Le bâillement, du réflexe à la pathologie
Le bâillement : de l'éthologie à la médecine clinique
Le bâillement : phylogenèse, éthologie, nosogénie
 Le bâillement : un comportement universel
La parakinésie brachiale oscitante
Yawning: its cycle, its role
Warum gähnen wir ?
 
Fetal yawning assessed by 3D and 4D sonography
Le bâillement foetal
Le bâillement, du réflexe à la pathologie
Le bâillement : de l'éthologie à la médecine clinique
Le bâillement : phylogenèse, éthologie, nosogénie
 Le bâillement : un comportement universel
La parakinésie brachiale oscitante
Yawning: its cycle, its role
Warum gähnen wir ?
 
Fetal yawning assessed by 3D and 4D sonography
Le bâillement foetal
http://www.baillement.com

mystery of yawning 

 

 

mise à jour du
24 février 2021
J Comp Psychol.
2021;135(1):98-113.
Facial displays in red-capped mangabeys
(Cercocebus torquatus):
Repertoire, social context, and potential intentionality
Aychet J, Blois-Heulin C, Palagi E, Lemasson A.

Chat-logomini

 
Primate communication relies strongly on the visual modality, notably through the production of a wide range of expressive facial signals. The authors investigated here the facial display repertoire of a relatively little-studied cercopithecid species, red-capped mangabeys (Cercocebus torquatus), and questioned whether their facial displays were dependent on social contexts and accompanied by indices of intentionality. Although the dual intentional and emotional use of apes' facial expressions has recently been suggested, the question of whether monkeys produce facial expressions intentionally to communicate remains open. They described 6 facial displays produced by captive red-capped mangabeys in social contexts. They are based on movements of the mouth, eyebrows, and ears, possibly graded in intensity and produced independently or in combination. They showed that most of the facial displays were produced preferentially in specific social contexts and that repertoires varied with subjects' characteristics, highlighting the communicative function of these displays. Moreover, behavioral markers of intentionality commonly used in gestural studies were found to accompany the production of some of the facial signals observed. Particularly, playful "open mouth" appeared strongly associated with intentionality indices, as previously noticed in ape species. All other facial displays, except yawns, did not exhibit all defined intentional indices but were, at least, directed toward a recipient. Interestingly, yawns presented different variants of intensity associated presumably with different social functions. Altogether, these results emphasize the communicative function of red-capped mangabeys' facial displays and provide a basis for further research on their intentional communication.
 
