Abstract
Chronic positive mood (CPM) has been shown to confer a wide variety of social, functional, and health benefits. Some researchers have argued that humans evolved to feel CPM, which explains why most people report better than neutral mood (the “positivity offset bias”) and why particularly happy people have particularly good outcomes. Here, we argue that the Duchenne smile evolved as an honest signal of high levels of CPM, alerting others to the psychological fitness of the smiler. Duchenne smiles are honest because they express felt positive emotion, making it difficult for unhappy people to produce them. Duchenne smiles enable happy people to signal and cooperate with one another, boosting their advantages. In our literature review, we found (a) that not all Duchenne smiles are “honest,” although producing them in the absence of positive emotion is difficult and often detectable, and (b) that the ability to produce and recognize Duchenne smiles may vary somewhat by a person’s cultural origin. In the final section of the article, we consider behavioral influences on CPM, reviewing research showing that engaging in eudaimonic activity reliably produces CPM, as posited by the eudaimonic-activity model. This research suggests that frequent Duchenne smiling may ultimately signal eudaimonic personality as well as CPM.
A “positive mood offset” bias is almost universally observed in mood data (Cacioppo & Berntson, 1994, 1999)—on average, people feel pretty good instead of feeling bad or merely neutral. But why is this the case? Diener, Kanazawa, Suh, and Oishi (2015) argued that generally positive mood may have been selected for during humans’ ancestral past (Owren & Bachorowski, 2001) because people in positive moods are, on average, better able to engage in behaviors that promote survival and reproductive success compared with people in neutral or negative moods (Diener et al., 2015). The functional benefits of positive mood include but are not limited to greater sociability, relationship success, fecundity, coping, job performance, health, and even longevity (Lyubomirsky, King, & Diener, 2005). Positive moods are also associated with eudaimonic (prosocial and growth-promoting) behaviors, as we will consider later in this article (Martela & Sheldon, 2019). Positive moods help people to travel fluidly through the world, living and performing at their best (Fredrickson, 2001).
Of course, positive mood is not always the most adaptive stance toward the world; negative moods and emotions can signal important problems to be solved (Forgas, 2017; Fredrickson, 2001; Nesse & Williams, 1994) and motivate taking steps toward their solution (Nesse & Ellsworth, 2009), whereas excessive positive emotions may at times hinder such adaptation (Forgas & East, 2008; Nesse, 1990). Still, the general functional benefits and positive outcomes of being in at least a mild positive mood are well known (Lyubomirsky et al., 2005).
Diener et al. (2015) confined their arguments to general or chronic positive mood (CPM). Moods are defined as relatively long-lasting affective conditions not attributable to or associated with a single identifiable event. In contrast, emotions (the other aspect of affect) are shorter lived, more intense, and usually associated with specific events or thoughts. Moods can be viewed as involving a person’s average or baseline level of feeling across a period of time—the general affective tone suffusing the person’s experience. In contrast, emotions can be viewed as involving a person’s transient fluctuations around a chronic baseline (Sheldon & Lyubomirsky, 2019). Of course, moods can also fluctuate at shorter time scales such that a person might be in a relatively good mood one day or in a relatively worse mood the next day. However, Diener et al. took a longer temporal view in focusing on CPM because CPM can presumably affect a person’s “total lifetime fitness consequences” (Cosmides & Tooby, 2000).
In this article, we propose that CPM is expressed by frequent Duchenne smiling and, moreover, that this linkage evolved as part of a social signaling system that helps happy people to assort and cooperate with each other. Later, we also consider how behavior influences CPM, positing that Duchenne smiles are ultimately derived from “eudaimonic” behaviors and personality styles, which cause CPM and also signal cooperative intent.
Duchenne Smiles
Duchenne smiles (DSs), first identified by the French neurologist Guillaume Duchenne, involve contraction of both the zygomaticus major muscle, which raises the corners of the mouth, and the orbicularis oculi muscle, which raises the cheeks and crinkles the eyes. DSs are associated with genuine mirth and enjoyment and are distinct from non-Duchenne smiles, which involve only the mouth and are often produced intentionally or without felt emotion to indicate politeness or acquiescence (Ekman & Friesen, 1982). DSs are difficult to produce intentionally because they typically occur spontaneously. Thus, our main contention is that DSs are “honest signals” that reliably communicate a positive affective condition and possibly much more. Facial expressions are generally accepted as parts of communication systems common to all human populations (Waller, Cray, & Burrows, 2008), and DSs are no exception. We consider the honest signal concept in greater detail below.
