Smiling

Abstract

Smiling is a type of facial expression: an upward pull of the lip corners by the zygomaticus major, often with contraction of the orbicularis oculi around the eyes, seen in enjoyment and many social situations. Cognitive psychology studies it on three fronts: whether eye-muscle involvement distinguishes a genuine enjoyment smile from a posed one, a contrast named for the physiologist Duchenne; whether smiling feeds back on felt emotion; and whether a smile expresses emotion or signals to an audience. The facial feedback hypothesis, whose famous pen-in-mouth demonstration failed a registered replication, now names a small and conditional effect. The signaling account instead treats smiles as communicative tools with distinct reward, affiliation, and dominance functions. The article and three interactive demonstrations present the smile as a muscular configuration, a disputed index of emotion, and an instrument of social communication.

Keywords: smiling, Duchenne smile, facial feedback hypothesis, emotional expression, social signaling

A smile is the most familiar of human expressions and the one most readily recognized across cultures, yet it is also among the most theoretically contested. In its narrowest physical description it is the drawing up of the mouth corners by the zygomaticus major; in its fullest emotional form it adds the contraction of the orbicularis oculi, which raises the cheeks and crinkles the skin around the eyes. For cognitive psychology the smile raises three questions that pull in different directions. Does a particular morphology reliably mark genuine enjoyment, so that a smile can be read as honest or false? Does smiling itself help cause the feeling it expresses, so that arranging the face into a smile makes a person happier? And is a smile really an expression of emotion at all, or is it better understood as a communicative signal, emitted strategically for an audience and serving purposes that have little to do with felt happiness? The three sections that follow take these in turn, and the recurring theme is that a behavior everyone recognizes resists a single, simple account.

What Smiling Is

A smile is, mechanically, the action of a small set of facial muscles. The defining movement is contraction of the zygomaticus major, which runs from the cheekbone to the corner of the mouth and pulls the lip corners upward and outward. In the Facial Action Coding System this is Action Unit 12, the lip-corner puller. When a smile accompanies genuine enjoyment it is typically joined by a second movement, contraction of the orbicularis oculi (Action Unit 6), the ring of muscle around the eye socket that raises the cheeks, narrows the eye aperture, and produces the characteristic creases at the outer corners. Paul Ekman, Richard Davidson, and Wallace Friesen reported that smiles including this eye-muscle activity accompanied self-reported enjoyment and a distinctive pattern of brain activity, whereas smiles without it did not, and they revived Duchenne's name for the enjoyment smile (Ekman et al., 1990). The nineteenth-century French physiologist Guillaume Duchenne de Boulogne had used electrical stimulation of the facial muscles to argue that the muscle around the eye is moved by the sweet emotions of the soul and does not obey the will, so that its presence betrays a true feeling a feigned smile cannot counterfeit.

That last clause is where the simple story breaks down. The claim that the eye-muscle contraction is involuntary, and therefore an infallible marker of sincerity, turns out to be too strong. Eva Krumhuber and Antony Manstead found that a substantial proportion of people can produce the orbicularis oculi contraction deliberately when posing a smile, so the Duchenne marker discriminates genuine from posed smiles only probabilistically, not perfectly (Krumhuber & Manstead, 2009). A later reconsideration by Jeffrey Girard and colleagues formalized and tested the hypotheses bound up in the Duchenne-smile idea and concluded that eye constriction does co-vary with the intensity of positive emotion, but as a graded correlate rather than a categorical on-or-off signature of authenticity (Girard et al., 2021). The useful summary is that the eye-muscle marker raises the probability that a smile reflects real enjoyment without ever guaranteeing it. The first demonstration lets the reader add or remove the eye-muscle component and the lip-corner pull independently and see how the resulting configuration reads.

