Yawning has a well documented contagious
effect: viewing or hearing a yawn&emdash;as well
as talking or thinking about yawns&emdash;causes
human subjects to yawn. While comparative
ethological and neurological accounts suggest
that yawning is a function of primitive
biological structures in the brain stem, these
analyses do not account for infectious yawning
caused by representational and semantic states.
Investigating the relationship between
perceptual and cognitive avenues of yawn
induction affords a unique opportunity to
examine how higher level cognitive faculties
interact with involuntary or automated
processing systems. In this paper, I examine
three distinct attempts to reconcile the
cognitive properties of contagious yawning with
its physiological basis&emdash;one neurological,
one philosophical, and one functional. None of
these accounts are unproblematic, and the most
plausible hypothesis for the evolution of
contagious yawning does not satisfactorily
explain the cognitive iterations of the
phenomenon.
I argue that the most likely explanation of
the contagion links perceptual elements of
witnessed yawns to conceptual representations.
More centrally, this kind of integrated account
has repercussions for general theories of human
thought and rationality, and suggests that
higher level representational states engage
neurophysiological structures in determining
human behavior.
Standard ethological accounts of innate
releasing mechanisms suggest a system of
stimulus feature detectors that are responsive
to a narrow range of behavioral signals. In
contrast to learned behaviors, which can be
tuned to a variety of features in the
environment, these mechanisms are rigidly
connected to specific perceptual cues. In
contagious yawning, these cues appear limited to
particular features of the auditory and visual
profile of witnessed yawns. However, the fixed
and inflexible nature of the stimuli feature
detectors that lead to contagious yawns stand in
stark contrast to the relative plasticity of
yawning contagion in thought. Discussing or even
thinking or reading about yawns also yields
bouts of contagious yawning. Moreover, these
episodes are not attributable to innate signal
releasing mechanisms in perception. There are
many ways one could come to think about yawning,
none of them rigidly tied to tell-tale
behavioral signals. As a consequence, it would
be difficult to account for the cognitive causes
of yawning in terms of innate signal detection
mechanisms. The signals are too variable across
time and different populations to function as a
trigger for this behavior.
The dual perceptual and cognitive aetiology
of infectious yawning speaks to the larger
difficulties in characterizing the neurological
and behavioral qualities of the act itself.
Yawning is common in the animal world,
suggesting a ''phylogenic antiquity'' (Provine,
1996) that comports well with neurological
studies that place yawning triggers in the lower
brain stem. However, the conceptual or
representational form of contagious yawning
indicates a social and cognitive role for
yawning that appears incompatible with a purely
noncognitive account of their causes. Even if
the proximate causes of yawning are ultimately
identified in the lower brain stem, the fact
that infectious yawning is responsive not merely
to perceptual cues but also to the content of
what we think and communicate suggests that the
causal profile of yawning is not limited to
lower level responses. This is an unexpected
detour for what appears at first to be a simple
reflex.
Consider, for example, how infectious
yawning compares with other involuntary or
reflexive behaviors. Inhaling pepper can cause a
sneeze, but the idea of pepper does not.
Chopping or slicing onions can cause tears, but
thinking about or describing chopped onions does
not. The difference lies with the nature of the
stimuli itself. Ideas of pepper and onions lack
the chemical properties (in these cases,
piperine and propanthial S-oxide respectively)
that underlie our characteristic physical
responses to both substances (see Sloop, 1995
for pepper and Imai et al., 2002 for onions).
But it does not come down to simple chemistry.
Other examples of infectious behavior do not
rely on chemical irritants. Being around an
anxious person can lead to heightened arousal or
anxiety amongst those who perceive that person's
discomfort. However, it is doubtful that reading
a clinical account of anxiety will similarly
increase the reader's blood pressure or pulse
rate. Studies of mood induction (see Martin,
1985) show that individuals can acquire moods
not only from observing behavioral (particularly
facial) expressions of those moods, but also
from considering the circumstances that
typically elicit them (sad stories can evoke
sadness, for example). However, the same is not
true for infectious yawning. One is not required
to imagine the circumstances of yawning to
experience this phenomenon.
A consideration of the concept itself, as
opposed to instances of boredom or fatigue, is
often sufficient to trigger episodes of
contagious yawning. The mere idea of anxiety (as
opposed to considering potential reasons for it)
appears decidedly less infectious than the
visual or perceptual awareness of its presence.
This shouldn't be surprising. The idea of
anxiety and its display are very different
things. However, this distinction dissolves in
the causes of infectious yawning. Whatever
property of the perception of yawns that is
responsible for its contagion effects, the same
or similar property appears to be present in
symbolic representations of the yawn itself.
