Why the same stories keep appearing

Read enough mythology from enough places and you start getting a nagging feeling you've been here before.

A great flood wipes out nearly everyone, and one warned family survives in a boat. A trickster steals fire, or the sun, or the secret of something, and gets punished horribly for it. Somebody goes down to the land of the dead to fetch back a person they love, and it goes wrong at the last moment. Twin brothers. A hero with an unusual birth. A world tree.

These turn up in traditions separated by oceans and by thousands of years, told by people who couldn't possibly have met. The resemblance is real and it's genuinely strange.

What you do with it is the hard part, and this is a field where the most satisfying explanations have tended to be the flimsiest.

There are only three answers, and they overlap

Strip out the mysticism and the possibilities come down to three.

Two stories can be similar because they're the same story inherited from a common ancestor, the way two languages share a word. They can be similar because one culture heard the other's version and adapted it. Or they can be similar because both groups of people were dealing with the same world — the same rivers, the same winters, the same deaths — and arrived at comparable stories on their own.

Shared descent, contact, independent invention. Every serious argument in comparative mythology is a fight about which of the three applies to a particular case.

And here's the awkward bit that gets left out of the popular treatments: telling them apart is genuinely difficult, often impossible, and the honest answer is frequently "we can't say".

Explanation one: the stories are cousins

Sometimes you can trace descent, because the language came down the same pipe.

The Indo-European family is the best case anyone has. Speakers of a single ancestral language spread across Europe, Iran and northern India, and their descendant languages are related in ways that can be demonstrated sound by sound rather than by impression. Where the language descent is documented, mythological descent becomes a testable claim rather than a hunch.

The clearest example is a name. Greek Zeus, Roman Jupiter — from an older Dies-piter — and Sanskrit Dyaus Pita all go back to the same reconstructed phrase, a sky father. That's not three cultures independently deciding the sky is paternal. That's one word, inherited three ways, with the sound changes behaving exactly as the rest of the vocabulary does.

Alongside it sits a recurring story of a hero or god killing a serpent that's blocking or hoarding water: Indra and Vritra in the Vedas, Thor and the world serpent, Zeus and Typhon, and others. The pattern is consistent enough, and tied tightly enough to inherited vocabulary, that most specialists accept some version of it.

Georges Dumézil pushed this much further, arguing that Indo-European societies organised their gods into three functions — sovereignty, force, and fertility or production — and that the scheme shows up from Rome to India. It's been enormously influential and it's also been picked apart, on the grounds that a three-way sort is easy to impose on almost any pantheon if you're willing to shuffle. Some of Dumézil stands. Not all of it.

Explanation two: stories travel, and they travel fast

Cultures in contact borrow relentlessly, and narrative moves at least as easily as pottery.

Trade routes carried stories the whole length of the Silk Road, around the Mediterranean, across the Indian Ocean and up the rivers of Central Asia. Tales from the Indian Panchatantra travelled into Persian, then Arabic, then Latin, and turn up in European collections centuries later, altered but traceable. Buddhist narratives moved across Asia with monks. Nobody had to write any of it down for it to spread.

Contact is the explanation with the strongest evidence when it applies, because it leaves a trail. You can often show the route, the direction and roughly the century. The borrowing version usually turns up later and keeps traces of the source — a name that doesn't fit the local language, a detail that made sense somewhere else, a geography that isn't local.

The trouble is that absence of a documented route doesn't prove there wasn't one. Most contact in human history left nothing behind.

Explanation three: people keep running into the same things

The third explanation is the least glamorous and probably does the most work.

Floods happen everywhere. Most early settlement clustered along rivers and coasts, because that's where the water and the transport and the fish are, and those are exactly the places that periodically get destroyed by water. A community whose river drowns the valley every few generations doesn't need to hear anybody else's story to produce one of its own.

