volume no. 2005 the quarterly review - semantic scholarthe quarterly review of biology, december...

Volume 80, No. 4 December 2005

395

The Quarterly Review of Biology, December 2005, Vol. 80, No. 4

Copyright � 2005 by The University of Chicago. All rights reserved.

0033-5770/2005/8004-0001$15.00

The Quarterly Reviewof Biology

THE EVOLUTION AND FUNCTIONS OF LAUGHTER AND HUMOR:A SYNTHETIC APPROACH

Matthew Gervais

Departments of Biology and Anthropology, Binghamton UniversityBinghamton, New York 13902 USA

e-mail: [email protected]

David Sloan Wilson

Departments of Biology and Anthropology, Binghamton UniversityBinghamton, New York 13902 USA

e-mail: [email protected]

keywordslaughter, humor, emotional expression, nonverbal communication,

contagion, mirror neurons, SMA, theory of mind, language, play, joy, multi-level selection, social neuroscience, positive psychology, evolutionary

psychology, Duchenne, coordination

abstractA number of recent hypotheses have attempted to explain the ultimate evolutionary origins of

laughter and humor. However, most of these have lacked breadth in their evolutionary frameworkswhile neglecting the empirical existence of two distinct types of laughter—Duchenne and non-Du-chenne—and the implications of this distinction for the evolution of laughter as a signal. Most ofthese hypotheses have also been proposed in relative isolation of each other and remain disjointed fromthe relevant empirical literature. Here we attempt to remedy these shortcomings through a synthesis ofprevious laughter and humor research followed by (i) a reevaluation of this research in light of theoryand data from several relevant disciplines, and (ii) the proposal of a synthetic evolutionary frameworkthat takes into account phylogeny and history as well as proximate mechanisms and adaptive signif-icance. We consider laughter to have been a preadaptation that was gradually elaborated and co-optedthrough both biological and cultural evolution. We hypothesize that Duchenne laughter became fullyritualized in early hominids between 4 and 2 mya as a medium for playful emotional contagion. Thismechanism would have coupled the emotions of small hominid groups and promoted resource-building

396 Volume 80THE QUARTERLY REVIEW OF BIOLOGY

social play during the fleeting periods of safety and satiation that characterized early bipedal life. Wefurther postulate that a generalized class of nonserious social incongruity would have been a reliableindicator of such safe times and thereby came to be a potent distal elicitor of laughter and playfulemotion. This class of stimuli had its origins in primate social play and was the foundation for formalhuman humor. Within this framework, Duchenne laughter and protohumor were well established inthe hominid biobehavioral repertoire when more cognitively sophisticated traits evolved in the hominidline between 2 mya and the present. The prior existence of laughter and humor allowed them to be co-opted for numerous novel functions, and it is from this process that non-Duchenne laughter and the“dark side” of laughter emerged. This perspective organizes the diversified forms and functions thatcharacterize laughter and humor today and clarifies when and how laughter and humor evolvedduring the course of human evolution.

He who approaches laughter upon science bentwill find it no laughing matter.

(McComas 1923)

Introduction

LAUGHTER AND HUMOR were ac-corded high evolutionary significance

by Darwin (1872) and have received increas-ing attention from biologists and psycholo-gists during the last 30 years. This attentionhas resulted in myriad empirical advancesand has left laughter and humor well char-acterized on multiple proximate levels (seeProvine 2000; Vaid 2002; Bachorowski andOwren 2003; van Hooff and Preuschoft 2003;Wild, Rodden et al. 2003). Laudably, this re-search has spawned a number of hypothesesattempting to explain the ultimate evolution-ary origins of laughter and humor (e.g., Eibl-Eibesfeldt 1989; Weisfeld 1994; Pinker 1997;Ramachandran 1998; Harris 1999; Miller2000; Provine 2000; Owren and Bachorowski2001; Caron 2002; Howe 2002; Jung 2003; Sto-rey 2003). Nevertheless, the scientific study oflaughter and humor is still in its infancy rela-tive to other comparable subjects in emotionsand communication research.

Many empirical questions about laughterand humor remain unanswered or neglected.For example, most researchers (e.g., Provine2000; Owren and Bachorowski 2003; Vettinand Todt 2004) have failed to make the im-portant distinction between Duchenne (stim-ulus-driven and emotionally valenced) andnon-Duchenne (self-generated and emotion-less) laughter (Keltner and Bonanno 1997;see also Wild, Rodden et al. 2003). Whilelaughter has recently been found to occurmost frequently during casual conversationand not following deliberate humor (Provine

1993; LaGreca et al. 1996; Vettin and Todt2004), researchers have yet to questionwhether such conversational laughter is dif-ferent in kind from that following humor.This oversight might well be the root cause ofthe widespread confusion concerning the di-versity of forms and functions that character-izes laughter today (Keltner and Bonanno1997).

As regards theory, the results of empiricalfindings of laughter and humor research re-main disjointed and only partially accountedfor by any one framework. In most cases, suchhypotheses are not mutually exclusive but po-tentially complementary, yet a synthesis re-mains unrealized. As a result, theoretical lim-itations abound. Most hypotheses were notinformed by crucial data in the first place andso leave no room for their integration. Therole of emotion in laughter and humor pro-cesses is an issue that has often been over-looked (e.g., Ramachandran 1998; Jung2003). Another, as mentioned, has been theneglect of a potential multiplicity of laughterkinds (e.g., Provine 2000; Owren and Bacho-rowski 2001, 2003). A different type ofshortcoming has been the utilization of im-poverished evolutionary frameworks. Someresearchers (e.g., Owren and Bachorowski2001) have neglected the multilevel nature ofevolution, while others (e.g., Weisfeld 1994)have failed to specify the evolutionary mech-anisms by which a novel trait emerges in thecontext of already existing traits. A relatedsource of problems follows from the confla-tion of current utility with historical genesis(Gould and Vrba 1982). The adaptive func-tions that have been proposed for laughterand humor range widely and are often con-tradictory (Keltner and Bonanno 1997), yetmany functions might be more properly char-

December 2005 397EVOLUTION OF LAUGHTER AND HUMOR

acterized as effects, or the functions of a traitthat are different from those for which thetrait was originally selected (Gould and Vrba1982). Laughter and humor may well haveevolved for one function while being later co-opted for others.

Laughter and humor researchers havelargely failed to interact with outside researchprograms, depriving both of potential in-sights. Such programs include social neuro-science (e.g., Rizzolatti and Craighero 2004),concerned recently with the neural bases ofgesture and expression initiation and percep-tion; positive psychology (e.g., Seligman andCsikszentmihalyi 2000), which emphasizesthe sources and facilitation of human well-be-ing and not merely the mitigation of dysfunc-tion; evolutionary psychology (e.g., Cosmidesand Tooby 2000), which provides a theoreti-cal toolbox for analyzing the evolution of hu-man behavior and emotion; and multilevel se-lection theory (e.g., Wilson and Sober 1994),which maintains that natural selection canoperate at multiple levels of the biological hi-erarchy. Laughter and humor have also beenomitted from most general accounts of hu-man evolution, taking a backseat to evolution-ary divas like bipedalism, encephalization,language, and culture.