La communication des primates repose fortement sur la modalité visuelle, notamment à travers la production d'une large gamme de signaux faciaux expressifs. Les auteurs ont étudié ici le répertoire d'affichage facial d'une espèce de cercopithécides relativement peu étudiée, les mangabeys à tête rouge (Cercocebus torquatus), et nous nous sommes demandé si leurs affichages faciaux dépendaient des contextes sociaux et étaient accompagnés d'indices d'intentionnalité. Bien que la double utilisation intentionnelle et émotionnelle des expressions faciales des singes ait récemment été suggérée, la question de savoir si les singes produisent des expressions faciales intentionnellement pour communiquer reste ouverte. Ils décrivent 6 affichages faciaux produits par des mangabeys à tête rouge captifs dans des contextes sociaux qui sont basés sur des mouvements de la bouche, des sourcils et des oreilles, éventuellement gradués en intensité et produits indépendamment ou en combinaison. Ils ont montré que la plupart des affichages faciaux étaient produits préférentiellement dans des contextes sociaux spécifiques et que les répertoires variaient selon les caractéristiques des sujets, mettant en évidence la fonction communicative de ces affichages. De plus, des marqueurs comportementaux d'intentionnalité couramment utilisés dans les études gestuelles se sont avérés accompagner la production de certains des signaux faciaux observés. En particulier, la "bouche ouverte" ludique est apparue fortement associée aux indices d'intentionnalité, comme cela a été remarqué précédemment chez les espèces de singes. Tous les autres affichages faciaux, à l'exception des bâillements, ne présentaient pas tous les indices intentionnels définis mais étaient, au moins, dirigés vers un receveur. Fait intéressant, les bâillements présentaient différentes variantes d'intensité associées vraisemblablement à différentes fonctions sociales. Dans l'ensemble, ces résultats soulignent la fonction communicative des affichages faciaux des mangabeys à coiffe rouge et fournissent une base pour d'autres recherches sur leur communication intentionnelle.
Introduction
Primate communication strongly relies on the visual modality (Liebal, Waller, Burrows, & Slocombe, 2014b; Redican, 1975) through a variety of gestural and postural signals. Moreover, compared to other mammals, primates produce a wide range of expressive facial movements, owing to their complex facial musculature (Burrows, 2008; Diogo, Wood, Aziz, & Burrows, 2009). Facial expressions can involve movements of the mouth and lips, eyelids and eyebrows, forehead and ears (van Hooff, 1967; Waller & Micheletta, 2013). Substantial variations are notable between primate genera and species regarding facial mobility (Santana, Dobson, & Diogo, 2014) and facial expression repertoires (Marler, 1965; van Hooff, 1967). Thus, the understanding of visual communication in primate species needs precise descriptions of their facial expression repertoire and the context of production as their communicative signals. Some authors include facial displays in the gestural domain (Altmann, 1962; Call & Tomasello, 2007; Chalmers, 1968a; Dube, 2013; Hesler & Fischer, 2007; Hinde & Rowell, 1962; Hostetter, Cantero, & Hopkins, 2001; Maestripieri, 1997, 2005; Poss, Kuhar, Stoinski, & Hopkins, 2006; Wolfheim & Rowell, 1972), whereas others consider facial displays as disen- tangled from gestures owing to differences in the definitions of signal intentionality (Byrne et al., 2017; Genty, Breuer, Hobaiter, & Byrne, 2009; Graham, Furuichi, & Byrne, 2017; Hobaiter & Byrne, 2011a; Liebal, Call, Tomasello, & Pika, 2004; Liebal, Pika, & Tomasello, 2006; Pollick & de Waal, 2007; Roberts, Roberts, & Vick, 2014). Indeed, the intentional use of facial expressions in primates is still a matter of debate (Arbib, Liebal, & Pika, 2008; Hopkins, Taglialatela, & Leavens, 2011; Slocombe, Waller, & Liebal, 2011).
 
Nonhuman primates seem to have a voluntary control of their facial movements, which is required for the flexible use of facial expressions. Regarding neuroanatomical data from humans and great apes, facial expressions are controlled by both voluntary and emotional pathways acting in parallel within the facial motor system that differs from the motor system controlling limb movements (Cattaneo & Pavesi, 2014; Mu_ri, 2016; Sherwood, Hollo- way, Erwin, & Hof, 2004; Sherwood, Hof, et al. 2005; Sliwa, Takahashi, & Sheperd, 2018). Moreover, some facial expression asymmetries are associated with attention-getting calls in chimpanzees, suggesting voluntary control of these facial movements (Hopkins et al., 2011; Reynolds-Losin, Russell, Freeman, Meguerditchian, & Hopkins, 2008). This possible voluntary con- trol may reveal that primates could use facial expressions intentionally as communication signals.
 
