Laughter, play faces and mimicry in animals: evolution and social functions

Laughter, play faces and mimicry in animals: evolution and social functions. Marina Davila-Ross and Elisabetta Palagi. Philosophical Transactions of the Royal Society B: Biological Sciences. September 21 2022. https://doi.org/10.1098/rstb.2021.0177

Abstract: Human laughter and laugh faces show similarities in morphology and function with animal playful expressions. To better understand primordial uses and effects of human laughter and laugh faces, it is important to examine these positive expressions in animals from both homologous and analogous systems. Phylogenetic research on hominids provided empirical evidence on shared ancestry across these emotional expressions, including human laughter and laugh faces. In addition, playful expressions of animals, in general, arguably have a key role in the development of social cognitive skills, a role that may help explain their polyphyletic history. The present work examines the evolution and function of playful expressions in primates and other animals. As part of this effort, we also coded for muscle activations of six carnivore taxa with regard to their open-mouth faces of play; our findings provide evidence that these carnivore expressions are homologues of primate open-mouth faces of play. Furthermore, our work discusses how the expressions of animal play may communicate positive emotions to conspecifics and how the motor resonance of these expressions increases affiliation and bonding between the subjects, resembling in a number of ways the important social–emotional effects that laughter and laugh faces have in humans.

3. Social use of open-mouth faces, laughter and other play vocalizations

(a) Play coordination, social cohesion and the development of skills

The differences in modality and anatomy of play expressions as well as their polyphyletic history indicate that these expressions have a complexity in both form and function. Carnivores and primates seem to modify their play expressions when they receive the attention of their playmates [38,45,58,80,81], and chimpanzees and bonobos are known to also modify them if the mothers of their infant playmates are nearby [82] or group members are attentive to the sender [83,84]. Multiple social functions of play expressions that are not necessarily mutually exclusive have been discussed.

As mentioned earlier in the present work, an important function of animal play expressions is to signal ‘this is play’, which helps to coordinate actions among playmates [46,85–87]. Probably most importantly, such signalling is likely to help avoid escalation into real fights during rougher play and, consequently, to prevent getting hurt, especially when the playmates are dissimilar in strength and do not have close social relationships [74,88–90].

Whereas mammals produce both types of open-mouth faces in both gentle and rough play [35,58], their upper-teeth exposure, which resembles wide-open mouth displays of submission and appeasement [39,54], tends to occur more often during rough play [37,60,91]. Similarly, play vocalizations seem to be predominantly produced in rough-and-tumble [74]. Thus, these types of expressions might signal to the recipient ‘this is just play’. It is also possible that the playmates widen their mouths further and expose their teeth owing to having to breathe more intensely and loudly during rough-and-tumble. Furthermore, the individuals producing these expressions could be in a state of high arousal and show more play biting [56]. The open-mouth faces without exposed teeth, on the other hand, seem to be less dependent on play intensity and have a more general application within play [37,56,60].

Consistent with the claim that play expressions signal ‘this is play’ or ‘this is just play’, empirical findings show that these expressions from rodents to primates may permit play actions and play bouts to be prolonged [58,84,86,87,89,92,93]. Furthermore, animals can sometimes produce open-mouth faces as part of the play invitation (e.g. when hitting the other playmate prior to play), and such signalling here is likely to help invite to play [60]. Such increase in playful interactions, key affiliative behaviours in social animals, is likely to have a notable impact on social bonding and, consequently, other behaviours among group members [43,94–98]. In humans, it is also known that laughter helps social cohesion [1,99]. Five-month-old infants already respond differently when hearing friends laughing together compared with strangers behaving this way [100].