The same goes for the rest of the recurring set. Everyone dies, so every tradition has to say something about where the dead go. The year turns, so seasonal disappearance and return is available to anyone who's noticed. Fire is transformative and dangerous, so its arrival demands an account. And people everywhere have had to explain to a child why the person who was here last winter isn't here now.

Shared circumstances, shared subject matter. That doesn't explain detailed structural matches, but it explains an awful lot of the broad ones.

The flood, taken apart properly

The flood is the standard exhibit for universal myth, and it's useful precisely because it splits across all three explanations rather than supporting any one of them.

Start with the case that's settled. In 1872 a former banknote engraver named George Smith, working through crates of cuneiform tablets at the British Museum, found a passage describing a man told to build a boat, a flood, and birds released to look for land. He read it aloud to the Society of Biblical Archaeology and the room reportedly emptied of composure. It's tablet eleven of the Epic of Gilgamesh, and the survivor is Utnapishtim.

Older Mesopotamian versions exist too — the Atrahasis story, and a Sumerian one with a survivor named Ziusudra. So it's not a single text but a regional tradition with several generations of retelling behind it.

The structural overlap with the account in Genesis is close: a warned survivor, a specified vessel, animals aboard, birds sent out to test for dry land, a sacrifice afterwards. These are neighbouring cultures in continuous contact over a very long period, and most scholars read the resemblance as a shared Mesopotamian background rather than coincidence. That's a claim about literary history, and it's held up.

Now look at flood stories from elsewhere — from Australia, the Americas, the Pacific, South Asia. Many of these traditions do have one. But once you compare the specifics, the resemblance often thins out to "there was a great flood and some people lived". There's no boat, or no warning, or no bird, or the flood is punishment for something completely different, or it's not the end of a world at all.

So the same body of material supports contact for one pair and independent origin for much of the rest. Treating "flood myths" as one phenomenon with one cause is what manufactures the overclaim.

People have also looked for a single real event behind all of it. The best-known candidate is the proposal that the Black Sea filled catastrophically from the Mediterranean around 7,000 years ago. It's a real hypothesis with real geologists on both sides, and the evidence for a sudden catastrophic filling has been contested since it was published. Even at its strongest it would explain one region, not a planet.

Where the universal-story idea gets oversold

The version most people have met is the monomyth, set out by Joseph Campbell in 1949: the claim that hero stories everywhere follow one underlying arc of departure, initiation and return.

Campbell wasn't first. Otto Rank had proposed a pattern for the birth of the hero decades earlier, and Lord Raglan had published a twenty-two-point hero template in 1936. Campbell's version was broader, better written and much more quotable, and it landed in Hollywood, where it's been openly used as a structural template ever since.

As a screenwriting tool it clearly works. As a description of the world's mythologies, folklorists have been fairly rough with it, and their objections are worth taking seriously.

The main one is that the pattern is elastic enough to fit almost anything, and that it was built by keeping the episodes that matched and setting aside the ones that didn't. Lots of myths aren't hero stories at all — they're origin accounts, ritual charters, explanations of a landscape feature, jokes. Lots of hero stories skip stages, run them out of order, or end with the hero dead and nothing restored. Alan Dundes, who spent a career on this material, was blunt about the method.

There's a second objection that matters more. Compressing genuinely different traditions into one shape throws away the differences, and the differences are the part that tells you something about the people who told the story. A theory that renders every mythology a variation on one plot isn't uncovering a universal. It's deleting the data.

Similar caution applies to broad archetype claims. That certain figures recur across traditions is an observation, and a decent one. That the recurrence proves an inherited psychological structure is a much bigger step, and the evidence for it is much thinner than the confidence with which it's usually asserted.

How folklorists actually compare stories

The discipline is far more careful than its pop version, and the tools are unglamorous on purpose.

There's a catalogue — the Aarne-Thompson-Uther index — that assigns numbers to recurring tale types, so a story can be identified as a version of a defined type rather than described by vibe. Type 425 covers the search for a lost husband, which is where Cupid and Psyche and Beauty and the Beast both live. Separately, Stith Thompson's motif index catalogues the components: the helpful animal, the impossible task, the youngest of three children who succeeds where the elders failed.

