Ingroup-Outgroup
Bias in Contagious Yawning by Chimpanzees
Supports Link to Empathy
Matthew W. Campbell, Frans B. M. de
Waal
Living Links Center, Yerkes
National Primate Research Center, Emory
University, Atlanta, Georgia, USA
-Campbell
M et al. Computer animations stimulate
contagious yawning in
chimpanzees Proceed Royal Soc Biol
2009:276(1676):4255-4259
-Campbell
MW, de Waal F. Ingroup-Outgroup Bias in
Contagious Yawning by Chimpanzees Supports Link
to Empathy. Plos One. 2011;6(4):1-4
-Campbell
M, de Waal F. Methodological Problems in the
Study of Contagious Yawning; In Walusinski O
(ed): The Mystery of Yawning in Physiology and
Disease. Front Neurol Neurosci. Basel, Karger,
2010, vol 28, pp 120-127
-Campbell M,
de Waal F. Chimpanzees empathize with group
mates and humans, but not with baboons or
unfamiliar chimpanzees. Proc. R. Soc. B 281:
20140013
Tara, a chimpanzee, yawns
while watching a video of chimpanzees from the
group yawning on an iPod. The chimps in the
study with Tara yawned 50 percent more
frequently in response to video of members of
their group yawning versus video of members of
another group yawning.
Introduction
Humans (Homo sapiens) favor other humans
seen as belonging to their own group (ingroup)
over humans seen as belonging to different
social groups (outgroup), even in absence of
explicitly stated bias [1,2]. Recently,
these biases have been extended to differential
brain activity during empathy for pain. Specific
brain areas, most notably the anterior cingulate
cortex (ACC) and anterior insula, activated
during functional magnetic resonance imaging
(fMRI) both when subjects experienced pain and
when another person present experienced pain,
whereas other areas activated only during the
direct sensation of pain [3]. Singer et
al. [3] interpreted these findings as
humans sharing the affective or emotional aspect
of pain with others, but not the physical
sensation of pain. Extending these findings to
bias, two studies presented visuals of painful
experiences to human ingroup and outgroup
members (as defined by race) while using fMRI to
examine brain activity [4,5]. Xu et al.
[4] found greater activity in the ACC in
response to ingroup empathy for pain than
outgroup, and Mathur et al. [5] found
differences in the medial pre-frontal cortex,
indicating a role of cognitive appraisal. During
transcranial magnetic stimulation (TMS) human
subjects watched videos of needles penetrating
the hand of ingroup or outgroup members, also
defined by race [6]. The subjects showed
a greater empathic response (in the form of
resonant corticospinal inhibition) to the
ingroup than outgroup stimuli. Most interesting,
subjects also saw needles penetrating a hand
that had been artificially colored violet,
removing race cues. The subjects responded with
a greater empathic response toward the violet
hand than the outgroup hand, yielding a pattern
of ingroup . violet . outgroup.
All three studies showed that humans have
differential empathic responses to pain based
upon group status, indicating ingroupoutgroup
bias. We wanted to explore whether
ingroup-outgroup bias is present in contagious
yawning. Lehmann [7] and Preston and de
Waal [8] both hypothesized that empathy
is the mechanism underlying contagious yawning.
The idea is that yawns are contagious for the
same reason that smiles, frowns, and other
facial expressions are contagious. The mechanism
that allows someone to reflexively mimic a smile
[9] is thought to also allow for
reflexive mimicry of yawns. In this article, we
use the definition of empathy supplied by
Preston & de Waal [8], in which
empathy is a term for a broad category of
resonant emotional responses comprising a
continuum from basic forms, such as emotional
contagion, to complex forms, such as cognitive
empathy.
The link between empathy and contagious
yawning has empirical support. Humans who
performed better at selfrecognition and
theory-of-mind, two abilities that contribute to
complex empathy, performed more contagious
yawning [10]. In gelada baboons
(Theropithecus gelada), the closer the social
bond between individuals, the more likely they
would yawn when the other yawned [11].
This finding is consistent with the observation
that empathy is more pronounced the closer the
relationship between individuals [8,12].
Also informative are the negative relationships.
Two conditions, schizotypy [10] and the
autism quotient [13,14], are associated
with decreased contagious yawning, possibly to
the point of being absent in autism.
Both of these conditions are associated with
atypical empathy functioning. Contagious yawning
has been documented in five mammalian species:
humans [10,13,14,15,16,17], chimpanzees
[18,19], stumptail macaques (Macaca
arctoides) [20], gelada baboons [11],
and domestic dogs (Canis familiaris)
[21,22], although some of the
interpretations differ. Because of its relevance
to human mental health, evolutionary biology,
and as a potential low-cost complement to other
measures, contagious yawning is a useful and
perhaps under-utilized tool for studying empathy
functioning.
Our hypothesis was that if empathy is the
mechanism underlying contagious yawning, then
contagious yawning should show the same biases
as other measures of empathy, specifically the
ingroup-outgroup bias. We tested two groups of
captive chimpanzees by showing them yawn and
control videos of their own group and the
strange group. Chimpanzees form communities that
are territorial and exclude neighboring
individuals and communities [23].