Honest Signaling Systems
Zahavi (1975) was the first to propose the idea of honest signals, defining them as structures or behaviors that convey accurate (“honest”) information to the receiver because they are costly to make and are not easily faked. For example, the exceptionally large and ornate tails of some male peacocks or the exceptionally brilliant coloring of some male cardinals are difficult to produce. Only a truly healthy and thriving animal could manifest such characteristics, and thus the inference “They are a superior specimen!” is likely to be correct. A related concept is the handicap principle, by which superior health or proficiency are advertised in a profligate way that might even impair the animal (as the peacock’s tail may impede his ability to escape predation). Such displays, evolved through sexual selection, can alert attuned perceivers to the desirable genes and condition of the signaler.
Smith and Harper (2003) provided a book-length analysis of the concept of signaling, using a game-theory approach to clarify terminological and theoretical confusions in the literature. We will adopt their approach to organize our literature review, pausing first to explain relevant ideas and terminology. Signals are structures or behaviors that communicate implicit messages between organisms. Signals, by definition, affect the behavior of others in ways that enhance cooperation between the two parties. Thus, signaling systems also operate in receivers via perceptual and decision-making systems evolved to receive those signals. Signaling systems can emerge independently of the qualities or characteristics being signaled because they are molded by different selection pressures. Signaling systems must be useful for both senders and receivers and must provide benefits to both, else they would disappear. For example, dominant male bonobos are especially alert to estrus swellings in female bonobos, which causes them to monopolize those females while also protecting them from impregnation by less desirable mates (Zinner, Nunn, van Schaik, & Kappeler, 2004). In the human case, some perceivers may be especially alert to the psychological state of joyful people, choosing to assort and cooperate with those people, to their mutual benefit (Sheldon, Sheldon, & Osbaldiston, 2000).
Of course, the interests of the signaler and the receiver are not always in perfect alignment; as in all cooperative arrangements, there can be conflicts of interest. In such cases, additional features can be added to the system in the manner of an arms race (Smith & Price, 1996). One such feature is the “manipulative” signal in which the signaler does not really have the signaled characteristic (Smith & Harper, 2003). An example is mimicry, as in the case of the Viceroy butterfly, which imitates the markings of the bad-tasting Monarch butterfly. On the receiver side, there is “mind-reading,” by which a receiver is able to penetrate a signaler’s disguise (Smith & Harper, 2003), or “cheater-detection,” by which potential victims of the dishonest signaler can discern the signaler’s true intentions (Tooby & Cosmides, 1990). Again, from a game-theory perspective, signaling systems are a type of cooperative interaction that benefits both parties to the extent that the interaction pattern can be maintained and cheating discouraged (Axelrod & Hamilton, 1981; Wilson, 2019).
A final important concept is the index signal. This is a special kind of honest signal “whose intensity is causally related to the quality being signaled, and which cannot be faked” (Smith & Harper, 2003, p. 15). The degree of brilliance of the cardinal’s red coat reflects the relative degree of physical health and vigor of the cardinal, just as the intensity or frequency of a person’s DSs may reflect their level of CPM.
Now we turn to our review of the DS research. In canvasing this material, we sought support for five different hypotheses, derived from the theoretical considerations above. The basic model we are proposing is that CPM enhances a person’s likelihood of experiencing momentary positive emotions, therefore making the person’s production of DSs more likely, therefore enhancing the likelihood of receiving cooperation from others. The hypotheses below address the various parts of this basic model, drawing on the extant DS body of research. Again, in a final section of the article, we review the behavioral sources of CPM and speculate that frequent DSs may ultimately signal that the person is living in a eudaimonic way. However, we do not put this forth as an explicit hypothesis because of the absence of data directly linking eudaimonic behavior to DSs.
Hypotheses
Hypothesis 1. DSs reliably signal positive emotion, especially in the presence of pleasant stimuli.
Hypothesis 2. DSs are index signals, directly linking the intensity or duration of the smile to the positive emotionality of the smiler.
Hypothesis 3. There is a reliable association between CPM and the presence, intensity, or duration of DSs produced.
Hypothesis 4. DSs are highly recognizable to receivers, influencing their perceptions of the smiler.
Hypothesis 5. DSs facilitate cooperation between signalers and receivers, producing benefits for both parties.
The Evidence
Hypothesis 1. DSs reliably signal positive emotion, especially in the presence of pleasant stimuli
Hypothesis 1 states that DSs are accurate signals of spontaneous positive emotion because they are difficult to produce in the absence of positive emotion. Another way in which DSs are honest is that they are best produced in the presence of actual pleasant events or stimuli. Thus, they must be a response to something genuinely enjoyable rather than a guise intentionally adopted.
