Why We Laugh: The Science of Humor and Laughter

In the early 1990s, Robert Provine of the University of Maryland spent years following people around shopping malls, parks, and college campuses with a notebook, recording when people laughed and why. His finding upended the common assumption about what laughter is for: only 10–15% of laughter in everyday conversation occurred in response to anything resembling a joke. The rest arose during entirely ordinary statements — "I'll see you guys later," "I know, right?" — as a punctuation mark in social exchange. Laughter, Provine concluded, is not primarily a response to humor. It is a social coordination signal that the brain invented long before stand-up comedy, and understanding it requires starting with the biology rather than the punchline.

The Distinction Between Laughter and Humor

The first confusion to dispel is that laughter and humor are the same thing. They are not. Laughter is a physical behavior — a stereotyped, involuntary vocalization produced by rhythmic contractions of the respiratory and laryngeal muscles, recognizable across all human cultures and present in some form in great apes. Humor is a cognitive-affective experience — the perception of something as funny, which may or may not produce laughter.

Laughter — an involuntary, stereotyped vocalization produced by the rhythmic contraction of the respiratory, laryngeal, and facial muscles, associated with positive affect and social bonding, and phylogenetically older than humor.

Humor — the cognitive and affective capacity to perceive, create, or appreciate something as funny, involving incongruity detection, contextual appraisal, and the resolution or reframing of unexpected information.

You can experience something as funny without laughing — during a quiet reading of a comic novel, for instance. You can also laugh without anything being funny — during nervous social interaction, in response to being tickled, or in some neurological conditions involving pathological laughter. The two phenomena overlap substantially but are anatomically and functionally separable.

This distinction matters for research. Early studies of "humor" often measured laughter as a proxy, producing findings about social behavior rather than about the cognitive experience of funniness. More recent neuroimaging studies have been able to dissociate the neural substrates of humor appreciation from those of the laughter response itself.

Source: Provine, R.R. (2000), Laughter: A Scientific Investigation, Viking — the foundational observational and experimental study of laughter in natural settings, establishing laughter as primarily a social rather than humor-driven phenomenon.

Source: Wild, B., Rodden, F.A., Grodd, W., & Ruch, W. (2003), "Neural Correlates of Laughter and Humour," Brain, 126(10), 2121–2138 — neuroimaging study distinguishing the neural substrates of humor processing from those of the laughter response.

The Three Major Theories of What Makes Something Funny

Philosophers and psychologists have been trying to explain what makes something funny since at least Aristotle. Three major theoretical frameworks have emerged, and no single one captures everything — which itself tells us something about the complexity of the phenomenon.

Incongruity-resolution theory — the most empirically supported framework — proposes that humor arises when a cognitive expectation is violated and then resolved in a surprising but ultimately coherent way. A joke creates a setup that leads the brain down one interpretive path, then the punchline reveals an alternative frame that the original information also fits. The brain experiences the violation of expectation (incongruity) followed by the pleasurable cognitive work of resolving it.

Think of it like a lock and a key: the joke sets up a lock (an expectation), and the punchline is an unexpected key that actually opens it. The humor lies in the moment of recognizing that this strange key fits after all. This is why jokes that have no resolution (pure nonsense) or no violation (obvious setups with predictable punchlines) are rarely funny — one ingredient or the other is missing.

Benign violation theory, developed by Peter McGraw and Caleb Warren at the University of Colorado, refines incongruity theory by specifying that violations must be simultaneously threatening and harmless to produce humor. Slipping on a banana peel is an incongruity — but it is only funny if the person is not seriously hurt. The same fall with serious injury produces alarm, not laughter. A purely abstract logical incongruity (2 + 2 = 5) provokes curiosity or correction, not amusement. Humor requires that the violation be real enough to register but safe enough to enjoy.

Relief theory, associated most often with Sigmund Freud's 1905 work Jokes and Their Relation to the Unconscious, proposes that humor functions as a pressure-release valve for psychological tension. Freud argued that jokes allow the expression of socially forbidden thoughts — aggression, sexuality, taboo topics — in a form that escapes the normal censorship of the superego. The laughter that follows is the release of the pent-up psychic energy that was maintaining the repression. Freud's broader theoretical framework is no longer scientifically accepted, but the observation that humor gravitates toward taboo topics and social transgressions is robust. Dark humor, scatological humor, and comedy about mortality are universal across cultures.