Thus a scientific understanding of laughterand humor has been hindered not only byempirical neglect and theoretical incom-pleteness but also by isolation from otherrelevant fields. Yet many plausible, supported,and even compelling ideas have been put for-ward and are integrated later in this review,and it is not our aim to deconstruct argu-ments or undermine past research. Our aimis instead to construct, to bridge those exist-ing islands through a synthesis of previous re-search and theorizing in an attempt to inte-grate multiple perspectives and broaden ourunderstanding of laughter phenomena.Methodologically, we utilize a balanced evo-lutionary perspective and multiple levels ofanalysis, while explicitly incorporating recentresearch from several previously disjointedscientific disciplines that lend insight to thesynthesis at hand. In doing so, we present aninherently interdisciplinary evolutionary hy-pothesis that accounts for the myriad typesand functions of laughter and humor as we

observe them today. While far from the finalword, this is an effort to provide a first ap-proximation of a comprehensive evolution-ary framework within which future laughterand humor research and theorizing can beconducted.

We begin by reviewing the relevant empir-ical data on laughter and humor, from whichwe distill a number of key points. We nextreview relevant theory and data from work onmirror neurons, positive emotions, animalplay, evolutionary psychology, and multilevelselection theory. This serves as the founda-tion for a two-part argument about the evo-lutionary trajectory traveled by laughter andhumor over the last 7 million years. The firstpart details the evolutionary elaboration ofhuman laughter and humor from ape ante-cedents between 4 and 2 mya. In the secondpart, we argue that laughter and humor wereco-opted post-2 mya by a battery of evolvinghominid traits, such as volitional oral-facialmuscle control and language, resulting in themyriad forms and functions of laughterfound today. Predictions and future direc-tions are discussed last.

An Empirical Characterization ofLaughter and Humor

biological foundationsLaughter is a universal human behavior

found in all cultures and virtually all individ-uals the world over (Apte 1985; Lefcourt2000; Provine 2000). Developmentally, laugh-ter is one of the first social vocalizations emit-ted by human infants (Deacon 1997) andemerges spontaneously at anywhere betweentwo and six months of age (Darwin 1872;McComas 1923; Sroufe and Waters 1976; Fo-gel et al. 1997; Scheiner et al. 2002). Evencongenitally deaf and blind children havebeen reported to laugh despite never havingperceived and learned from the laughter ofothers (Provine 2000). Although recent re-search has revealed a great deal of acousticvariation in the form of laughter (Bacho-rowski and Owren 2001; Bachorowski andSmoski 2001; Kipper and Todt 2001; Vettinand Todt 2004), few deny that laughter is dis-tinctive and universally recognizable (Ed-monson 1983; Eibl-Eibesfeldt 1989; Deacon


1997; Provine 2000; Kipper and Todt 2001;van Hooff and Preuschoft 2003). Experi-ments have shown that a wide range of repe-tition intervals are unambiguously perceivedas laughter (Kipper and Todt 2001; Vettin andTodt 2004). Laughter thus has a characteris-tic, stereotyped form (Grammer and Eibl-Eibesfeldt 1990; Provine 1991) and can evenbe considered a ritualized, species-specific,fixed- action pattern (Grammer and Eibl-Eibesfeldt 1990). Taken together, these factssuggest that every normal human being isstrongly genetically predisposed to developthe ability to produce and perceive laughter(Darwin 1872; McGhee 1979; Weisfeld 1994;Provine 2000; Caron 2002).

Comparative and phylogenetic analysesprovide additional support for consideringlaughter an evolved instinct. Within the con-text of social play, numerous primate speciesexhibit a phylogenetically conserved facial ex-pression, the relaxed open-mouth or “play”face (Preuschoft and van Hooff 1997). In thegreat apes, particularly Pan, this face is oftenaccompanied by a pant- like vocalization dur-ing tickling and chasing (Darwin 1872; Fry1994; Provine 2000; Gamble 2001; Caron2002). Such “play panting” is instrumental infacilitating continued playful interaction(Flack et al. 2004; Matsusaka 2004) and ismore salient to chimpanzees than is the con-comitant facial expression (Parr 2004). It hasbeen convincingly argued that human laugh-ter is homologous to this great ape play pantcorrelate of the primate play display (vanHooff 1972; Provine 2000; van Hooff andPreuschoft 2003; Matsusaka 2004). The evi-dence therefore suggests that a rudimentaryprecursor to human laughter has distant evo-lutionary origins in the common primate an-cestor of humans and the other great apes atleast 6.5 mya (Fry 1994).

The evolved basis of laughter is furtherhighlighted by the cross-cultural and cross-species similarities in the stimuli that elicitDuchenne laughter. Formal humor is ofcourse the most familiar laughter stimulus,and theorists have debated its nature for atleast 2,000 years (Provine 2000). More re-search has been devoted to adult humor thanto any other type of laughter stimulus (Prov-ine 1996), and what emerges from the liter-

ature is something of a consensus that incon-gruity and unexpectedness underlie almostall instances of formal laughter-evoking hu-mor. This is insofar as the perceived inconsis-tency between one’s current and past expe-riences involves both a nonserious or playfulframe and an alternate type of intelligibility,that is, a meaningful interpretation of somestimulus or event that is different from thatwhich was initially assumed (Shultz 1976;McGhee 1977, 1979; Apte 1985; Weisfeld1994; Ramachandran 1998; Lefcourt 2000;Gamble 2001). These prerequisites providethe basis for a remarkably wide range of hu-mor-related laughter stimuli, such as jokes,puns, and slap-stick mishaps.

Although adult humor has received a dis-proportionate amount of attention relative toother laughter phenomena (e.g., infantlaughter, tickling, and rough-and tumbleplay), there is an undeniable relationship—acontinuity of form, structure and context—that connects all Duchenne laughter-induc-ing stimuli (Provine 2000). Human infantsspontaneously laugh when they perceive un-expected stimuli in nonserious contexts(McGhee 1976; Sroufe and Waters 1976), forinstance the peek-a-boo faces of their moth-ers, specifically if those stimuli are novel rela-tive to the infants’ developing mentalschema. The same stimuli that elicit laughterin infants can evoke crying if the situation isnot perceived as safe or if the incongruity istoo intense (Sroufe and Waters 1976). Tick-ling likewise involves the unpredictable stimu-lation of vulnerable areas of the body by an-other person (as we are unable to tickleourselves; Weisfeld 1994; Harris 1999; Prov-ine 2000), and must come across as unthrea-tening and playful or else actual attack or ag-gression is perceived (Provine 2000). Ticklingin conjunction with mock aggression andchasing is the prime elicitor of ape laughter(Fry 1994; Provine 2000; Gamble 2001) aswell as laughter in human infants (Sroufe andWaters 1976) and children (Harris 1999;Provine 2000). Social contexts facilitatelaughter (Devereux and Ginsburg 2001), andlaughter is 30-times more likely to occur in asocial context than when alone (Provine andFischer 1989). It has even been argued that


laughter requires another person to stimulateit (Howe 2002; Jung 2003).