The evidence for this initial hypothesis is quite strong. In one of the first articles on this topic, Ekman, Davidson, and Friesen (1990) showed that the presence of DSs correlated positively with rated enjoyment of films and with momentary positive emotion (see Ekman & Friesen, 1982). DSs were also more likely to occur when pleasant rather than unpleasant films were watched, further attesting to the honesty of the pleasure signals produced. M. G. Frank and Ekman (1993) thus asserted that DSs are reliably associated with felt positive emotion, whereas other types of smiles are not. Likewise, Soussignan (2002) showed that participants assigned to express DSs reported more positive experience, but only when pleasant scenes and humorous cartoons were presented. In addition, Johnson, Waugh, and Fredrickson (2010) showed that DSs correlate with self-reported positive emotion and occur more often during positive emotion inductions than during neutral or negative emotion inductions. Likewise, Zaalberg, Manstead, and Fischer (2004) found that only funny jokes predicted stronger positive affect and longer DSs. Further afield, Gonzaga, Keltner, Londahl, and Smith (2001) showed that DSs correlate with self-report and partner-report estimates of love within romantic relationships. DSs even penetrate sleep; Clé et al. (2019) showed that DSs during dreaming are associated with happy dreaming scenarios, reflecting a “true inner mirth.” Taken together, these findings support the hypothesis that DSs arise in the presence of felt positive emotions.
Hypothesis 2. DSs are index signals, directly linking the intensity or duration of the DS to the intensity of the positive emotionality within the smiler
There is good evidence to support the index signal hypothesis, that the intensity or duration of DSs is related to the intensity or duration of felt positive emotion. M. G. Frank, Ekman, and Friesen (1993) showed that DSs expressing greater enjoyment are marked by smoother zygomaticus major actions of more consistent duration than the zygomaticus major actions of non-DSs. Hess, Banse, and Kappas (1995) found that the intensity of a DS was predicted by the intensity of the emotional stimulus as well as by the social context and the relationship between the signaler and the audience. In a meta-analysis, Gunnery and Ruben (2016) found that participants were worse at telling the difference between DSs and non-DSs when the smiles were artificially matched for intensity. This suggests that the intensity of DS facial actions is an important linear cue to the intensity of the underlying emotional condition. In a related work, Gunnery and Hall (2015) summarized emerging research on the importance of smile intensity in signaling the strength of the underlying affective state. Krumhuber and Manstead (2009) found experimental evidence showing that observers distinguished between DSs of varying intensity, or smiles that involved more compared with less orbicularis oculi activation. Intuitively, it makes sense that a person whose face beams with delight would be perceived differently than a person whose face expresses a more contained type of still genuine pleasure, such as amusement.
Furthermore, Dapelo et al. (2015) showed that participants suffering from anorexia nervosa produced DSs of lower intensity and duration in response to humor compared with control participants, consistent with the inference that people with eating disorders experience less CPM. Zlochower (2002) found that deliberate and spontaneous smiles were discriminated on the basis of smile intensity, especially the intensity of eye narrowing in the spontaneous smiles. Krumhuber and Kappas (2005) reported that the perceived genuineness of smilers varied monotonically with the duration of the smiles presented. Fairbairn (2015) showed that peoples’ shared DSs were of significantly longer duration when they were under the influence of alcohol, indicating the presence of extraexuberant emotions. Thus, the collective evidence suggests that the intensity or duration of DSs is a reliable indicator of the intensity of a person’s momentary positive emotion.
Hypothesis 3. There is a reliable association between CPM and the presence, intensity, or duration of DSs produced
We assume CPM facilitates the easy production of DSs, making it possible to produce streams of DSs as a mere consequence of living a satisfying and enjoyable life. Of course, there are other influences on DSs besides CPM, such as the appeal of the particular stimulus to which the signaler is responding and the kind of relationship the signaler has with the receiver or receivers (Hess et al., 1995). But the ongoing facilitation of DSs by a platform of general good feeling is likely ever relevant.