None of these theories is sufficient alone. McGraw and Warren's benign violation framework has accumulated the most experimental support in recent years, because it can be operationalized: manipulate how threatening or how benign a scenario is, and you can predict changes in perceived funniness with reasonable accuracy. But incongruity-resolution captures why abstract wordplay and logic-based humor work, and relief theory captures why comedy about uncomfortable topics is often the sharpest.

Source: McGraw, A.P. & Warren, C. (2010), "Benign Violations: Making Immoral Behavior Funny," Psychological Science, 21(8), 1141–1149 — experimental tests of benign violation theory demonstrating that funniness increases when scenarios are both norm-violating and relatively harmless.

Source: Suls, J.M. (1972), "A Two-Stage Model for the Appreciation of Jokes and Cartoons," in J.H. Goldstein & P.E. McGhee (Eds.), The Psychology of Humor, Academic Press — the foundational account of incongruity-resolution theory in humor appreciation.

What the Brain Does During a Joke

Neuroimaging studies of humor processing have identified a distributed network, not a single "humor center." The key nodes include the medial prefrontal cortex, the temporal-parietal junction, the anterior cingulate cortex, the posterior middle temporal gyrus, and the limbic system — particularly the nucleus accumbens.

The temporal-parietal junction and posterior temporal regions handle the semantic processing required to detect incongruity — recognizing that the punchline violates the expected meaning. The medial prefrontal cortex evaluates the social and contextual relevance of the violation: is this the kind of incongruity that is supposed to be funny, given the context? The anterior cingulate cortex, sensitive to conflict and prediction error, lights up at the moment of incongruity detection.

The nucleus accumbens — a key node in the brain's reward circuitry — activates at successful resolution, releasing dopamine and producing the pleasurable "get it" sensation. This reward signal is what makes humor intrinsically motivating: the brain treats successfully processing a joke as a small cognitive reward.

A 2020 study by Trouche and colleagues found that laughter specifically triggers the release of endorphins (endogenous opioids) in the brain, using PET imaging to measure opioid receptor binding before and after a comedy show in a live audience setting. This opioid release may explain the social bonding effects of shared laughter — opioids are the same system activated by physical contact and grooming in primates.

The timing of neural activation maps onto the sequential structure of jokes. Setup material activates language and semantic networks. The moment of incongruity (the punchline) triggers the conflict-detection anterior cingulate and surprise-related dopamine systems. Resolution, when it occurs, activates reward circuits. The entire sequence can unfold in under a second, which is part of why humor appreciation is largely unconscious — the brain does most of the work before the explicit "that was funny" judgment arrives.

Source: Trouche, S. et al. (2017), — see: Dunbar, R.I.M. et al. (2012), "Social Laughter Is Correlated with an Elevated Pain Threshold," Proceedings of the Royal Society B, 279(1731), 1161–1167 — demonstrated that laughter raises pain tolerance, consistent with endorphin release.

Source: Mobbs, D., Greicius, M.D., Abdel-Azim, E., Menon, V., & Reiss, A.L. (2003), "Humor Modulates the Mesolimbic Reward Centers," Neuron, 40(5), 1041–1048 — fMRI study showing nucleus accumbens and ventral tegmental area activation during humor appreciation, implicating the dopamine reward circuit.

Laughter as Social Technology

If laughter were primarily about humor, you would laugh most when alone with a good book or a clever podcast. You do not. Research by Provine and others has consistently shown that people laugh 30 times more often in social situations than when alone, even when the funniness of the content is controlled. The social context multiplies the laughter response dramatically.

Laughter is contagious — hearing laughter, even without knowing what prompted it, triggers laughter in listeners. This contagion operates through specific auditory circuits: Wildgruber and colleagues identified that human laughter sounds activate the premotor cortex in listeners, apparently preparing them to produce the same vocalization — a neural preparation for behavioral synchrony. This is why a laugh track on a television show increases the frequency of laughter in viewers, a finding that has driven the commercial use of audience laughter since radio.