Thus, for chimpanzees, strangers are
outgroup by default. Evidence for an
ingroup-outgroup bias would be if chimpanzees
yawned more in response to watching familiar
individuals yawning than strangers. Studying
chimpanzees also allows us to test whether human
ingroup-outgroup empathy bias is rooted in
evolved mechanisms assessing social closeness,
familiarity, and group status.
Discussion
The chimpanzees yawned more in response to
the familiar yawn video than the familiar
control, demonstrating contagious yawning.
However, the video of unfamiliar chimpanzees had
no detectable effect, as the difference in
yawning between the yawn and control videos was
nonsignificant. Critically, the chimpanzees
yawned more in response to the familiar yawns
than the unfamiliar yawns, demonstrating
ingroup-outgroup bias. This bias supports the
hypothesis that empathy is the mechanism
underlying contagious yawning. The link between
empathy and contagious yawning is further
supported by our data on attention. The
chimpanzees actually watched the videos of
unfamiliar individuals more than the videos of
familiar individuals.
They attended more to the unfamiliar yawns,
but yawned more to the familiar yawns. This
finding rules out attention per se as a
mediating factor and supports the idea that
social identification with the stimuli
influenced the rate of contagion. Even though
all of the ingroup videos were presented before
the outgroup videos, we can think of no a priori
reasons for an order effect. The attention data
show that the chimpanzees did not lose interest
in the videos since they watched the outgroup
videos more than the ingroup videos. There is no
evidence nor are there suggestions in the
literature that contagious yawning is transient
and fluctuates over time. These same subjects
were previously tested and showed contagious
yawning [19], so contagion seems to be
an enduring behavior. The rate of yawning toward
all of the control videos has remained the same
over three years (2007- 2010) and three
different stimuli, suggesting no change in
baseline rates of yawning. The more pertinent
order effect would be between the yawn and
control videos within a stimulus type (i.e.,
ingroup or outgroup), but these were always
counter-balanced.
In contrast to chimpanzees, humans
[10,13,14,15,16,17] and dogs
[21,22] have shown contagion in response
to watching unfamiliar individuals yawn. Some
different variables may explain this. First, we
cannot rule out that our sample size, large by
chimpanzee standards, was too small to detect a
significant difference. Chimpanzees may indeed
yawn contagiously in response to unfamiliar
individuals, but if so the magnitude of the
effect is probably small and would require more
subjects to detect statistically. A similar
situation occurred in the first study of yawn
contagion in chimpanzees [18], which had
too small of a sample size to detect contagious
yawning at the population level (the significant
effects were at the individual level). In
addition, Anderson et al. [18] did test
for ingroup-outgroup bias, but since they could
not detect a population-level effect for
contagious yawning overall, they did not detect
a difference between these stimuli. Larger
samples of chimpanzees have shown
populationlevel contagious yawning and an
ingroup-outgroup effect ([19] and the
present study). It may take an even larger
sample than the one we had available to detect
yawn contagion in response to unfamiliar
chimpanzees.
We should also be mindful of social
structure, as we may have two different factors
at work: familiarity and group membership.
Chimpanzees are territorial and aggressive
toward neighboring communities [23].
Since all members of a community know each
other, for chimpanzees, unfamiliar individuals
are by definition outgroup individuals. Humans,
at some point in our evolution, gained the
ability to include unfamiliar individuals in our
ingroup. Therefore, humans do not necessarily
view strangers as belonging to an outgroup. Pet
dogs are accustomed to interacting with
unfamiliar humans, and sometimes unfamiliar
dogs, in positive ways.
Possibly, we artificially selected dogs to,
like us, have disassociated familiarity and
group status, but this needs testing. Exposed to
artificial stimuli that transcend the
ingroupoutgroup distinction, chimpanzee yawn
contagion shows patterns similar to those of
brain imaging studies of empathy. Chimpanzees
yawned in response to 3D computer-animated
chimpanzees yawning [19]. These
animations were not familiar individuals, yet
they stimulated contagious yawning. Chimpanzees
seem to process animations the same way they
process pictures of chimpanzees [25],
but the inherent artificiality of the animations
may have prevented them from being processed as
outgroup individuals. This finding is similar to
the greater empathy of humans to pain inflicted
on a hand artificially colored purple than a
hand of an other-race individual present in
society [6].
Thus, animations and artificial stimuli may
allow us to distinguish between and test the
variables of familiarity and group status, in
humans and nonhumans alike. Contagious yawning
in humans has not yet been tested for biases,
including social closeness [11] and
ingroup-outgroup bias, but we would expect
similar responses. Contagious yawning has
several advantages as a measure of empathy given
its low cost, high portability, and
applicability to multiple species, which may
make it a useful complement to physiological,
questionnaire, and mental health diagnostic
based measures of empathy. Given that
chimpanzees exhibit both altruism [26]
and extreme violence [23] toward others,
studying how and when empathy is engaged may
tell us about how humans switch between these
two extremes as well.