However, the direct evidence for this hypothesis is somewhat limited because few researchers have published associations between measured CPM and the frequency or quality of a person’s DS production (but see the end of this section for a report of recently collected data relevant to this hypothesis). Instead, researchers have focused, as summarized above, on showing that momentary positive emotions are associated with situational DSs. Still, Auerbach (2017) showed that patients higher in trait cheerfulness, a direct analogue of CPM, produced more DSs in response to a hospital clown visit than patients low in cheerfulness. Extraversion is often used as an indicator of CPM (Lucas & Fujita, 2000); Abe and Izard (1999) showed that the positivity of emotion displayed by children in strange situations was associated with maternally rated extraversion of the children (see also Fairbairn et al., 2015). Relevant to the question of whether people can infer chronic affect (long-term mood) from momentary affect (current emotions), Hall, Gunnery, Letzring, Carney, and Colvin (2017) showed that people could accurately judge affective personality traits on the basis of state emotional information. For example, participants could judge that a target person was high on extraversion on the basis of their display of positive emotion, just as depression could be judged on the basis of the display of distressed emotion.
Additional corroboratory evidence for our third hypothesis is supplied by studies of people with negative psychological conditions. As mentioned above, Dapelo et al. (2015) showed that people diagnosed with anorexia nervosa, who are presumably low in CPM (O’Brien & Vincent, 2003), produced DSs of shorter duration and lower intensity while watching a humorous film clip compared with control participants. Keltner and Bonanno (1997) showed that increases in the number of DSs in bereaved persons were associated with greater enjoyment and reduced anger over time, presumably indexing their return to their CPM set points (see also Bonanno & Keltner, 1997). Likewise, Bonanno et al. (2007) found that the number of DSs in sexual abuse survivors was associated with better social adjustment 2 years later. More indirectly, Harker and Keltner (2001) found that the coded quality of the smiles displayed by college women in their yearbook photos, presumably indexing CPM at that time, was associated with personal well-being up to 30 years later. So although there is relatively little direct evidence of a link between CPM (dispositional positive affect) and the presence, intensity, or duration of DSs, there is at least indirect evidence to support the idea.
In an attempt to remedy the paucity of direct evidence linking CPM and smiling, we recently collected relevant data in our lab. A sample of 162 introductory psychology students from the University of Missouri (54 men and 108 women; Mdn age = 19 years; 74% White) completed a measure of positive and negative mood (Emmons, 1991; Sheldon, Corcoran, & Prentice, 2019) and a measure of life satisfaction (Brunstein, 1993). Before doing this, they were asked to upload a photograph of themselves, with the following instructions: One question we are studying is the relationship of peoples’ physical appearance to their survey responses. Therefore, we would like you to upload a picture of your own head and shoulders right now! You can either browse for such an image that you already have, or you can take a new picture now and upload it.
We used photo-editing software to standardize the uploaded images to show only faces, which displayed smiles of varying intensity.
Next we asked three research assistants to independently rate each photograph in terms of four dimensions: overall happiness of the person (“To what extent is their life characterized by strong positive moods?”), genuineness of the smile (“How authentic and real is the smile?”), smiling in the mouth (“How much are the corners of their mouths turned up?”), and smiling in the eyes (“How much are their cheeks raised and their eyes crinkled?”). Ratings were made on 7-point Likert scales. The ratings of the three research assistants were highly correlated within each of the four rating dimensions, so we averaged across the three raters (Cronbach’s αs = .85, .80, .89, and .84, respectively; on the 7-point scale, means = 4.41, SD = 1.38; 4.00, SD = 1.52; 4.04, SD = 1.70; and 4.00, SD = 1.66; respectively). The four resulting variables were also highly correlated, so we averaged them into a single aggregate measure but also examined them separately. We observed strong sex differences in the five variables: Female participants were rated higher in all five tests (ts range = 3.81–4.23). Thus, we controlled for sex in our hypothesis tests.
Specifically, we conducted 15 regression analyses: five for positive mood, five for negative mood, and five for life satisfaction, while controlling for sex. Within each set of five analyses, four analyses employed the four rating dimensions taken separately, and the fifth analysis employed the aggregate variable.
In the negative affect and life satisfaction analyses, none of the rated variables were significant. That is, raters’ estimations of participants’ overall happiness and the intensity and genuineness of their smiles (based on their pictures) did not predict participants’ self-rated negative affect and life satisfaction. However, in the positive mood analyses, the aggregate variable was a significant predictor (β = 0.18, p = .028). This was also the case for ratings of the genuineness of the smile (β = 0.17, p = .038), the intensity of the smile in the mouth area (β = 0.19, p = .020), and the intensity of the smile in the area of the eyes (β = 0.19, p = .020). Ratings of the overall happiness of the person did not reach significance (β = 0.14, p = .090), perhaps because the inference of happiness in the target is more indirect than the inference that the target has an intense and authentic smile. Note that 15 of the pictures were identified as having no trace of a smile. To ensure that this group did not account for the effects, we removed them from the data and reran the analyses. Results were essentially unchanged, and indeed, the effects were slightly larger than those found in the full analysis.
