Why Do We Laugh? Mostly Not at Jokes

The findings were consistent and largely unexpected:

• •Only about ten to twenty percent of laughter followed anything resembling a joke or a formally funny remark
• •Speakers laughed roughly forty-six percent more often than their audience, which is the reverse of what a humour-response model predicts
• •Laughter almost never interrupts the structure of speech, arriving at the ends of phrases and clauses as punctuation does, which indicates it is governed by the speech system rather than competing with it
• •It is around thirty times more frequent in company than alone, which is the single strongest signal that its function is social
• •It is contagious, spreading through a group without anyone knowing why, which is the basis of the laugh track invented for American television in 1950
• •Women laugh more than men on average in mixed conversation, and both laugh more at men, a pattern that appears across cultures and in personal advertisements

Three explanations of what makes something funny have competed since antiquity. The superiority theory, held by Plato, Aristotle and Hobbes, says we laugh at others' misfortune or foolishness from a sense of our own advantage, which accounts for mockery, slapstick and a good deal of comedy and not for puns. The relief theory, associated with Herbert Spencer and Freud, holds that laughter discharges built-up nervous tension, which fits nervous laughter at funerals and the way taboo subjects generate humour, and which rests on a model of psychic energy nobody now accepts. The incongruity theory, favoured by most modern researchers and traceable to Kant and Schopenhauer, says humour arises from a mismatch between what is expected and what occurs, resolved in a way that makes unexpected sense, which is exactly the structure of a joke's setup and punchline. The refined version adds that the incongruity must be benign, since a violation that is genuinely threatening is not funny, which explains why the same event is a joke at a distance and a horror close up.

Laughter is a stereotyped vocalisation, a series of short vowel-like notes about seventy-five milliseconds apart, produced on the outbreath and largely outside voluntary control, which is why deliberate laughter sounds different and why people can usually tell. Two distinct pathways produce it. Spontaneous laughter runs through older subcortical structures including the brainstem and limbic system, while voluntary laughter is driven from the motor cortex, and brain imaging distinguishes them; listeners' brains also respond differently to the two, with more effort devoted to interpreting the forced kind. The dissociation is visible in neurological conditions, since damage in certain areas produces pathological laughing or crying with no corresponding feeling, and gelastic seizures produce laughter as a symptom. The physiological effects are modest and real: a drop in stress hormones, a rise in pain tolerance, which a 2011 study attributed to endorphin release during genuine social laughter, and short-term improvements in blood vessel function.

It is older than speech and not unique to humans. Chimpanzees, bonobos, gorillas and orangutans all produce a panting vocalisation during play and tickling, and acoustic analysis published in 2009 mapped these onto the primate family tree in a pattern matching known evolutionary relationships, indicating a common origin between ten and sixteen million years ago. Rats emit ultrasonic chirps at fifty kilohertz when tickled and when playing, will chase a hand that tickles them, and show individual differences in how much they like it. The likely original function is a play signal, telling a partner that the rough behaviour under way is not an attack, which is why it is so tightly linked to tickling and chasing in both children and animals. Human babies laugh from around three or four months, before they can speak, and deaf and blind children laugh without having heard or seen it, which indicates it is built in rather than learned.

The current consensus treats it as a social tool with several jobs. It signals that an interaction is safe and playful. It builds and marks bonds, and studies of conversation find that listeners can tell from laughter alone whether two people are friends or strangers, across cultures and languages. It smooths awkwardness, which is why so much of it follows minor embarrassment. It signals agreement and inclusion, and its absence at the wrong moment is an unmistakable social message. Robin Dunbar has proposed that it functions as a low-cost form of grooming that can be directed at several people at once, allowing humans to maintain larger social groups than primates who must groom each other individually, and the endorphin release during shared laughter fits that account. None of this requires anything to be funny, which is why the shopping centre transcripts contained so few jokes and so much laughter.

Field recordings of natural conversation found that only a small minority of laughter follows anything joke-like, that speakers laugh more than their listeners, and that it is thirty times more common in company than alone, which makes it a social signal rather than a humour response. It is produced by two distinct brain pathways, spontaneous and voluntary, appears in great apes and in tickled rats, and probably began as a play signal meaning that rough behaviour is not an attack.