In primates, yawn contagion (the yawning
response elicited by others' yawn) is variably
influenced by individual (e.g., sex, age) and
social factors (e.g., familiarity) and possibly
linked to interindividual synchronization,
coordination, and emotional contagion. Two out
of three studies on yawn contagion in bonobos
(Pan paniscus), found the presence of the
phenomenon with mixed results concerning the
effect of familiarity and no replication on its
modulating factors. To address this puzzling
issue, the authors recorded all occurrences data
on yawn contagion in a captive bonobo group
(March-June 2021; 18 individuals; La
Vallée des Singes, France). Contrary to
chimpanzees and humans, the number of triggering
yawns increased contagion, possibly owing to a
higher stimulus threshold. This aspect may
explain the interindividual variability observed
in yawn contagion rates. In subjects under
weaning, they did not detect yawn contagion and,
as it occurs in certain human cohorts, yawn
contagion declined with age, possibly due to
reduced sensitivity to others. Females responded
more than males and elicited more responses from
females when showing sexual swelling. As
reproductive females are central in bonobo
society, our results support the hypothesis
that-as in other Hominini-the most influential
sex can influence yawn contagion. The
relationship quality (measured via
grooming/play) did not affect yawn contagion,
possibly due to bonobos' xenophilic nature.
Overall, this study confirms the presence of
yawn contagion in bonobos and introduces new
elements on its modulating factors, pointing
toward the necessity of cross-group
studies.
Chez les primates, la contagion du
bâillement (la réponse au
bâillement provoquée par le
bâillement des autres) est
influencée de manière variable par
des facteurs individuels (par exemple, le sexe,
l'âge) et sociaux (par exemple, la
familiarité) et peut-être
liée à la synchronisation
interindividuelle, la coordination et la
contagion émotionnelle. Deux des trois
études sur la contagion des
bâillements chez les bonobos (Pan
paniscus) ont révélé la
présence du phénomène avec
des résultats mitigés concernant
l'effet de la familiarité et aucune
confirmation de ses facteurs modulateurs.
Pour répondre à cette
question, les auteurs ont enregistré
toutes les occurrences de contagion des
bâillements dans un groupe de bonobos en
captivité (mars-juin 2021 ; 18 individus
; La Vallée des Singes, France).
Contrairement aux chimpanzés et aux
humains, le nombre de bâillements
déclencheurs augmente la contagion,
peut-être en raison d'un seuil de stimulus
plus élevé. Cet aspect peut
expliquer la variabilité
interindividuelle observée dans les taux
de contagion des bâillements. Chez les
sujets en cours de sevrage, nous n'avons pas
détecté de contagion des
bâillements et, comme cela se produit dans
certaines cohortes humaines, la contagion des
bâillements a diminué avec
l'âge, peut-être en raison d'une
sensibilité réduite aux autres.
Les femelles répondaient davantage que
les mâles et suscitaient davantage de
réponses de la part des femelles
lorsqu'elles montraient un gonflement sexuel.
Comme les femelles reproductrices jouent un
rôle central dans la société
des bonobos, nos résultats soutiennent
l'hypothèse selon laquelle, comme chez
les autres hominidés, le sexe est le plus
influant peut moduler la contagion du
bâillement. La qualité de la
relation (mesurée par le toilettage/le
jeu) n'a pas affecté la contagion du
bâillement, peut-être en raison de
la nature xénophile des bonobos. Dans
l'ensemble, cette étude confirme la
présence de la contagion du
bâillement chez les bonobos et introduit
de nouveaux éléments sur ses
facteurs modulateurs, soulignant la
nécessité d'études
intergroupes.
INTRODUCTION
While spontaneous yawning is not dependent
on the detection of others' yawns, contagious
yawning occurs when the yawn emitted by an
individual (hereafter trigger) works as a
releasing stimulus (sensu Tinbergen &
Perdeck, 1950) and induces yawning in another
in- dividual (hereafter responder) (Provine,
1989). Spontaneous yawning (or a yawning_like
morphological pattern) is likely a plesiomorphic
display because it is present in a wide range of
vertebrates (Baenninger, 1987), including human
(Homo sapiens, Provine, 1986, 2012) and
non_human primates (Anderson, 2020).