In accordance with Barbara Fredrickson's Broaden-and-Build Theory of Positive Emotions, play expressions may contribute to the development of a range of skills that are central for individuals living in social groups, including social-cognitive skills [4]. Supporting this claim, play may consist of cooperative and competitive behaviours, where young individuals can practise with low risk a range of behaviours and further explore the impact these behaviours have on their conspecifics [93,101], which may become more relevant at a later stage in their development [102]. Such functions are consistent with the notion that positive expressions, such as laughter and other play expressions, do not necessarily need to constantly have immediate benefits, and perhaps their range in function makes them different from negative expressions, where it can be crucial to respond quickly in a risky situation [4,5].

Despite overlapping contexts, play vocalizations and open-mouth faces are at least to some extent likely to differ in function. Play vocalizations seem to be more limited to the context of play than open-mouth faces ([84]; for functional flexibility, see [103,104]). Although open-mouth faces occur predominantly during play bouts, they have also been observed shortly prior to them in order to invite a conspecific to play [105]. On a few occasions, they have been observed fully outside of play. One such incident took place at the Serengeti Park Hodenhagen, where Pia, a juvenile chimpanzee, was unsuccessfully play-inviting her father by pulling his hair (see [33]). As he did not budge, Pia left the scene, laid down to relax for a while, and suddenly started producing open-mouth faces (for a video footage, see [33]). Such rare incidents, where open-mouth faces that occur after nonaggressive violations of expectations resemble the use of human laughter linked to benign violations and humour [106,107], can already be observed in humans during infancy [108].

Perhaps the main difference in playful expressions between human and nonhuman animals lies in their occurrence. Human laughter and laugh faces with their sophisticated volitional as well as spontaneous forms are characteristic components of human everyday social interactions that may certainly vary in function and express, for instance, politeness, embarrassment, mocking and Schadenfreude [9,11,109,110]. They show a level of control that has, to our knowledge, not been found for animal play expressions, at least thus far.

(b) Mimicking and why it may be important for animals

The matching of expressions has a special role in animal play, where the expression of one playmate induces the same expression in another playmate. It has been mainly studied in the form of mimicry (e.g. dogs–horses [46], carnivores [58,62,111,112], primates [26,34,86,92,113]). Mimicry involves an automatic response system that is perhaps most easily observable as rapid mimicry, with a response latency of 1 s or less [114,115]. Rapid mimicry within short-distance communication has been predominantly observed in playful contexts, perhaps because they represent a platform for acquiring a range of social, emotional and cognitive skills [4,5].

The matching of animal play expressions, however, also comes in other types. For instance, delayed matching responses have been reported for primate open-mouth faces and play vocalizations [86,92,112]. Although caution is necessary when discussing why these responses were slower than rapid mimicry, it is interesting to note that humans sometimes respond more slowly when the behaviour is volitional, because additional neural processes are then involved compared with rapid mimicry (see [114,116,117]). Furthermore, the matching of play expressions among animals may range from being exact, i.e. with the same variant matched (e.g. open-mouth faces with upper lips raised [58,112]), to being distinct, i.e. with a different variant of the same expression matched (e.g. long laugh bouts seem to induce short laugh bouts [86]). Interestingly, previous studies have examined only dyadic constellations, so that research is needed to quantitatively explore if triadic facial expressions can occur in primates and other animals.

Thus, the matching of play expressions comes in various types, suggesting that they take up important functions among animals. Such matching is likely to heightened advantages that already come with spontaneously producing play expressions. Owing to its facial and vocal feedback component, it may serve even further as a social glue than spontaneous play expressions and may also contribute more to modulating interactions among playing animals [58,92]. In lowland gorillas, for instance, Bresciani and colleagues [118] found that such matching is prolonged when the facial response of the receiver mirrors the facial constellation of the playmate.

Although it can be problematic to link behavioural actions consistently with emotional states [119–121], expressions of play seem to be, in general, closely associated with positive affect in both nonhuman animals and humans (see [33]). Perhaps the context of gentle solitary play shows its link to positive states in animals most readily. For example, expressions produced by a young animal playing by her/himself are unlikely to have an interactive application value, making it reasonable to argue that such expressions are positive emotional outbursts. Such a link to affect may certainly be sustained during social interactions. Consequently, the mirror effect of play expressions may well be linked to elevated valence arousal states among playmates.