Taken together, these data provide corroboration for the central claims of this article, that people who experience CPM signal this condition in part via their smiles, in this case through self-selected photographs. There were no effects involving negative mood and life satisfaction, the two other components of subjective well-being (Diener, 1994). These three variables are often combined into an aggregate subjective well-being (SWB) variable without examining the individual components. However, if we had done so in this case, the unique, and predicted, effects involving CPM would have been obscured.
An important limitation of these data is that we do not know whether participants uploaded a photograph taken during the assessment or whether they instead uploaded a previously taken photo. Most of the aforementioned DS studies relied on spontaneous smiling, and it is possible that some participants in our study, who felt neutral or negative in the moment, chose positive past photos for self-presentational purposes. If so, our raters were still able to predict the participants’ CPM despite this potential source of error. Presumably, CPM (a trait-like variable) has effects whenever or however photographs are taken of the person. Thus, we suggest this limitation does not detract from our central arguments. However, future research should take into account whether the photo was taken in the moment so the possible influence of this factor can be empirically evaluated.
Hypothesis 4. DSs are highly recognizable to receivers, influencing their perceptions of the smiler
To reiterate, the concept of an honest signaling system requires that the system should operate in both signalers and receivers, enabling both to participate in a temporal sequence of cooperative behavior. Thus, perceivers should (a) readily recognize DSs and (b) evaluate their makers positively.
There is good evidence that DSs are highly recognizable. Surakka and Hietanen (1998) showed that perceivers’ facial electromyograph (EMG) reactions differentiated between neutral faces and DSs but not between neutral faces and non-DSs. Malek et al. (2019) showed that images of DSs were looked at longer by perceivers and were rated as more intense and sincere. As noted above, Zlochower (2002) found deliberate and spontaneous smiles were discriminated on the basis of the intensity of eye narrowing in the spontaneous smiles. Williams, Senior, David, Loughland, and Gordon (2001) found evidence that humans respond automatically to the orbicularis oculi activation that is characteristic of DSs, presumably because of the important communicative function of DSs. These findings suggest that DSs are not only very recognizable but also may have evolutionary roots, as evidenced by automatic responses in the brain.
There is also good evidence that perceivers can evaluate DSs for authenticity. Surakka and Hietanen (1998) showed that DSs induced pleasure in perceivers in addition to eliciting different EMG reactions, and Malek et al. (2019) showed that DSs were rated as more intense and sincere in addition to being looked at longer. In a meta-analysis, Gunnery and Ruben (2016) showed that DSs are associated with more positive ratings by perceivers (greater authenticity, genuineness, realness, attractiveness, and trustworthiness), especially when smiles were elicited naturally rather than through posing paradigms. More peripherally, Ilicic, Kulczynski, and Baxter (2018) showed that DSs rehabilitate the negative image of celebrities via consumers’ perception of genuineness in the celebrity. Auerbach (2017) found that hospital clowns elicited more DSs than non-DSs and that DSs were associated with greater perceived funniness of the clowns. Carroll and Russell (1997) showed that movie actors judged as happy in a scene displayed DSs 74% of the time in those instances. Grandey, Fisk, Mattila, Jansen, and Sideman (2005) showed that DSs produced by service providers boosted customer satisfaction only if the service was good. Shore and Heerey (2011) found that participants were willing to sacrifice the chance of a monetary reward to receive a genuine smile and produced inflated estimates of the value of genuinely smiling faces, suggesting that DSs “may be useful social reinforcers and therefore important in the control of social behavior on a moment-to-moment basis during interaction” (p. 169). Canadas and Mast (2017) even found a third-party effect in which perceivers prefer objects that have been gazed at by other people displaying DSs. In summary, people enjoy and are drawn toward stimuli associated with DS smiling.
Hypothesis 5. DSs facilitate cooperation between signalers and receivers, producing benefits for both parties
This hypothesis derives from Smith and Harper’s (2003) definitional requirement that honest signals affect the behavioral choices of receivers, not just receivers’ perceptions of senders. Again, Smith and Harper took a game-theory perspective on such systems, viewing them as enabling cooperation between interdependent social actors. Note that Owren and Bachrowski (2001) made a similar argument: “that applying the basic tenets of selfish-gene selection to the evolution of communication in early hominids points to the emergence of honest signaling of positive affect as a valuable mechanism for forming and maintaining cooperative relationships” (p. 176).