There is thus an intuitive family resem-blance among the different proximate causesof Duchenne laughter, such that they can becharacterized as sharing a single form orstructure: a sudden unexpected change inevents that is perceived to be at once not se-rious and in a social context—that is, nonser-ious social incongruity. We will hereafter usethis phrase to refer to the essential and at rootsingular stimulus that elicits Duchenne laugh-ter in apes and humans. Our use of the termis not meant to constitute a formal definitionof humor or of laughter stimuli, but is insteadmeant to serve two functions: (1) to simplifythe discussion by meaningfully categorizingthe stimuli of Duchenne laughter without be-coming swamped by the philosophical quag-mire of what is “humor;” and (2) to give thisreferential term some descriptive power. Inregards to the first point, we still use forms ofthe term “humor,” especially where this wasthe term used by researchers or theorists intheir original studies. “Humor” is generallyused for formal attempts at eliciting laughterin adults, while “protohumor” refers to otherforms of nonserious social incongruity, suchas rough-and-tumble play, tickling, physicalmishaps, and pleasant surprise in apes andinfants. In relation to the second point, we donot mean to imply that any nonserious socialincongruity is a necessary and/or sufficientcause of laughter, but we feel that the termnonetheless works to convey their evolution-ary continuity.

This underlying connectedness among theproximate causes of Duchenne laughter hasnot gone unnoticed. Darwin (1872) likenedhumor to “tickling of the mind,” while Mc-Ghee (1979:103) called humor the “logicalresult of an extension of playful forms of be-havior to the more abstract intellectualsphere of ideas.” Gamble (2001) referred tothe laughter-evoking stimuli of infants as“primitive forms of humor,” while arguingthat the laughter-eliciting stimuli of adultsigning apes, most notably Koko the gorilla,has essentially the same incongruity-basedstructure as human humor (see also McGhee1979). Provine (2000) in turn referred to tick-ling as “protohumor,” while Weisfeld

(1994:151) stated that “tickling is a rathersimple, rudimentary form of humor.” Re-search has even shown a significant correla-tion between individuals who laugh whentickled and those who are prone to laugh atand use humor (Fridlund and Loftis 1990),suggesting indirectly that both are triggeredby the same mechanism.

The claim that all elicitors of Duchennelaughter share a similar structure finds sup-port in the suggestion that there is a “laugh-ter-coordinating” center in the dorsal upperpons that serves as the common ground forthe different types of Duchenne laughter(Wild, Rodden et al. 2003). This assertion waspresaged by Provine (1996), who noted thatdifferent forms of laughter- inducing stimuliappear to access a single “laugh generator”despite the fact that they are perceivedthrough different modalities. Taken together,the laughter of nonhuman primates, thespontaneous laughter of human infants, tick-ling, and formal adult humor all share whatis essentially a phylogenetically and ontoge-netically-conserved structure and context(Darwin 1872; McGhee 1979; Provine 2000),here referred to as nonserious social incon-gruity.

the role of culture and learningDespite the evident biological foundation

of laughter and humor, there is little doubtthat laughter-evoking contexts and howlaughter is interpreted are fundamentally in-fluenced by cultural norms and learning.Goodson (2003) has called this “learning(cultural) overlay,” or the codetermination ofhuman actions by biology and culture (seeMaryanski et al. 1997), and noted how it canconfound analyses of the evolutionary foun-dations of behavior, including laughter. Burl-ing (1993) has also argued that no humanbehavior is impervious to learning and thatthe human “gesture-call system,” of whichlaughter is a component, is no exception.The impact of learning on human laughterand humor is perhaps best illustrated by thecultural variation that exists not in the struc-ture or “theme” (Ekman 1999) of the elicitorsof laughter but in the subject matter that ful-fills such structure (Apte 1985; Lefcourt


2000). These specifics can vary from toilet-and sex-based humor to political humor andspan a wide spectrum of cultural institutionsand customs. The contexts, frequency, and in-tensity of laughter are also variable in differ-ent cultures, which indicates that laughtercan be inhibited and elicited in accordancewith display rules and varying norms and cus-toms (Apte 1985; Weisfeld 1994).

The role of learning in laughter produc-tion is also apparent in the ways that laughtertypically occurs in everyday life. Contrary tothe idea that laughter is always a reaction tononserious social incongruity, studies haveshown that the speaker in a conversational in-teraction produces up to 46% more laughterthan his or her listeners (Provine 1993; La-Greca et al. 1996; Vettin and Todt 2004). Suchlaughter appears spontaneous and functionsto smooth interactions and promote positivefeelings between interlocutors (Provine 1996;Vettin and Todt 2004). Similar laughter canbe expressed simply in response to an aware-ness of stress (i.e., nervous laughter) and isemployed to signal that anxiety is being ex-perienced (Keltner and Bonanno 1997). Theprevalence of conversational laughter amongstrangers suggests that norms (or nervous-ness) play a role in leading people to utilizesuch laughter to avoid misunderstandings orto facilitate friendliness (Devereux and Gins-berg 2001; Vettin and Todt 2004). These in-stances of laughter are all arguably examplesof strategic non-Duchenne laughter, which isin effect a learned facsimile of Duchennelaughter, and is thus different in kind from it.As Provine (1993) has noted, such conversa-tional laughter is characterized by a “punc-tuation effect,” where laughter never inter-rupts speech but is used instead to“punctuate” statements. This phenomenon isbest explained by positing that such conver-sational laughter is being used strategicallylike speech as a metacommunicative marker,as opposed to being “uncontrollable” like Du-chenne laughter. An exception to this, wherenon-Duchenne laughter does co-occur withspeech, is the learned hybrid of laughter andspeech dubbed “laugh-speak” that is utilizedby talk show hosts and salespeople to influ-ence the attitudes and behaviors of others(Provine 1996).

Neuropsychological studies support the po-sition that laughter can sometimes be con-trolled by learning despite having ancientphylogenetic roots, as researchers have re-vealed the existence of two partially dissocia-ble neural pathways underlying two differenttypes of laughter (for a review, see Wild, Rod-den et al. 2003). One is emotionally-drivenand involuntary, arising in subcortical, limbic,and brainstem areas and culminating in a“laughter-coordinating” center in the dorsalupper pons; the other is a voluntary motorpattern that originates in frontal premotor ar-eas and directly influences the motor cortex.The most salient evidence for this dissocia-tion comes from two types of facial paralysis:emotional facial paresis, in which subjectshave no spontaneous laughter response tohumorous stimulation yet can volitionallymimic laughter on command; and volitionalfacial paresis, in which subjects have an intactautomatic laughter response to nonserioussocial incongruity yet cannot consciouslymimic such laughter (Wild, Rodden et al.2003). These findings have been supportedby brain imaging data showing that sponta-neous stimulus-induced laughter and voli-tional laughter do indeed arise from separateneural systems (Iwase et al. 2002). The exis-tence of two types of laughter, one involun-tarily elicited and one consciously controlled,is paralleled by the existence of two separatepathways for spontaneous and volitional fa-cial expressions (Gazzaniga and Smylie 1990;Maryanski et al. 1997) and by the observedbehavioral and functional difference betweenDuchenne and non- Duchenne laughter(Keltner and Bonanno 1997).

