Why traditions survive
Somewhere today, a group of people are doing something elaborate and slightly inconvenient for reasons none of them can fully explain. Carrying an image through streets. Throwing water. Wearing a particular colour. Eating a specific dish on a specific date, prepared a specific way, because that's how it's done.
Ask why and you'll get one of three answers. It's what we've always done. It's what my grandmother did. Or a shrug, and a sort of amused acknowledgement that nobody actually knows any more.
All three answers are more informative than they look, and the third one is the most interesting of the lot.
A lot of ancient tradition is younger than the railway
In 1983 Eric Hobsbawm and Terence Ranger published a collection called The Invention of Tradition, and it landed like a brick through a window.
Their argument was straightforward and irritating. A great many practices that present themselves as immemorial are recent — often nineteenth century, often deliberately constructed, and frequently assembled during periods when new nations, new institutions or newly powerful groups needed a past that suited them. The antiquity isn't incidental to the practice. It's part of the product.
The book's examples were mostly British, which spared the authors accusations of pointing at other people. Hugh Trevor-Roper's essay on Highland dress argued that the modern short kilt dates to the 1720s and is attributed to an English industrialist, that the system of clan-specific tartans was largely assembled in the nineteenth century, and that a foundational tartan manual published in 1842 was a forgery by two brothers claiming Stuart descent. George IV's visit to Edinburgh in 1822, choreographed by Walter Scott, did an enormous amount to fix the whole ensemble in place.
Scottish historians have since pushed back hard on Trevor-Roper, and reasonably — he overstated, he enjoyed overstating, and older tartan and Highland dress traditions certainly existed. But the general shape of the finding survived the criticism.
David Cannadine's contribution went after British royal ceremonial and found something similar. The great state occasions of the early nineteenth century were often badly organised, poorly attended affairs where the clergy lost their place and the crowds behaved appallingly. The seamless, gleaming version that reads as timeless is substantially a creation of the period from about 1870 onwards, developed alongside the technology to photograph and later broadcast it.
Much of the modern European Christmas is Victorian too. The decorated tree spread rapidly through Britain after an illustration of the royal family around one appeared in 1848. The commercial Christmas card dates to 1843, as does A Christmas Carol. The white wedding dress owes its dominance largely to Victoria's wedding in 1840.
None of which makes any of it fake. A tradition invented in 1848 and practised continuously since is a real tradition by any sensible definition. It just isn't the age it claims.
Why anyone bothers inventing one
The obvious question is why people go to the trouble, and the answer is that invented traditions do work that arguments can't.
A new state needs its citizens to feel something about it, and reasoned explanation of constitutional arrangements doesn't produce feelings. A song does. A parade does. A shared date on which everybody does the same thing at the same time does. Nineteenth-century nation-building ran on exactly this, all across Europe, and the products were national anthems, flag protocols, commemorations and folk revivals assembled from whatever material was to hand.
Institutions do the same at a smaller scale. Regiments, schools, universities, clubs and companies all manufacture ceremonial, and the manufacture is usually most intense when the institution is newest or least secure.
The claim of age isn't a lie so much as a load-bearing element. A practice presented as an arbitrary recent decision invites the question "why this?" A practice presented as ancient inheritance doesn't, and questions that don't get asked don't need answering.
Ritual is a storage format
Here's the part that changes how the whole subject looks.
Before writing was widespread, everything a society needed to keep — genealogy, law, navigation, seasonal timing, technique — had to live in human memory and survive transmission across generations without a copy to check against. Ordinary prose is terrible for this. It degrades fast, and each retelling introduces drift.
So the material gets encoded in forms that resist degradation. Rhythm and metre, which reject a wrong word because it doesn't scan. Rhyme, which constrains what can be substituted. Melody, which carries text and flags omissions. Fixed phrases used in fixed metrical positions. Repetition with small variations. Physical action tied to sequence.
Milman Parry and Albert Lord established a great deal of this by studying living oral poets in the Balkans in the 1930s and comparing the results with Homeric epic. What they found was a compositional system, not memorisation — singers built performances in real time out of stock formulae sized to fit the metrical line, which is why the Homeric epics are full of repeated epithets. The formula is the technology.
Practices with a physical component are sturdier still, because the body remembers sequences the mind can't articulate. Ise Jingū in Japan has been ritually rebuilt on an adjacent site at roughly twenty-year intervals for many centuries, the most recent rebuilding completed in 2013. The stated purpose is religious. A consequence, whether intended or not, is that the specialist carpentry involved gets transmitted from one generation of craftsmen to the next by doing it, on a cycle short enough that nobody who knows how ever dies before teaching someone.