The evidence is good for the link between DS and cooperation. R. H. Frank (1988) was perhaps the first to apply signaling theory to human emotion and communication, marshaling the evidence of that time in favor of his argument. Research testing Fredrickson’s (2001) broaden-and-build theory of positive affect has shown that positive emotion displays build social as well as cognitive resources (Fredrickson & Kurtz, 2011). Gonzaga et al. (2001) showed that DSs correlated with self-report and partner estimates of love and also with commitment-enhancing processes when the relationship is threatened. Bernstein, Young, Brown, Sacco, and Claypool (2008) found that social rejection improved peoples’ ability to discriminate between DSs and non-DSs, presumably because rejected people are more attentive to social cues relevant to reestablishing affiliation and belongingness. As mentioned above, Grandey et al. (2005) found that server smiling affected customer satisfaction only when the service was good (i.e., when the server cooperated well with the customer). DSs help to initiate and cement mutually beneficial partnerships.
There are also several high-quality studies of behavioral cooperation conducted in the laboratory, which we describe in some detail so that interested readers can appreciate the rigor of the findings.
Centorrino, Djemai, Hopfensitz, Milinski, and Seabright (2015) tested the idea that smiles perceived as honest serve as a signal that evolved to induce cooperation in situations requiring mutual trust. Participants recorded video clips of themselves and watched each other’s clips before responding to a trust game. Smiles rated as more genuine predicted higher judgments about the trustworthiness of other participants and greater willingness to send them money. Participants rated as smiling genuinely also returned more money on average to senders, showing that both parties benefited in the exchange (Smith & Harper, 2003).
Likewise, Mehu, Grammer, and Dunbar (2007) tested the idea that smiles advertise cooperative dispositions and increase the likelihood that a potential social partner invests resources in the relationship. Participants were covertly filmed while interacting with a friend in either a sharing or a control condition. DSs were displayed at higher rates in the sharing situation compared with the control situation, and non-DSs were unaffected by the type of interaction. Mehu et al. concluded that DSs “could therefore be an important signal in the formation and maintenance of cooperative relationships” (p. 415).
In the same vein, Reed, Zeglen, and Schmidt (2012) investigated specific positive and negative facial actions displayed among strangers immediately following verbal promises of commitment within an unrestricted acquaintance period and anonymous decisions of cooperation or defection in a one-shot, two-person prisoner’s dilemma game after the acquaintance period. They found that facial actions related to enjoyment were predictive of cooperative decisions within dyads and that participants were able to accurately predict their partner’s decisions following the acquaintance period. Therefore, the authors concluded that “facial actions may function as honest signals of cooperative intent” (p. 200).
Finally, Krumhuber et al. (2007) used an animation methodology to manipulate film clips of potential cooperation partners before asking participants which partner they would like to play with and what their choices in the game would be. They found that inferences about the chosen partner’s trustworthiness mediated the effects of manipulated facial dynamics on cooperative behavior. In other words, potential partners in social dilemmas perceived to display genuine smiling were more often selected as actual partners and received more cooperative behavior from participants because the partners were perceived as honest and trustworthy. In short, people do cooperate more with others who display DSs, as predicted by their honest signaling perspective.
In the section below, we summarize two other relevant patterns of findings that became apparent during our evaluation of the body of DS research. These patterns were not addressed by our five hypotheses, but we describe them for the sake of completeness.
Duchenne smiles are not always honest
As noted earlier, honest signaling systems are describable in game-theory terms (Smith & Harper, 2003) because they involve communications between two or more social actors embedded within a social dilemma. Thus, there are always temptations to “cheat” on the arrangement to exploit would-be cooperators (Wilson, 2019). Although cheating threatens the evolutionary stability of the signaling system, it can provide cheaters with strong benefits in the short term. Given this, we would expect that at least some percentage of the population would be able to produce DSs intentionally or on demand, potentially degrading the value of the signal.