Two distinctive pathways that may lead to such an elevated state have received notable research attention [122–125]. First, the matching response is induced on a motoric level, a pathway that has been discussed in relation with behavioural contagion [126,127] as well as motor mimicry [122,128]. In this case, a spontaneous play expression triggers the same expression in the other playmate. Especially for motor mimicry, it has been argued that the motor resonance may trigger in the recipient the same emotional state experienced by the playmate [34,128,129]. However, emotions do not necessarily need to be involved when a behaviour is matched. The matching of behaviours may indicate, for instance, that the playmates are already in comparable states, perhaps in elevated positive states, which could help to prolong play. Interestingly, studies on animal play have shown that rapid facial mimicry and delayed laugh responses are linked to longer play bouts [34,58,62,86,92]. Whereas not all matching expressions in play must be linked to affect, it seems reasonable to argue that this association will strengthen over time, especially since young animals typically experience myriad playful events with familiar conspecifics.

Second, the matching response is affect-induced [122,130]. Here, a spontaneous play expression of a playmate causes an elevated positive state in the other playmate, a state that then induces the behavioural response. Whereas the two mentioned pathways may both result in elevated positive emotional states that are likely to benefit the playmates in multiple ways (for benefits of play, see [4,5]) it is difficult to determine which is the underlying path for the various forms of matching in animal play. What we know with more certainty is that any involved emotion state changes are likely to be minimal if the studied animals are already playfully interacting, i.e. in the same social–emotional context.

To systematically test for positive emotional contagion, where an emotional state spreads across individuals, it is important to have subjects socially and emotionally disconnected and, preferably, to examine them beyond a dyadic level [131,132]. An interesting study by Schwing and colleagues [75] on keas, a playful parrot species, demonstrated that played-back recordings of play vocalizations induce play behaviours in conspecifics previously involved in other behaviours. This study supports the notion that positive emotional contagion might not be a human-unique phenomenon. Interestingly, similar playback approaches carried out with chimpanzee laughter recordings did not show a comparable outcome in their conspecifics [133,134]. Unlike humans [135], chimpanzees do not seem to produce laughter merely based on hearing such vocalizations.

Mimicking and other types of behavioural matching within the context of play are also likely to be important for socially learning and practising a wide range of behaviours in humans and nonhuman animals [4,5,136]. In support of this notion, there is evidence that animals match the exact variant of the same expression of their playmates [58,112] and that the matching of play expressions may differ in form and function between social groups [86,137,138]. This brings us back to the Power Asymmetry Hypothesis [59], which could be extended to colony differences. Colonies may differ in the degree in which they show tolerance and aggression [139,140] and it seems reasonable to argue that clearer forms of communication may be essential when there is more risk of getting hurt (see [59]). Furthermore, the absence of the exposed upper teeth in the laughing chimpanzees at Burgers' Zoo, mentioned by Jan van Hooff in his pioneering work [41], and its presence in other chimpanzee colonies (see [26]) might indicate group differences regarding this facial feature. Exact matching of facial variants could help explain such potential differences. However, this topic requires further research.

Interestingly, there is some indication that the upper-teeth exposure develops at a later stage in immature primates [57,105], so that its occurrence throughout the developmental trajectory could depend on the exact matching mechanism and the social environment. More research is, however, needed on this topic. In a recent psychoacoustic study, Kret et al. [141] played back human infant laughter to adult participants, who were asked to determine the airflow direction from the recordings. The researchers found that the infants produced laughter increasingly as an egressive vocalization (i.e. a vocalization that is produced during the exhalation phase only; see [22,142]) over time and that this acoustic trait was perceived to be more positive by the adult listeners [141]. Such developmental findings could indicate that human infants already adjust laughter to their acoustic environment via social feedback.