — that living things are not reproduced from a master file but performed anew each time, and that no lineage is working from another’s blueprint. Convergence is the opposite of duplication: it is independent arrival. What the octopus and the vertebrate have in common is not descent and not contact. It is the problem. Light behaves in exactly one way; the physics of forming an image permits only so many solutions, and a lens focusing rays onto a light-sensitive surface is one of the few that work, so any lineage that gets good at seeing is pushed toward it. Water resists a moving body according to fixed laws, and so the tuna, the shark, the dolphin, and the extinct ichthyosaur — four utterly separate lineages — all converged on the same torpedo shape. The constraints are not in the animals. They are in the world. And this is the point I most want to make: the constraint is real. The optics, the hydrodynamics, the acoustics of ranging by sound — these are relationships that hold whether or not any creature is there to be shaped by them, and they are the shared cause that two lineages who never met nonetheless had in common. Convergent evolution is the world imposing its relationships on anything that must live inside them. The lineages never met; the ocean they were solving did.

The strongest objection: replay the tape

The most serious case against everything I have just said was made by one of the finest scientific essayists of the last century, and it deserves to be stated at full strength. In Wonderful Life (1989), the paleontologist Stephen Jay Gould proposed a thought experiment that has organized the debate ever since. Rewind the tape of life to some early moment, he said, and let it play again. Would you get anything like the world we have? Gould’s answer was an emphatic no. Evolution, on his view, is dominated by contingency — by frozen accidents, mass extinctions that spare the lucky rather than the fit, chance events that send whole histories down one branch rather than another. Replay the tape and humans never appear; very likely nothing recognizable does. If Gould is right, then my convergences are survivorship stories after all: we notice the eye built forty times and forget the countless solutions that were tried once and lost, and the “real relationships” I am pointing to dissolve into the noise of what merely happened to endure.

This objection is answered, I think, by the very evidence Gould’s metaphor invites us to imagine — and the answer came most forcefully from the paleontologist Simon Conway Morris, who had worked on the same fossils as Gould and drew the opposite conclusion. In Life’s Solution (2003), Conway Morris pointed out that we do not have to imagine replaying the tape, because nature runs the experiment for us, constantly, on separate continents and in separate lineages. Every instance of convergent evolution is a partial replay — the tape restarted from a different point, in a different organism, with different raw material — and it would be highly improbable to observe similarities if evolution were indeed fully guided by chance events. You cannot dismiss forty independent eyes as an accident; forty accidents that all arrive at a lens are not accidents but a law wearing the costume of chance. Gould and Conway Morris are each describing the same process from different vantage points; one emphasizing chance as a driver variability in the biosphere, and the other noting that variability alone cannot explain the obvious organization of the biosphere. Contingency governs the path: which lineage, at which moment, through which particular gene mutations, are the accidents moving towards a destination. Constraint, the physical and mathematical laws of life, governs the destination: what counts as a working eye, a swimming body, an echolocating ear is fixed in advance by the world; the accidents, however wild, are accidents within a space whose shape they cannot change. The space of distinct configurations of life that support survival is simply not infinite and may actually be much smaller than we imagine.

This does not mean that any particular outcome was fated, that humanity was written into the first cell, or that some purpose was steering the whole from behind the curtain. It is a mathematical claim tied to distributions, not specific instances or species. There is no planner here and no destiny; the constraint is not a mind and not an overarching goal. It is simple, locally applied rules giving steadily rise to global organization and not pre-programmed global organization — optics that permit some eyes and forbid others, a chemistry that allows some molecules to carry information and not others. And I leave open, as one always should, that the space of solutions may be wider than we currently imagine as well, and that somewhere the tape has found answers we would not recognize as answers at all. As human uniqueness may be increasingly challenged by AI technologies, the space that evolution could not explore may help use redefine human nature — that our uniqueness was not in any one skill (chess, art, trivia, medical diagnostics, etc.) but our capacity to ask questions, to synthesize, bringing the various differences that dominant our lives and seemingly separate us, and achieve what evolution could not: coherence.

When the disciplines converged

Return now to the human beings, because the same logic that governs the octopus governs the seminar. Between 1946 and 1953 an extraordinary group met periodically in New York under the sponsorship of the Josiah Macy Jr. Foundation: the mathematician Norbert Wiener, the neurophysiologist Warren McCulloch, the anthropologists Gregory Bateson and Margaret Mead, the engineer Julian Bigelow, and others from a half-dozen fields that had, until then, no reason to speak to one another. What drew them together was that each had independently run into the same object from a different side. The engineer met it in the feedback governor that keeps a machine from tearing itself apart; the physiologist met it in the reflex arc; the anthropologist met it in the self-correcting rituals of a culture. It was the same object — a circle of control and communication, a system that senses its own output and adjusts — and Wiener gave it a name in 1948, cybernetics. The disciplines converged not because they borrowed from one another but because feedback is a real relationship that appears wherever there are systems holding themselves together, and enough separate observers had walked far enough into their own subjects to bump against it.

I have argued in earlier essays that intelligence is a relationship rather than a thing, and that a fact is not an object we stumble upon but a belief that a community is compelled to hold. Convergence is the hinge that joins those two claims, and it is worth making the connection explicit. What we dignify with the word fact is precisely what independent observers, taking independent routes with no way to coordinate, are nonetheless forced to affirm. A single observer can be fooled; a single method can flatter a single illusion. The world shows its hand only when separate paths, which could not have colluded, land on the same result — when the chemist’s measurement and the astronomer’s calculation and the geologist’s strata all point one way. Geocentrism, which I have written about before, did not fall to a single decisive blow; it fell because line after independent line — the phases of Venus, the mechanics of falling bodies, eventually the parallax of the stars — converged somewhere else and left the old picture with nowhere to stand. Objectivity, in practice, is not a single mind seeing clearly. It is many minds, by many roads, being driven to the same place. Convergence is the mechanism by which the merely believed becomes the shared, and the shared becomes what we are willing to call true.

The serendipity is the evidence

Which returns us, at the end, to Darwin at his desk in June 1858, feeling his originality about to be smashed. He had it exactly backward, and the correction is the whole of what I have been trying to say. The coincidence that unsettled him was not a threat to his achievement; it was the strongest confirmation he could possibly have received that natural selection was not a thing he had invented but a thing he had found. If the mechanism had been his private construction, a clever story fitted to the facts, then no stranger on a distant island should have been able to build the identical story from the identical facts. That Wallace did is the proof that the pattern was not in Darwin. It was in the living world, in the relationships between organisms and the conditions they must survive, and it was there to be read by anyone whose eye had been trained to the right things. The two men were not rivals for a possession. They were two witnesses to the same fact, and the value of a second independent witness is that it turns a claim into knowledge.

This is why I distrust the word serendipity when it is used to wave away these meetings as happy accidents. The serendipity is the evidence. When separate lines of inquiry, or separate lineages of life, or separate disciplines that share no vocabulary, keep arriving at the same structure, that convergence is not luck decorating the surface of things; it is the world telling us, in the only language it has, which of our ideas are actually about it. We are living now through a season of convergences of exactly this kind — independent laboratories, independent theories, independent traditions of thought closing in on the same questions about mind and machine and matter from directions that were supposed to have nothing to do with each other. The temptation is to read this as fashion, a bubble of collective attention that will pass. I think it is the opposite. Where the paths are genuinely independent and they meet anyway, the meeting is data about reality, and the task before us is not to dismiss it as coincidence but to do the harder and more hopeful thing: to ask what it is about the world that keeps forcing so many roads to the same door. The convergence was never the accident. The world was always there, waiting for the paths to find it.