Content uploaded by Robin Dunbar
Author content
All content in this area was uploaded by Robin Dunbar on Dec 23, 2014
Content may be subject to copyright.
LAUGHTER AS SOCIAL LUBRICANT
Laughter as Social Lubricant:
A Biosocial Hypothesis about the Pro-social Functions
of Laughter and Humor
Mark Van Vugt, Charlie Hardy, Julie Stow, & Robin Dunbar
Please do not cite or circulate without permission from the authors
1
LAUGHTER AS SOCIAL LUBRICANT
Abstract
We test the hypothesis that laughter acts as a social lubricant to enhance a sense of
group membership through the release of endorphins. Using a public good game, we
show that laughter increases social bonding and cooperation between strangers and
that it increases the subjective experience of positive affect, while reducing negative
affect. We also show that laughter stimulates the release of endorphins (as indexed by
levels of pain tolerance) and that endorphin release is approximately proportional to
the amount of laughter performed. We propose a biosocial hypothesis of laughter
which addresses both the ultimate and proximate functions of laughter for connecting
individuals within groups.
2
LAUGHTER AS SOCIAL LUBRICANT
Laughter as Social Lubricant:
A Biosocial Hypothesis about the Pro-social Functions
of Laughter and Humor
“In laughter we emit sounds and express emotions that come from deep within our
biological being—grunts and cackles from our animal unconscious. But what do these
vocalizations signify? More than 2,000 years of the contemplation of laughter by
some of history’s great philosophers, writers, scientists, and physicians certify the
importance of the question but provide only the vaguest of answers.”
--- R. Provine (2000; p. 2)
Humans are a uniquely social species. They are able to empathize and form deep
emotional connections with non-kin (Baumeister & Leary, 1995). They come to the
rescue of complete strangers in need (Becker & Eagly, 2004). And, they are capable
of sacrificing themselves for the welfare of others in sometimes very large groups
(Van Vugt & Van Lange, 2006). Humans are also unique in their capacity to laugh,
talk, sing and dance (Dunbar, 2004a). These social capacities are not unrelated. The
scale and intensity of human cooperation requires mechanisms to facilitate
communication and affective bonding between people.
Here we propose one such bonding mechanism, human laughter. We develop a
biosocial theory of laughter in which we argue that laughter promotes psychological
well-being, which in turn enables individuals to cooperate and function better in
groups. The effects of laughter are complex and diverse. At the basic physiological
level, we suggest that laughter is involved in the release of endorphins, which create
the experience of positive affect. At the individual psychological level, we predict that
laughter elicits positive mood states while decreasing their negative mood. At the
3
LAUGHTER AS SOCIAL LUBRICANT
social psychological level, we predict that laughter promotes group identification as a
direct consequence of this shift in affect. At the behavioral level, we predict that
laughter improves group functioning by increasing altruistic group contributions.
Our biosocial theory of laughter is rooted in both the (evolutionary) biological
and (social) psychological literatures. We will first review the empirical literature on
laughter to summarize a number of key empirical points. Based on the literature
review, we discuss the various proximate functions of laughter and embed them
within an evolutionary hypothesis about the origins of human laughter. Our biosocial
hypothesis leads to a number of specific predictions concerning the effects of
laughter, some of which we test here. Our primary hypothesis is that laughter induces
positive affect and this enables individuals to feel and function better in groups.
A Review of the Laughter Literature
Universality of Human Laughter
Humans have long been characterized as the “laughing animal” (McComas, 1923)
and for some very good reasons. The scientific study of laughter and humor,
gelotology, has produced a number of important insights. Laughter is a universal
human behavior (Dunbar, 2004a; Eibl-Eibelsfeldt, 1999; Provine, 2000). It is
universally recognizable and occurs in a characteristic, stereotyped form (Eibl-
Eibelsfeldt, 1999). Ontogenetically, laughter emerges spontaneously in children as
young as 17 days old and it is one of the first social vocalizations (Deacon, 1997;
Kawakami et al., 2006). Laughter has also been observed in children who are blind
and deaf at birth and who therefore cannot possibly have learnt it from others
(Provine, 2000). This suggests that humans are strongly genetically disposed to
produce laughter.
4
LAUGHTER AS SOCIAL LUBRICANT
The evolved basis of laughter is further strengthened by phylogenetic and cross-
cultural comparisons. Some nonhuman primates, notably the chimpanzee (Van Hooff,
1972; Waller & Dunbar 2005), display facial expressions and emit vocalizations
during social play that resemble (and are homologous with) human laughter. Some
theorists have argued that laughter supplements grooming as a bonding device in
primates (Dunbar, 2004a). This suggests that human laughter may have an ancient
evolutionary basis and that a primordial form of laughter may have already been
present in our common ancestor, from which humans and chimpanzees departed some
six million years ago. Cross-culturally, there are important similarities in the stimuli
that elicit laughter. Humor is perhaps the most familiar stimulus, but others include
social play and tickling. Interestingly, Darwin (1872) considered laughter to be the
“tickling of the mind.”
In the literature, there is a general consensus that laughter occurs mostly in
response to unexpected non-serious social events – the stuff that, for example, jokes
are made of (Gervais & Wilson, 2005). We ought to draw a distinction here between
Duchenne and non-Duchenne forms of laughter (Keltner & Bonanno, 1997). The first
is the spontaneous, emotional, stimulus-driven laughter which we are interested in
explaining. The second is primarily self-generated and emotionless, such as when
people laugh out of appeasement, embarrassment or in response to derision. Brain
imaging data show that these two forms of laughter are supported by different neural
systems (Iwase et al., 2002). Duchenne laughter emerges in the subcortical, limbic
system and is connected to a laughter-coordination centre in the dorsal pons (Wild,
Rodden, Grodd, & Ruch, 2003). In contrast, the non-Duchenne laughter originates in
the premotor region and directly influences the primary motor cortex (Wild et al.,
2003).
5
LAUGHTER AS SOCIAL LUBRICANT
Although Duchenne laughter is presumably hardwired, it is not insensitive to
cultural factors. Cultural norms can dictate what people laugh about and in what
situations laughter is appropriate or inappropriate. For example, in some cultures there
is a greater emphasis on sexual and toilet humor in eliciting laughs, whereas in others
the focus is more on, say, political humor (Provine, 2000). Furthermore, the intensity
and context of laughter are also variable in different cultures (Provine, 2000). This
indicates that laughter can be reinforced or inhibited in accordance with local social
norms and customs (Apte, 1985).
The Functions of Laughter
Laughter theorists have proposed a diversity of proximate, psychological
functions of laughter. First, laughter has been linked to positive emotional
experiences. Brain imaging studies show that laughter involves affective reward
networks in the brain (Iwase et al., 2002), while several clinical studies have revealed
that significant health benefits can be derived from laughter, for example, in immune
system protection (Rosner, 2002). Also, social-psychological studies have found that
humor and laughter are associated with stress relief and a reduction in negative affect
while experiencing the stress of, for example, bereavement (Keltner & Bonanno,
1997). A recent study shows that as little as one minute of laughter increases people’s
self-reported mood (Neuhoff & Schaeffer, 2002). The mood effects of laughter have
been attributed to the release of “feel-good” neuropeptides like endorphins but this
still awaits investigation (Provine, 2000).
Another proposed function of laughter is communicative. Laughter has been
conceptualized as a signal designed to communicate the subjective emotional state of
a sender to receivers. The meaning of laughter is considered something like “that was
social play” to indicate that the situation is not serious (Gervais & Wilson, 2005). A
6
LAUGHTER AS SOCIAL LUBRICANT
related signaling function of laughter is to induce positive affect in others (Owren &
Bachoroskwi, 2003). There is support for this theory in a study showing that
particular forms of laughter (like voiced laughter) elicits positive affect in others
(Bachorowski & Owren, 2001). Furthermore, an fMRI study shows that hearing
laughter activates the amygdala, a key emotion area in the brain (Sander & Scheich,
2001).