Contagious yawning between conspecifics is
possibly an apo- morphic phenomenon described so
far in a limited array of species (Palagi et
al., 2020). From an adaptive point of view, yawn
contagion can promote synchronization and
coordination of activities within social groups
(Palagi et al., 2020). Moreover, it can be the
expression of interindividual physiological
resonance (Prochazkova & Kret, 2017) and
possibly emotional contagion, a powerful driver
of prosocial behavior (de Waal & Preston,
2017).
Experimental and naturalistic studies on
chimpanzees (e.g., Anderson et al., 2004;
Campbell & Cox, 2019; Campbell & de
Waal, 2011) and humans (e.g., Bartholomew &
Cirulli, 2014; Chan & Tseng, 2017; Norscia
& Palagi, 2011; Norscia et al., 2021a;
Provine, 1986, 1989) have consistently found
intraspecific yawn contagion. In bo- nobos the
situation is not as much clear. Amici et al.
(2014) examined whether yawning was subject to
response facilitation triggered by videorecorded
yawns from conspecifics. They found that
chimpan- zees (14 subjects) but not bonobos (4
subjects) yawned significantly more while or
after watching a familiar conspecific yawning on
video. On the other hand, on a larger sample (25
subjects), Tan et al. (2017) found that bonobos
showed evidence for involuntary, contagious
yawning in response to videos of yawning
conspecifics. Finally, Demuru and Palagi (2012)
also reported yawn contagion in captive bonobos
(12 subjects) based on ethological observations
under nat- uralistic conditions. Hence, yawn
contagion as a social signal might have been
present in the last common ancestor between Pan
and Homo.
Beyond Hominini, it is not possible to
associate the emergence of yawn contagion with a
single common ancestor. Yawn contagion was not
detected in lowland gorillas (Gorilla gorilla
gorilla; Amici et al., 2014; Palagi et al.,
2019) but it was found in orangutans (Pongo
spp.; van Berlo et al., 2020) which separated
earlier from the human line (Groves, 2018).
Interestingly, lowland gorillas show low
affiliation levels (Palagi et al., 2019) whereas
orangutans do not form social groups but
orangutans might have been more social in the
past (Harrison & Chivers, 2007). In
non_hominid primates, yawn contagion studies
show mixed results (cf. geladas, Theropithecus
gelada, Gallo et al., 2021; Palagi et al., 2009;
Tonkean macaque, Macaca tonkeana, Palagi &
Norscia, 2019; but see: stump_tailed macaques,
Macaca arctoides: Paukner & Anderson, 2006;
Japanese macaque, Macaca fuscata, Palagi &
Norscia, 2019). Finally, no evidence of yawn
con- tagion was found in strepsirrhines (Lemur
catta and Varecia variegata, Reddy et al., 2016)
even though contagious yawning is present in non
primates (Gallup et al., 2015; for review:
Palagi et al., 2020). Hence, yawning might have
been co_opted as a communicative signal mul-
tiple times over the course of the evolution, in
relation to the type of sociality.
When present, yawn contagion in primates
usually occurs in the few minutes following the
yawning stimulus (hereafter trig- gering yawn)
with a peak in the first minute in Hominini
(e.g., humans: Palagi et al., 2014; chimpanzees,
Pan troglodytes: Campbell & Cox, 2019; and
bonobos: Demuru & Palagi, 2012). In humans,
perceptual factors may influence the yawning
response probability (Massen & Gallup, 2017;
Norscia et al., 2020). How- ever, the distance
between trigger and responder and/or the number
of observed triggering yawns were not found to
affect yawn contagion (humans: Norscia &
Palagi, 2011; chimpanzees: Campbell & Cox,
2019; geladas: Palagi et al., 2009).