What makes yawning unusual is not simply that it
is an involuntary physical response to the
content of a thought. Anxiety and other moods
can be generated this way, as well as laughter.
What is surprising in this case is the
suggestion that both the perceptual and
conceptual variations of this contagion result
in the same behavioral responses.
This parallel response between reflex and
representation is not the norm: thinking about a
punch doesn't trigger our blink reflex when
ducking an actual one does. Blinking to avoid a
punch (or minimize its impact) makes a kind of
sense with actual punches that doesn't carry
over to notional ones. This functional
dissimilarity is augmented by the fact that the
processes involved, in terms of perceptual and
cognitive processing, are physiologically and
functionally distinct. However, something
similar might be said about the two causes of
contagious yawning. A perceptual input that
causes an involuntary response can be likened to
a reflex, a simple unthinking (and hence lower
brain stem) response to immediate input.
Perceptual reflexes like shivering in response
to cold, or puckering in response to sourness
largely if not entirely bypass conscious
awareness or cognitive appraisal. Tickling may
cause laughter in a subject, but rarely does
thinking of tickling generate the same response.
Catching yawns from reading about them, on the
other hand, seems to involve a set of processes
distinct from those used in perceptual input and
reflex processing. Interpreting a symbol or
contemplating ideas engages higher level
processes located not in the brain stem, but the
cortex itself.
The apparent tension between these two
causes of contagious yawning suggests a
promising avenue for research into the relation
between the innate feature detectors that
subserve the perceptual component of infectious
yawning and the cognitive or intentional
features that trigger it in communication and
thought. Since analyses of perceptual reflexes
typically proceed independently of functional
analyses of mental representation and cognition,
they often fail to address links between innate
feature detectors and the determinants of
voluntary behavior. Recent studies (see, e.g.,
Gallese, Fadiga, Fogassi, & Rizzolatti,
1996) that link the perception and performance
of particular behaviors to brain activation
within what has been called the ''mirror''
neuron system have emphasized voluntary
responses without addressing the most
mirror-like of human responses, contagious
behaviors.1 Hence, contagious yawning offers a
valuable opportunity to examine the neurological
and perceptual models of involuntary and
contagious behaviors in the context of these
discoveries and against the background of
causal/functional models of mental content, with
the potential to shed light on both. As Robert
Provine writes, ''in studying contagious yawning
. . . we move seamlessly from the neural to the
social level of analysis'' (1996, p. 180). The
bridge is a psychological one.
Yawning ......
Achitectural approaches to yawning
......
Empathy
It is a curious feature of infectious
yawning that despite the robust relationship
between depictions and discussions of yawning
and increased yawning in subjects, many
individuals are relatively unresponsive to the
presence of any yawning stimulus. A theory of
yawning contagion should explain not merely why
yawns are contagious, but also why, for some
people, they are not. Recent studies by Steven
Platek and his colleagues (2005) suggest that
this variation can best be explained in terms of
differing levels of empathetic response amongst
distinct individuals: the more empathetic a
person is, the more likely she will respond to
the yawning stimulus. Platek found that the
better subjects were able to recognize and
attribute (as well as make inferences about)
mental states, the more likely they were to
experience infectious yawning. Conversely,
subjects who scored highly on a schizotypal
personality index and thus show decreased
mind-reading abilities were less likely to
exhibit contagious yawning than those who did
not. Platek hypothesizes that contagious yawning
depends on one's capacity to engage with and
imagine the mental states of another. This
empathetic identification leads to the
reproduction of another individual's yawns in
oneself
Platek's theory presupposes a strong
connection between empathetic response and a
theory of mind. A theory of mind is the
mechanism by which individuals read and
attribute mental states to others. On this view,
empathy requires the ability not merely to take
on the feelings of another, but also to be able
to predict and explain other people's behavior.
Empathetic identification would thus break down
in cases where an individual's theory of mind is
diminished or absent. Platek's model suggests
that in those cases we would likewise encounter
an absence of contagious yawning. Recent studies
with autistic children and in animal ethology
seem to confirm this point. Autism has long been
associated with decreased capacities to predict
and explain the behavior of others. Simon
Baron-Cohen (1997) has characterized this
condition as a sort of mind blindness. In
accordance with Platek's findings, we would
expect autism to effectively immunize affected
individuals from infectious yawning. Senju et
al. (2007) offer compelling evidence that this
is indeed the case. They found that children
with autistic spectrum disorder were
significantly less likely to yawn when viewing
video presentations of yawning adults than
developmentally normal children (even though
they are equally likely to yawn at a neutral
stimulus).