Does Smiling Cause Happiness

If the Duchenne account asks whether a smile reveals happiness, the facial feedback hypothesis asks the reverse: whether producing a smile causes happiness. The idea traces to Darwin and William James and holds that the sensory feedback from one's own facial muscles contributes to emotional experience, so that arranging the face into a smile should intensify felt positive emotion. Its most celebrated demonstration was devised by Fritz Strack, Leonard Martin, and Sabine Stepper, who had participants hold a pen either between the teeth, which covertly engages the smiling muscles, or between the lips, which prevents a smile, under the cover story of a task for disabled people. Those holding the pen in their teeth rated cartoons as funnier, an effect obtained without any participant knowing they were being made to smile (Strack et al., 1988). The pen-in-mouth study became a textbook illustration of embodied emotion.

It did not survive direct replication at scale. A Registered Replication Report coordinated by Eric-Jan Wagenmakers assembled seventeen laboratories running the pen procedure under a preregistered common protocol and found, in aggregate, no reliable difference in rated amusement between the teeth and lips conditions (Wagenmakers et al., 2016). The failure was one of the salient episodes of psychology's replication crisis. But a failed replication of one paradigm is not the same as a dead hypothesis, and the question was reopened more carefully. Nicholas Coles, Jeff Larsen, and Heather Lench conducted a meta-analysis of nearly a century of facial feedback research across many paradigms and concluded that the effect of facial feedback on emotional experience is real but small and variable, with the pen-in-mouth cartoon-rating design among the weaker and least reliable ways to elicit it (Coles et al., 2019). The matter was then put to a direct adversarial test: the Many Smiles Collaboration, with Coles and colleagues including the original proponents and critics, ran a preregistered multi-laboratory study whose design the disputants agreed on in advance, and found that posing a smile does produce a small increase in felt happiness when it is elicited by mimicry or direct instruction rather than by the pen (Coles et al., 2022). The second demonstration traces this arc, letting the reader see how the estimated effect size shifts from the original single study through the registered replication to the pooled modern estimate.

Smiles as Social Signals

The debates above share a hidden assumption: that a smile is fundamentally an expression of an inner emotional state, which it either reveals (the Duchenne question) or amplifies (the feedback question). An influential tradition rejects that premise. Alan Fridlund argued from a behavioral-ecology standpoint that facial displays are not read-outs of emotion but social tools, emitted to influence others, and that their occurrence tracks the social situation rather than private feeling. His key demonstration showed that people smile more while watching an amusing film when they believe a friend is co-viewing or will see the film, even when physically alone, so that smiling is potentiated by an implicit audience rather than driven by felt amusement alone (Fridlund, 1991). Smiling, on this view, is addressed to someone, actual or imagined.

A reconciling position treats smiles as embodied signals whose meaning observers recover by simulation. Paula Niedenthal's Simulation of Smiles model proposes that a perceiver interprets a smile partly by covertly mimicking it and reading the resulting feedback, so that the eye contact and facial mimicry exchanged during a smile are integral to decoding which kind of smile it is (Niedenthal et al., 2010). Building on this, Jared Martin, Niedenthal, and colleagues argued that smiles are best understood as multipurpose social signals serving several distinct functions rather than one (Martin et al., 2017). Magdalena Rychlowska and colleagues gave the functional taxonomy empirical form, identifying three physically distinguishable kinds: a reward smile that reinforces behavior, an affiliation smile that signals social bonds and nonthreat, and a dominance smile that conveys superiority or contempt, each with a characteristic morphology that observers reliably distinguish (Rychlowska et al., 2017). The third demonstration presents these three functional smiles and the facial features that separate them.

Table 1

Three Theoretical Framings of the Smile

Note. The three framings are not mutually exclusive; a single smile can index an emotion, feed back on it, and serve a social function at once.

Development and Individual Differences

Smiling is present and functional long before it could be strategic. Daniel Messinger's microanalytic work on infant smiling showed that early smiles are not uniform but graded: smiles that combine the lip-corner pull with cheek raising and, especially, with mouth opening index greater positive emotion, so that even in the first months of life the morphology of a smile scales with its intensity in a lawful way (Messinger et al., 2001). The infant smile is thus a graded signal from the outset, continuous with the adult distinction between fuller and slighter smiles rather than a separate phenomenon.

