Why do we laugh: most laughter has no joke

When did you last laugh out loud entirely on your own? Not a small smile at a video on your phone, but a real laugh, in an empty room, with no one around. Most people search their memory longer than they expect to.

The psychologist Robert Provine turned the study of laughter upside down simply by choosing where to do it. Instead of a laboratory he sent his team into shopping malls, university corridors and bars to record who laughed, when, and after what. He summed the results up in his 2000 book Laughter, and one figure is worth keeping: laughter is roughly thirty times more likely in company than alone.

Laughter belongs to the sentences between the jokes

The second finding is stranger still. Provine's team logged more than a thousand moments when somebody laughed and wrote down the line immediately before each. Jokes, punchlines and funny stories made up a minority. The overwhelming majority followed perfectly ordinary remarks like "see you around", "do I know you?" or "yeah, that'll be him".

Tom and Sarah reach the elevator at the same second on a Friday morning, both holding a coffee cup. Sarah says "you were up this early as well?" and both of them laugh. Nobody said anything funny. The laugh is not there by accident either: without it, the line would land like a complaint.

The same recordings produced one more regularity. Laughter in speech almost never lands in the middle of a sentence. It arrives at the ends, in the places where a written text would put a comma or a period, as though it had its own reserved slot in the running order. Talking and laughing do not compete, they take turns.

Look at laughter that way and it stops being a response to comedy and becomes a conversational tool. A signal that says nothing here needs defending and that you can both keep going. Research also suggests that the person telling the story tends to laugh more than the people listening. If laughter measured how funny a line was, that would make no sense.

The same category covers what anyone who has sat in a meeting with their boss has noticed. A quip from whoever makes the decisions gets a laugh out of almost everybody, while the same quip from the new hire gets very few. Laughter here is not measuring comedy, it is measuring relationships.

What the body is doing meanwhile

Laughter starts as an interrupted out-breath. The diaphragm and the muscles between the ribs push air out in short bursts and the vocal folds give each burst a sound, producing a run of syllables faster than you could repeat on purpose. That is why laughter is so hard to fake. Ordinary speech runs on a different engine.

The rest of the body joins in. Heart rate and blood pressure climb, the muscles of the face and abdomen get involved, and in hard laughter the eyes fill and muscle tone drops far enough that people have to lean on something. Afterwards comes a dip below the starting level, that loose-limbed stretch when nobody feels like getting up.

No single place is in charge of that sequence. During brain surgery, stimulating a particular patch of the frontal lobe has repeatedly set off laughter when there was nothing to laugh at, and the patient promptly invented a reason for it: the picture on the wall was funny, or the surgeon's face was. The body laughed first and the explanation caught up behind it.

The most interesting piece came from the anthropologist Robin Dunbar and his colleagues in 2011. They had people watch either a comedy video or a neutral documentary, once alone and once in a group, and measured before and after how much discomfort they could take. The ones who had laughed together lasted longer. The explanation points toward endorphins, the body's own chemistry for damping pain, which also comes with the feeling of closeness.

From there it is a short step to the idea that laughter heals. It is not that simple, and the line between the documented and the invented gets a text of its own on how laughter and health fit together. For this article a more modest note will do: the effect showed up in shared laughter, not in laughter on your own.

Why laughter is contagious

The sound of laughter does not behave in the brain like an ordinary sound. Hearing it activates not only auditory areas but motor ones as well, the ones that run the facial muscles, as though the body were getting ready to join in. Hence that awkward moment when a group of students bursts out laughing on the bus and the corner of your mouth twitches even though you have no idea why.

Television has been using this for decades. A laugh track under a sitcom works even on a weak sketch, and research repeatedly finds that viewers rate identical material as funnier when laughter is running underneath it. The recording persuades nobody of anything. It simply signals that others are laughing, and the body acts on that before reason gets a say.