Yawn contagion can be influenced by
individual and social fac- tors (Palagi et al.,
2020). The age of the responder can affect yawn
contagion rates in some cohorts of humans
(Anderson & Meno, 2003; Bartholomew &
Cirulli, 2014; Helt et al., 2010; Hoogenhout et
al., 2013) and chimpanzees (Madsen et al.,
2013). No study so far has addressed this issue
in bonobos. Moreover, in Hominini the yawning
response can vary depending on the sex of the
responder or the trigger. For example, women may
respond more to others' yawns (Chan & Tseng,
2017; Norscia et al., 2016a, 2016b), although
this does not occur in all cohorts (Bartholomew
& Cirulli, 2014; Norscia & Palagi,
2011). Moreover, in the Pan genus yawning
response can vary in relation to the trigger's
sex, possibly depending on the social role that
each sex has in different species (Demuru &
Palagi, 2012; Massen & Gallup, 2017).
Finally, yawn contagion was found to be
influenced by the level of familiarity between
subjects in humans (Norscia & Palagi, 2011;
Norscia et al., 2020), chimpanzees (Campbell
& de Waal, 2011), and in one out of two
groups of bonobos (cf. Demuru & Palagi,
2012; Tan et al., 2017), with highest yawn con-
tagion rates being recorded between particularly
familiar subjects.
In sum, two out of the three independent
studies on the presence of yawn contagion in
bonobos detected the phenomenon (cf. Amici et
al., 2014; Demuru & Palagi, 2012; Tan et
al., 2017) and yawn contagion was higher between
closely bonded (compared to weakly bonded) group
mates (Demuru & Palagi, 2012) but not
between group mates when compared to non_group
mates (Tan et al., 2017). To better understand
the phenomenon, we investigated yawn contagion
in yet another group of bonobos. We formulated
the following predictions.
Prediction 1: Presence and distribution of
yawn contagion. Based on previous findings on
the presence of yawn contagion in two bonobo
groups (Demuru & Palagi, 2012; Tan et al.,
2017), we expected to find the phenomenon also
in our study group (Prediction 1a). Demuru and
Palagi (2012) found the maximum yawn contagion
rates in the first minute after the triggering
stimulus. Hence, we expected to find a similar
result in our study group (Prediction 1b).
Because yawn contagion was not found in all
bonobos (Amici et al., 2014), we expected to
find a high contagion variability across
subjects (Prediction 1c).
Prediction 2: Perceptual
factors&emdash;Possibly due to the high visual
acuity of anthropoid primates (Fleagle, 2013),
the spatial distance from trigger and responder
was found to have no effect on yawn contagion in
chimpanzees (Campbell & Cox, 2019) and
geladas (Palagi et al., 2009). Hence, we
expected to find no in- fluence of
trigger_responder distance on yawn contagion in
bonobos (Prediction 2a). Moreover, in humans and
chimpanzees observing several yawns in a row
does not seem to raise the chance of yawn
contagion (humans: Norscia & Palagi, 2011;
chimpanzees: Campbell & Cox, 2019). Hence,
we expected a similar result in bonobos owing to
their phylogenetic closeness with humans and
chimpanzees (Prediction 2b).
Prediction 3: Individual and social factors.