Some researchers have argued that oral traditions can preserve information across far longer spans than historians once assumed. Patrick Nunn and Nicholas Reid proposed in 2016 that certain Aboriginal Australian accounts of coastlines that are now underwater may encode sea level rise from the end of the last glacial period, which would put the transmission in the range of thousands of years. The claim is contested and hard to test. It's also not absurd, given what oral form is capable of when it's structured for endurance.
The cost is doing something
Traditions tend to be expensive, and the expense looks like waste until you ask what it's for.
Fasting, feasting, dietary restriction, distinctive dress, long ceremonies, pilgrimage, days of lost labour, elaborate gifts. All of it consumes time and resources that could obviously be spent on something else. Economists find this puzzling; anthropologists mostly don't.
The costly signalling account says that a requirement which is hard to fake is a reliable signal of commitment, and reliable signals of commitment are what let strangers cooperate. Anyone can say they belong. Only someone who actually intends to stay will absorb the cost of the practice year after year.
Richard Sosis and Eric Bressler tested this against an unusual data set: nineteenth-century American communes, both religious and secular, with records of how long each survived. The religious communes lasted substantially longer. More pointedly, among the religious ones, the number of costly requirements imposed on members predicted longevity — the more demanding the group, the longer it held together. The same relationship didn't appear among the secular communes, which suggests the mechanism has something to do with requirements being unquestionable rather than merely burdensome.
Which explains something that otherwise makes no sense. Groups that demand more of people often retain them better than groups that demand less.
Rare and overwhelming, or frequent and dull
Harvey Whitehouse noticed that ritual practices around the world cluster into two very different patterns, and that each is suited to a different kind of group.
One mode is high-frequency and low-arousal. Daily or weekly observance, repetitive, calm, verbally rich, easy to perform correctly. Repetition builds the content into ordinary semantic memory, the kind of knowing that doesn't come attached to any particular occasion. This scales enormously well — it can hold together a group far too large for everyone to know each other, because the shared material is standardised and transmissible.
The other mode is low-frequency and high-arousal. Rare, intense, sometimes frightening or painful, often involving ordeal or seclusion, and undergone with a specific set of people. Emotionally charged events get remembered episodically, with unusual vividness and detail, and the people who went through it together are bound to each other rather than to an abstraction. This doesn't scale at all. It produces small, fiercely committed groups.
Most large traditions run the first mode with occasional bursts of the second — an ordinary weekly rhythm punctuated by a once-in-a-lifetime rite. That combination gets you both reach and intensity, and it's noticeably common across otherwise unrelated cultures.
Children copy the parts that don't work
There's an experimental finding that explains more about tradition than any amount of theory, and it involves a puzzle box.
Show a young chimpanzee a sequence of actions on a box that yields a reward — tap the top three times, poke a stick in a hole, then open a flap. If the box is opaque, the chimp copies the lot. If the box is transparent, so the animal can see that the tapping and the poking do nothing mechanically, it skips the useless steps and goes straight for the flap.
Human children, tested the same way, copy everything. Even with the box transparent. Even when the pointlessness of the extra actions is plainly visible. Andrew Whiten and Victoria Horner reported this contrast in 2005, and Derek Lyons and colleagues followed it up in 2007, finding that children kept reproducing the redundant steps even when they were explicitly told to skip anything unnecessary.
Over-imitation looks like a bug. It's much more likely to be the central feature.
A child copying only the steps it understands would be limited to knowledge it can personally verify, which for most of the technology humans actually use is nothing at all. Detoxifying certain foods requires a sequence in which several steps produce no visible effect and skipping any of them can be lethal. Whoever copies the whole procedure without demanding an explanation survives. Whoever insists on understanding first has to rediscover the entire thing from scratch, and won't.
So humans are built to acquire procedures wholesale, including the bits with no function. That's how a technique survives a thousand years without anyone knowing why it works. It's also, unavoidably, how the decorative and the useless survive right alongside it, because the mechanism can't tell the two apart.
Uncertainty manufactures ritual
Bronisław Malinowski noticed something in the Trobriand Islands in the 1910s that has been rediscovered repeatedly ever since.
Fishing inside the lagoon was safe, predictable and productive, and it was done with technique and no ceremony. Fishing in the open sea beyond the reef was dangerous and its results unpredictable, and it was surrounded by elaborate magical practice. The ritual didn't appear where the work was hard. It appeared where the outcome was uncertain and beyond control.