Contagion also has a form that nobody would think to laugh at. In 1962, fits of laughter spread through a girls' school in what was then Tanganyika far enough that the school closed for weeks and the fits carried over into nearby villages. There was nothing cheerful about it. It was a mass reaction in a group under strain.

The spread follows a rule. A group gets going more easily when the first sound comes from someone the others rate, and it stays quiet when the same laugh in the same room comes from an outsider. Laughter travels along the same lines as trust, so who joins in and who does not says a fair amount about a group without a word being spoken.

On a smaller scale you meet this at a funeral, in an exam or in a meeting that has broken down, where one person snorting is enough to finish everybody off. Laughter spreads at moments like that not because the situation is funny but because the group needs to let the tension out, and laughter is the fastest valve within reach.

Real laughter and the other kind

Everybody knows laughter can be faked. The more interesting question is why it can only be faked halfway. The French physician Guillaume Duchenne spent the nineteenth century stimulating facial muscles with electrodes and described two separate muscle groups behind a genuinely delighted expression: one lifts the corners of the mouth, the other tightens the area around the eyes. The first obeys your will, the second joins in only reluctantly on command. Paul Ekman picked the finding up a century later, and the term Duchenne smile stuck.

Polite laughter deserves neither contempt nor scare quotes. It is a fully functional social tool for showing that you are on the same side, and people who never use it come across as remote. The difference between the two lies elsewhere: one happens to you, the other is sent.

Decoding it in other people is trickier than it looks and rarely produces anything useful. The more interesting move is to turn it on yourself: when did you last laugh purely to keep a conversation moving? That kind of laugh almost never registers, because it costs nothing, and that is precisely its job.

The ear is a better judge here than the eye. From sound alone, people separate acted from spontaneous laughter surprisingly often, though not every time, and they do best with people they know well. Sarah can tell over the phone whether Tom is laughing because of her or because he wants the call to move along.

Why you cannot tickle yourself

Try running your fingers along your own ribs now in a way that makes you laugh. It will not work, and the brain is specifically to blame. For every movement you make yourself, the cerebellum works out in advance what sensation it will cause and damps that prediction down, so that in the flood of signals from your own body you can still tell what is arriving from outside. Your own hand cannot surprise you in principle.

Tickling therefore needs a second person, and one you trust at that. A stranger's hand or a bad mood turns it into an unpleasantness within a second. That condition is a clue to where laughter most likely came from.

A play signal older than speech

Chimpanzees and other great apes make a peculiar panting sound during rough play, chases and mock fights. It sounds nothing like human laughter, yet it appears in exactly the same situations: when two animals go at each other for show. The sound carries a single message. This is not an attack, this is a game.

That it is equipment rather than a learned habit is given away by children. They start laughing somewhere around the third or fourth month, long before the first word, and children born deaf and blind laugh too, so they cannot have copied it from anyone. Speech takes years to learn. Laughter we can do before we can say anything at all.

Laughter as a safety signal explains plenty of small things that otherwise make no sense. Why we erupt when somebody trips and gets straight back up. Or why a group laughs harder after a near miss than after a well-told joke. Among people you feel uneasy with, hardly anyone laughs, even though the jokes are exactly the same ones.

One question is left that this text has been walking around. If most laughter is not aimed at comedy, what separates people who laugh differently? The answer is not in what strikes whom as funny, but in what each of them uses humor for. One person joins people together with it, another keeps their own troubles at arm's length, a third enjoys a well-placed jab, and a fourth makes a target of themselves. The psychologist Rod Martin and his team set that four-way split out in 2003, and it is written up in the overview of humor styles.

None of the four is better or worse. They are habits rather than verdicts, and they shift depending on who you are sitting at the table with. Which one your laughter tends to trigger is what the short humor styles test is for; treat the result as a starting point for your own observation, not as a label.

One more thing to take home. Tomorrow, spend the whole day noticing what actually makes you laugh and jot each one down in your phone in a word or two. In the evening, go through the list and count how many times a joke appears on it. Most people stop at that number and read back through what they wrote down.