The distinction between intentional behaviors and intentional communication was previously addressed by Bretherton and Bates (1979) in their developmental studies in children, based on the implication of social agents within intentional sequences of behavior. Then, the philosophical stance of intentional communication as a goal-directed process was formulated by Dennett (1983), who defined different degrees of intentionality. The first degree, as studied in animal communication, supposes that an intentional signal would be controlled and produced voluntarily by a signaler and directed toward a recipient to trigger the appropriate response by the receiver (Hobaiter & Byrne, 2013; Roeder & Gosset, 2001). The goal-directedness discriminates intentional communicative signals from mere informative signals, which convey messages without being voluntarily emitted by the signaler (Poggi & D'Errico, 2012). One example of "informative signal" could be piloerection, which conveys information about the individual's internal state (Benedek & Kaernbach, 2011; Dettling, Pryce, Mar- tin, & Do_beli, 1998) without appearing voluntarily emitted. Bard (1992) introduced Bates' definition of intentional communication to the study of primate communication, distinguishing intentional communication from other intentional behaviors in orangutan mother-infant interactions. In the same line, several behavioral indices of intentional communication inherited from child devel- opmental studies were then applied to nonhuman gestural studies to assess intentionality (see, for reviews, Byrne et al., 2017; Leavens, Russell, & Hopkins, 2005; Liebal, Waller, Burrows, & Slocombe, 2014c; Schel, Bono, Aychet, Pika, & Lemasson, 2020; Townsend et al., 2017). Townsend et al. (2017) recently gathered these indices in a general framework implying three main criteria to attribute intentionality to communicative signals: The signal has to be directed toward a recipient, produced to reach a particular social goal, and has to be followed by a response from the recipient that is consistent with the goal that is deduced for the signaler. This last criterion, related to the communicative value of the signal (Altmann, 1968; Green, 1975; Marler, 1967), would imply that the signal is followed by a consistent change in the recipient's behav- ior, consistency with signaler's intention corresponding to an ac- ceptance or a refusal of signaler's goal. Following this framework, appropriate behavioral markers have thus to accompany the pro- duction of the signal so that it could be associated with intention- ality. These behavioral markers may have validity limitations to admit a signal's intentionality, as each of them can also be ex- plained by unintentional processes, such as learning or changes in signaler's emotional state (Graham, Wilke, Lahiff, & Slocombe, 2019; Liebal, Waller, Burrows, & Slocombe, 2014a). It has been argued that the more indices observed, the more we get confidence that the criteria are fulfilled and so that the signal might be intentional (Schel, Townsend, Machanda, Zuberbu_hler, & Slocombe, 2013), considering that several convergent indices are more likely to be explained by a single cognitive mechanism (intentionality) rather than a series of lower level cognition expla- nations (Byrne & Bates, 2006; Liebal et al., 2014a; Townsend et al., 2017). However, these markers have to be used cautiously when studying animal signals, with due consideration to all possible cognitive mechanisms underlying signal production.
 