Across adulthood, smiling shows robust group and individual differences. Marianne LaFrance, Marvin Hecht, and Elizabeth Paluck conducted a meta-analysis of sex differences in smiling and found that women smile more than men on average, a moderate difference strongly moderated by context — it is largest when people know they are observed and smallest in solitary or absorbing situations — which fits the signaling account that smiling tracks the social situation (LaFrance et al., 2003). And the intensity of a person's habitual smile carries long-range predictive weight. LeeAnne Harker and Dacher Keltner coded the smiles in women's college-yearbook photographs and found that the strength of the Duchenne, eye-involving smile at around age 21 predicted favorable personality, marital stability, and well-being decades later, an illustration that a brief expressive measure can forecast measurable life outcomes (Harker & Keltner, 2001).

Worked Example

The modern verdict on facial feedback is that the effect is small, and reading that claim correctly requires translating a standardized effect size into something concrete. Effects in this literature are reported as Cohen's d, the difference between two group means expressed in standard-deviation units. A transparent way to feel the size of a d is the common-language effect size: the probability that a randomly chosen person from the smiling condition reports more positive emotion than a randomly chosen person from the control condition. For two normal distributions of equal variance this probability is Φ(d / √2), where Φ is the standard-normal cumulative distribution.

Take the small effect the meta-analysis describes, an illustrative d = 0.20. Then d / √2 = 0.20 / 1.414 = 0.141, and Φ(0.141) ≈ 0.56. So even when facial feedback is present, a smiling person is more positive than a control person only about 56% of the time, barely above the 50% of no effect at all. Halve the effect to d = 0.10 and the probability falls to Φ(0.0707) ≈ 0.53; raise it to a moderate d = 0.40 and it rises only to Φ(0.283) ≈ 0.61. The arithmetic makes vivid why a real effect can still be a small one: an influence that moves the odds from 50% to 56% is genuine, reproducible under a careful adversarial protocol (Coles et al., 2022), and yet far too weak to warrant the folk advice that one can simply smile oneself happy. The illustrative values span the small-to-moderate range the literature reports; the point is the conversion, which turns an abstract d into an honest statement about how often the manipulation actually changes which person feels better.

Discussion

The three framings of the smile correspond to three eras of thinking about emotional expression. The index-of-emotion view, inherited from Darwin and sharpened by Ekman, treats the face as a window onto discrete inner states and the Duchenne marker as the window's seal of authenticity; it remains useful but has been qualified into a probabilistic cue rather than a guarantee. The cause-of-emotion view, the facial feedback hypothesis, was for a time oversold on the strength of a single striking paradigm, and its correction is a model of how the field now works: a failed registered replication, a careful meta-analysis, and an adversarial collaboration together converged on a small, real, and conditional effect instead of either the original large claim or its wholesale rejection. The social-signal view reframes the entire question, holding that a smile is addressed to an audience and that asking what emotion it expresses can be the wrong question for a behavior whose job is to influence others.

These views are complementary rather than exclusive. A single smile can reflect genuine enjoyment, slightly deepen that enjoyment through feedback, and simultaneously signal affiliation to a companion. What the modern literature discourages is the reductive shortcut that any one of these exhausts the behavior — that a smile is simply true or false, simply a route to happiness, or simply manipulation. The smile is a compact illustration of why emotional expression resists tidy single-function accounts, and of how registered, multi-laboratory methods have reshaped the standard of evidence for claims about it.