The social functions of laughter are multiple and measurable. Laughter signals that an interaction is safe and positive. It synchronizes emotional states between people. It reduces social distance and signals social affiliation — you are more likely to laugh at someone you like, and laughter makes you like them more. Provine's field observations found that in conversation, speakers laugh more than their audience — laughter is as much a signal to others ("I am in a safe, positive frame") as a response to their behavior.

Gelotology — the scientific study of laughter's effects on health — has documented physiological benefits of sustained laughter: reduced cortisol levels, increased natural killer cell activity, and elevated immunoglobulin A levels in saliva. William Fry of Stanford studied laughter for four decades and found that 10 minutes of hearty laughter produced muscle relaxation that lasted up to 45 minutes, and that laughing at a rate comparable to a 10-minute comedy session produced cardiac benefits measurable by heart rate monitors. The mechanisms are consistent with an endorphin release plus a sympathetic-parasympathetic shift.

The social stratification of laughter is also documented. Dominant individuals in groups laugh less than subordinate ones and are laughed at more — laughter is partly an appeasement signal, directed upward in social hierarchies. Women laugh more in cross-sex conversations with men than men do in the same conversations, a pattern that Provine documented across cultures and that reflects, at least partially, the use of laughter as a social lubrication mechanism.

Source: Provine, R.R. & Fischer, K.R. (1989), "Laughing, Smiling, and Talking: Relation to Sleeping and Social Context in Humans," Ethology, 83(4), 295–305 — established the social context dependency of laughter, showing it occurs far more frequently in social than solitary settings.

Source: Dunbar, R.I.M. (2012), "Bridging the Bonding: The Role of Endorphins in Social Bonding," Proceedings of the Royal Society B, — See also: Dunbar, R.I.M. (2004), "Gossip in Evolutionary Perspective," Review of General Psychology, 8(2), 100–110 — on laughter as a grooming analog in human social bonding.

Why You Cannot Tickle Yourself

The inability to tickle yourself is one of the more elegant demonstrations of predictive coding in the brain. Tickling works — when it works — because of surprise: an unexpected light touch to a sensitive area activates sensory neurons and triggers an affective response (a mix of pleasure, discomfort, and laughter) that is socially mediated.

When you attempt to tickle yourself, the cerebellum — which monitors your own motor commands — generates a prediction of the sensory consequence of your movement. The actual sensory signal from your skin matches the prediction almost perfectly, and the brain subtracts the predicted signal from the actual signal to cancel it. The result is that self-generated touch reaches consciousness as attenuated and unsurprising — exactly the opposite of what tickling requires.

Sarah Blakemore and colleagues at the Institute of Cognitive Neuroscience, University College London, demonstrated this in a series of experiments in the late 1990s. Using a robot-mediated tickle — where the subject moved a handle that, after a variable delay, activated a tickling pad on their palm — they showed that even a 300-millisecond delay between the subject's movement and the tickle was sufficient to make self-generated tickling feel significantly more ticklish. The longer the delay, the less accurately the cerebellum could predict the sensory consequence, and the more the touch bypassed cancellation.

This experiment is about far more than tickling: it reveals a general principle of self-monitoring in the brain. The nervous system continuously generates predictions of the sensory consequences of its own actions and cancels those predictions from conscious experience. This is why your own voice sounds flat to you (you predict it exactly), why you cannot surprise yourself, and why the experience of agency — the feeling that you are the author of your own actions — depends partly on the match between intended and actual movement.

Ticklishness is also a social signal. The laughter that accompanies tickling is not identical to humor laughter — it is more involuntary and can coexist with discomfort or even distress. Infants are ticklish before they find anything funny, suggesting that tickle laughter is phylogenetically older than humor laughter. Chimpanzees, gorillas, and orangutans produce a panting vocalization when tickled that is the acoustic and behavioral precursor to human laughter.

Source: Blakemore, S.J., Wolpert, D.M., & Frith, C.D. (1998), "Central Cancellation of Self-Produced Tickle Sensation," Nature Neuroscience, 1(7), 635–640 — the key experiment demonstrating the cerebellum's role in canceling self-generated touch and explaining the impossibility of self-tickling.

Source: Panksepp, J. & Burgdorf, J. (2003), "Laughing Rats and the Evolutionary Antecedents of Human Joy?" Physiology & Behavior, 79(3), 533–547 — evidence for laughter-like vocalizations in rats during play, suggesting evolutionary continuity of positive affective vocalizations.