Third, laughter has been proposed to have an emotional contagion function
(Provine, 2000). Research strongly suggests that laughter activates laughter, which is
why many TV-comedies use canned laughter (Provine, 2000). Thus, there might be a
mimicking effect of laughter which results in a convergence of the emotional states
and behaviors of people, the chameleon effect (Chartrand & Bargh, 1999). This shared
affect (as a result of laughter) might be a good way of coupling the emotional states of
individuals within groups enabling them to better coordinate their actions and
experiences.
A final proposed function of laughter is affiliative. Laughter has been proposed to
foster the integration of new individuals into existing groups (Gamble, 2001). Some
research suggests that laughter plays a role in delineating ingroup and outgroup
boundaries by establishing exclusionary group identities and by fostering aggression
towards members of outgroups (Platow et al., 2005). Others propose that laughter, as
a function of humor, increases the cohesiveness and performance of goal-oriented
groups (Graetbach & Clark, 2003).
Laughter as Social Lubricant: A Biosocial Hypothesis
The evolutionary perspective is premised on the assumption that traits which
enhanced the survival and reproductive success of our ancestors have spread through
the population at the cost of alternative, less successful traits (Barrett, Dunbar, &
7
LAUGHTER AS SOCIAL LUBRICANT
Lycett, 2002). By this logic, a legitimate question to ask is what function laughter
would have served our hominid ancestors in order for it to become a design feature of
modern humans.
Our evolutionary hypothesis links laughter to individual well-being and the
welfare of their social groups through its impact on positive affect. Fredrickson’s
(1998) Broaden-and-Build hypothesis suggests that positive emotions “serve to
broaden an individual’s momentary though-action repertoire, which in turn has the
effect of building that individual’s physical, intellectual, and social resources”
(Fredrickson, 1998, p. 300). Emotions such as joy, happiness, and pleasure that are
elicited by laughter foster the urge of people to engage in social play and these play
activities help build resources that can be relied upon when individuals or groups face
a variety of different challenges.
Life in ancestral times is thought to have been extremely stressful. The transition
of hominids to more open savannah type environments meant increased daily travel
times to gather food and water, and increased risks of being exposed to predators and
hostile outgroups (Foley, 1996). These threats, in turn, increased the pressures on
groups to expand in size -- in some cases by adding non-kin group members -- and
develop a more complex social organization. These demographic changes would
inevitably have added further stress, not least because large groups increase the extent
to which free-riders -- those who take the benefits of sociality without paying all the
costs -- can prosper at the expense of other group members (Dunbar, 2004ab; Van
Vugt & Van Lange, 2006).
During this transition, there would have been selection for an adaptive mechanism
that could quickly and effectively release positive emotions in people that had the
effect of increasing group cohesiveness and social cooperation. Laughter could
8
LAUGHTER AS SOCIAL LUBRICANT
potentially have served this particular purpose. Additionally, if positive emotions
could be easily and efficiently transmitted between individuals – through a
mechanism such as laughter -- then even those with a negative mood could be
recruited into social activities through the inducement of positive affect. Laughter as
well as the stimuli to produce it, notably humor, could potentially have functioned in
this fashion. Furthermore, the contagious effect of laughter could have facilitated the
social interactions between individuals, allowing them to operate in larger social
networks with minimal costs to group cohesiveness and performance.
Thus, laughter may have been selected for in early humans as a mechanism
whereby individuals, through social play, could build their individual and group
resources to cope with a variety of physical and social dangers. Although the
phylogenetic evidence suggests that laughter was already present in the common
ancestor of humans and chimpanzee -- where it might have played a role in social
bonding and grooming (Dunbar, 2004ab) -- it might not have become a strong
selection force until hominids started to live and move around in large social groups
(Gervais & Wilson, 2005).
The Experiments
Here we propose a biosocial theory of laughter, which integrates the evolutionary
and proximate, psychological functions of laughter into a single coherent theoretical
framework. Evidence from multiple sources, including for example social-
psychological, cognitive-psychological, neuro-endocrinal and neuro-imagining
studies should help build a comprehensive picture of the functions and consequences
of human laughter.
The primary hypothesis derived from this theory is that laughter acts as social
lubricant by facilitating group cohesiveness and social cooperation through eliciting
9
LAUGHTER AS SOCIAL LUBRICANT
positive affect. This hypothesis makes a number of specific predictions about the
individual and group-level effects of laughter, some of which will be tested in this
study. In the experiments, we first consider the behavioral effects of laughter by
examining the effects of a laughing experience on decision-making in a public good
dilemma, a group game measuring people’s cooperative inclinations towards groups
(Fehr & Gaechter, 2002; Van Vugt, De Cremer, & Janssen, 2007). We also examine
the extent to which laughter promotes group cohesiveness more directly by measuring
its effect on people’s identification with their group. We then turn our attention to the
role of laughter in inducing positive affect and reducing negative affect. Finally, we
consider the basic physiological processes underlying the social lubricant effect,
particularly the role of laughter in the release of endorphins, the “feel good”
neuropeptides (Keverne, Martensz, & Tuite, 1989). More specifically, the laughter-as-
social lubricant hypothesis generates several predictions.
The first prediction concerns the behavioral effects of laughter. We predict that,
after a laughing experience, people will be more inclined to act altruistically when
interacting with strangers. Cooperation, especially between strangers, can be quite
risky because it is not clear whether one’s contributions will be matched by others in
the group (Van Vugt & Van Lange, 2006). In contexts where reciprocity is the only
basis for balancing the altruism equation, the risk of free-riders reneging on their
debts is greatly increased. Laughter might make people more optimistic so that they
are more willing to increase their cooperation levels by taking others on trust.
A second prediction from our model concerns the affiliative consequences of
laughter. If the social lubricant hypothesis is correct, we would expect an increase in
the social cohesiveness of a group after laughter. One way to measure this is to
10
LAUGHTER AS SOCIAL LUBRICANT
examine the extent to which people identify with a group of strangers after
undergoing a laughter experience (Hogg, 1992).
Third, laughter should increase people’s positive affect on a range of specific
subjectively experienced emotions, while decreasing their negative affect. There has
been much speculation on this relationship, but not a great deal of evidence that
laughter is directly responsible for inducing positive mood (for an exception, see
Keltner & Bonanno, 1997). We therefore induce laughter experimentally in people
and examine the resulting mood change by administering the validated and well-
established emotional state scale, PANAS, to measure temporary positive and
negative affect states (Watson, Tellegen & Clark, 1980). Ideally, we should see an
increase in positive emotions such as joy and happiness and a reduction in negative
emotions like sadness which would mediate the effects of laughter.
Laughter and Endorphins
Our final prediction concerns the neurophysiological basis of laughter. If laughter
facilitates social cooperation through its effects on positive affect, then we might
expect laughter to correlate with various physiological indices of positive affect. This
prediction arises from our hypothesis that in the human lineage laughter might have
emerged as a “social grooming” device to connect individuals to groups. Grooming is
known to trigger endorphin release from the hypothalamus (Keverne et al., 1989).
Endorphins are associated with the pain control mechanisms, one of whose by-
products is to make people feel happy and allow them to cope better with aversive
stimuli (Pert & Snyder, 1973). Endorphins create a state of mild euphoria which, for
example, enables individuals to maintain physically stressful physical activities (e.g.,
the famous runners high; see Howlett et al., 1984). Their release during social
grooming produces a sense of mild euphoria which appears to be instrumental in
11
LAUGHTER AS SOCIAL LUBRICANT
creating the internal psychological environment conducive for social cooperation. At
the basic neuro-endocrinal level, laughter is therefore expected to be linked to
endorphins, creating the same kind of “high” (Dunbar, 2004b).
Unfortunately, measuring endorphins is a challenge. Because endorphins are
released in the brain, assaying their production can only be done by taking spinal fluid
samples using lumber puncture (Mosby, 1994). However, experts agree that a good
proxy is provided by pain tolerance, since increased pain thresholds are one functional
consequence of endorphin release (Provine, 2000; Zillmann, Rockwell, Schweitzer, &
Sundar, 1993). Therefore, in the final experiment, we induce laughter in participants
and subsequently examine how they cope with a painful task. We predict that laughter
increases pain tolerance levels and argue that this effect is likely caused by the release
of endorphins as a direct result of laughter.