A signal is considered as recipient-directed if it is produced in the presence of potential receivers (i.e., "socially used") and physically directed to another individual (Call & Tomasello, 2007; Leavens & Hopkins, 1998; Leavens, Hopkins, & Bard, 1996; Leavens, Hopkins, & Thomas, 2004; Liebal, Pika, & Tomasello, 2004; Liebal et al., 2014a; Schel et al., 2020; Tomasello et al., 1985; Tomasello, Call, Nagell, Olguin, and Carpenter, 1994). To discriminate socially dependent signals from others, supplemen- tary indicators should be taken into account, such as the fact that the signaler first checked the audience's attentional state before producing the display (Genty et al., 2009; Graham et al., 2017; Hobaiter & Byrne, 2011a; Schel et al., 2020), at least in case of signals that have to be perceived on the visual channel. Further- more, the sensitivity to the attentional state of the recipient, both regarding the rate or the modality of the signal, is often studied as an indicator of intentional use of gestures (Bourjade, Meguerditchian, Maille, Gaunet, & Vauclair, 2014; Demuru, Ferrari, & Palagi, 2015; Leavens, Hostetter, Wesley, & Hopkins, 2004; Liebal, Pika, et al., 2004; Maille, Engelhart, Bourjade, & Blois- Heulin, 2012; Meunier, Prieur, & Vauclair, 2013; Poss et al., 2006). The second dimension of intentional communication, that is, the goal-directedness of the signal, could be checked by monitoring response-waiting from the signaler after it produced the display (Call & Tomasello, 2007; Cartmill & Byrne, 2010; Gra- ham et al., 2017; Hobaiter & Byrne, 2011a; Liebal, Pika, et al., 2004; Molesti, Meguerditchian, & Bourjade, 2020; Pika, Liebal, & Tomasello, 2003; Roberts et al., 2014; Roberts, Vick, Roberts, Buchanan-Smith, & Zuberbu_hler, 2012; Schel et al., 2020; Toma- sello et al., 1985, 1994). Moreover, goal-directedness is recognized when the signaler stops producing the signal once the presumed goal has been reached, and repeats and/or elaborates the signals in case the goal has not been met yet (Cartmill & Byrne, 2007, 2010; Genty, Clay, Hobaiter, & Zuberbu_hler, 2014; Graham et al., 2017; Gupta & Sinha, 2019; Hobaiter & Byrne, 2011b; Hostetter et al., 2001; Leavens et al., 2005; Liebal, Call, Tomesella, Pika, 2004; Liebal et al., 2014a; Roberts, Vick, & Buchanan-Smith, 2013).
Based on these indices, the intentionality of great apes' com- munication gestures has been admitted, notably from Tomasello et al. (1994)'s work on chimpanzee sensitivity to the receiver's attentional state when gesturing, followed by numerous field and lab studies on great apes (see, for reviews, Byrne et al., 2017; Call & Tomasello, 2007; Liebal et al., 2014c). Intentional gestures have been explored recently in monkey species such as macaques (Macaca mulatta, tonkeana, and radiata, Canteloup, Bovet, & Meunier, 2015a, 2015b; Deshpande, Gupta, & Sinha, 2018; Gupta & Sinha, 2019), olive baboons (Papio anubis, Bourjade et al., 2014; Meunier et al., 2013; Molesti et al., 2020), red-capped mangabeys (Cercocebus torquatus, Aychet et al., 2020; Maille et al., 2012; Schel et al., 2020), squirrel monkeys (Saimiri sciureus, Anderson, Kuroshima, Hattori, & Fujita, 2010). However, studies of potential intentionality markers accompanying the production of primates' facial expressions are still rare.
 
Some observational studies seem to indicate the social dependency of apes' facial expressions. The facial expressions of five gibbon species were shown to last longer when the signaler and the receiver were facing each other in social contexts (Scheider, Waller, On_a, Burrows, & Liebal, 2016). This social directedness and sensitivity to receiver's attentional state were also observed in orangutans (Pongo pygmaeus) playful facial expressions as they produce more intense play faces when facing a visually attentive play partner (Waller, Caeiro, & Davila-Ross, 2015), and in bono- bos (Pan paniscus), who display this facial expression more during social than solitary play, and more in the presence of a visually attentive recipient (Demuru et al., 2015). Thus, the audience effect supports the communication function of the play faces in these species and their social-directedness may reveal both their emo- tional and intentional use. Moreover, four of siamangs' (Sympha- langus syndactylus) facial expressions, mostly performed in combination with tactile or visual gestures, were found to be recipient-directed and accompanied by response waiting or persistence (Liebal, Pika, et al., 2004). Assumptions of intentional facial displays were also made for orangutans (Pongo pygmaeus), with the same recipient-directedness and goal-directedness indices found for both gestures and facial expressions (Cartmill & Byrne, 2010).
 
Although observations of apes suggest the dual intentional and emotional use of facial expressions, the question whether monkeys produce facial expressions intentionally to communicate remains open. To our knowledge, even though the facial expressions of numerous catarrhine and platyrrhine monkey species have been described (e.g., Macaca genus: Hesler & Fischer, 2007; Hinde & Rowell, 1962; Maestripieri, 2005; Partan, 2002; Thierry et al., 2000; gray-cheeked mangabeys: Lophocebus albigena, Chalmers, 1968a; hamadryas baboons: Papio hamadryas, Dube, 2013; Talapoins: Miopithecus talapoin, Wolfheim & Rowell, 1972; capu- chin monkeys: Cebus apella, De Marco & Visalberghi, 2007; Visalberghi, Valenzano, & Preuschoft, 2006; Weigel, 1979; and see comparative studies of Redican, 1975; van Hooff, 1967), there is no evidence that nonhominoid monkeys produce facial expressions in association with behavioral markers of intentional communication, as reported for gestures. The investigation of this topic is necessary to fill the gap in understanding both monkeys' facial signal functions and the evolutionary history of intentional communication, which is of prime interest in the study of language origin (Fro_hlich, Sievers, Townsend, Gruber, & van Schaik, 2019; Hauser, 1996; Hauser, Chomsky, & Fitch, 2002).
 