While neuropsychological studies revealtwo types of laughter, one spontaneous andone volitional, some laughter that appearsspontaneous (i.e., occurring in the flow ofconversation) is in fact not. We agree withKeltner and Bonanno (1997) and Wild, Rod-den et al. (2003) that there are two types oflaughter and suggest that spontaneous non-Duchenne laughter is simply volitional laugh-ter that has achieved a degree of automaticity,and so appears spontaneous. Duchenne andnon-Duchenne laughter should not be distin-guished in terms of spontaneity but insteadin terms of antecedent stimuli, emotional va-


lence, or physical characteristics. Vettin andTodt (2004) recently discovered that thephysical characteristics of conversational(non-Duchenne) and humor-driven (Du-chenne) laughter differ insofar as conversa-tional laughter consists of fewer elements andhas a lower average fundamental frequency.The simpler physical characteristics of non-Duchenne laughter might be indicative of thelimited ability of people to volitionally utilizelaughter. The automaticity of this non-Du-chenne laughter is further suggested by Vet-tin and Todt’s (2004) claim that people un-derreport the frequency of theirconversational laughter because they are notaware of how often they utilize it. Non-Du-chenne laughter here is akin to other learnedcomplex motor acts that become awkwardand unnatural the more we think about thembut that we can nonetheless become quiteadept at through practice and the condition-ing of unconscious motor programs.

McComas (1923:50) was one of the firsttheorists to acknowledge that laughter can besubjected to learning and utilized for multi-ple ends. He said, Certainly the pleasureswhich grow out of the sensations, activitiesand instincts cannot account for all forms oflaughter. They have no part in the laughterof derision, for example, or the laughter ofembarrassment. Such laughter is an artifice.It is allied to language. It is not a natural out-come of a pleasurable experience. It is rathera means of conveying an idea, a form of state-ment.

After drawing this distinction, however,McComas (1923:50) noted that “it is not easyto say where spontaneous laughter ends andcontrolled laughter begins.” This sentiment isreflected in our characterization of sponta-neous non-Duchenne (or conversational)laughter, and leads us to insert here an im-portant caveat. Terms such as “spontaneous,”“automatic,” “involuntary,” and “noncon-scious” have often been used interchange-ably, while “intentional,” “volitional,” “volun-tary,” and “conscious” have been used as theirantonyms. “Strategic” behaviors are usuallycharacterized by the latter set of terms. Spon-taneous non-Duchenne laughter presents aproblem because it exists in varying degreesacross this conceptual divide, being strategic

and learned but largely beyond consciouscontrol. Our caveat is that empirical studiesof laughter have either not been looking todistinguish Duchenne from non-Duchenne(e.g., Vettin and Todt 2004), or have not yetachieved a grain fine enough to detect thedifference (e.g., Provine 1993). Even a recentdiscussion of spontaneous and pseudo-spon-taneous nonverbal behavior (Buck andVanLear 2002) ignored the potential autom-atization of pseudo-spontaneous nonverbalcommunication, consistently referring to it as“intentional.” While it is beyond the scope ofthis paper to untangle the empirical natureof the problem, we feel that this issue will be-come central to future discussions of laughterand nonverbal behavior, especially as a morerealistic, neurobiologically- grounded pictureemerges of the interplay of genes and expe-rience in shaping human behavior. To be asclear as possible, we will hereafter use “non-Duchenne” to refer to spontaneous conver-sational laughter that occurs in the absenceof attempts at humor.

emotion, contagion, andmultifunctionality

The distinction between Duchenne andnon-Duchenne laughter highlights a key fea-ture of Duchenne laughter: its inherent linkwith emotional experience. The emotionalconcomitant of Duchenne laughter has vari-ously been called mirth or amusement (seeMcGhee 1979), joy (Darwin 1872; Panksepp2000; Panksepp and Burgdorf 2003), exhila-ration (Ruch 1993), or positive affect (Bach-orowski and Owren 2001), while the commontheme has been the decidedly positive feelingwith which it coincides. Duchenne laughterappears to be intrinsically linked to the an-cient brain circuit that underlies mammalianrough-and-tumble play, which is mediated atleast in part by pain-reducing opioids (Pank-sepp 1998). This should come as no surprisesince Duchenne laughter was already de-scribed as having been derived from primateplay signals (van Hooff 1972; Provine 2000).The rewarding neurochemical cocktail un-derlying Duchenne laughter has not beenfully described (Panksepp 1998), but thequick positive conditioning of infants to their


mothers and other individuals suggests its ex-istence (Panksepp and Burgdorf 2003). Brainimaging studies have revealed that humor ap-preciation—Duchenne laughter’s phenome-nological correlate—does involve affective re-ward networks in the brain (Goel and Dolan2001; Shibata and Zhong 2001; Iwase et al.2002; Moran et al. 2004). Behavioral studieshave also found humor and laughter associ-ated with increased positive affect and im-proved mood (Neuhoff and Schaefer 2002)and reduced negative affective responses tostressful events (Keltner and Bonanno 1997;Kuiper and Martin 1998; Fuhr 2002), whilenumerous clinical fields have reported signifi-cant health benefits derived from laughter(see Rosner 2002).

Owing to these associations, laughter hastraditionally been conceptualized as a signaldesigned to communicate the subjective emo-tional state of a sender to receivers, such thatthe receivers cognitively decode the infor-mation of the signal and infer the originalmental state of the sender (Russell et al.2003). Within such a paradigm the meaningof laughter has typically been labeled some-thing comparable to “this is play” or “that wasplay” (see van Hooff 1972; Grammer 1990),which has parallels in the “play signals” ofother species (e.g., Bekoff 1995; Pellis andPellis 1996; Kipper and Todt 2002; Pankseppand Burgdorf 2003). Recently, however, someresearchers have reconceptualized animalcommunication and emotional communica-tion in humans (Rendall and Owren 2002;Bachorowski and Owren 2003; Owren andBachorowski 2003). An affect-induction ap-proach has been proposed (Owren and Bach-orowski 2003) that abandons the traditionalidea that communication carries encoded in-formation (see Hauser 1996) and instead em-phasizes the emotional and behavioral im-pact that expressions have on those whoperceive them (Krebs and Dawkins 1984;Bachorowski and Owren 2003). Laughter hasbeen at the heart of this proposal. Laughterproduction is still tied to some degree of emo-tional arousal (Bachorowski and Owren2003), but laughter’s function stems not fromthe communication of emotional states as in-formation but instead from the accentuationor induction of positive affect in others. This

occurs either directly by laughter’s acousticproperties or indirectly through conditionedexperience (Owren and Bachorowski 2003).This conception of laughter finds support instudies showing that heard “voiced” (or song-like) laughter elicits positive affect in others(in contrast to “unvoiced” or grunt-like laugh-ter; Bachorowski and Owren 2001; see Bach-orowski and Owren 2003) and from a fMRIstudy that found the amygdala, a key emotionand stimulus relevance area in the brain, ac-tivated by the perception of laughter (Sanderand Scheich 2001).