In each of these experiments, we induce laughter by exposing people to carefully
selected comedy clips for a 10-15 minute period, before measuring the main
dependent variables of interest. Consistent with other laughter studies (Provine,
2000), we also examine potential sex differences in our experiments.
Experiment 1: Laughter and Cooperation with Strangers
Does laughter induce cooperation between strangers? To put this suggestion to the
test we exposed people to a brief laughing experience – watching a comedy video clip
-- before examining their decision-making in a cooperative group game. In order to
separate the laughter effects from social bonding that occurs through spending time
together, we compared a condition in which people, before playing the group game,
watched the clip together with their fellow group members (acquaintances condition),
versus a condition where people first watched the clip individually before playing the
game with others (strangers condition). If laughter acts as social lubricant, we
12
LAUGHTER AS SOCIAL LUBRICANT
predicted that after a laughter experience people would cooperate more easily with
strangers. We also predicted that this effect would be related to the degree of laughter.
Method
Design and Participants
One hundred and twelve students, 78 women and 34 men, from the University of
Kent participated in this study. Their age range varied from 18 to 51 years, with a
median age of 20. Participants were randomly assigned to one of four experimental
conditions, with a 2 (clip: neutral vs. comedy) x 2 (grouping: strangers vs.
acquaintances) between subjects design. There were 28 experimental sessions in total,
each session containing four individuals each. One group was removed from the
experiment (session 10), because their data were incomplete due to a technical failure,
thus leaving 27 group sessions.
Procedure
Participants were recruited via the Department of Psychology research
participation scheme and the student webpages at the University of Kent, and signed
up individually for the experiment. Four individuals arrived simultaneously at the
meeting room of the social psychological laboratory, where they were welcomed by
two experimenters and completed an informed consent form.
Manipulation of grouping. To watch the clip, depending upon the grouping
condition, participants were placed either in separate soundproof experimental
cubicles (strangers condition) or in a larger room with three other participants with
whom they formed a group (acquaintances condition). Each room had a single
desktop computer with a 15 inch screen.
Manipulation of clip. Windows media player was used to play the video clips.
The neutral clip showed a 10-minute excerpt from the BBC-series Walking with
13
LAUGHTER AS SOCIAL LUBRICANT
Dinosaurs (the actual episode was from Series 1, the “Giant of the Skies”), filmed in
the style of a nature documentary. The comedy clip consisted of a 10-minute excerpt
of a home video comedy show called You Have Been Framed The show contains a
compilation of viewers’ funny home videos, which commonly stimulate Duchenne
laughter. For instance, one clip shows a man jumping on a trampoline when seconds
later he falls straight through the canvas. The show uses canned laughter to maximize
laughter in viewers.
Public good game. After watching the video, individuals were led back to the
meeting room where they played a one-shot public good task with the three other
participants with whom they had already interacted (acquaintances) or not (strangers).
We used a public goods game without a step-level (for a similar procedure, see De
Cremer & Van Vugt, 1999), and the instructions were given on paper. Each member
received an endowment of £2.50 (approximately $5) in an envelope and was told that
they could invest any amount in a group fund or a private fund. They could keep
whatever they put in their private fund. Whatever ended up in the group fund would
be doubled and divided equally among the four of them, regardless of their individual
group contribution. We ran a few practice trials to enhance their understanding of the
game and its pay-offs. The participants were then asked to split the money between
their private fund and the group fund and complete a short questionnaire about the
experiment.
Subsequently, participants received a careful debriefing about the experiment
and its manipulations from one experimenter, while the other calculated the amount
each person had earned. They were then paid off and asked not to convey the purpose
of the experiment to any other students.
Results
14
LAUGHTER AS SOCIAL LUBRICANT
Manipulation Checks
We asked the participants several questions about the clip they had watched, for
example, “How funny did you find the clip?” (1 = not funny at all, 5 = very funny).
This was analyzed with a 2 (Sex) x 2 (Clip) x 2 (Grouping) between-subjects design.
It produced a significant main effect for clip, F(1, 100) = 49.80, p <.001. The means
showed that the comedy clip was rated considerably funnier (M = 3.05, SD = 0.96)
than the neutral clip (M = 1.67, SD = 0.96). Further tests revealed that in absolute
terms the comedy clip was rated moderately funny (i.e., the mean did not differ from
the scale midpoint), t(49) = .40, p =. 69, whereas the neutral clip was judged to be
rather unfunny, t(47) = -9.70, p <.001.
We also asked them to assess how much they had laughed while watching the
clip: “How much did you laugh while watching the clip” (1 = not at all, 5 = very
much). As predicted, participants thought they laughed more watching the comedy
clip (M = 2.38, SD = 0.97) than the neutral clip (M = 1.31, SD = 0.62), F(1, 100) =
31.62, p <.001. There were no other main effects or interaction effects with sex or
grouping on the amount of laughter.
Thus, our manipulation of laughter was successful. The comedy clip was rated
funnier and produced more (self-reported) laughter.
Group Cooperation
Our main dependent variable was the contribution to the group fund in the public
good game (0 to 2.5 GBP). We analyzed this using a 2 (Sex) x 2 (Clip) x 2 (Grouping)
design. We predicted that people would be more likely to cooperate with strangers
after watching a comedy clip than a neutral clip. The two-way interaction between
clip and grouping was marginal, F(1, 100) = 2.69, p <.10.
15
LAUGHTER AS SOCIAL LUBRICANT
Looking at the conditions separately shows that in the strangers condition people
contributed more after watching the comedy clip (M = 1.34, SD = 0.60) than the
neutral clip (M = 1.04, SD = 0.57), F(1, 48) = 4.00, p =.05. There was no difference
between the two clips (M’s = 1.41 vs. 1.32, SD’s = 0.69 and 0.63) in the acquaintances
condition, F(1, 52) = 1.24, p =.27. Furthermore, the contribution level in the
comedy/strangers condition was virtually the same as in the two acquaintances
conditions in which people had already interacted with their fellow group members
before playing the game (both p’s >.05).
There was also an unexpected three way interaction between sex, clip and
grouping, F(1, 100) = 3.94, p =.05. The means associated with this interaction are
given in Table 1. They suggest that this three-way interaction was caused by males
being particularly cooperative towards strangers after watching the neutral clip (M =
1.73, SD = 0.61) and that this, in turn, was mainly due to one particular group of three
males and one female (session 20) in which the three men contributed an unusually
large amount of money to the group fund (M = 1.83). This suggests that the sex
composition of the group might be an important factor, something which we will
address later. (Incidentally, when this particular group was eliminated from the
analysis, the two-way interaction between clip and grouping became significant, F(1,
96) = 3.55, p <.05).
Laughter as Mediator
We wanted to know if laughter would mediate the effects of watching a comedy
clip on cooperation among strangers. Hence, we used the manipulation check (“How
much did you laugh while watching the clip?”) in a covariance analysis with group
contribution as dependent variable.
16
LAUGHTER AS SOCIAL LUBRICANT
As expected, there was a positive correlation between laughter and contribution in
the strangers condition (r = .33, p <.02), but not in the acquaintances condition (r = .
10, p =.47). Thereafter, we performed an ANCOVA on group contribution within the
strangers condition only using a 2 (Sex) x 2 (Clip) design and adding laughter as
covariate to the design
The covariate was significant, F(1, 47) = 4.86, p <.04. Moreover, the effect of clip
was completely wiped out after partialing out the impact of laughter, F(1, 47) < 1,
(i.e., in the original analysis: F[1, 48] = 4.00, p =.05), Sobel’s z = 2.45, p <.02.
Summary of Results
Experiment 1 supports the social lubricant hypothesis in three separate respects.
Watching the comedy clip induced laughter and enhanced cooperation among groups
of strangers in particular (to the same level as for acquaintances). Furthermore, the
rise in group contributions -- due to the effect of watching a comedy clip -- was
accounted for by the amount of (self-assessed) laughter. Finally, we encountered an
interesting sex difference: Men contributed relatively more when they were among
acquaintances watching the neutral clip, but this could be attributed to one session
with three males and a single female (we will address this in the General Discussion).