George Gmelch applied the same lens to American baseball in 1971 and found the pattern intact. Batting and pitching, where outcomes are chancy no matter how good you are, attract dense personal ritual — routines, lucky items, sequences performed identically before every attempt. Fielding, where a competent professional succeeds almost every time, attracts almost none.
Any occupation with high stakes and imperfect control develops the same growth. Sailors, farmers, soldiers, surgeons, actors, gamblers, students before examinations.
Whether the ritual helps is a separate question from why it exists, and the answer is probably that it sometimes does — not by influencing the outcome, but by lowering the anxiety of the person about to perform, which is a real effect on a real variable.
The traditions that mark thresholds
Arnold van Gennep published Les rites de passage in 1909 and identified a structure that turns up nearly everywhere, in practices with no historical contact between them.
Ceremonies marking a change of status — birth, adulthood, marriage, ordination, death — tend to have three phases. Separation, in which the person is removed from their former position, often physically and often with some marking of the body or clothing. A middle phase, in which they belong to neither the old category nor the new one. And incorporation, in which they're returned to the community under the new status, usually with something public and irreversible.
Victor Turner spent much of his career on the middle phase, which he called liminality, and argued that its strangeness is functional. Ordinary rules are suspended. Hierarchy flattens. People going through it together develop an intense levelling solidarity he named communitas, which is why cohorts who endured something odd together remain bonded for decades afterwards.
The reason this structure is so widespread is probably that social status is a shared fiction requiring shared updating. A private change of category doesn't stick. Everyone has to witness the transition, at the same time, in a form nobody can later claim to have missed. Ritual is how a group changes its collective mind on the record.
What happens when the reason goes
G. K. Chesterton set out the useful principle in 1929, and it's usually cited under the name of the fence.
You find a fence across a field with no obvious purpose. The impatient reformer says it's pointless and pulls it down. Chesterton's answer is that you don't get to remove it until you can explain why it was put there — not because it's necessarily still needed, but because you can't judge that until you know the original problem.
Plenty of traditions are fences whose fields have changed. Some encoded a genuine solution to a problem that no longer exists. Some encoded a solution to a problem that still exists but is now handled another way. Some were never a solution to anything.
Separating those categories after the fact is much harder than people assume, and the folk explanations offered for surviving practices are wrong more often than they're right. The story that standard railway gauge derives from the wheel spacing of Roman chariots is a lovely example — it circulates endlessly and it isn't true. The measurement comes from the wagonways of northern English collieries, carried into railway practice by the engineers who came from that world.
Once a practice is established, ordinary lock-in keeps it in place regardless of merit. Coordinated behaviour is costly to change even when everyone agrees the current arrangement is imperfect, because the cost of switching falls on whoever moves first and the benefit only arrives when everyone has moved. Which side of the road to drive on is the cleanest illustration. There's no argument that one side is better. There's an overwhelming argument for everyone doing the same thing.
They change constantly and insist they haven't
Living traditions are not stable, and the claim of stability is itself part of the practice.
Recipes drift with available ingredients. Music adopts new instruments. Ceremonies shorten as attention spans and working hours change. Dress adapts to available fabric. Practitioners absorb these changes and continue to describe what they're doing as unchanged, and they're not being dishonest — the continuity they're asserting is of identity and meaning, not of detail.
Jaroslav Pelikan drew the distinction as neatly as anyone: tradition is the living faith of the dead, and traditionalism is the dead faith of the living. A practice that's still being adapted by people who care about it is in a different condition from one being preserved exactly because nobody wants to be blamed for altering it.
Which raises a real problem for preservation efforts. UNESCO's convention on intangible cultural heritage, adopted in 2003, exists to protect living practices — performance, craft, ritual, oral tradition — rather than buildings and objects. It has done a great deal of good, and it carries a hazard that its own practitioners discuss openly: documenting a practice tends to fix it. The recorded version becomes the authoritative one, variants get treated as errors, and a thing that survived for centuries by quietly changing acquires an official form it now has to match.
The English folk revival ran into this a century earlier. Collectors from the early 1900s onwards travelled the countryside writing down songs and dances that were on the point of disappearing, and without them a huge amount would simply have been lost. They also selected, tidied and occasionally bowdlerised, and the versions they published became the versions everybody learned. What got saved was real, and it was a snapshot presented as the thing itself.
That tension doesn't resolve. A tradition needs enough rigidity to stay recognisable and enough slack to keep being worth doing, and no one has ever found the correct setting. Most of them are still adjusting it, in public, right now, under the firm impression that nothing has changed at all.