In addition to inducing positive affect inothers, Duchenne laughter is contagious(Provine 1992); heard laughter is a sufficientstimulus to elicit laughter. With this property,laughter falls within a larger category of prim-itive emotional contagion (Hatfield et al.1994) or the phenomena by which the actionsand emotions of interlocutors are coupled.The more general underlying tendency of in-teraction partners to adopt the same pos-tures, gestures, and mannerisms has recentlybeen dubbed the “chameleon effect” (Char-trand and Bargh 1999). These processes ofbiobehavioral coupling are predictive of thefinding that group size has a facilitative effecton laughter responses to humor (Chapman1983), with increases in group size increasingaudience laughter levels in response to hu-morous material (Levy and Fenley 1979;Butcher and Whissell 1984). The contagious-ness of laughter has therefore been impli-cated in coupling the emotions and behaviorsof individuals within a group (Provine 1992;Deacon 1997). This possibly explains thefinding that laughter, as a function of humor,increases cohesiveness and cooperationwithin goal- oriented groups (Banning 1987;Vinton 1989; Greatbatch and Clark 2003), aconclusion consistent with evidence of the co-operative value of the chameleon effect(Lakin et al. 2003) and, more generally, thespread of affect through a group (Spoor andKelly 2004).

Laughter today functions in so many waysand contexts that its multifunctionality hasbeen deemed “awe inspiring” (Fry 1977).Many of these putative functions remain spec-ulative and will require additional research tobe established or falsified. As already men-


tioned, Duchenne laughter has been foundto increase positive affect and improve mood(Neuhoff and Schaefer 2002) and mitigatenegative affective responses to stressful events(Keltner and Bonanno 1997; Kuiper and Mar-tin 1998; Fuhr 2002), while promoting myriadother health benefits (see Rosner 2002).Laughter can similarly be used as a defensemechanism that trivializes stressful or psycho-logically overly demanding circumstances(Panksepp 2000). This is perhaps best rec-ognized as nervous laughter, or a type of non-Duchenne laughter (Keltner and Bonanno1997), and has been subsumed into a general“laughter-as-dissociation” hypothesis (Keltnerand Bonanno 1997). Duchenne laughter andits associated positive affect also enable in-fants to maintain interactions with novel,mildly-stressful stimulation so that they cancultivate world knowledge and develop socialcompetence (McGhee 1976; Sroufe and Wa-ters 1976). During conversational interaction,laughter acts as a social lubricant to lightenthe mood and can make listeners more re-ceptive, as well as increase in-group feelings(Provine 1993, 2000). Laughter can likewisefunction to manipulate the emotions of oth-ers to the benefit of the laugher (Owren andBachorowski 2003) and mitigate problems orsocial ambivalence within a group (Caron2002), particularly by dampening friction andviolent competition between individuals andestablishing dominant and subordinate rela-tionships (appeasement laughter, Fry 1977;superiority laughter, van Hooff 1972). Laugh-ter can promote the integration of new indi-viduals into an already-present group struc-ture (Gamble 2001), but can also play a rolein delineating in- group and out-groupboundaries (laughter’s “dark side;” Panksepp2000) by establishing “exclusionary groupidentities” and by being directly aggressive to-wards members of the out- group (Eibl-Eibesfeldt 1989; Pinker 1997). Laughter canalso function to recruit playful interactionand fitness-enhancing stimulation (Weisfeld1994). Lastly, as discussed, the contagious na-ture of laughter can function to couple theemotions and coordinate behaviors of the in-dividuals within a group (Provine 1992; Dea-con 1997).

Key Empirical PointsA number of key propositions can be dis-

tilled from the above discussion that providethe basis for our synthesis. (1) There are twodistinct types of laughter, each with a partiallydissociable neural pathway. Duchenne laugh-ter is stimulus-driven and emotionally val-enced, originating (save for initial sensoryanalysis) in subcortical and brainstem re-gions. Non-Duchenne laughter is uncon-nected (except perhaps via facial feedback)to emotional experience and originates inprefrontal premotor and motor areas. Onecan learn to automatically employ non- Du-chenne laughter for various functions, whichmanifest themselves in conversational laugh-ter, nervous laughter, appeasement laughter,and aggressive or derisive laughter, amongothers. (2) Duchenne laughter is geneticallypredisposed to develop. Every human devel-ops the laughter “program” and the signalconstitutes a ritualized, species-specific fixed-action pattern. (3) Duchenne laughter is de-rived from the primate relaxed open-mouthdisplay (“play face”) and its associated pant-like vocalization, and so has homologues inthe other great apes. Rudimentary laughterevolved originally at least 6.5 mya, before thehominid line split from the other hominoids,and most likely resembled the pant-likelaughter of chimpanzees. (4) The major elic-itors of Duchenne laughter—safe surprise ininfants, tickling and play in apes and humanchildren, and incongruity-based conceptualhumor in human adults and signing apes—all share a common structural and contextualground (nonserious social incongruity) bywhich they each access a common “laugh gen-erator” in the brainstem. This is despite theirdifferent sensory modalities and the variationin initial processing that must occur to senseor understand them. (5) Human laughter, de-spite its hardwired basis, is amenable to cul-tural overlay and learning and is thus code-termined by biology and culture. Not only docultural norms determine the subject matterof humor and the contexts in which laughteris appropriate (laughter can be inhibited),but individuals can also learn to strategicallyemploy non-Duchenne laughter in conversa-tion. (6) Duchenne laughter is inherently


linked to positive emotional experience. Feel-ings of joy or playful mirth occur concomi-tantly with such laughter, and this playful neu-robiological cocktail functions not only toimprove positive affect and key aspects ofhealth but also to mitigate negative affect andstressful situations. (7) Laughter, especially ofthe Duchenne type, has a congruent emo-tional effect on others and appears to be amedium for emotional contagion. Senderlaughter induces or enhances positive affectin receivers while simultaneously elicitinglaughter itself. In this way, laughter is conta-gious. (8) Laughter serves myriad functionsin modern human society. These range fromfacilitating emotional stability and health,curtailing negative emotion and stress, andpromoting socialization, to lubricating socialinteraction, easing tensions and competition,delineating and maintaining group identities,and coordinating the emotions and behaviorsof a group. Laughter thus functions on mul-tiple levels, from individual physiology andpsychology to dyadic and group-based inter-actions.

Integrating Other FieldsThe above characterization of laughter can

be further developed by integrating findingsand ideas from several scientific disciplinesthat have hitherto remained largely isolatedfrom laughter research. This interdisciplin-ary approach provides insights to the func-tionality and evolution of laughter and dem-onstrates the possibilities of integrating thebiological and social sciences (Maryanski etal. 1997; Rozin et al. 1997) and even the in-tegration of knowledge as a whole (Wilson1998). The following discussion concentrateson Duchenne laughter, as only it was involvedin the initial elaboration of laughter and hu-mor in hominids. Non-Duchenne laughterbecomes important in the later discussion oflaughter and humor as co-opted preadapta-tions.

the view from social neuroscienceRecent research in cognitive neuroscience

provides a mechanistic framework for under-standing the contagiousness and affect-induc-ing properties of laughter, and hence the evo-

lution of laughter as a signal. While thisframework has been mostly applied to the vi-sual perception of action, it is equally relevantto vocal signals like speech (Rizzolatti and Ar-bib 1998; Stamenov and Gallese 2002; Wat-kins et al. 2003; Wilson et al. 2004) and toother aspects of auditory perception (Kohleret al. 2002; see Rizzolatti and Craighero2004).