Experiment 2: Laughter, Positive Affect, and Social Bonding
The second experiment had three aims. First, we wished to replicate the results
from the first study in order to give added support to the laughter-as-social lubricant
hypothesis by measuring actual laughter instead of self-assessed laughter. The social
lubricant hypothesis predicts that laughter facilitates social cooperation among
strangers through inducing positive mood. Therefore, a second aim was to determine
whether the amount of laughter is directly proportional to an increase in positive
affect and decrease in negative affect (this link has often been suggested in the
17
LAUGHTER AS SOCIAL LUBRICANT
literature, but not directly tested; but see Neuhoff & Schaefer, 2002). Hence, we
included the PANAS to measure people’s affective states. A third aim was to
determine whether laughter would increase social cohesiveness in groups of strangers.
To do this, we added a group identification scale (De Cremer & Van Vugt, 1999) to
measure the extent to which members felt emotionally attached to their group during
the game.
We used different clips in this experiment, for two reasons. First, although the
comedy clip in Experiment 1 was rated funnier than the neutral clip, it was not rated
very funny in absolute terms. Second, there was an unexplained sex difference in the
rating of the neutral clip in the first study. Although we were not quite sure what
caused this, we nevertheless decided to use a different clip.
Method
Design and Participants
Eighty students, 46 women and 34 men, from the University of Kent participated
in this study. Their age range varied from 18 to 28 years, with a median age of 20.
Participants were randomly assigned to one of four experimental conditions,
following a 2 (clip: comedy vs. neutral) x 2 (grouping: strangers vs. acquaintances)
between-participants design. There were 20 experimental sessions, containing four
individuals each.2
Procedure
The procedure was essentially the same as in the first experiment with a few
notable exceptions. First, we used different video clips. For the comedy clip, we opted
for a 10 minute clip of an episode from the Simpsons (episode 3F31; “The Simpsons
138th episode spectacular”), which the experimental assistants (who were themselves
18
LAUGHTER AS SOCIAL LUBRICANT
university students) rated as funny. For the neutral clip, we took a clip from a sports
documentary entitled “Go Tigers”, which the assistants rated as boring.
Second, we measured actual laughter during the clip. We used audio-recorders,
because we did not want the participants to watch the clip in the same room as the
experimenters. After each experimental session, two experimental assistants listened
to the audio-tape and wrote down independently the number of laughs they heard
during the 10 minute clip. Their ratings were highly consistent (Cohen’s kappa = .88).
For analysis, we averaged their scores in those cases where there were discrepancies.
Once again, in the strangers condition we placed people in separate
soundproof cubicles and in the acquaintances condition they were placed in a room
together with their fellow group members to watch the clip.
After watching the clip, they were brought together in a meeting room. Before
playing the group game, we asked the participants to rate their feelings by completing
the 20 item PANAS (e.g., “At this moment I feel happy, …sad, …excited, …upset; 1
= very slightly or not at all, 5 = extremely).
Thereafter they played the group game, which was essentially the same as in the
previous experiment with one exception. In order to determine whether the results in
Experiment 1 were due to a ceiling effect, we increased the stakes of the game, giving
each participant a £5 endowment which they had to divide between their private fund
and the group fund. After this, they completed a 10 item group identification scale
(adopted from Brown et al., 1986) asking them how they felt about the group with
whom they played the game: Items included “I am a person who is glad to belong to
this group” “…who feels strong ties to the group” (1 = never, 5 = very often).
19
LAUGHTER AS SOCIAL LUBRICANT
After some final questions about the experiment (including some manipulation
checks), the participants were debriefed by one experimenter while the other
calculated the amount each individual had earned and paid it out to them.
Results
Manipulation Checks
We asked participants the question: “How funny did you find the clip?” (1 =
not funny at all, 5 = very funny). This was analyzed with a 2 (Sex) x 2 (Clip) x 2
(Grouping) design. This analysis only revealed a significant main effect for clip, F(1,
72) = 44.63, p <.001, with the comedy clip (M = 3.53, SD = 0.88) being rated funnier
than the neutral clip (M = 2.18, SD = 0.93). Each of the means differed significantly
from the scale midpoint, respectively t(39) = 3.79, p <.001, and t(39) = -5.61, p <.001.
There were no other main or interaction effects.
The actual laughter was measured via audio-recorders. To standardize the
measure between the two groups, we compared the number of laughs for each
individual in the strangers condition with the average of the group laughter in the
acquaintances condition (i.e., n laughs/4). The maximum number of laughs for any
individual was 23 and the minimum was zero with a median of 4 laughs. As predicted,
there was a main effect for clip, F(1, 72) = 128.87, p <.001. People laughed
considerably more when watching the comedy clip (M = 11.88, SD = 6.02) than the
neutral clip (M = 1.05, SD = 1.24).
There was also a marginal interaction effect between clip and sex, F(1, 72) =
3.86, p =.06, revealing that, relative to females (M = 10.05, SD = 5.84), males laughed
more when watching the comedy clip (M = 14.11, SD = 5.61) with no difference in
the neutral clip (respective M’s = 1.00 and 1.13, SD’s = 1.31 and 1.22). There were no
further main or interaction effects on laughter.
20
LAUGHTER AS SOCIAL LUBRICANT
Taken together, it seems that the laughter manipulation was successfully
induced. The comedy clip was considered funnier and produced considerably more
laughter than the neutral clip.
Group Cooperation
The main dependent variable was the contribution to the group fund (0 - 5 GBP)
in the public good game. We analyzed this using a 2 (Sex) x 2 (Clip) x 2 (Grouping)
design. Recall that we expected group contributions to increase after watching the
comedy clip, particularly among groups of strangers. We found two main effects, for
clip, F(1, 72) = 10.52, p <.01, and for grouping, F(1, 72) = 8.02, p <.01, which were
qualified by the predicted two-way interaction between clip and grouping, F(1, 72) =
6.82, p <.02. The means are depicted in Figure 1. Further testing shows that group
contributions were considerably higher in the strangers condition after watching the
comedy clip (M = 2.45, SD = 1.47) than the neutral clip (M = 0.80, SD = 0.83), F(1,
36) = 18.12, p <.001, but there was no difference between these clips in the
acquaintances condition (respective M’s = 2.55 and 2.40, SD’s = 1.13 and 1.47), F(1,
36) <1. In fact, the group contributions in the comedy-strangers condition were
roughly equal to the two acquaintances conditions, Fs(1, 36) <1. There were no main
effects or interaction effects for sex in this experiment.
Group Cohesiveness
We created a single identification score for each participant based on the average
score across the 10 items of the scale to measure the cohesiveness of the group (α = .
91). We then subjected this to an ANOVA with the complete three factor design. This
yielded two main effects, for clip, F(1, 72) = 30.92, p <.001, and grouping, F(1, 72) =
69.77, p <.001, as well as the predicted interaction between clip and grouping, F(1,
72) = 21.96, p <.001. The means are displayed in Figure 2. Paralleling the results on
21
LAUGHTER AS SOCIAL LUBRICANT
social cooperation, participants identified more strongly with a group of strangers
after watching the comedy clip (M = 3.19, SD = 0.86) than the neutral clip (M = 1.65,
SD = 0.32), F(1, 36) = 54.12, p <.001. In this condition, they identified as much with
their group as in the neutral acquaintances condition (M = 3.62, SD = 0.54; p =.22)
but not as much as in the comedy-acquaintances conditions (M = 3.75, SD = 0.77; p
<.05) – the latter two means did not differ from each other (p >.05).
Positive and Negative Emotions
We examined the effects of laughter on the mood of the participants with the
positive and negative subscales of the PANAS. An exploratory factor analysis
revealed that the affective states could be neatly grouped into two main factors,
positive affect, which explained 37.6% of the variance in items, and negative affect,
which accounted for an additional 14.4% of item variance. The following items
scored highest on the Positive Affect Scale (PAS): Excited (.74), happy (.72), joyful
(.71), enthusiastic (.78), and inspired (.72). Items scoring highest on the Negative
Affect Scale (NAS) were: Upset (.67), sad (.64), uptight (.85), lonely (.58), and bored
(.55). Both PAS and NAS were reliable scales (respective α’s = .91 and .77) and the
correlation between the scales was moderate in size (r = -.52).