In the Hominini, the trigger's rank and sex can
have an influence on yawn contagion rates, with
individuals responding mostly to men in certain
cohorts of hu- mans (for yawns that are heard
but not seen; Norscia et al., 2020) and
chimpanzees (dominant males especially; Massen
& Gallup, 2017) and to females in bonobos
(Demuru & Palagi, 2012). While males are
central in chimpanzee dominance relationships
(Bray et al., 2021; Lewis et al., 2021), in
bonobos reproductive females are central in
determining group dynamics (e.g., Furuichi,
2011). Hence, we expected that trigger's rank
and sex, especially adult females, could play a
major role in eliciting the yawning response
(Prediction 3a). As concerns the effect of age,
no study on bonobos has addressed this factor on
yawn contagion so far. However, age appears to
have an effect in humans (Anderson & Meno,
2003; Bartholomew a& Cirulli, 2014; Helt et
al., 2010) and in chimpanzees (Madsen et al.,
2013), with yawn contagion being higher in
adults than in immature sub- jects. In certain
cohorts of adult humans, yawning decreases with
aging (Bartholomew & Cirulli, 2014). This
aspect has not been in- vestigated in
chimpanzees. Owing to the phylogenetic closeness
of bonobos to humans and chimpanzees (Pru_fer et
al., 2012), we expected that age might have a
similar effect on yawn contagion in our study
group (Prediction 3b). In humans and
chimpanzees, familiarity between individuals has
been reported to increase yawn contagion rates
(humans: Norscia & Palagi, 2011; Norscia et
al., 2016a; chimpanzees: Campbell & de Waal,
2011). In bonobos, no familiarity effect was
found between non_group members in an
experimental setting (using video trials; Tan et
al., 2017) but it was found within known
subjects in naturalistic conditions, with yawn
contagion being highest between closely bonded
group mates (Demuru & Palagi, 2012). Thus,
we expected to find a positive effect of
familiarity on yawn contagion in our bonobo
group, observed under naturalistic conditions
(Prediction 3c).
DISCUSSION
Presence of yawn contagion
Yawn contagion was present in our study
group because it was more likely that bonobos
yawned after seeing a yawn (PY condition)
compared to when they did not observe any
previous yawn (MC condition; Prediction 1a
supported; Figure 3). Hence, yawn contagion may
be present at the population level, as it has
been found so far in three different groups
(present study; Demuru & Palagi, 2012; Tan
et al., 2017). Yawn contagion is also present in
different cohorts of other Hominini
(chimpanzees: Anderson et al., 2004; Campbell
& Cox, 2019; Campbell & de Waal, 2011;
humans: e.g., Bartholomew & Cirulli, 2014;
Chan & Tseng, 2017; Cordoni et al., 2021;
Norscia & Palagi, 2011; Provine, 1989) and,
as a form of autonomic contagion, can increase
interindividual synchronization and coordination
(Casetta et al., 2021; de Waal & Preston,
2017; Prochazkova & Kret, 2017).
We found that the phenomenon was present
only in the first minute after the yawn stimulus
(Prediction 1b supported), when we detected a
significant difference between PY and MC
conditions (Figure 3b). On one hand, this result
is in line with previous reports showing a peak
of yawn contagion in the first minute, compared
to following minutes, in bonobos (Demuru &
Palagi, 2012) and the other Hominini
(chimpanzees: Campbell & Cox, 2019; humans:
Palagi et al., 2014). On the other hand, our
result introduces an element of novelty because
it shows that yawn contagion occurred only (not
just maximally) in the first minute (or up to
the second minute, if we consider the
nonsignificant trend as the basis for further
investigation).
In our naturalistic study we found
significant variability in the ICTs across
subjects (Prediction 2c supported) and one adult
male did not show contagion (yawning more in MC
than in PY condition). Even though previous
studies did not specifically focus on yawn
contagion interindividual variation (Amici et
al., 2014; Demuru & Palagi, 2012; Tan et
al., 2017), Amici et al. (2014) found that in
four bonobos yawning was not triggered by video
stimuli of yawning conspecifics. Interindividual
variability may explain at least in part why
contagion is not expressed in all subjects. In
healthy humans, 40%&endash;60% of sub- jects did
not show yawn contagion under laboratory
conditions (Platek et al., 2003; Provine, 1986,
1989) and susceptibility to others' yawns
appears to be stable across contexts
(Bartholomew & Cirulli, 2014). Analogously
in bonobos, yawn contagion can be context in-
dependent (e.g. resting/relaxing vs social
tension contexts; Demuru & Palagi, 2012).