Studies with nonhuman primates have re-vealed the existence of now-famous mirrorneurons, a class of visuomotor neurons thathave the remarkable characteristic of beingactive not only during particular actions butalso during the perception of those actions asperformed by others (Rizzolatti et al. 2002;Rizzolatti and Craighero 2004). Mirror neu-rons were first discovered in monkey premo-tor cortices by researchers utilizing single-cellrecording techniques and have been found tobe involved in many actions including reach-ing and grasping as well as ingestive and com-municative mouth actions like lip smacking(Ferrari et al. 2003).

This work has led to a search for mirrorneurons, or mirror matching systems, withinthe human brain. Imaging technologies, likePET and fMRI, and neurophysiological tech-nologies, such as EEG, MEG, and TMS, havebeen utilized to uncover a wider range of mir-ror matching systems in humans than single-cell recording techniques were able to revealin monkeys (Rizzolatti et al. 2002). Humanpremotor areas for hands, feet, and mouth allappear to be at least partly constituted by neu-rons that are active during action, the mentalrehearsal of that action, the observation of asimilar action by others, and the imitation ofthat action. Importantly, the mirror system inhumans responds to the perception of intran-sitive actions and not only object-oriented ac-tions, as is the case in monkeys (Rizzolatti andCraighero 2004).

Even more recent brain imaging researchon humans has tentatively revealed a neuralnetwork via which the perception of others’emotional expressions leads to covert behav-ioral synchrony and congruent emotion (orprimitive emotional contagion; Hatfield et al.1994). Although empathy has received morediscussion than emotional contagion per se,the results of several studies (Carr et al. 2003;


Wild, Erb et al. 2003; Wicker et al. 2003; Les-lie et al. 2004) have suggested that perceivedemotional expressions are first matched to areceiver’s own representations of those ex-pressions in their premotor cortices. This ac-tivity in premotor and connected motor mir-ror systems leads to emotional contagion viaactivity in limbic regions that is congruentwith the emotion corresponding to the orig-inally perceived emotional expression. Thecrucial pathway is proposed to be the insularcortex (Carr et al. 2003), which potentiallyconnects the mirror circuits to emotion cen-ters by means of the limbic system. Generally,this process of emotional contagion throughmotor matching is best viewed as a covertsimulation process (Goldman and Sripada2005).

On a theoretical level, this research has ledto a number of hypotheses about the functionof mirror matching circuits. It is thought thatmirror neurons function to facilitate an un-derstanding of the actions of others, as theycreate a direct link between senders and re-ceivers and allow the simulation of perceivedactions through extant representations ofthose actions in the observer (Rizzolatti andCraighero 2004). The “shared manifold of in-tersubjectivity” hypothesis proposed by Gal-lese (2003) views mirror neurons as the con-stituent elements of a system for the empathicunderstanding of others. This hypothesis pos-its that mirror-matching systems are a basicorganizational principle of the human brainand human society, creating an intersubjec-tive social space where individuals can expe-rience and appreciate the movements, sen-sations, and emotions of others.

As discussed, Duchenne laughter co-occurswith positive affect, and its perception elicitspositive affect and even laughter in receivers.Laughter has even been explicitly acknowl-edged in discussions of primitive emotionalcontagion (e.g., Hatfield et al. 1994), al-though only in passing. These qualities makelaughter highly amenable to explorationwithin a mirror system paradigm. Yet surpris-ingly, laughter has received little attentionfrom mirror system researchers. However,laughter has been characterized as a “reso-nance behavior of the first type,” or a behav-ior that is prone to be immediately repro-

duced by others because its perceptiondirectly activates neurons that generate mo-tor movements identical to those perceived(Rizzolatti et al. 1999).

This proposal has received hitherto unrec-ognized support from other neuroscientificstudies. A somatotopically organized area ofthe brain, Brodmann’s Area 6, or the supple-mentary motor area (SMA), has been impli-cated in all the requisite aspects of laughterphenomena that would be predicted from amirror system account. The SMA is active dur-ing: (i) laughter production, both in Du-chenne and non-Duchenne laughter (Iwaseet al. 2002); (ii) imagined or internal laugh-ter (Shibata and Zhong 2001; Osaka et al.2003); (iii) laughter perception, although tovarying degrees (K Sander personal commu-nication February 23, 2005; D Shibata, per-sonal communication November 16, 2004);and (iv) humor appreciation (Iwase et al.2002: Mobbs et al. 2003). The human SMAhas been implicated in some more generalaccounts of mirror system activity involvingthe observation of grasping movements (seeRizzolatti et al. 1999), the specular (left/rightreversed) imitation of hand movements (Ko-ski et al. 2003), and facial expression obser-vation and imitation (Carr et al. 2003; Leslieet al. 2004). In these studies, there was in-creased SMA activation during imitation,which parallels the results with laughter (Shi-bata and Zhong 2001). Most significantly,electrical stimulation of points in the SMAhas been found to elicit Duchenne laughter(Fried et al. 1998). The authors noted thatthis same area of the SMA is associated withspeech initiation and stuttering, phenomenaalready being discussed in terms of mirror sys-tem activity (Kalinowski and Saltuklaroglu2003). Supporting the proposal of Carr et al.(2003), the insula and amygdala have beenfound to be active during laughter percep-tion (Sander and Scheich 2001) and humorappreciation (Goel and Dolan 2001; Mobbset al. 2003; Moran et al. 2004).

Together this evidence suggests that theperception of laughter involves a mirror cir-cuit centralized in the SMA, whose activityfeeds into limbic areas via the insula, and inso doing elicits positive affect congruent withthe perceived laughter. Such a mechanism


could be considered within the establishedframework of a perception-action model (seePreston and de Waal 2002) and would for-malize the hypothetical laughter contagionmechanisms that Provine (1996) referred toas a “laugh detector” and a “laugh generator,”although in this framework they would essen-tially depend on a more general and singularmirror mechanism.

Laughter and yawning have traditionallybeen seen as embodying the best examples ofcontagious behaviors in humans (Provine2000). While this might well be the case, wethink it is worth distinguishing them. The keydifference is that laughter evolved in relationto a mirror system as a signal, whereas yawn-ing did not. Laughter is unique to the greatapes and uniquely elaborated in humans.While Panksepp and colleagues (e.g., Pank-sepp and Burgdorf 2003) have suggested thatrats laugh as well, rat laughter is surely anal-ogous and not homologous to ape laughteras the many mammalian species intermediateto rats and apes have never been documentedto laugh. Yawning, in contrast, is an ancientreflex ubiquitous to vertebrates that evolvedto regulate individual arousal irrespective ofsocial contexts (Baenninger 1997). Yawningis only contagious in great apes (Anderson etal. 2004) and humans (Provine 1986) andonly insofar as it interacts with novel ape cog-nitive abilities. The point of this contrast isthe assertion that the neural bases of laughterand yawn contagion (e.g. Platek et al. 2005;Schurmann et al. 2005) should not necessar-ily be the same since laughter is a more com-plex and variable signal that evolved in rela-tion to a mirror system and is thus muchmore likely to involve such a system than isyawning. The two behaviors deserve to beevaluated on their own terms and only thencompared.