A MANOVA including the full factorial design and PAS and NAS as dependent
variables revealed significant main effects for clip, F(2, 71) = 21.74, p <.001, and
grouping, F(2, 71) = 10.23, p <.001and a significant interaction between these factors,
F(2, 71) = 7.78, p <.01 with univariate effects on PAS, F(1, 72) = 6.89, p <.001, and
NAS, F(1, 72) = 13.32, p <.001. The means of this interaction are displayed in Table
2. They suggest that watching the comedy clip increased people’s positive mood,
while reducing negative mood, particularly in the strangers condition when people
were watching the clips alone. The affective state of people in this condition was
22
LAUGHTER AS SOCIAL LUBRICANT
virtually the same as in the two acquaintances conditions where people had interacted
with each other before.
This analysis also revealed a multivariate main effect for Sex, F(2, 71) = 4.65, p
<.05, with a univariate effect on NAS only, F(1, 72) = 6.89, p <.02. The means show
that males reported greater negative affect (M = 2.04, SD = 0.75) than females (M =
1.68, SD = 0.60), but this effect was not further influenced by our manipulations.
Does Laughter Induce Social Cooperation and Positive Affect?
Finally, we tested whether the beneficial effects of watching a comedy on affect
and behavior could actually be explained by the amount of laughter. The social
lubricant theory asserts that laughter is responsible for these effects. To investigate
this, we conducted analyses of covariance on the measures of social cooperation,
group identification, and the two affective scales with the individual laughter score as
covariate.
Encouragingly, across the entire sample there were positive correlations of
laughter with social cooperation (r = .27, p <.02), group identification (r = .24, p
<.04) and positive affect (r = .49, p <.001), and a negative correlation with negative
affect (r = -.43, p <.001). Thus, across conditions, laughter seems to have positive
effects on people’s affective state and within group bonding.
We expect the beneficial effects of laughter to be particularly strong in the groups
of strangers. Thus, we first performed an ANCOVA on group contribution within the
strangers condition only using a 2(Sex) x 2(Clip) design and adding laughter as
covariate to the design
The covariate was significant, F(1, 35) = 4.86, p <.04. Furthermore, adding
laughter significantly reduced the effect of the clip, F(1, 35) = 11.91, p <.01 (i.e., in
the original analysis: F[1, 36] = 18.12, p <.001, a 33% reduction), Sobel’s z = 2.40, p
23
LAUGHTER AS SOCIAL LUBRICANT
<.02. For group identification, a similar pattern was obtained. Laughter was a
(marginally) significant covariate, F(1, 35) = 3.51, p =.06. Adding laughter to the
analysis reduced the impact of watching a comedy clip on group identification, F(1,
36) = 30.51, p <.001 (i.e., in the original analysis; F(1, 36) = 54.12; a reduction of
nearly 45%), Sobel’s z = 3.66, p <.001. For both measures this suggests that laughter
was partially (but not completely) responsible for the beneficial effects of watching a
comedy on social bonding between strangers.
Finally, when we looked at the effects of laughter on the mood of people, a
similar pattern emerged. Adding laughter as covariate, F(1, 35) = 11.58, p <.001,
completely wiped out the effect of watching a comedy clip on positive affect, F(1, 35)
= 0.56, p =.46 (a highly significant effect in the original analysis; F(1, 36) = 38.99, p
<.001), Sobel’s z = 6.63, p <.001. The same happened for the negative affect scale:
laughter, F(1, 35) = 8.41, p <.01, statistically eliminated the effect of comedy on (a
reduction in) negative affect, F(1, 36) = .91, p =.35 (in the original analysis; F(1, 36)
= 36.97, p <.001), Sobel’s z = 5.49, p <.001.
Summary
Experiment 2 replicated and extended the social lubricant hypothesis of laughter.
First, watching the comedy clip induced laughter. Second, watching the comedy clip
enhanced cooperation rates among groups of strangers up to the level as in groups of
acquaintances. This was paralleled by an increase in the cohesiveness of groups of
strangers after watching comedy. Furthermore, watching a comedy clip induced a
positive mood among individuals, while reducing their negative mood. Finally, as
predicted by the social lubricant hypothesis, we found that the effects of watching
comedy on the affective and behavioral measures could be attributed, either entirely
or partly, to laughter.
24
LAUGHTER AS SOCIAL LUBRICANT
Experiment 3: Laughter and Pain Tolerance
In the third experiment, we examine the hypothesis that laughter stimulates the
production of endorphins, arguing that it is the relaxing effect of these endorphins that
is responsible for the increased positive affect and willingness to cooperate in groups.
Since direct assays for endorphin production can only be done using lumber puncture
(a process that is itself painful, as well as being fraught with adverse side-effects:
Mosby, 1994), we followed conventional practice (e.g., Zillman et al., 1993) and used
pain tolerance as an index of endorphin production. In this study we used slightly
longer clips (15 rather than 10 minutes), because we were not sure how quickly the
impact of endorphins would set in. Furthermore, we used a more refined measure of
laughter, measuring whether or not the participant was laughing on instantaneous scan
samples taken at 30 second intervals, which we used to calculate a percentage score of
laughter (out of 30 events).
Method
Design and Participants
This experiment employed a between subject design and consisted of 32
participants (14 female and 18 male) who were recruited from the student population
at the University of Liverpool. Half of them watched a comedy clip and half watched
a neutral clip. All participants completed a questionnaire before the experiment which
provided background information on their attitudes to humor, and also provided
confirmation of their willingness to take part in the project. Participants ranged in age
from 18-60 years.
Procedure
For each session, participants undertook a pain tolerance assay (see below), then
watched a 15-minute video, and repeated the pain tolerance assay. Participants were
25
LAUGHTER AS SOCIAL LUBRICANT
told that they were taking part in a study of pain tolerance (which had to be measured
twice), but not about the relationship between pain and endorphins or the relationship
between laughter and endorphins. They were fully debriefed after the experiment.
While they were watching the video, the experimenter recorded whether or not each
participant was laughing at 30-sec intervals.
Participants watched the clips in groups of 2-6 individuals, and were asked not
to talk to each other while watching the clips. We used a variety of comedy clips such
as Mr Bean, You’ve Been Framed, Eddie Izzard, Father Ted, Friends and The
Simpsons, all of which were regular comedy slots on UK television which had been
selected by the experimental assistants who were students themselves. The neutral
clips were either from a News programme, Eastenders (a TV soap), Songs of Praise
(a Sunday evening religious slot), a science documentary and a sports programme (on
golf).
Pain tolerance was examined using a frozen vacuum wine cooler sleeve,
frozen to -16oC for the start of each test. Participants were asked to put a sleeve over
their forearm and to keep it there for as long as they could stand it (but, in any case,
for a maximum of 180 seconds to avoid skin damage). A stopwatch was used to
determine (to the nearest complete second) how long they kept the sleeve on. This
procedure was approved by the University’s strict ethical guidelines for conducting
research with human participants.
Results
We assessed endorphin production by the change in pain tolerance (indexed as
the number of seconds for which the frozen wine cooler could be kept on the arm)
before and after watching the clip. Figure 3 plots the difference in pain tolerance
(after minus before) against the proportion of laughs (0-100%) for individual
26
LAUGHTER AS SOCIAL LUBRICANT
participants. Laughter was a significant predictor of the difference in pain tolerance,
t(31) = 3.12, p <.01, with more laughter associated with an increase in pain tolerance
(beta = .50).
We employed an ANOVA to assess the impact of laughter on pain tolerance.