The theoretical connection between
contagious yawning and the theory of mind could
also be questioned if we could find an example
of contagious yawning in some species with a
doubtful claim to a theory of mind.6 However, as
Anderson, Myowa- Yamakoshhi, and Matsuzawa
(2004) have shown, there is very little evidence
for contagious yawning outside of the higher
primates. Whether in fact higher nonhuman
primates have a theory of mind remains
contentious, but there have been significant
experimental results to suggest that they do.7
Even so, it would be difficult to draw strong
conclusions from the spotty evidence we
currently possess. A recent study by Paukner and
Anderson (2006) suggests that yawning contagion
can also be found in stumptail macaques, which
would complicate any straightforward equation of
theory of mind with contagious yawning. A full
resolution of these issues will ultimately
depend on further empirical data on yawning in
the animal world as well as a more complete
account of theories of mind among monkeys and
other animals.
Platek has also uncovered a new source of
evidence for the empathetic nature of contagious
yawning in fMRI studies of the brain. This
research suggests that the neural substrates
involved in contagious yawning are the same as
those that have been connected to ''identifying
words that describe oneself, retrieval of
autobiographical events/memories, engaging in
self-generated actions and selfmonitoring, and
discriminating between theory of mind stories
and 'physical' stories'' (Platek et al., 2005,
p. 450). However, Platek did not find activation
in other areas associated with self-reference or
conscious self-recognition in the forebrain. He
suggests that this allows us to draw a line
between the conscious (and intermittent)
processing of self-awareness and empathy with
what he calls ''a primitive form of empathetic
modeling'' (2005, p. 450). Contagious yawning
finds its home in the latter.
The differences between the two are
instructive. In contagious yawning, the
bilateral thalamic and parahippocampal gyrus are
also activated (Platek, 2005, p. 451). These
substrates have been tied to visual and facial
sensory processing. The strong (and as Provine
has pointed out, profoundly specific) perceptual
sensitivity of contagious yawning suggests a
sensory link between visual encounters with
yawns and the production, in a primitive level
empathetic response system, of a yawning
reflex.8 Noticeably absent from this account,
however, is the involvement of higher level
processing mechanisms. In distinguishing between
two levels of empathetic response, one higher
level and more theoretical, and another
primitive, lower level system, Platek risks
offering a split account of the causes of
contagious yawning. The epidemiology of
perceptual encounters with yawns falls under the
lower level response system, while conceptual
precursors of the contagion are processed
outside the primitive response system. As a
consequence, Platek's model cannot explain how
higher level representations can engage the
response system implicated in perceptual
instances of yawning. This problem is
exacerbated by the fact that Platek's
experiments exclusively employ visual
presentations of yawns as triggering stimuli for
the contagion. Understanding the activation of
neural substrates during the semantic variant of
the contagion could help shed some light on how
or whether the two releasing mechanisms for
contagious yawning are related. The connection
Platek and his colleagues have discovered
between yawning and primitive empathetic
processing suggests a social role for contagious
yawning. Since contagious yawning appears to
require some rudimentary theory of mind,
neurological structures that subserve this
system are central to Platek's account.
Surprisingly, the structures most often
implicated in the formation and development of
empathy, mirror neurons (see Gallese &
Goldman, 1998), do not appear to play a role in
the empathetic response that governs episodes of
contagious yawning. Platek's (2005) fMRI studies
identify little or no activation in the cortical
regions associated with mirror neurons during
episodes of contagious yawning. What role mirror
neurons play in mind-reading and empathy remains
obscure, but these results suggest that
contagious yawning might not share their
primarily imitative function.
Perception
Platek's studies indicate a link between
empathy and the perceptual exemplars of yawning
contagion.....
Yawning and Adaptive Purposes
While contagious yawning appears connected
to one's capacity to read or empathize with
others, Platek's accounts offers little insight
into the adaptive purpose of yawns. That is,
contagious yawning might depend on empathy, but
this gives us little clue about why. This
problem is made more difficult by the fact that
yawns might be used to play different roles in
different species. Yawning has been associated,
particularly in primate species, with aggression
and dominance displays (Deputte, 1994; van
Lawick-Goodall, 1968). Male patas monkeys use
yawns to bare their canine teeth in facing down
perceived threats (see Trois, Aureli, Schino,
Rinaldi, & De Angeli, 1990 on male threat
displays)&emdash;a pointless act for humans
without similarly large canine teeth.12 This
role in aggression might help explain both the
predominance of male yawners in the animal world
(though there is no gender discrimination in
human yawning, see Schino & Aureli, 1989)
and its connection to sexuality. Researchers
have found that in rats chemicals that induce
yawns also induce erections (see Melis,
Stancamiano, Lai, & Argiolas, 1995), while
testosterone injections in female rhesus monkeys
elicits increased yawning. Interestingly, even
though anti-depressants typically suppress
sexual desire and receptivity in humans, some
have been known to cause orgasms in human
subjects during yawns (Provine, 2005). Yawning
has also been linked to increased stretching and
abrupt changes in levels of attention or arousal
(Provine & Hamernik, 1986; Provine,
Hamernik, & Curchack, 1987).