Increasing data support the existence of such an ability. Gunnery and Hall (2015) wrote, “until recently, scientists believed that a smile only reaches the eyes when the person was feeling genuinely happy” (p. 114). They then uncovered several strands of evidence countering this idea. They found that DSs could be deliberately produced by instructions, although the intentionally produced DSs were persuasive only when the producer was exposed to pleasant stimuli (Gunnery & Hall, 2014). Gunnery, Hall, and Ruben (2013) showed that substantial minorities of people can produce DSs on demand, as shown across a series of role-playing tasks and imitation tasks. Krumhuber and Manstead (2009) also instructed participants to produce DSs and found that some participants could do this with at least some degree of success. Crivelli, Carrera, and Fernández-Dols (2015) analyzed the expressions of winners of judo matches and found that although all reported feeling very happy after winning, DSs were produced only during actual social interactions with audiences following the match. That is, DSs were “tools of social interaction, rather than readouts of basic emotions” (p. 52). The authors interpreted this to mean that DSs are not a necessary concomitant of happiness and that they may serve primarily as social displays rather than as happiness displays. We agree with these conclusions and argue that the display of happiness to others is perhaps the main point of DSs (because it signals and elicits cooperation), and so it is logical that DS effects would occur in the context of positive contexts and social interactions.
Duchenne smile sending and receiving are affected by cultural upbringing
One implication of the fact that some people can produce DSs “on demand” is that different cultural contexts, via the differing emotion display rules they teach (Matsumoto, 1990), might enhance or discourage peoples’ productions of or responses to DSs. For example, Elfenbein, Beaupré, Lévesque, and Hess (2007) showed that when Canadian participants pose smiles, they are more likely to show DS markers than Gabonese participants. Thibault, Levesque, Gosselin, and Hess (2012) further showed that Gabonese participants did not use DS markers when judging smile authenticity. Mui, Goudbeek, Swerts, and Hovasapian (2017) found that children playing together displayed more DS markers than children playing alone; however, this was qualified by an interaction in which the difference was not observed for Dutch children playing together but was observed for Chinese children playing together. Likewise, Tsai, Chentsova-Dutton, Freire-Bebeau, and Przymus (2002) showed that Hmong Americans showed fewer non-DSs when posing happiness compared with European Americans. However, Chentsova-Dutton and Tsai (2007) did not find Hmong compared with European differences in smiling while reliving happy experiences. Sheldon et al. (2017) examined context effects on happiness expression in U.S. and Russian college students and found that Russians inhibit the expression of happiness, but only toward strangers, not toward friends and family members. Sheldon et al. explained this in terms of different display rules regarding positive emotions in Russia compared with the United States.
The influence of cultural upbringing on ideal affect is another consideration to make within this context. Ideal affect concerns peoples’ beliefs and values regarding the “best” kinds of emotional experiences. Previous research has shown that more collectivist cultures value low-arousal positive emotions more than high-arousal positive emotions, whereas for more individualistic cultures, the opposite is true (Tsai, Knutson, & Fung, 2006). Cultures and individuals that value low-arousal positive emotions are also less likely to see individuals in high-arousal positive states as affiliative (Tsai et al., 2019).
Taken together, there are clear differences in cultural expressions of happiness and valuing of certain emotional expressions. Some cultures express less DSs, have differing display rules on expression, and value emotions differentially. In cultures in which low-arousal positive emotions are valued more than high-arousal positive emotions, DSs may not serve as honest signals as detailed here. Thus, cultural values shape both the DS signaling contexts and the target audience for DS signals. Future research into DS expressions may also uncover differences in expression rules and expectations for other groups or individuals (e.g., marginalized groups or individuals). In sum, the studies that exist suggest that further research might uncover other cultural and within-cultures differences.
This concludes our review of the DS body of research. Below we turn to a very relevant question, mentioned earlier: the ultimate sources of CPM. If CPM facilitates momentary positive emotions, which produce momentary DSs, then what facilitates CPM?
Eudaimonic living may be the ultimate source of CPM and frequent DSs
The main contention of this article thus far is that DSs are an honest index signal of CPM, which both facilitates and marks psychological fitness (Diener et al., 2015). In this view, perceivers want to know more than “Is this person experiencing positive emotion right now?” Rather, they want to know, “Is this a person who can be relied on in the long term? Do I want to get involved with this person? Will I enjoy the relationship?” When receivers observe that, over time, a potential partner experiences frequent, intense, and/or long-lasting positive emotions and consequent DSs, they infer a high level of CPM within the partner and begin to respond more affirmatively to him or her.
But this leads to a potentially important question: What causes CPM, that is, what leads a person to be generally happy? This has been the subject of a large body of research. Although early thinking suggested that CPM is quite stable and is strongly determined by genetics (i.e., by a person’s inherited happiness tendency or set point; Lykken, 1999), researchers in the past 2 decades have clearly shown that people have considerable latitude to affect their own levels of CPM via their motivated behaviors, life choices, values, and goals (see Sheldon & Lyubomirsky, 2019). In other words, it is possible for people to live their lives in ways that elevate their CPM to levels higher than they would experience otherwise. For example, one monozygotic twin may be demonstrably happier than the other twin because of the more adaptive behavioral life choices made by the first twin (Sheldon & Lyubomirsky, 2019). Bearing this in mind, we can ask the following: Are the behavioral dispositions that lead to CPM the same dispositions that promote cooperation and social success?