Hence, a 2(Sex) x 2(Clip) ANOVA with pain tolerance (i.e., the difference between
the preclip and postclip pain measures) as dependent measure revealed a significant
main effect for clip, F(1, 28) = 11.36, p <.001, which was qualified by an interaction
with sex, F(1, 28) = 5.67, p <.03. It suggests that there was an increase in pain
tolerance after watching the comedy clip (M = +23.91, SD = 38.81) versus the neutral
clip (M = -9.38, SD = 26.52), but this effect was considerably larger for males (Mdiff
= +62.99; respective M’s +44.87 vs. -18.12, SD’s = 44.12 and 38.25) than for females
(Mdiff = +10.59; respective M’s = +8.20 vs. -2.39, SD’s = 27.34 and 8.27).
An ANOVA on the laughter proportion (0-100%) revealed a significant main
effect for clip, F(1, 28) = 3099.85, p <.001, and a significant interaction between clip
and sex, F(1, 28) = 10.36, p <.01. There was more laughter in the comedy condition
(M = 54.06, SD = 3.51) than in the neutral condition (M = 1.67, SD = 2.63).
Furthermore, males laughed more frequently watching the comedy clip than the
neutral clip (respective M’s = 56.70 vs. 0.84, SD’s = 2.37 and 3.12) relative to females
(respective M’s = 52.09 vs. 2.33, SD’s = 3.54 and 2.75).
Finally, we conducted an ANCOVA on pain tolerance with laughter as
covariate and this covariate was marginally significant, F(1, 27) = 3.35, p =.07.
Furthermore, both the main effect for clip, F(1, 27) < 1, Sobel’s z = 3.04, p <.01, and
the interaction between clip and sex disappeared, F(1, 27) = 2.90, p =.10, Sobel’s z =
2.42, p <.02, when laughter was added as covariate, suggesting that there is a direct
27
LAUGHTER AS SOCIAL LUBRICANT
link between the laughter produced by the comedy clip and the increase in pain
tolerance of participants.3
Summary
Thus, watching a comedy clip seems to be associated with a significant
increase in serum endorphin titre (as indexed by pain tolerance), and the magnitude of
this effect is a linear function of the degree of laughter among participants. There was
an interesting sex difference with males showing an increased pain tolerance after
watching a comedy clip, which was related to the proportion of laughter (we address
this in the General Discussion).
General Discussion
This research yields convergent support for the idea that laughter acts as social
lubricant by enhancing a sense of group membership via the release of endorphins.
Laughter was induced through exposure to a comedy (vs. neutral) clip. We then
measured participants’ physiological, affective, and behavioral reactions in various
experimental tasks. Our findings revealed that laughter – through watching a comedy
– induces positive affect, while reducing negative affect. Furthermore laughter
facilitates social bonding and social cooperation particularly among strangers.
Laughter also significantly increases pain tolerance, which suggests that laughter is
associated with endorphin release. Our research is one of the first to show that
laughter is directly responsible for these effects. Below we discuss the implications of
our findings, attempting to integrate them into a broader biosocial hypothesis about
the functions of laughter in groups.
Social Lubricant Theory of Laughter
The social lubricant theory is inspired by social, evolutionary, and positive
psychological approaches to the study of laughter. It assumes that laughter fosters
28
LAUGHTER AS SOCIAL LUBRICANT
people’s attachment to groups through the release of mood-inducing endorphins.
Laughter serves to enhance a sense of social identity which facilitates cooperation
among strangers. Our theory connects different levels of analysis about the proposed
functions of laughter. At the proximate level, we can look at the immediate functions
of laughter, the way it affects individual well-being and group welfare.
Consistent with our theory, our research reveals a myriad of beneficial effects of
laughter. In terms of neurophysiology, our research is among the first to show the
effects of laughter on endorphin production. Endorphins are neuropeptides that create
a mild sense of euphoria in organisms which is believed to have a range of benefits
for health and mood (Provine, 2000). Laughter is used, for example, in health
intervention programs to overcome depression (Rosner, 2002) and cope with pre-
surgery anxiety (Vagnoli, 2005). Procedures to measure endorphins directly in
humans are quite invasive (via a spinal puncture) and so we obtained a reliable
indirect measure in the form of pain tolerance during a physical task (Zillman et al.,
1993). Laughter significantly predicted pain tolerance.
Mirroring the endorphin effects, our research also shows that laughter affects
people’s subjective mood state. Laughter increased positive affect while reducing
negative affect. Participants who watched a comedy clip were happier, more excited,
more inspired, less sad and less bored. Furthermore, their mood changes could be
attributed to the amount of laughter. These data are consistent with other research
showing, for example, that a one-minute laughter experience enhances mood
(Neuhoff & Schaeffer, 2002)
The social lubricant theory proposes that laughter (through the above effects)
facilitates social cooperation between strangers. For this purpose, we utilized a public
goods game to measure the impact of laughter on the convergence between individual
29
LAUGHTER AS SOCIAL LUBRICANT
and group interests (Van Vugt et al., 2007). Our research shows that people who
watched a comedy became more cooperative toward strangers, thus improving the
performance of their group. As expected, the amount of laughter was directly
responsible for these effects. Paralleling these results, we also found that laughter
positively affected group cohesiveness by raising the extent to which people identified
with their group.
This suggests that the effects of laughter are primarily the result of a
transformation of motives, whereby the collective group interest starts to weigh more
heavily than self-interest (Brewer & Kramer, 1994; De Cremer & Van Vugt, 1999).
Interestingly, in the acquaintances conditions, laughter did not produce a significant
increase in cooperation. The second experiment (in which endowments were doubled)
revealed that this was not simply due to a ceiling effect. The most probable reason is
that, in these conditions, participants already trusted each other through talking to
each other. This suggests that laughter and language serve as complimentary
mechanisms for inducing cooperation, possibly with laughter having a stronger effect
on group-based affect and language on group-based trust (Dawes, McTavish, &
Shaklee, 1977). This awaits further investigation.
This brings us to speculate about the ultimate functions of laughter. We
hypothesize that the capacity for laughter originates in the importance of group living
for hominids. Group living enabled our ancestor to solve many important problems of
survival and reproduction such as food sharing, communal child care, and coalitional
defense. As group life became more socially complex – perhaps due to an increase in
group size (Dunbar, 2004a) – this development created a need for mechanisms that
could foster the integration of individuals within groups. Laughter may be one such
mechanism. Because laughter quickly releases positive emotions, it is an effective
30
LAUGHTER AS SOCIAL LUBRICANT
way to smoothen social interaction between individuals (especially in potentially
conflicting relations between strangers), thus enabling them to create and operate in
highly cooperative units necessary for survival. In this view, laughter might have
evolved as an adaptive mechanism to build up social reserves that groups could rely
upon in times of crises, a view that is consistent with positive psychology (cf.
Fredrickson, 1998).
Albeit somewhat speculative, this evolutionary scenario is not a “just-so story.”
Experts agree that the human capacity for laughter has very ancient origins, and is
probably older than our language ability itself (Gervais & Wilson, 2005). Laughter is
found across all human cultures, and is one of the first vocal expressions of babies
(Provine, 2000). Neuro-imagining studies strongly suggest the presence of a
“laughter-coordination centre” in the human brain (Wild et al., 2003). And, a
rudimentary form of laughter is present among our closest primate cousin, the
chimpanzee, which is elicited during social play and tickling (Van Hooff, 1972;
Waller & Dunbar 2005). Human laughter may have co-evolved with the language
capacity to create typically human phenomena such as humor, a very potent stimulus
to elicit Duchenne laughter, as well as non-Duchenne laughter, a form of laughter that
lacks a strong emotional component (e.g., the nervous laugh; Keltner & Bonanno,
1997).
Sex Differences
There were various sex differences in our studies and they deserve some special
attention. The research on sex differences in laughter is a mixed bag. In social
contexts, there seem to be few differences between males and females in laughter
production (Owren & Bachorowski, 2003). However, it is quite likely that males and
females differ in laughter as a function of differences in what they consider to be
31
LAUGHTER AS SOCIAL LUBRICANT
funny (Provine, 2000). For example, the comedy clips that were used in Experiments
2 and 3 such as the Simpsons produced more laughter from men than from women.