Here, we investigated the facial display repertoire of a relatively little-studied cercopithecid species, red-capped mangabeys (Cercocebus torquatus), and questioned whether these facial displays were dependent of social contexts and accompanied by indices of intentionality. Red-capped mangabeys are semiterrestrial primates (Mitani, Call, Kappeler, Palombit, & Silk, 2012) that originate from West African rainforest coastal regions (Cooke, 2012; Gautier-Hion, Colyn, & Gautier, 1999; Jones & Sabater-Pi, 1968; Orimaye, 2017). They have colored hair and skin and white eyelids that contrast with their black face (see Figure 1), presumably in accordance with a communication function of facial expressions (Gautier-Hion et al., 1999; Hill, 1974; Redican, 1975). Their gestural communication has been described recently (Schel et al., 2020), and suggests, together with data on their multimodal inter- actions (Baraud, Deputte, Pierre, & Blois-Heulin, 2016), a wide use of visual signals for communication. Experimental studies highlighted the ability of red-capped mangabeys to use a learnt begging gesture intentionally, as a function of visual attentional state of a human receiver (Aychet et al., 2020; Maille et al., 2012). Moreover, spontaneous intraspecific gestures produced by red- capped mangabeys were associated with key-markers of intentional communication (Schel et al., 2020). Thus, we hypothesized that the intentional use of visual displays as communication signals by this species would encompass facial expressions. Our first aim was to describe the repertoire of facial displays produced by red-capped mangabeys in social situations and test how this repertoire depends on individual characteristics. Moreover, in accor- dance with a presumed communication function, we expected that facial displays would be performed under particular social con- texts, whereas their production in nonsocial contexts would con- note a predominant emotional component underpinning the dis play. Then, we hypothesized that facial displays would be part of this species' intentional dyadic communication, and thus that they would be accompanied by behavioral markers of intentionality, similarly to gestures.
  
Discussion
This study explored the repertoire of six facial displays of a captive population of red-capped mangabeys, and tested whether these displays were produced socially and accompanied by indices of intentional communication. This preliminary assessment of first-order intentionality in mangabeys' facial displays was based on behavioral markers similar to those used in gestural studies, to evaluate whether signals were (a) directed toward a recipient, (b) produced in order to reach a specific goal, and (c) led to a change in the recipient's behavior (Townsend et al., 2017). The social condition of mangabeys' facial displays, as well as the effects of individual characteristics on the repertoire, supports their communication function. The playful "open mouth" displays were produced exclusively in social contexts and in most cases in the direction of an attentive recipient, with all defined markers of intentional communication. We discuss later the significance of these results regarding the potential intentionality of these dis- plays, given the current limitations of the behavioral method use to assess first-order intentionality in animal communication (Graham et al., 2019). Although other cognitive mechanisms may explain our observations, the present results constitute a first step to address the potential intentionality of red-capped mangabeys' facial displays, particularly the "open mouth." Although not meeting criteria for goal-directed signals for all occurrences, other facial displays ("raise eyebrows," "ears back," lipsmacks and "stick tongue out"), except yawns were at least recipient-directed in most cases. Finally, yawns seemed not associated with dyadic communication, and thus did not fulfill the defined criteria for intentional communication. Interestingly, these displays presented different variants of intensity, which were presumably associated with different functions.
 