Future studies on interindividual fluctuations
can shed light on within_population
variability.
Perceptual factors affecting yawn
contagion
The spatial distance between trigger and
responder had no significant effect on yawn
contagion (Table 1; Prediction 2a confirmed).
Con- sistently, no influence of
trigger_responder distance was found in
chimpanzees (Campbell & Cox, 2019) and
geladas (Palagi et al., 2009). This is not
surprising because anthropoid primates possess
high visual acuity and mainly rely on
stereoscopic vision to orient themselves in the
world (Fleagle, 2013).
In our study group, yawn contagion
probability increased as the number of
triggering yawns increased (Prediction 2b not
supported; Figure 4a; Table 1). This is in
contrast with the situation found in humans and
chimpanzees, in which no such effect was found
(Campbell & Cox, 2019; Norscia & Palagi,
2011). Interestingly, Norscia et al. (2021b)
found that in domestic pigs both trigger_
responder spatial distance and the number of
(non_vocalized) yawning stimuli affected yawn
contagion rates possibly due to the scarce
visual acuity of the species. It is possible
that bonobos, compared to humans, possess a
higher yawn contagion threshold and that the
yawning response is most likely primed after
observing multiple yawns. This possibility may
contribute to the interindividual variability
observed in bonobo yawn contagion and might
point to- wards possible neurobiological
differences in stimulus processing. Future
cross_species studies are necessary to clarify
this issue.
4.3 | Individual and social factors
modulating yawn contagion
Compared to males, females were not overall
more effective as triggers even though a
previous study fount that adult females ten- ded
to induce others' yawns more than males (Demuru
& Palagi, 2012). This difference may be due
to the fact that our female sample included
females with and without a swelling cycle, which
allowed us to test for this variable (not tested
before). We found that females with a swelling
cycle elicited more yawning responses from other
females compared to females without swelling
cycle (Table 1 and Figure 4d). In this respect,
Prediction 3a can be at least partially
confirmed. Swelling in bonobos is an important
communicative signal not just for males but also
for females (Demuru et al., 2020) and can
contribute to determining their social status by
favoring female_ female socio_sexual
interactions and alliances (Furuichi, 2011;
Moscovice et al., 2019). Analogously, in
chimpanzees _ in which males form alliances to
control resources (Bray et al., 2021; Lewis et
al., 2021)&emdash;males seem to be most powerful
in eliciting yawn contagion, especially if
dominant (Massen & Gallup, 2017). Rank per
se had no significant influence on yawn
contagion in bonobos possibly due to the high
tolerance level of the species (Furuichi, 2011;
Hare & Kwetuenda, 2010). Indeed, in our
bonobo group hierarchy showed relatively low
steepness, which indicates rather shallow
hierarchy. Interestingly, females showed the
highest yawn contagion rates (Table 1 and Figure
4b), which may related to their central role in
bonobo groups. Such a role may require an
enhanced sensitivity to social signals, such as
yawning, which may favor interindividual syn-
chronization and social cohesion. In humans, an
increased yawning response of women has been
observed in some cases (Chan & Tseng, 2017;
Norscia et al., 2016a) but not in others
(Bartholomew & Cirulli, 2014; Norscia &
Palagi, 2011). The socio_cultural influence
characterizing different human cohorts makes it
hard to single out an unambiguous effect of
gender on yawn contagion (Palagi et al.,
2020).
We detected no yawn contagion (as
responders) in the two infants (aged 4 months
and 4 years old) and our statistical analysis on
subadults and adults showed that yawn contagion
decreased with age (Table 1 and Figure 4c;
Prediction 3b confirmed). The responder's age
seems to affect yawn contagion also in other
Hominini. In chimpanzees yawn contagion was
found in adult subjects, but absent in infant
subjects (Madsen et al., 2013). In humans, yawn
contagion is absent, reduced or differently
age_modulated in infants (Anderson & Meno,
2003; Cordoni et al., 2021; Helt et al., 2010;
Millen & Anderson, 2011). In human and
non_human mammals, the increase of yawn
contagion with age (especially from the immature
phase to adulthood) has been associated with
possible maturation of socio_ cognitive
abilities and/or neural pathways that decode
social cues and with the ontogenetic variation
in the ability to identify the in- ternal states
of others (Cordoni et al., 2021; Madsen &
Persson, 2013; Norscia et al., 2021b).