Regardless of the details, the most impor-tant implication of mirror system research forlaughter is that the neural network involvedin the production of a given action is also ac-tivated during the perception of that actionin others. This idea suggests an inherent linkbetween senders and receivers such as thatalready found in laughter. Significantly, it hasbeen shown that only Duchenne laughter hasthe capacity to undo negative emotions and

stress (Keltner and Bonanno 1997), whileonly “voiced,” or fully vocalized, laughter(Bachorowski and Owren 2001), that is emo-tionally-driven (Keltner and Bonanno 1997),has the capacity to elicit positive affect in oth-ers. Also, Smoski and Bachorowski (2003b)showed that those laughs that are moststrongly associated with the induction of posi-tive affect are also most facilitative of antiph-onal (or laughter-elicited) laughter. Gram-mer and Eibl-Eibesfeldt (1990) also foundthat only voiced laughs were correlated withan interest in seeing one’s experimental part-ner again, indicating the association of suchlaughs with positive affect. This is logical fromthe perspective of mirror circuits, as only theperception of Duchenne laughter would havebeen matched by the activation of motor pat-terns that influence emotion pathways.

If Gallese (2003) and colleagues are cor-rect that there is a shared manifold, in whichthe same system is used to perceive and senda signal, then laughter should be viewed as aconstituent part of such a manifold, and thefunctions of laughter should be reexaminedin light of the manifold’s overarching func-tion(s). We want to emphasize here a putativefunction for laughter that others have sug-gested—namely, coupling the emotions ofgroup members and coordinating their ac-tions (Provine 1992; Deacon 1997). Deacon(1997:429) referred to laughter and crying as“vehicles for coordinating the emotions of asocial group” and considered them “a sort ofpreestablished harmony,” while Provine(1992:3) asserted that “ . . . contagious laugh-ter . . . is an example of a social coupling pro-cess that synchronizes the biological and be-havioral state of a group.” The mirror-systemperspective independently suggests such aninherently social function for Duchennelaughter as it benefits all involved individualsby being a medium for playful emotional con-tagion. The mirror-system perspective goesone step farther by providing a novel andplausible neural mechanism. In this view, mir-ror neurons are the proximate basis (Prestonand de Waal 2002) of the shared manifold,while the resulting emotional synchrony (orsocial understanding) requires a complemen-tary ultimate explanation in terms of group-based functional integration.


the view from positive psychologyand play research

The field of positive psychology focuses onthe determinants of optimal human perfor-mance (Seligman and Csikszentmihalyi 2000)in contrast to the traditional focus in psy-chology on dysfunction. The work of Freder-ickson (1998) is most relevant to our discus-sion of laughter and humor.

Utilizing an adaptationist perspective,Fredrickson formalized the Broaden-and-Build model that accounts for the form andfunctions of positive emotions, such as joy, in-terest, contentment, and love. According toFrederickson, these positive emotions “serveto broaden an individual’s momentarythought-action repertoire, which in turn hasthe effect of building that individual’s physi-cal, intellectual, and social resources” (Fred-rickson 1998:300). Fredrickson additionallyproposes that positive emotions function to“undo” the effects of negative emotional ex-perience. A recent study has shown that posi-tive emotions do indeed promote psycholog-ical resilience in the face of stress (Tugadeand Fredrickson 2004). Fredrickson proposesthat joy specifically fosters the urge to play, inthe broadest sense of the word, and that thisplay builds resources while promoting socialcohesion, cooperation, and even altruism.

Though not elaborated by Fredrickson(1998), research on animal and human playlargely corroborates her account. It is becom-ing established that the social and develop-mental advantages of play outweigh its im-mediate costs (Sharpe et al. 2002; de Oliveiraet al. 2003; Fagen and Fagen 2004). The mostwidely accepted function for play entails thefacilitation of physical, cognitive, and emo-tional development (Byers and Walker 1995;Spinka et al. 2001), particularly in juvenile or-ganisms (Fagen 1993). Play functions to pro-mote synaptogenesis and muscle differentia-tion (Byers and Walker 1995) while trainingindividuals to deal with unexpected physical,cognitive, and emotional stressors (Spinka etal. 2001; Donaldson et al. 2002). Social playcan also decrease aggression (Drea et al.1996), increase fairness and cooperation (Be-koff 2001), and it can be used as a tool forsocial assessment and for promoting the es-

tablishment and maintenance of social bonds(Pellis and Iwaniuk 2000). Social play inchimpanzees has been shown to mitigate ten-sions arising in potentially competitive con-texts, such as just before feeding (Palagi et al.2004). Social play in humans, in turn, hasbeen shown to facilitate competitiveness with-out aggression (Grave-Ardec et al. 2003)while serving both training and social assess-ment functions (Fry 2005). And as discussed,social play appears to have a distinct circuit inthe mammalian brain (Panksepp 1998) thatis mediated, at least in part, by pain-reducingopioids (Panksepp 1998; Van den Berg et al.2000; Guard et al. 2002). The functions ofplay are thus not unlike those subsumedwithin Fredrickson’s model of positive emo-tion.

Our account of laughter and humor ishighly compatible with Frederickson’sBroaden- and-Build model and the literatureon play. As we have shown, Duchenne laugh-ter is inherently tied to positive emotional ex-perience and functions in essentially all of theways that Fredrickson proposed: to create aplayful mood that “undoes” negative emo-tions while promoting individual and group-level functioning. Laughter appears to func-tion in many of the same ways as social play,particularly since laughter often follows fromsocial play and facilitates playful interaction(Weisfeld 1994; Matsusaka 2004). Insofar associal play is functional, so too is Duchennelaughter. As Panksepp (1998:287) has argued,“it is reasonable to suppose that the sourcesof human laughter go back to ancient socialengagement systems that first mediated mam-malian playfulness.” It is noteworthy that hu-mans play across the lifespan and to a degreenot seen in other animals, such that somehave even referred to humans as Homo ludens,or “playful man” (see Bjorklund and Pelle-grini 2002). Laughter (and the joy associatedwith it) does indeed appear to be linked withplay, not only in children but also in apes andhuman adults. This supports our argumentthat all genuine laughter-eliciting stimulishare a common evolutionary and mechanis-tic basis derived from social play in primates.

Fredrickson’s (1998) Broaden-and-Buildmodel of positive emotions has the potentialto formalize the functions served by Du-


chenne laughter. It has been suggested thatthe contradictory functions that are often at-tributed to laughter (i.e., promoting play, andshowing aggression) arise from the simplefact that Duchenne laughter is typically notdistinguished from non-Duchenne laughterin discussions of laughter’s functionality(Keltner and Bonanno 1997). We maintainthat this is indeed the case and suggest thatthe functional model proposed by Fredrick-son (1998) be used as a foundation for con-sidering the functions of Duchenne laughter.