Current Directions

Three strands are active. The first is the adversarial, multi-laboratory study of facial feedback itself: the Many Smiles Collaboration is a template for how disputed effects are now adjudicated, with proponents and skeptics agreeing a protocol in advance so that the result binds both sides, and its small-but-real conclusion is being extended to which elicitation methods and which emotions the effect generalizes to (Coles et al., 2022). The second is the functional-taxonomy program, which seeks to ground the reward, affiliation, and dominance smiles in distinct physical signatures, social consequences, and possibly physiological systems, moving the field from a single enjoyment-versus-fake contrast toward a repertoire of socially specialized signals (Rychlowska et al., 2017). The third is the formal re-examination of the Duchenne marker with coded, high-resolution measurement of eye constriction, testing the long-standing authenticity claim quantitatively rather than accepting it by tradition (Girard et al., 2021). Across all three the methodological trend is the same as elsewhere in emotion science: preregistration, pooled multi-site data, and objective coding in place of a single memorable demonstration.

Common Misconceptions

A real smile can always be told from a fake one by the eyes.

The eye-muscle (orbicularis oculi) marker raises the probability that a smile is genuine but does not guarantee it: many people can contract that muscle deliberately when posing, so the Duchenne marker is a graded, probabilistic cue, not an infallible test of sincerity (Krumhuber & Manstead, 2009; Girard et al., 2021).

Smiling on purpose can make a person happy.

Facial feedback is real but small: the famous pen-in-mouth demonstration failed a large registered replication, and the best current estimate is an effect that nudges felt emotion slightly rather than reliably lifting mood (Wagenmakers et al., 2016; Coles et al., 2019).

A smile is simply the expression of feeling happy.

People smile more when an audience is present or imagined even at constant amusement, and smiles serve distinct reward, affiliation, and dominance functions, so a smile is a social signal addressed to others as much as a read-out of private emotion (Fridlund, 1991; Rychlowska et al., 2017).

Glossary

Affiliation Smile.

A functional smile that signals social bonds, nonthreat, and goodwill, distinct in form from reward and dominance smiles.

Cohen's d.

A standardized effect size expressing the difference between two means in standard-deviation units; facial feedback effects are small values of d.

Common-Language Effect Size.

The probability that a random member of one group exceeds a random member of another; a concrete way to read a small d.

Dominance Smile.

A functional smile that conveys superiority, contempt, or social control, often asymmetric and blended with non-enjoyment features.

Duchenne Smile.

An enjoyment smile that combines the lip-corner pull with contraction of the orbicularis oculi around the eyes, named for Guillaume Duchenne.

Emotion.

An affective state with physiological, expressive, and experiential components; the smile is a candidate expression and cause of positive emotion.

Facial Action Coding System.

An anatomically based scheme for coding facial movement as Action Units; the lip-corner pull is AU12 and the eye-ring contraction is AU6.

Facial Expression.

The configuration of facial muscles that accompanies or signals an emotional or social state; smiling is one of its most recognizable forms.

Facial Feedback Hypothesis.

The proposal that sensory feedback from one's own facial muscles contributes to emotional experience, so that smiling can intensify felt happiness.

Orbicularis Oculi.

The ring of muscle around the eye socket (AU6) whose contraction raises the cheeks and creases the outer eye corners, marking the enjoyment smile.

Registered Replication Report.

A preregistered, multi-laboratory, direct replication of a key study under a common protocol; the one on the pen-in-mouth effect found no reliable effect.

Reward Smile.

A functional smile that reinforces another's behavior, broadly corresponding to the enjoyment smile and typically symmetric and eye-involving.

Simulation of Smiles (SIMS) Model.

Niedenthal's account in which observers decode a smile by covertly mimicking it and reading the resulting embodied feedback, with eye contact and mimicry integral to interpretation.

Social Perception.

The processing by which observers read others' states and intentions; smiling recruits it when a perceiver infers which kind of smile is being shown.

Zygomaticus Major.

The muscle running from the cheekbone to the mouth corner (AU12) whose contraction pulls the lip corners up and out, present in every smile.

