Three hypotheses have attempted to explain
the phenomenon of contagious yawning. It has
been hypothesized that it is a fixed action
pattern for which the releasing stimulus is the
observation of another yawn, that it is the
result of non-conscious mimicry emerging through
close links between perception and action or
that it is the result of empathy, involving the
ability to engage in mental state attribution.
This set of experiments sought to distinguish
between these hypotheses by examining contagious
yawning in a species that is unlikely to show
nonconscious mimicry and empathy but does
respond to social stimuli: the red-footed
tortoise Geochelone carbonaria.
A demonstrator tortoise was conditioned to
yawn when presented with a red square-shaped
stimulus. Observer tortoises were exposed to
three conditions: observation of conditioned
yawn, non demonstration control, and stimulus
only control. We measured the number of yawns
for each observer animal in each condition.
There was no difference between conditions.
Experiment 2 therefore increased the number of
conditioned yawns presented. Again, there was no
significant difference between conditions. It
seemed plausible that the tortoises did not view
the conditioned yawn as a real yawn and
therefore a fmal experiment was run using video
recorded stimuli. The observer tortoises were
presented with three conditions: real yawn,
conditioned yawns and empty background. Again
there was no significant difference between
conditions. We therefore conclude that the
red-footed tortoise does not yawn in response to
observing a conspecific yawn. This suggests that
contagious yawning is not the result of a fixed
action pattern but may involve more complex
social processes
Contagious yawning is well documented in
humans, however, little is known about its
function and prevalence in the animal kingdom or
the brain mechanisms underlying it. The function
of yawning itself is also poorly understood.
Yawning has been observed in a number of
vertebrate taxa and, though it is likely that
such a prominent and widespread behaviour serves
a biological function, the nature of this
function remains unclear (Guggisberg et al.,
2007). It has been suggested that yawning may
cause an increase in arousal which will reduce
the probability of sleep. This is something
which is likely to be important for vigilance in
all animal species (Walusinski and Deputte,
2004, cited by Guggisberg et al., 2007). Another
hypothesis suggests that yawning is a form of
communication used to synchronize group behavior
(Daquin et al., 2001) this could be for a
variety of reasons, those postulated have
included communicating drowsiness, social stress
or even boredom (Guggisberg et al., 2007). This
would potentially serve an important social
function and it is possible that contagious
yawning may have evolved as a result of
this.
Experimental analysis of contagious yawning
in humans has revealed that it occurs in 40%-60%
of participants when they see (videos of) a
yawning person (Platek et al., 2005). However,
the mechanisms underlying contagious yawning
remain poorly understood (Nahab et al., 2009). A
number of hypotheses have been proposed to
account for the occurrence of contagious yawning
and current evidence to support or refute them
is equivocal at best. It has been suggested that
contagious yawning may simply be the result of a
fixed action pattern for which the releaser
stimulus is the observation of another yawn
(Provine, 1986; Yoon and Tennie, 2010). Evidence
to support this hypothesis comes from the fact
that yawns follow a highly stereotyped pattern
(Provine, 1986). Further, yawning in humans can
be triggered by observing a yawn, hearing a yawn
(Arnott et al., 2009) or even thinking about
yawning (Provine, 1986). This hypothesis
predicts that contagious yawning may be observed
in all vertebrates that exhibit yawning
behavior.
A second hypothesis suggests that
nonconscious social mimicry, the tendency to
adopt postures, gestures and mannerisms of an
interaction partner (also known as the chameleon
effect; Chartrand and Bargh, 1999) may control
contagious yawning behavior (Yoon and Tennie,
2010). Non-conscious mimicry is assumed to
reflect close links between perception and
action. Many studies in macaque monkeys and
humans have shown that when an individual
observes another perform a particular action,
corresponding action representations in the
observer's action repertoire are activated
(Rizzolatti and Sinigaglia, 2010). Nonconscious
mimicry may occur when inhibitory processes that
normally keep us from executing observed actions
(Brass et al., 2005) are overridden. It has been
shown to be modulated by specific social
motivations e.g. the desire to affiliate with
the social partner. This is well documented in
humans and there is some evidence of this
phenomenon in primates (Rizzolatti and
Sinigaglia, 2010) but no research has directly
examined this in terms of contagious yawning.