We described six facial displays produced by red-capped mang- abeys, which involve movements of the ears, eyebrows or mouth, and can be produced in combination or independently one from the other. One noticeable point of the present repertoire is the description of unitary ear displays used socially by red-capped mangabeys. Although nonhuman primates' ear musculature is controlled by facial nerves (Mu_ri, 2016), few studies of facial displays describe ear movements (but see in rhesus and long-tailed macaques, Macaca mulatta and fascicularis: Chance, Emory, & Payne, 1977; Hinde & Rowell, 1962; Partan, 2002). Although most facial displays of the repertoire are discrete, we distinguished variants for two of them, corresponding to different intensities of mouth displays. The repertoire size and this possible gradualness of the morphology of some facial displays suggest that red-capped mangabeys' communication relies strongly on complex visual cues (Dobson, 2012; Freeberg, Dunbar, & Ord, 2012).
 
Thus, the "open mouth" displays, highly related to play con- texts, appeared graded. Some facial displays are indeed graded signals that can be associated with different motivations of the signaler (Marler, 1965; Parr, Cohen, & De Waal, 2005; Preuschoft & van Hooff, 1996; Waller & Micheletta, 2013). In primates, the variants "open mouth fully" and "half" have been largely de- scribed as playful facial displays (Chevalier-Skolnikoff, 1994; van Hooff, 1967). Moreover, subjects' age influenced the production of "open mouth fully" and "half," as younger subjects were in- volved in playful interactions more frequently and were possibly more emotionally involved in play than adults (Demuru et al., 2015). Although primate gestures are used in variable contexts (Byrne et al., 2017; Call & Tomasello, 2007; Hobaiter & Byrne, 2013; Liebal et al., 2014c), the fact that facial displays are more context-specific allows us to hypothesize that their function is to help elucidate the meaning of associated signals used in multi- modal communication. These facial displays may act as "meta- communicative signals" (Altmann, 1967; Bateson, 1955), notably because play might involve agonistic-like gestures (Demuru et al., 2015; Palagi, 2008; Waller & Cherry, 2012).
 
Red-capped mangabeys generally produced "open mouth fully" and "open mouth half" in association with all previously defined behavioral markers of intentionality, and significantly more in situations when the receiver was visually attentive. Therefore, our results on red-capped mangabeys' directed "open mouth" are con- sistent with previous observations of great apes. Bonobos' (Pan paniscus: Demuru et al., 2015) and orangutans' (Pongo pygmaeus: Waller et al., 2015) homologous facial displays were suggested to be dependent on audience attentional state, indicating that these facial expressions are not an automatic response to play.
 