Key Researchers

Nicholas A. Coles

. Research psychologist at the University of Florida; he led the meta-analysis establishing that facial feedback effects are small and variable and organized the Many Smiles Collaboration, the adversarial multi-laboratory test of the hypothesis. ORCID

Guillaume Duchenne

(1806-1875). French physiologist whose electrostimulation studies of the facial muscles identified the eye-muscle contraction of the genuine enjoyment smile, which now bears his name. Wikipedia

Paul Ekman

(1934-2025). American psychologist, co-developer of the Facial Action Coding System; his work revived the Duchenne distinction and identified the orbicularis oculi as the marker of enjoyment smiles. Wikipedia - Wikidata

Alan J. Fridlund

. Psychologist at the University of California, Santa Barbara; his behavioral-ecology view recast facial displays as social tools and showed that solitary smiling is potentiated by an implicit audience. Faculty Page

Marianne LaFrance

. Professor emerita of psychology at Yale University; her meta-analysis quantified sex differences in smiling and showed how strongly they depend on social context. Faculty Page

Daniel S. Messinger

. Professor of psychology at the University of Miami; his microanalytic studies established that infant smiling is a graded signal whose morphology scales with positive emotion. ORCID

Paula M. Niedenthal

. Professor of psychology at the University of Wisconsin–Madison; her Simulation of Smiles model explains how observers decode smiles through embodied simulation, and her functional account treats smiles as multipurpose social signals. Wikipedia - Wikidata

Fritz Strack

(b. 1950). Social psychologist and professor emeritus at the University of Würzburg; his pen-in-mouth study was the definitive demonstration of the facial feedback hypothesis and the focus of its later registered replication. Wikipedia - Wikidata

Frequently Asked Questions

What is the difference between a real and a fake smile?

A genuine enjoyment, or Duchenne, smile combines the lip-corner pull of the zygomaticus major with contraction of the orbicularis oculi around the eyes, which raises the cheeks and creases the outer eye corners; a posed smile often lacks that eye involvement. The marker is a reliable cue on average but not infallible, because some people can produce the eye contraction deliberately (Ekman et al., 1990; Krumhuber & Manstead, 2009).

Does forcing a smile actually increase happiness?

Slightly, at most. The facial feedback effect is real but small: arranging the face into a smile produces a modest increase in felt positive emotion under careful conditions, far weaker than the popular advice implies, and its most famous demonstration failed to replicate (Wagenmakers et al., 2016; Coles et al., 2022).

What was the pen-in-mouth study?

Strack, Martin, and Stepper had people hold a pen in their teeth, covertly engaging the smiling muscles, or in their lips, preventing a smile; the teeth group rated cartoons funnier, which was taken as evidence that smiling intensifies amusement (Strack et al., 1988). A seventeen-laboratory registered replication later found no reliable effect (Wagenmakers et al., 2016).

Why is the smile named after Duchenne?

Guillaume Duchenne de Boulogne used electrical stimulation of the facial muscles in the nineteenth century to show that the eye-ring muscle moves in genuine joy and resists voluntary control, so Ekman and colleagues named the enjoyment smile the Duchenne smile in his honor (Ekman et al., 1990).

Are there different kinds of smile?

Yes. Beyond the enjoyment-versus-posed contrast, smiles serve at least three social functions with distinct physical forms: a reward smile that reinforces behavior, an affiliation smile that signals bonds and nonthreat, and a dominance smile that conveys superiority (Rychlowska et al., 2017).

Do people smile only when they feel happy?

No. People smile more when an audience is present or merely imagined, even at constant amusement, which shows that smiling is a social signal addressed to others rather than an automatic read-out of felt emotion (Fridlund, 1991).

Do men and women smile differently?

On average women smile more than men, a moderate difference that is largest when people know they are being observed and smallest in private or absorbing situations, consistent with smiling tracking the social situation (LaFrance et al., 2003).

Can a smile predict anything about a person's life?

The intensity of the Duchenne smile in young-adulthood yearbook photographs predicts more favorable personality, marital stability, and well-being decades later, showing that a brief expressive measure can forecast long-range outcomes (Harker & Keltner, 2001).

References