These links hardly point to a definitive
purpose for yawning, but together they are
suggestive of the most plausible adaptive
account of their function. Building on a
suggestion made by Arthur Myrberg (1972) and
Provine, Hamernik, and Curchack's (1987)
pioneering studies, Ronald Baenninger (1987)
argues that yawning plays a role in
transitioning between different states of
wakefulness or arousal. In particular, he
contends that the purpose of yawning is to help
individuals remain alert in poor or weak
stimulus conditions. The connection between
yawning and arousal is well-documented. Yawning
is most common in the hours immediately after
waking and in those immediately preceding sleep
(Provine et al., 1987). Studies have shown that
individuals yawn more frequently in the hour
before they go to bed than they do while lying
in bed waiting to fall asleep (Greco,
Baenninger, & Govern, 1993). If yawning were
simply an indicator of fatigue, one would expect
yawning to continue to increase as one
approaches sleep. In Greco's (1992) study, she
found that subjects who were discouraged from
yawning during a simulated driving task made
more mistakes than those that were not (even
though many subjects actively suppressed yawns).
This suggests that active yawning helps maintain
higher levels of awareness and function in
fatigued or stimulus deprived individuals
(Baenninger, 1997, p. 201; Provine, Hamernik,
& Curchack, 1987).
Similar results can be obtained in the
animal world. Zoo animals yawn more frequently
while waiting for feedings than they do
afterwards (Baenninger, 1987). Baenninger also
notes that carnivores, for whom the cycle of
daily activity and arousal is typically more
varied, tend to yawn more than herbivores (1997,
p. 200). In these cases yawning appears to
function primarily as a noncognitive mechanism
for increasing an individual's level of
attention or arousal in contexts where
actionpreparedness is important but difficult to
sustain.
While we might be able to make a plausible
case that arousal is the adaptive purpose of
yawning, it is difficult to establish this claim
in the absence of an account of how yawns serve
this function. That is, it is hard to say that
this is what yawning is for if we don't know the
mechanism used to sustain or increase levels of
awareness in individuals who yawn. However,
there are some theories of the physical
structure of yawning that seem congenial to this
view. If yawning really did increase oxygen
uptake, then it might as a consequence lead to
more alert individuals. Though this theory has
been largely discredited, other so far
unsubstantiated accounts fit the same mold.
Askenasy (1989) argues, for example, that
yawning may dilate arteries in the brain, thus
increasing cerebral blood flow. This process may
yield increasing levels of awareness. Neither
theory, however, explains the contagion effect
of yawns. If yawns are primarily a means of
regulating oxygen uptake or cerebral blood flow,
then why are yawns any more contagious than
liver function? Why should my need to increase
cerebral blood flow have any influence on
yours?
The answer might lie in the social nature of
animal communities. There are situations in
which maintaining group arousal can accrue a
significant evolutionary advantage to members of
the group. Consider, for example, the advantage
of maintaining high levels of alertness against
predators in feeding situations. The population
is more secure in circumstances in which each of
its members is in a position to issue an alarm
call. As one group member begins to succumb to
fatigue or boredom, the ensuing yawn might help
other members of the group maintain their
vigilance against attack. Similarly, a
sensitivity to yawning stimuli might help arouse
or awaken a group in the early morning or after
a rest. Yawning might be used, in these
contexts, to co-ordinate the timing and vigor of
activities within a community.
Whether in fact yawning plays a similar role
in human communities is unclear. Robert Provine
(2005) tells the story of the Bakairi people of
central Brazil who signal the end of their
evening conversations with a yawn that passes
through each member of the group. This is
perhaps a more explicitly cultural adaptation of
the yawning reflex or contagion that might serve
a similar purpose in other groups. Provine
evidently thinks so: ''the chain reaction of
contagious yawning synchronizes the behavioral
as well as the physiological state of our
tribe'' (2005, p. 532). The contagious nature of
yawning in human (and other animal) communities
might have emerged, then, as a prelinguistic
attempt to harness an arousal mechanism to
engage and manipulate group activities.