This hypothesis would predict the presence of
contagious yawning in species in which
perception and action rely on common neural
representations and social relations are of
import.
The majority of recent research has
attempted to explain the phenomenon of
contagious yawning in terms of mental state
attribution and, in particular, empathy (e.g.
Platek et al., 2005). There are many different
defmitions of empathy (Vignemont and Singer,
2006), but in this context empathy is considered
to be the understanding of another's feelings
based on the capacity to infer others' mental
states (Baron-Cohen et al., 2005). According to
this view contagious yawning should only be
observed in those species that possess mental
state attribution, and thus we would expect to
see little evidence of contagious yawning
outside the higher primates.
Contagious yawning has been observed in
non-human primates (chimpanzees Pan troglodytes,
Anderson et al., 2004; stump-tailed macaques
Macaca arcto ides, Paukner and Anderson, 2006;
gelada baboons Theropithecus gelada, Palagi et
al., 2009) and dogs (Canis familiaris,
Joly-Mascheroni et al., 2008, but see Han et
al., 2009). These studies interpret their data
in terms of empathy. However the data presented
in these papers do not allow the other
hypotheses to be dismissed (see Yoon and Tennie,
2010 for further details) as the studies have
focused on those species that may possess the
ability to engage in some aspects of mental
state attribution or fit the criteria for
nononscious mimicry.
This study aimed to discriminate between the
possible mechanisms controlling contagious
yawning by asking whether contagious yawning is
present in a species that is unlikely to show
empathy or nonconscious mimicry: the redfooted
tortoise Geochelone carbonaria. To our knowledge
there is no evidence of social mimicry, mental
state attribution or empathy in this species.
Though naturally solitary there is evidence that
this species possesses a sensitivity to visual
social cues (Auffenberg, 1965), that it can
follow the gaze direction of a conspecific
(Wilkinson et al., 2010a) and can learn to
access an otherwise inaccessible goal by
observing the behavior of a conspecific
(Wilkinson et al., 2010b). Further, research
suggests that this species is highly visual and
when available will use visual cues over both
olfactory cues (Wilkinson et al., 2007) and over
a highly successful response based behavior
(Wilkinson et al., 2009). Taken together this
makes them ideal subjects for examining this
question. If contagious yawning is simply the
result of a fixed action pattern for which the
releaser stimulus is the observation of another
yawn, then we would expect to observe it in this
species; however, if it is controlled through
social processes such as nonconscious mimicry or
empathy then we would expect it to be
absent.
3 Discussion
The results of the three experiments
presented in this paper suggest that the
red-footed tortoise does not yawn in response to
observing a conspecific yawn. Experiment 1
examined whether tortoises would yawn more when
observing a conspecific perform a conditioned
yawn than in other control conditions. They did
not. The results revealed that there was no
overall difference in responding across
conditions suggesting that tortoises do not
possess the ability to yawn contagiously. This
suggests that contagious yawning may not be the
result of a fixed action pattern for which the
releaser stimulus is a yawn (Provine, 1986; Yoon
and Tennie, 2010) but rather supports the idea
that higher level mechanisms such as
nonconscious mimicry (Yoon and Tennie 2010) or
empathy (Anderson et al., 2004; Paukner and
Anderson, 2006; Palagi et al., 2009;
Joly-Mascheroni et al., 2008) may control this
behavior. However, examination of the individual
data revealed an overall low level of
responding. Interestingly, of the three animals
that did respond two responded more in the yawn
condition than in the control conditions. This
suggests that the tortoises may have the ability
to yawn when they observe a conspecific yawning
but it is possible that a single conditioned
yawn per trial was not enough to evoke
convincing evidence of contagious yawning in
this species. The majority of research in this
area has used multiple yawns (up to 19,
Jolie-Mascheroni et al., 2008) as stimuli. It is
therefore plausible that, under experimental
conditions, multiple yawns are necessary for
contagious yawning to occur.