In certain human cohorts, yawn contagion can
decline with age (over 40; Bartholomew &
Cirulli, 2014) possibly due to a decreased
sensitivity to others' states (Palagi et al.,
2020). Yawn contagion, possibly mediated by
bottom_up cognitive processes (Palagi et al.,
2020), might also decrease with age as the
result of the increased top_down mechanisms in
emotional processing. Interestingly, in humans
aging seems to be associated with a switch from
bottom_up to top_down processes in emotion
appraisal (Petro et al., 2021; Reed &
Carstensen, 2012). Further neuroethological
studies are necessary to verify these
hypotheses.
Finally, the affiliation levels between
group mates (a social attachment indicator;
Dunbar, 1991) did not affect the likelihood of
yawn contagion (Table 1; Prediction 3c not
confirmed). Social attachment (informed by
affiliation levels, kinship and/or group
membership) can increase yawn contagion rates
(Palagi et al., 2020). Such effect has been
observed in humans (Norscia & Palagi, 2011;
Norscia et al., 2016a), chimpanzees (Campbell
& de Waal, 2011) and other mammals (e.g.,
domestic pigs, Norscia et al., 2021b; wolves,
Romero et al., 2014). The presence of the
so_called 'familiarity bias' suggests that
emotional contagion may influence the phenomenon
of yawn contagion (de Waal & Preston, 2017).
In bonobos, the situation is puzzling because no
effect of group membership (group vs. non_group
members) was experimentally found in one group
(Tan et al., 2017) whereas a positive effect of
social bond between group mates was found in
another group via a naturalistic approach
(affiliation rates and kinship were combined;
Demuru & Palagi, 2012). At the very
proximate level, the familiarity bias on yawn
contagion may be dampened in our study colony by
the fact that individuals had been together in
the same group&emdash;with no fission_ fusion
management&emdash;for a long time
(min&endash;max range: 4&endash;12 years).
Affiliation rates occurring in the short term
may not reliably inform on long_term
familiarity. At the ultimate level, the
xenophilic nature of bonobos (showing
affiliation between group residents and non_
residents, high intergroup tolerance and food
sharing with strangers; Furuichi, 2011; Idani,
1991; Lucchesi et al., 2020; Tan & Hare,
2013; Tan et al., 2017) may have contributed to
reducing the adaptive value of familiarity. The
lack of familiarity bias was also found in an
opposite situation. Particularly, van Berlo et
al. (2020) found the presence of yawn contagion
in captive orangutans with no effect of
familiarity was detected. Wild orangutans do not
live in social groups but show dispersed
sociality (with occasional en- counters). Here,
the effect of familiarity may have a reduced
adaptive significance because individuals do not
form preferential social bonds or alliances. The
opposite cases of bonobos (Demuru & Palagi,
2012; Tan et al., 2017; present study) and
orangutans (van Berlo et al., 2020) converge in
indicating that the familiarity bias may be
related to interindividual cohesion (proximate
level) and type of sociality (ultimate level).
In contrast with previous reports
(Joly_Mascheroni et al., 2008; Romero et al.,
2013; Silva et al., 2012), a meta_analysis
showed that familiarity seems not to affect
interspecific yawn contagion between dogs and
humans (Neilands et al., 2020). A similar
approach could help disentangle the familiarity
issue in bonobos, especially if by including
data collected with the same methodologies on
different colonies. Once again, owing to the
differences observed across study groups and
sites, we stress the importance of expanding the
dataset on yawn contagion to account for
intergroup differences and clarify what factors
can modulate the phenomenon at the population
level.