This model fits well with Platek's data. In
supposing that yawning plays a social role in
coordinating group behavior, the connection
between contagion and empathy suggests an avenue
by which this coordination could be achieved. By
hardwiring the response to perceived yawns into
our theory of mind, signals can be easily and
reliably transmitted throughout a population.
However, this account shares the limitations of
Platek's account. While actual yawns are
elicited in circumstances in which group arousal
falls below optimum levels, talking about
yawning is only intermittently associated with
fatigue or boredom. While actually yawning might
lead us to consider and think about yawns, such
a connection would be redundant as a social
signal designed to increase awareness.
Presumably the yawn that started us thinking
about yawning will have already served that
purpose. However, as any researcher of yawning
knows all too well, any consideration of yawning
can lead to the increased production of yawns
whether they are situationally appropriate or
not. Moreover, it is not clear that
circumstances of fatigue or boredom are causally
related to thoughts about yawning in the same
way that they are related to yawns
themselves&emdash;being tired doesn't make one
think about yawns in the same way that it
appears to induce actual yawns. As a
consequence, many of the regulatory benefits we
might associate with the production of yawns do
not apply to thoughts about yawns. In fact,
thinking about yawning might lead one to engage
arousal mechanisms when they are a hindrance to
sleep or other biologically necessary
functions.
The social arousal model of contagious
yawning, that the propagation of yawns through
an animal community can function to regulate and
synchronize patterns of activity in animal
populations, is currently the best candidate to
explain why yawns cause others to yawn. Such an
explanation is limited, however. It does not
extend to the semantic or cognitive forms of the
contagion. Since the semantic form of the
contagion does not generate situationally
appropriate regulative interactions, it seems
unlikely to play any significant social role
within a community. These considerations do not
suggest that the social theory is false, but
rather that the best avenue to explore semantic
contagion will come not from an understanding of
what yawning is for, but rather from a theory
that connects conceptual understanding to the
perceptual precursors of contagion. On this
view, the semantic element of yawning contagion
is best seen as a side effect or carryover from
the perceptual cases.
Accepting this link between perception and
higher level representational states does not,
however, resign us to the details of the
Prinz/Barsalou account. On this view, the
relation between the conceptual elements of
contagion and their root perceptual causes is
not necessarily constitutive. Instead, a weaker
relation between perceptual and semantic
components could be imagined. This suggestion is
supported by evidence that the perceptual form
of contagion is more robust than its cognitive
counterpart. This gives us reason to suppose
that cognitive contagion emerges from an echo
chamber effect that is ultimately triggered by
perceptual exemplars of the contagion's
cause.
Robert Provine has suggested a hybrid
releasing mechanism that tunes higher level
cognitive functions to more basic or lower level
cognitive/behavioral responses. He suggests that
contagious laughter conforms to this model:
''Humour and joking are modern cognitive and
linguistic triggers of laughter, recently
evolved cortical bells and whistles superimposed
on laughter's ancient neurological core'' (2000,
p. 190). The semantic form of contagion, Provine
suggests, may reflect a similar reprogramming of
innate feature detectors towards a more flexible
set of releasing mechanisms. Unlike the
radically empiricist account, the connection
between perceptual and semantic forms is not a
by-product of other cognitive processes, but
would have been modeled to perform a particular
task. Nevertheless, it remains unclear why or
how the new signaling mechanism became grafted
onto the older or more basic action pattern.
Further study into this area would benefit from
a consideration of the priming literature both
on yawning and other involuntary behaviors like
blinking or coughing and their influence on
contagious behaviors. Tracing out the path of
this contagion can thus begin to uncover the
representational structures at the heart of
cognition.
The links between perception and cognition
that lead to the production of both forms of
contagious yawning offer an opportunity to
better understand the mechanisms that underlie
all representational states. In pursuing an
account of the semantic version of yawning
contagion, we can gain a unique perspective on
the connection between lower levels of cognition
and their relation to higher level
representational states. In so doing, we can
begin to develop an account of cognition that
takes into account not merely higher level
cognitive processes, but also their relations to
basic level neurobehavioral responses.
As a consequence, we have the opportunity to
develop an account of mental representations
that is neither top heavy nor structurally
uninformed. I hope these considerations show not
merely that yawning contagion requires further
empirical research, but that the study of
borderline cognitive states can illuminate more
general theories of human thought and
rationality.