A story becomes describable as a particular arrangement of specified motifs. That sounds bureaucratic. It's the whole point.

"Both cultures have a hero who leaves and comes back changed" is compatible with essentially every story ever told, so it can't be wrong and therefore can't be informative. "Both are type 425, sharing four motifs in the same sequence, in adjacent regions with a documented trade route, with the older attestation on the side you'd expect" is a claim somebody can go and check, and occasionally knock down.

More recently, researchers have borrowed methods from evolutionary biology — building family trees of story variants the way you'd build them for species. Jamie Tehrani's work on the Little Red Riding Hood group, published in 2013, is the best-known attempt, treating each variant's traits as characters and testing which branching structure explains the spread best. Whether the analogy between genes and story elements holds up well enough for this is debated. It's still a considerable improvement on counting resemblances by eye.

The archive itself is a problem

This one rarely gets mentioned and it undercuts a lot of confident comparison.

The corpus of world mythology available to an English-speaking reader wasn't gathered evenly or neutrally. Enormous amounts of it were written down by missionaries, colonial officials and nineteenth-century folklorists, working through translators, decades or centuries after those communities had already been in contact with Christianity, Islam or European traders.

That matters in a specific way. A collector who arrived expecting to find a flood story, asking questions shaped by the one he already knew, recording the answer in a language with its own vocabulary for the sacred, can produce a text that looks reassuringly familiar for reasons that have nothing to do with ancient descent. Some recorded resemblances are almost certainly artefacts of how and when the recording happened.

Add the obvious survivorship problem. Traditions whose custodians were killed, displaced or forbidden to practise are underrepresented or gone. What's in the anthologies isn't a sample of human mythology. It's a sample of what particular outsiders were able and inclined to write down.

Living traditions, meanwhile, aren't specimens. The stories in these collections belong to communities who in many cases are still telling them, and who often have their own scholars and their own view of what the material means.

Why certain stories survive being retold

There's one more effect worth adding, and it's about transmission rather than origin.

Stories that get passed on by mouth are filtered hard. Every retelling is a chance to forget something, and what survives is what people find easy to remember and satisfying to repeat. Over enough generations that's a serious selection pressure, and it doesn't care where the story came from.

Cognitive researchers have argued that a particular kind of content survives this filter best: mostly ordinary, with a small number of pointed violations. A person, but he doesn't die. A tree, but it speaks. A statue that weeps. Fully realistic accounts are forgettable, and wildly incoherent ones are too hard to hold onto, so the middle ground wins.

That's an interesting third factor, because it predicts convergence without requiring either contact or descent. Different cultures aren't just facing similar floods. They're running their stories through similar human memory, generation after generation, and similar filters produce similar output.

How to work out which one applies

The useful move is to ask what each explanation would predict, rather than picking whichever feels most profound.

Shared descent predicts that resemblance tracks known lines of linguistic descent, and fades as you move away from them. That's testable, and it's been tested, which is why the Indo-European sky father is on firmer ground than most claims in this field.

Contact predicts a route and a chronology — the borrower's version appearing later, with traces of the source still showing. Also testable, and it's why the Gilgamesh case is accepted and a lot of others aren't.

Independent invention predicts no route and no descent, but a shared circumstance strong enough to prompt the story unaided, plus resemblance that stays generic. Two flood stories that share only "the water came" are behaving exactly as independent invention predicts.

Often the honest verdict is a mixture, or that the evidence doesn't reach. That's less satisfying than a single grand pattern behind everything, and it's closer to what the material supports.

The stories aren't diminished by it. A flood myth that arose because one people lived beside one particular dangerous river tells you something real about them, their water and their fears. A claim that all myths are secretly the same story tells you nothing about anyone.