Experiment 2 thus examined whether
contagious yawning would be observed if the
demonstrator performed multiple conditioned
yawns. The results revealed that this was not
the case. The tortoises were equally as likely
to respond in the control conditions as they
were in the experimental conditions. The
combined results of Experiment 1 and 2 suggest
that the tortoises do not yawn after observing a
conspecific yawn. Again, the data contradict the
hypothesis that contagious yawning is the result
of a fixed action pattern (Provine, 1986; Yoon
and Tennie, 2010) and suggests that higher
processes such as empathy may be involved
(Anderson et al., 2004; Anderson and Matsuzawa,
2006; Paukner and Anderson, 2006; Palagi et al.,
2009; Joly-Mascheroni et al., 2008) or
nonconscious mimicry (Yoon and Tennie
2010).
Experiment 3 examined the possibility that
the lack of differential responding observed in
the first two experiments was because the
conditioned yawn did not appear as a yawn to the
tortoises. The results revealed that the animals
appeared to respond more in both the yawn and
the conditioned yawn conditions than when they
were presented with the background control.
However, this apparent difference was not close
to reaching statistical significance. Thus, the
findings of Experiment 3 support those of
Experiments 1 and 2 which together reveal that
the red-footed tortoise does not yawn in
response to observing a conspecific yawn. It is
possible that the tortoises did not perceive the
video stimuli as a real tortoise and that the
experimental stimulus (the conditioned yawn)
with which the tortoises were presented may have
lacked some elements which, though not apparent
to humans, were essential for contagious yawning
to take place. For example, the demonstrator was
trained to express a simulated yawn by opening
its mouth wide and turning its head up. However,
a yawn also involves the movement of air and
this is something which was not simulated in
either our conditioned yawn experiments or in
the video playback. It is possible that a real
yawn is necessary to stimulate the observer
tortoise. Yet, video stimuli have successfully
stimulated yawns in a variety of species
(Anderson et al., 2004; Paukner and Anderson,
2006) and video stimuli have produced
appropriate responses to social stimuli in the
red-footed tortoise (Wilkinson et al.,
unpublished data). However, the use of video
stimuli to elicit behavior in animals is
controversial because it is not clear what the
animals perceive on the screen. This may account
for the differences seen between Joly-Macheroni
et al.'s (2008) study in which dogs observed a
real-life human demonstrator yawning and that of
Harr et al. (2009) in which the yawns were
presented as video stimuli. Little work has
directly investigated pictureobject recognition
in reptiles. However, there is evidence that
reptiles, including the red-footed tortoise,
respond to video stimuli of conspecifics as if
they were the real animals (e.g. Ord and Evans,
2002; Ord et al., 2002; Van Dyk and Evans, 2008;
Wilkinson et al., unpublished data).
Overall, our findings are more consistent
with the suggestion that tortoises do not yawn
in a contagious manner and that thus suggest
that contagious yawning is not simply the result
of a fixed action pattern and releaser stimulus,
as if this mechanism controlled the behavior it
would be predicted that contagious yawning would
be present in all vertebrates that yawn. We
suggest that contagious yawning may be
controlled through social processes such as
nonconscious mimicry or empathy, neither of
which would have predicted the presence of
contagious yawning in the red-footed tortoise.
This finding indirectly suggests that, rather
than increasing arousal, yawning may be a form
of communication that evolved to synchronize
group behavior (Daquin et al., 2001). However,
the type of information that it might
communicate or behavior that it might promote
remains unclear. Numerous researchers have
suggested that contagious yawning may be an
indicator of empathy; however, results in
experiments with humans have been equivocal
(Platek et al., 2003; Schurmann et al., 2005).
The fmdings of this study suggest that
contagious yawning may be controlled by higher
level social processes as it is believed that
tortoises do not possess nonconscious mimicry or
empathy. However, the current data do not allow
us to determine whether contagious yawning is a
result of nonconscious mimicry or empathy. The
nonconscious mimicry hypothesis predicts the
presence of contagious yawning in species in
which perception and action rely on common
neural representations we therefore might expect
to observe it in animals living in complex
social groups. The empathy hypothesis predicts
that we would expect to see little evidence of
contagious yawning outside the higher primates
and (possibly) domesticated dogs, species
believed to be capable of empathy
(Joly-Mascheroni et al., 2008). Further research
is needed to determine which of these social
processes may be involved in controlling
yawning.