Much of the research studying behavioral effects on well-being is describable by the eudaimonic-activity model (EAM; Martela & Sheldon, 2019; Sheldon, 2013, 2016; Sheldon & Lyubomirsky, 2019). The EAM applies the currently popular concept of “eudaemonia” (the condition of flourishing or of living well, according to the Ancient Greeks and Aristotle in particular) to specify a general process model by which higher levels of CPM are produced. Figure 1 summarizes the model.
The eudaimonic-activity model.
The EAM asserts that subjective well-being (including CPM) is ultimately produced by “eudaimonic” intentional behavior, involving virtuous or high-quality values, motivations, goals, and practices, rather than by mere “hedonic” behavior, which does not produce deep satisfaction. Eudaimonic goals and practices tend to bring about basic psychological need satisfaction (Ryan & Deci, 2017; Tay & Diener, 2011), which in turn bring about subjective well-being, composed prominently by CPM. The EAM accommodates the fact that striving directly for enhanced CPM/SWB tends not to work (Sheldon et al., 2019; van Zyl & Rothman, 2014). The model also accommodates findings that striving for integrity, connection, and growth are instead more reliable routes to higher CPM (summarized below). The EAM does not formally attempt to define “eudaimonic motives and activities,” instead viewing this as an empirical question; eudaimonic motives and behaviors are the ones that provide basic need satisfactions, therefore providing CPM as a kind of side effect.
The list of personality constructs that have already been shown to fit the three-phase framework of the EAM is indeed broadly consistent with Aristotle’s concept of eudaimonic living. For example, valuing growth, community, and high-quality relationships rather than valuing money, status, and image predicts longitudinal changes in need satisfaction, explaining longitudinal changes in CPM (Niemiec, Ryan, & Deci, 2009; Sheldon & Elliot, 1999; Sheldon & Kasser, 1998). Striving for self-concordant goals that challenge one and that express one’s deep values and potentials have been shown to facilitate goal attainment and to produce boosted need satisfaction, and thereby CPM, when attained (Sheldon & Elliot, 1999). In addition, increases in measured life balance also predict boosted need satisfaction, leading to greater CPM (Sheldon, Cummins, & Kamble, 2010). The same is true of being psychologically authentic (Sheldon, Gunz, & Schachtman, 2012) and of engaging in prosocial or other-serving behavior (Martela & Ryan, 2016).
To summarize, the personality characteristics that fit the EAM are all evenly actively conducive to the manifestation of CPM as well as the creation of high-quality and presumably highly cooperative relationships with others. Thus, when receivers infer that a person is high in CPM from the person’s frequent or intense DSs, perhaps they are also inferring that “this happy person is a good person, somebody I can trust and rely on.” Future research will be necessary to fully test this idea.
Conclusion
In this article, we have summarized research showing that DSs reliably signal positive emotions, in part because they occur more readily in the presence of genuinely enjoyable stimuli and experiences (Hypothesis 1). We have also shown that DS intensity is linearly related to the intensity of self-reported positive emotional emotion in the smiler (Hypothesis 2) and that DSs are associated with higher CPM (Hypothesis 3). Furthermore, DSs are highly recognizable to perceivers (Hypothesis 4) and also facilitate warm and cooperative interactions with perceivers (Hypothesis 5). Thus, DSs seem to fit the concept of evolved index signals of fitness (Smith & Harper, 2003) despite the fact that their display is somewhat influenceable by individual differences and by cultures.
We also considered individual difference or personality variables that can produce, increase, or maintain CPM, concluding that these tended to be eudaimonic variables, such as prosociality, authenticity, balance, and compassion. Such virtues have long been theorized as aspects of a “life well-lived” (Waterman, 2013), and the current review suggests that such virtues, by conducing to the generation of CPM, also serve in the production of DSs ultimately facilitating cooperative interactions and warm relationships with others. Conversely, people leading less virtuous or more exploitative lives presumably derive less CPM. This makes it more difficult for them to produce DSs, making it more difficult for them to experience congenial relations with others. Through the DS signaling system, happier people are alerted regarding the mental state of other happy people, whom they might then choose to approach, to their mutual benefit.