However, we should be careful before drawing the conclusion that these "open mouth" displays are produced intentionally, given some alternative explanations to audience effect. For instance, the production of "open mouth" preferentially in front of an attentive recipient could be the result of a learnt discrimination. Individuals may be more likely to produce a facial display in these conditions because they have learnt that seeing the face of a recipient is necessary to obtain a response from their signal (Graham et al., 2019). Moreover, the social use of "open mouth" and the sensi- tivity to the recipient's attentional state cannot rule out an emotional-based production of these displays. Eye gaze is an emotional stimulus in primates (Emery, 2000), and arousal is prone to be higher in the presence of others compared to alone (Zajonc, 1965). Thus, these factors could induce changes in the signaler's emotional state, resulting in the production of the facial expressions. Besides, observations in apes suggest that the produc- tion of play faces has a strong emotional component. This expres- sions can be observed during bonobos' (Pan paniscus) solitary play (Demuru et al., 2015), and the example of a female gorilla (Gorilla gorilla) hiding her play face (Tanner & Byrne, 1993) suggests that she was not able to inhibit this facial display, as it is the case of humans experiencing intense emotions (Porter, ten Brinke, & Wallace, 2012). We found concomitant markers of other criteria of intentional communication (i.e., waiting for a response from the signaler and recipient's consistent behavioral response) accompanying red-capped mangabeys' "open mouth," but one particular limit of the present observations is that "response wait- ing" should be cautiously interpreted as a goal-directedness marker per se (Graham et al., 2019; Liebal et al., 2014a). As a first insight in the intentionality of facial displays in red-capped mangabeys, the present framework about "goal-direction" did not include markers related to goal-dependent cessation of communication or persistence and elaboration, because these aspects of signal's "aboutness" require that the observer is knowledgeable of the signal function and thus presumed signaler's intention displaying it. Thus, we must remain cautious when concluding about the possible intentionality of the signals studied, given that criteria of recipient-directedness and goal-directedness were both assessed based on markers related to visual attention only. Moreover, while analyzing intentionality at the signal-level, this study did not address yet the question of interindividual variability in intentional communicative facial displays. In this sense, the present results will need to be supplemented by further investigations of the goal-directedness of "open mouth fully" and "half," which pre- sumably corresponds to an invitation to play or to maintain a playful interaction. We believe that a multimodal approach would be beneficial in this regard, addressing goal-persistence by taking into account associated vocalizations and gestures.
Overall, there are current validity limitations in the behavioral method used to assess first-order intentionality in animal commu- nication (see the abovementioned examples of other cognitive processes possibly underlying the behaviors identified as inten- tionality markers). The framework proposed by Townsend et al. (2017) has the advantage to be widely applicable to diverse species and communication modalities, although it does not permit to conclude with certainty that signals are intentionally produced. Graham and colleagues (2019) suggested some perspectives in light of this methodological problem, such as experimentally ad- dress second-order intentionality (which implies mental-state at- tribution, see Dennett, 1983), or complete observational studies with valid measures of arousal during signal production. Until such approaches are applicable, behavioral assessment of first-order intentionality still permits to gather meaningful information, such as the present results, which we believe constitute a first step in assessing the intentionality of red-capped mangabey facial dis- plays, particularly for "open mouth."
 
Other facial displays "raise eyebrows," "ears back," lipsmacks and "stick tongue out" appeared only partially accompanied by defined behavioral markers of intentionality. However, these dis- plays were at least recipient-directed in most of the cases, thus indicating their communicative function (Call & Tomasello, 1994; Leavens et al., 1996). Eyebrow-raising and ears back were exclusively produced in social contexts, that is, agonistic and affiliative contexts, respectively, and mostly directed to a visually attentive recipient. Lipsmacks were produced more in social than nonsocial contexts, particularly in affiliative and grooming situations, but not more in front of an attentive than an inattentive recipient, as found for "stick tongue out." Although lipsmacks' production was not dependent on recipient's visual attention, we may hypothesize that this signal can also be acoustically perceived (as the "audible smacking sound" produced by macaques: Maestripieri, 2005; Mae- stripieri & Wallen, 1997), and therefore visual attention would not be necessary. We note that most of the above-cited facial displays described, except "stick tongue out," were preferentially or exclusively produced in social contexts. Lipsmacks were produced in affiliative and grooming contexts, as found for instance in olive baboons (Papio anubis: Easley & Coelho, 1991), but also in nonsocial contexts, especially during auto-grooming. This enabled us to hypothesize that lipsmacking may be an automatic response to grooming actions and, as for nonsocial production of yawns, might indicate a prominent emotional component driving the pro- duction of this facial display.
 
Yawns, as opposed to the other facial displays, seem to be disentangled from dyadic interactions, as even in social contexts we found almost no signs of recipient-directed use. The high frequency of yawns in nonsocial context is in agreement with the hypothesis that yawns are strongly linked to internal states of the signaler (Deputte, 1994). Moreover, regarding their production in social conditions, yawns may have an informative function, for conspecifics, on the signaler's emotional state during agonistic and alarm contexts (Bolwig, 1959; Deputte, 1994). An interesting point concerning this particular display lies in the relationship between its form and function, as red-capped mangabeys produced different types of yawns, with regard to mouth opening and teeth uncover- ing, in different contexts. These variations are known and characterized for gray-cheeked mangabeys (Lophocebus albigena: Deputte, 1994), as well as for macaque species (Macaca fascicu- laris: Deputte, 1994; Zannella, Stanyon, & Palagi, 2017), geladas (Theropithecus gelada: Palagi, Leone, Mancini, & Ferrari, 2009), and chimpanzees (Pan troglodytes: Vick & Paukner, 2010).
 