But there were no differences in laughter for the Have you Been Framed clips (funny
home videos). The fact that the effects of laughter itself did not differ between the
sexes is encouraging for our theory. The final experiment showed that differences in
pain tolerance between males and females were fully accounted for by the sex
differences in laughter production. Nevertheless it is important for future research to
be aware of sex differences when developing stimulus materials (Azim, 2005).
Another issue to consider in future research is the sex composition of the groups.
A consistent finding in the literature is that females tend to laugh more around males
in conversational interactions (Owren & Bachorowski, 2003). The results of the
second experiment confirm this: Females laugh more in mixed sex groups (5.3 laughs
on average) than in all female groups (3.5 laughs). The other group composition effect
we obtained was for cooperation rates in Experiment 1. Within a complicated three-
way interaction there was evidence for increased male group contributions in one
particular condition. This happened to be a group in which three males were paired
with one female (we avoided this grouping in the second experiment). There is
evidence to suggest that males sometimes use generosity as a strategy to attract the
interest of a female (sexual selection theory; Buss, 1994; Hardy & Van Vugt, 2006),
which is what may have happened in this case.
Limitations of the Present Research and Directions for the Future
The first obvious criticism concerns the ecological validity of our findings. How
important is laughter in real life? We can only speculate about this. Obviously we only
examined relative small personal sacrifices from participants, particularly in the first
experiment (up to $5). Although such amounts are quite common in economic games
32
LAUGHTER AS SOCIAL LUBRICANT
experiments (Komorita & Parks, 1994), the question is whether laughter promotes
more substantial group sacrifices. The short answer is: We don’t know. But
interestingly, once we doubled the endowment in Experiment 2, laughter produced
similar pro-social effects, suggesting that the amount of personal sacrifice makes little
difference.
Setting aside these results, it is noteworthy that one of the most desirable traits in
potential partners is sense of humor (Buss, 1994), suggesting the importance of
laughter for mating relationships. In addition, many experts on the human condition
from the Bible to Shakespeare to modern day health professionals have reminded us
that “laughter is the best medicine.” In support of this, various health programs use
laughter therapy to help people cope with various stressors like bereavement, surgery,
and fertility treatment (Rosner, 2002). Yet, as Provine (2000) suggests, many studies
suffer from poor experimental designs and further research into the health benefits of
laughter is desperately needed.
A second potential criticism is the absence of any direct neurophysiological
evidence for the impact of laughter. For example, our claim that endorphins are
involved in laughter is based on indirect evidence. It is therefore important to collect
further neurobiological data on laughter. For example, via saliva and urinary samples
it is currently possible to measure oxytocin, a neuro-hormone that is believed to be
involved in attachment processes (Taylor et al., 2000). It may well be that laughter
stimulates the release of oxytocin in addition to endorphins. In addition, neuro-
imagining studies may be useful to examine the neuropsychological correlates of
laughter production. So far, the constraints of imaging technology mean that it is
virtually impossible to study laughter in any detail: Accurate imaging requires the
33
LAUGHTER AS SOCIAL LUBRICANT
subject to remain very still, but full-blown Duchenne laughter commonly involves
gross head and body movements.
Another recommendation for research is to compare the effects of Duchenne with
non-Duchenne laughter. Our theory predicts beneficial effects from laughter, but our
research has focused on Duchenne laughter, that is, spontaneous stimulus-driven
laughter which is elicited when people are exposed to humor. What are the effects of
non-Duchenne laughter, which occurs in the absence of attempts at humor? Because
this type of laughter lacks a strong emotional component, we expect it will be less
effective in social bonding. In a similar set-up as the present studies, we could ask
people to laugh out loud for a certain duration and contrast the effects with the
Duchenne laughter generated by humor.
Conclusions and Implications
The main goal of this research has been to introduce a new theoretical perspective
on laughter, the laughter-as-social-lubricant hypothesis, and provide some data in
support of this hypothesis. This hypothesis integrates findings from many different
disciplines that have studied laughter, including social psychology, positive
psychology, evolutionary psychology, cognitive neurosciences, and gelotology (the
scientific study of laughter). Through controlled experimental research we have
shown that laughter enhances a sense of group membership which enables individuals
to engage in profitable social exchanges with strangers. Our research suggests a role
for endorphins in producing these effects of laughter. More research is obviously
needed to test different elements of the social lubricant theory of laughter and we have
given some recommendations for future inquiry. Laughter is a serious topic of
scientific inquiry and deserves more attention from researchers interested in human
social behavior.
34
LAUGHTER AS SOCIAL LUBRICANT
References
Apte, M. L. (1985). Humor and laughter: An anthropological approach. Ithaca
(NY): Cornell University Press.
Azim, A., Mobbs, D., Jo, B., Menon, V., & Reiss, A. (2005). Sex differences in brain
activation elicited by humor. Proceedings of the National Academy of
Sciences, 102, 16496-16501.
Bachorowski, J. A., & Owren, M. J. (2001). Not all laughs are alike: Voiced but not
unvoiced laughter readily elicits positive affect. Psychological Science, 12,
252-257.
Baumeister, R. F., & Leary, M. (1995). The need to belong: desire for interpersonal
attachments as a fundamental human motivation. Psychological Bulletin, 117,
497-529.
Barrett, L., Dunbar, R., & Lycett, J. (2002). Human evolutionary psychology. London:
Palgrave.
Becker, S., & Eagly, A. H. (2004). The heroism of women and men. American
Psychologist, 59, 163-178.
Brown, R., Condor, S., Mathews, A., Wade, G., & Williams, J. (1986). Explaining
intergroup differentiation in an industrial organization. Journal of Occupational
Psychology, 59, 273-286.
Buss, D. (1994). The evolution of desire: Strategies of human mating. New York:
Basic Books
Chartrand, T. & Bargh, J. (1999). The chameleon effect: The perception behavior link
and social interaction. Journal of Personality and Social Psychology, 76,
893-910.
35
LAUGHTER AS SOCIAL LUBRICANT
Darwin, C. (1872). The expressions of the emotions in man and animals. London:
John Murray.
Dawes, R. M., McTavish, J., & Shaklee, H. (1977). Behavior, communication, and
assumptions about other people’s behavior in a commons dilemma situation.
Journal of Personality and Social Psychology, 35, 1-11.
De Cremer, D., & Van Vugt, M. (1999). Social identification effects in social
dilemmas: A transformation of motives. European Journal of Social
Psychology, 29, 871-893.
Deacon, T. W. (1997). The symbolic species.: the co-evolution of language and the
brain. New York: W. W. Norton.
Dunbar, R.I.M. (1999). Culture, honesty and the freerider problem. In: R.I.M.
Dunbar, C. Knight & C. Power (eds) The Evolution of Culture, pp. 194-213.
Edinburgh University Press: Edinburgh.
Dunbar, R.I.M. (2004a). Gossip in evolutionary perspective. Review of General
Psychology, 8, 100-110.
Dunbar, R.I.M. (2004b). The Human Story. Faber & Faber: London.
Dunbar, R.I.M. (in press). Mind the bonding gap: constraints on the evolution of
hominid societies. In: S.Shennan (ed) Pattern and Process in Cultural
Evolution.
Eibl-Eibelsfeldt, I. (1989). Human ethology. New York: Aldine De Gruyter.
Fehr, E., & Gaechter, S. (2002). Altruistic punishment in humans. Nature, 415,
137-140.
Foley, R. (1996). An evolutionary and chronological framework for human social
behavior. In W. Runciman, J. Maynard-Smith, & R. Dunbar, Evolution of social
behavior patterns in primates and man; A joint discussion meeting of the royal
36
LAUGHTER AS SOCIAL LUBRICANT
society and the British Academic (pp. 95-117). Oxford: Oxford University press.
Gervais, M. & Wilson, D. S. (2005). The evolution and functions of laughter and
humor: A synthetic approach. The Quarterly Review of Biology, 80, 395-430.
Fredrickson, B. (1998). What good are positive emotions. Review of General
Psychology, 2, 300-319.