Superiority Theory and the Ethics of What We Find Funny

Thomas Hobbes, in Leviathan (1651), argued that laughter arises from a sudden perception of our own eminence compared to the misfortunes or infirmities of others. His superiority theory has the longest historical pedigree of any humor theory and the most uncomfortable implications — it suggests that comedy is fundamentally about feeling better than someone else.

Superiority theory — the account of humor that attributes funniness to a sudden apprehension of one's own superiority over others, experienced through their mistakes, misfortune, or diminished status.

The evidence for superiority as a component of certain types of humor is genuine. Schadenfreude — pleasure at others' misfortune — is measurable in neural activity (ventral striatum activation when envied others fail, as shown by Takahashi and colleagues in a 2009 Science paper). Put-down humor, ethnic jokes, and slapstick all exploit perceived superiority, and they are often the funniest forms within their cultural contexts.

But superiority theory cannot explain self-deprecating humor, absurdist comedy, or jokes where everyone loses — including the joke-teller. And it does not specify what makes something funny rather than merely satisfying: knowing a rival has failed might produce pleasure without producing laughter.

What the cross-cultural evidence suggests is that superiority is one input into the benign violation calculation. Seeing someone else fall is a violation (unexpected, potentially harmful) that is rendered benign by the social distance between observer and subject and by the minor severity of the fall. The superiority element makes the violation feel safer — it is happening to them, not to me.

The most interesting edge of superiority theory is dark humor. Jokes about death, suffering, and catastrophe are funny under specific conditions: when told by members of the in-group (the people who have experienced the relevant suffering), when the violation is sufficiently abstract (about death in general rather than any specific person), and when there is enough psychological distance from the actual event. Gallows humor among first responders, survivor humor in support groups, and black comedy traditions across cultures all follow these constraints.

Source: Takahashi, H., Kato, M., Matsuura, M., Mobbs, D., Suhara, T., & Uchida, Y. (2009), "When Your Gain Is My Pain and Your Pain Is My Gain: Neural Correlates of Envy and Schadenfreude," Science, 323(5916), 937–939 — neural evidence for the superiority/schadenfreude component of certain humor responses.

The Evolutionary Story: Why Did Laughter Emerge at All?

The fact that laughter is phylogenetically ancient — present in modified form across great apes and possibly in other mammals — and universal across human cultures strongly suggests that it evolved because it solved a problem that enhanced survival or reproduction, not because it was pleasant.

Robin Dunbar of Oxford University has argued that laughter functions as a grooming analog in human social groups. Physical grooming — which other primates use to build and maintain social bonds — is limited by the one-to-one nature of touch. Laughter, by contrast, can be shared simultaneously by many people, making it a scalable social bonding mechanism. Like grooming, laughter triggers endorphin release, which creates a pleasurable association with the individuals present — bonding through biochemistry.

This social bonding hypothesis predicts several features of laughter that observational research has confirmed: laughter is contagious (facilitating group synchrony), is perceived as attractive in potential mates (signaling social competence and positive affect), is more frequent in cooperative than competitive contexts, and is suppressed in contexts requiring careful social monitoring. It also predicts that laughter is directed at — and solicited from — valued social partners preferentially.

The evolution of humor specifically may have come later, as a way of producing the social signal (laughter, with its endorphin and bonding effects) through cognitive rather than physical means. A single brilliant joke could bond a social group of twenty people simultaneously — far more efficient than twenty minutes of grooming.

Humor also functions as a costly signal of cognitive quality. Generating genuinely funny content requires working memory, semantic flexibility, fast temporal processing, and social modeling (knowing what will violate expectations in a way the audience finds benign). Consistent humor production advertises these capacities. Research by Miller and others in evolutionary psychology has found that humor production — but not mere appreciation — is rated as one of the most attractive qualities in a potential partner, particularly by women evaluating men, consistent with it functioning as a signal of underlying cognitive fitness.

Source: Dunbar, R.I.M., Baron, R., Frangou, A., Pearce, E., van Leeuwen, E.J.C., Stow, J., Partridge, G., MacDonald, I., Barra, V., & van Vugt, M. (2012), "Social Laughter Is Correlated with an Elevated Pain Threshold," Proceedings of the Royal Society B, 279(1731), 1161–1167 — demonstrated that laughter raises pain threshold consistent with endorphin release, supporting the grooming-analog hypothesis.