Yawns of the first type, with mouth less open, were used only in nonsocial contexts and might correspond to physiological regulators during transition between activity and resting or during feed- ing anticipation (Deputte, 1978). Some yawns of the second and third types, with mouth more open and teeth more visible were associated with alarm calls, as described also for gray-cheeked mangabeys (Lophocebus albigena: Deputte, 1994), apparently be- cause of high arousal and social tension. Moreover, yawns of the third type were mainly produced in agonistic contexts, as visible canines are hypothesized to play the role of threat or prethreat (Altmann, 1967; Redican, 1975). Yawns of the third type were produced more frequently by middle-aged and old adult males, while individual characteristics did not affect the production of yawns of the first type. In our sample, the males were not equally distributed in the social groups, most of them living in all-male groups. However, as sex influenced some variants of yawns and not the others, we assume that male group membership did not bias the analysis of the effect of sex on facial display repertoire. These results on sex effects are consistent with the hypothesis of a threat value of the yawns with visible canines (Leone, Ferrari, & Palagi, 2015; Zannella et al., 2017), particularly because red-capped mangabeys exhibit an important sexual dimorphism in canine size that becomes evident with sexual maturity (Deputte, 1986; Hill, 1974). This hypothesis is supported by similar sexual differences of yawn production by macaque species (Macaca fuscata and tonkeana: Zannella et al., 2017), taken that this difference was not found for primate species such as lemurs (Propithecus verreauxi and Lemur catta: Zannella, Norscia, Stanyon, & Palagi, 2015) or humans (Homo sapiens: Schino & Aureli, 1989), with absent or slight sexual dimorphism in canine-size. Moreover, the age effect on yawn production may be associated with testosterone level changing with sexual maturity since testosterone was correlated with yawn production in rhesus macaques (Macaca mulatta: Phoe nix & Chambers, 1982; Robinson, Scheffler, Eisele, & Goy, 1975; Wallen & Goy, 1977). The relationship between yawn morpho- logical variability and function, highlighted by the contexts of production and by the effect of individual characteristics, is in agreement with previous results for geladas (Theropithecus ge- lada: Leone et al., 2015). In macaques, this was associated with communication redundancy observed in species showing more tolerant social styles (Zannella et al., 2017), which also exhibit larger meaningful communicative repertoires (Dobson, 2012).
 
Moreover, we found no dominance or kinship effects on any of the facial expressions displayed, presumably in relation with the sociality of red-capped mangabeys, which have been shown to adopt both despotic and egalitarian behaviors (Dolado & Beltran, 2012). Despite a steep hierarchy within their groups, red-capped mangabeys have indeed relatively dynamic dominance relation- ships (Dolado & Beltran, 2011) with some propensity to counter- attack, as do tolerant macaque species (De Waal & Luttrell, 1989; Thierry, 2007).
 
Conclusion
In sum, the repertoire of red-capped mangabeys' facial displays is composed by six different displays that could be graded in intensity. We characterized the social-dependent production of facial expressions by red-capped mangabeys, and brought to light behavioral indices of intentionality accompanying facial displays. Particularly, playful "open mouth" appeared strongly associated with some intentionality indices, as previously noticed in ape species. We believe that the present repertoire provides a basis for further research on function and intentionality of red-capped mangabeys' facial expressions. As the gestural, vocal, and facial signals are now described in this species, we also believe that it will constitute a useful tool for a multimodal approach for the study of catarrhine monkeys' intentional communication, a key feature for evolutionary research on language origins.