Graetbach, D., & Clark, T. (2003). Displaying group cohesiveness: humor and
laughter in public lectures of management gurus. Human relations, 56, 1515-1544
Hardy, C. L., & Van Vugt, M. (2006). Nice guys finish first: The competitive altruism
hypothesis. Personality and Social Psychology Bulletin,
Hogg, M (1992). The social psychology of group cohesiveness: From attraction to
social identity. London: Harvester Wheatsheaf.
Howlett, T.A., Tomlin, S., Ngahfoong, L., Rees, L.H., Bullen, B.A., Skrinar, G.S.,
& MacArthur, J.W. (1984). Release of β-endorphin and met-enkephalin
during exercise in normal women in response to training. British Medical
Journal, 288, 295-307.
Iwase, M., Ouchi, Y., Okada, H., Yokoyama, C., & Nobezawa, S. (2002). Neural
substrates of human facial expression of pleasant emotion induced by comic film:
A PET study. Neuroimage, 17, 758-768.
Keverne, E.B., Martensz, N. & Tuite, B. (1989). Beta-endorphin concentrations in
cerebrospinal fluid of monkeys are influenced by grooming relationships.
Psychoneuroendocrinology 14: 155-161.
Kramer, R., & Brewer, M. (1984). Effects of group identity on resource use in a
simulated commons dilemma. Journal of Personality and Social Psychology,
46, 1044-1057.
Mosby, (1994). Medical, Nursing and Allied Dictionary. Mosby: London.
37
LAUGHTER AS SOCIAL LUBRICANT
McComas, H. C. (1923). The origin of laughter. Psychological Review, 30, 45-55.
Keltner, D., & Bonanno, G. A. (1997). A study of laughter and dissociation: Distinct
correlates of laughter and smiling during bereavement. Journal of Personality
and
Social Psychology, 73, 687-702.
Komorita, S., & Parks, C. D (1994). Social dilemmas. xxx
Neuhoff, C. C., & Schaeffer, C. (2002). Effects of laughing, smiling and howling on
mood. Psychological Reports, 91, 1079-1080.
Owren, M. J., & Bachoroskwi, J. A. (2003). Reconsidering the evolution of
non-linguistic communication: The case of laughter. Journal of Nonverbal
Behavior, 27, 183-2000.
Provine, R. R. (2000). Laughter: A scientific investigation. New York: Viking.
Rosner, F. (2002). Therapeutic efficacy of laughter in medicine. Cancer investigation,
20, 434-436.
Kawakami, K. et al., (2006). Origins of smile and laughter: A preliminary study.
Early Human Development, 82, 61-66.
Platow, M., Haslam, A., Both, A., Chen, I., Cuddon, M., Goharpy, N., Maurer, J.,
Rosini, S., Tsekouras, A., & Grace, D. M. (2005). It is not funny if they are
laughing: self-categorization, social influence, and responses to canned laughter.
Journal of Experimental Social Psychology, 41, 542-550
Sander, K., & Scheich, H. (2001). Auditory perception of laughing and crying
activates human amygdala regardless of attentional state. Cognitive Brain
Research, 12, 181-198.
Taylor, S. Taylor, S., Klein, L. C., Lewis, B. P., Gruenewald, T., Gurung, R. A.,
Updegraff, J. A. (2000). Biobehavioral responses to stress in females: Tend-and
38
LAUGHTER AS SOCIAL LUBRICANT
-befriend, not flight or fight. Psychological Review, 107, 411-429.
Van Hooff, J. A. R.A. M. (1972). A comparative approach to the phylogeny of
laughter ands smiling. In R. Hinde (Ed.), Nonverbal communication (pp.
209-243). Cambridge: Cambridge University press.
Vagnoli, L., Caprilli, S., Robiglio, A., Messeri, A., (2005). Clown doctors are a
treatment for preoperative anxiety in children: A randomized prospective study.
Paediatrics, 116, 563-567.
Van Vugt, M., De Cremer, D., & Janssen, D. (2007). Gender differences in
cooperation and competition: The male warrior hypothesis. Psychological
Science, 18, 19-23.
Van Vugt, M., & Van Lange, P. (2006). Psychological adaptations for prosocial
behavior: The altruism puzzle. In M. Schaller, D. Kenrick, & J. Simpson,
Evolution and Social Psychology (pp. 237-261). Psychology Press.
Waller, B. & Dunbar, R.I.M. (2005). Differential behavioral effects of ‘smiling’ and
‘laughing’ in chimpanzees (Pan troglodytes). Ethology, 111, 129-142.
Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief
measures of positive and negative affect. Journal of Personality and Social
Psychology, 54, 1063-1070.
Wild, B., Rodden, F. A., Grodd, W., & Ruch, W. (2003). Neural correlates of laugher
and humor. Brain, 126, 2121-2138.
Zillman, D., Rockwell, S., Schweitzer, K. & Sundar, S. (1993). Does humor facilitate
coping with physical discomfort. Motivation and Emotion, 17, 1-21.
39
LAUGHTER AS SOCIAL LUBRICANT
Footnotes
1There was a three-way interaction between sex, clip and grouping, F(1, 100) =
5.08, p <.03. Inspection of the means suggest that the difference lies in the neutral clip
condition. Males considered the neutral clip particularly unfunny in the strangers
condition (M = 1.00, SD = 0.00), whereas females considered it particularly unfunny
in the acquaintances condition (M = 1.45, SD = 0.89).
2Following the unexpected sex effects of Experiment 1, we ensured that there were
no sessions with three males and one female.
3 We conducted a different set of ANOVAs adding group size (2-6) to the design,
but this factor was not significant for pain tolerance, F(1, 19) = 1.40, p =.27, nor for
laughter F(1, 19) = 1.38, p =.28, and there were no interactions with sex or clip. This
suggests that our measures were not affected by the number of people per session.
40
LAUGHTER AS SOCIAL LUBRICANT
Tables
Table 1. Cooperation as a function of sex, clip, and grouping (Experiment 1)
_____________________________________________________________________
Strangers Acquaintances
_____________________________________________________________________
Neutral
Male 0.92a (.62) 1.73b (.61)
Female 1.08a (.57) 1.16b (.57)
Comedy
Male 1.52b (.68) 1.32b (.76)
Female 1.30b (.59) 1.48b (.64)
_____________________________________________________________________
Notes: cooperation ranges from 0-2.5 GBP; standard deviations are in between
parentheses; means with a different superscript differ from each other; p <.05.
41
LAUGHTER AS SOCIAL LUBRICANT
Table 2. Affect as a function of clip and grouping (Experiment 2)
_____________________________________________________________________
PAS NAS
_____________________________________________________________________
Strangers
Comedy 3.24a (.68) 1.58b (.68)
Neutral 1.89b (.36) 2.65a (.39)
Acquaintances
Comedy 3.01a (.68) 1.48b (.44)
Neutral 2.51a (.95) 1.65b (.48)
_____________________________________________________________________
Notes: scales range from 1 (not at all) to 5 (extremely); standard deviations appear in
between parentheses; means with a different superscript differ from each other; p <.05
(column wise comparison).
42
LAUGHTER AS SOCIAL LUBRICANT
Figure Captions
Figure 1. Cooperation as a function of clip and grouping; Experiment 2
Figure 2. Group identification as a function of clip and grouping; Experiment 2
Figure 3. Relationship between laughter and pain tolerance, r = .50; paindiff is the
difference in pain tolerance between posttest and pretest; Experiment 3
43
Control Group Experimental Group
LAUGHTER AS SOCIAL LUBRICANT
0
0.5
1
1.5
2
2.5
3
3.5
4
Strangers Acquaintances
Cooperation
(in GBP)
Neutral
Come dy
1
1.5
2
2.5
3
3.5
4
4.5
5
Strangers Acquaintance s
Group
identification
Neutral
Comedy
44
LAUGHTER AS SOCIAL LUBRICANT
1
1.5
2
2.5
3
3.5
4
4.5
5
Strangers Acquaintance s
Group
identification
Neutral
Comedy
45
LAUGHTER AS SOCIAL LUBRICANT
60.0050.0040.0030.0020.0010.000.00
laughs
90.00
60.00
30.00
0.00
-30.00
-60.00
-90.00
paindiff
comedy
neutral
Clip
46