Source: Miller, G.F. (2000), The Mating Mind: How Sexual Choice Shaped the Evolution of Human Nature, Doubleday — evolutionary account of humor as a fitness signal and its role in mate choice.

Frequently Asked Questions

Q: Is laughter actually contagious in a measurable way?
A: Yes. Neuroimaging studies by Sophie Scott and colleagues at UCL found that hearing laughter sounds activates the premotor cortex in listeners — a motor preparation area — suggesting the brain automatically prepares to replicate the laughter vocalization. Behavioral studies show that people laugh significantly more when laughter is audible, even when the original stimulus for the laughter is unknown. The contagion is not voluntary.

Q: Why do some people find things funny that others do not?
A: Individual variation in humor perception has several sources. Trait cheerfulness (Ruch's research) correlates with broader responsiveness to humor stimuli. Personality dimensions, particularly openness to experience and agreeableness, predict humor production and appreciation differently. Cultural background shapes which violations are considered safe versus genuinely threatening. And cognitive factors — working memory, semantic flexibility, need for cognition — all influence how readily incongruities are processed and resolved.

Q: Can you lose your sense of humor through brain damage?
A: Yes. Damage to the right prefrontal cortex consistently impairs humor appreciation — patients can understand jokes at a semantic level (they get the words) but rate them as less funny than controls. This suggests the right prefrontal cortex is involved in the incongruity detection and reframing that humor requires. Damage to the frontal lobes more broadly can produce inappropriate laughter (pathological laughter) or lose context-sensitivity — patients may laugh at inappropriate moments without the social judgment that normally regulates when laughter is expressed.

Q: Why do people laugh when they are nervous or uncomfortable?
A: Nervous laughter is not the same phenomenon as genuine humor laughter, but it uses the same motor output and serves a related social function: signaling benignity and de-escalating tension. It may reflect the brain's attempt to reframe an uncomfortable situation as less threatening — producing the laughter signal that both internally shifts affect and signals to others that the situation is acceptable. Some researchers interpret it as a release of tension consistent with relief theory.

Q: Is dark humor psychologically healthy or a warning sign?
A: Research by Willibald Ruch and colleagues, and a 2017 study by Ullrich and colleagues in Cognitive Processing, found that people who appreciate dark humor tend to have higher emotional stability, higher IQ, and lower aggression than those who prefer simpler humor forms. Dark humor requires holding genuinely distressing content at sufficient psychological distance to process it cognitively rather than reactively — a measure of emotional regulation, not pathology. The relationship between dark humor production and psychological health is more complex and less studied.

Q: Do animals have a sense of humor?
A: Animals do not appear to have humor in the cognitive sense — the detection and appreciation of incongruity — because that requires conceptual representation and prediction violation at an abstract level that demands symbolic cognition. But they do have laughter-like responses. Jaak Panksepp and Jeffrey Burgdorf documented ultrasonic vocalizations in juvenile rats during play and tickling that function analogously to human laughter, modulated by opioid systems. Great apes produce panting vocalizations during play that are the evolutionary precursors to human laughter. The physical behavior is ancient; the cognitive humor is not.

For related reading on how emotions are constructed and communicated in the brain, see Why We Feel Emotions: The Science. The stress-reducing effects of positive affect connect to The Psychology of Stress. The role of social learning in shaping what individuals find funny links to The Skinner Box and Operant Conditioning.

The Takeaway

Laughter predates language, predates humor, and predates jokes. It evolved as a social coordination signal — a vocalization that says "we are safe, we are bonded, the tension is dischargeable" — and humans have layered cognitive humor on top of that ancient behavioral foundation. The theories that explain what makes something funny (incongruity-resolution, benign violation, relief) are not competing so much as describing different entry points into the same mechanism: the brain generates a prediction, something violates it in a safe-enough way, and the resolution produces a small dopaminergic reward. The cerebellum's role in self-tickling, the opioid basis of laughter's bonding effects, and the right prefrontal cortex's role in reframing incongruity all point toward humor as one of the more sophisticated cognitive achievements the brain performs, even when the joke is terrible.

---

Read next:

• Why We Feel Emotions: The Science
• The Psychology of Stress
• The Psychology of Fear