Taken together, the literature of social neu-roscience and positive psychology suggeststhat Duchenne laughter functions to pro-mote group cohesion and resource-buildingsocial play by spreading playful emotions.This idea is similar to the function of playpanting in chimps (Flack et al. 2004; Matsu-saka 2004) and to Weisfeld’s (1994) accountof laughter in humans, although where heemphasized the musical quality of laughterand cognitive mediation, we emphasize mir-ror-neuron based emotional contagion. Thenext point to be addressed is the nature ofthe stimuli (nonserious social incongruity)that elicit Duchenne laughter.

the view from evolutionarypsychology

While the perspective provided by Fred-rickson (1998) takes valuable strides towardsformalizing the value inherent in positiveemotions, and thus in Duchenne laughter,there is also an element that is missing. Fred-rickson (1998:303) discusses how emotionsevolved “because they promoted specific ac-tions in life-threatening circumstances andthereby increased the odds of the ancestors’survival.” This applies most obviously to neg-ative emotions, which coordinate the body toact adaptively in the face of a specific stimu-lus, such as a fight-or-flight response whensomething dangerous suddenly approaches.But positive emotions do not occur in life-threatening situations, and so presumablywould not have been selected to promotesuch specific, necessary “action-tendencies”(Fredrickson 1998). However, this ignores thenature of human laughter, which is a specificbodily response to a very specific stimulus

(i.e., nonserious social incongruity). Thismeans that in order for laughter to be un-derstood, one must consider the reason whyit would have been adaptive for an individualto playfully influence others in the face ofnonserious social incongruity. This position ismade most intelligible from the perspectiveof evolutionary psychology (Cosmides andTooby 1997).

Modern emotion theorists (e.g., Ekman1999) treat emotion as a response to the per-ception of a specific stimulus that adaptivelycoordinates the physiological, experiential,and behavioral aspects of an individual in thecontext indicated by that stimulus. An evolu-tionary psychology perspective in particular(e.g., Tooby and Cosmides 1990; Cosmidesand Tooby 2000) takes this insight and askswhy, in adaptationist terms, a given coordi-nated emotional response is conjoined with agiven stimulus. One guiding assumption isthat the stimulus was a recurrent componentof the human ancestral environment whilethe coordinated behavior evolved in relationto that stimulus, such that those individualswho responded appropriately to the stimuluswould have had relatively increased fitness.The coordinated reaction would thus haveevolved in the human environment of evolu-tionary adaptation (EEA) in conjunction withthe ecological, social, and morphologicalcharacteristics of our ancestors. The stimulus(or class of stimuli) would have constituted aspecific adaptive problem and been the selec-tive agent driving the evolution of the emo-tional reaction. By this logic, one must ques-tion what function the reaction would haveserved our ancestors.

Duchenne laughter fits well into the gen-eral emotions framework outlined above. It isbest seen as the behavioral component of apositive emotional reaction to a specific stim-ulus, nonserious social incongruity. Laugh-ter’s accompanying physiological response issympathetic arousal (Fry 1994; Panksepp1998), while the experience is “joy” (Pank-sepp 2003) or “exhilaration” (Ruch 1993). Inthis framework, the affect of joy and the be-havior of Duchenne laughter occur simulta-neously in response to the same stimulus. Thesignificance of the reaction lies not in laugh-ter occurring when we reach a certain level


of arousal, as many others have maintained(e.g., Owren and Bachorowski 2001), but inevolution having conjoined the affective andbehavioral components in the face of a com-mon stimulus. Laughter is also a signal, notjust a behavioral response, a fact highlightedboth by the dependence of laughter on a so-cial context and its effect on receivers. Thereis a distinct difference between fighting orrunning and laughing: the former work to in-fluence the individual experiencing the emo-tion, the latter functions to influence otherindividuals (Owren and Bachorowski 2003).

Why in the face of nonserious social incon-gruity is it adaptive for an individual to berendered playful, and for him or her to elicitin others positive affect and playful tenden-cies? This question will be central to our laterdiscussion and follows logically from the pro-gram of evolutionary psychology. However, atthe outset, we wish to make clear that whilewe find value in this approach to emotion andto laughter and humor in particular, it doeshave its limitations. Evolutionary psychologyhas been increasingly criticized for relyingtoo heavily on adaptationist thinking, em-ploying unrealistic assumptions about brainmodularity, and ignoring comparative evi-dence and the many homologues among hu-mans and other animals (see Panksepp andPanksepp 2000; Lloyd and Feldman 2002). Aswe have argued, laughter and humor have ob-vious homologues in other great apes and de-rive their form from a distant primate heri-tage. The affective circuit underlyingDuchenne laughter is also ancient, existingubiquitously among mammals and formingthe foundation for many affiliative and socialplay behaviors (Panksepp 1998). For this rea-son it would be misguided to formulate a hy-pothesis that recounts laughter and humorevolving de novo. Instead, we are interestedin the process by which laughter and humorinitially changed in degree (not kind) fromwhat is found in other great apes. Phyloge-netic considerations will therefore take cen-ter stage in our account despite typically be-ing accorded secondary importance byevolutionary psychologists (e.g., Cosmidesand Tooby 1997). Ancestral environmentalconditions and uniquely human selectivepressures will be invoked, but only to describe

why laughter and humor were elaborated inhumans relative to other apes. In identifyingan “adaptive problem,” we are thus not fol-lowing the modularity assumption that eachadaptive problem was solved by the evolutionof a novel piece of cognitive equipment (e.g.,Cosmides and Tooby 1997). Rather we focuson the phylogenetic origins of laughter andhumor and argue that an extant trait was co-opted and elaborated in the face of novel se-lection pressures. The mirror system perspec-tive (Gallese 2003) suggests such an ancientand general mechanism while also being ev-olutionarily and neurobiologically plausible.Invoking the historical logic of evolutionarypsychology we try to stay as close to the em-pirical facts as possible while avoiding theadaptationist excesses of mainstream evolu-tionary psychology (see Panksepp and Pank-sepp 2000).

the view from multilevelselection theory

Many have argued that laughter is adaptive,but few have thought carefully about the levelat which laughter is adaptive. This questionhas been neglected in part because of generalskepticism about group selection since the1960s, causing all adaptations to be inter-preted as forms of individual or genetic self-interest. Multilevel selection has recently re-turned to respectability as a theoreticalframework, especially in the case of humanevolution, where social processes can radi-cally alter the partitioning of phenotypic vari-ation within and among groups (Sober andWilson 1998; Boehm 1999; Hammerstein2003; Richerson and Boyd 2004; Wilson2004). Since laughter is quintessentially a so-cial process that alters the behaviors of othersin addition to the self, it is worth examiningits fitness effects from a modern multilevelperspective.

To begin with a brief tutorial, consider apopulation that varies for a trait such aslaughter and is divided into a large numberof social groups. There will be variationwithin groups (some individuals who laughmore than others or who get others to laughmore) and also variation between groups(groups in which laughter occurs more than


in others). Suppose that laughter is selectivelyadvantageous and evolves in the total popu-lation. This can happen in two ways. First, in-dividuals who laugh can be more fit than in-dividuals in the same group who do notlaugh. Second, groups in which laughter oc-curs can be more fit than groups in whichlaughter does not occur. The differential con-tribution of groups to the total gene pool canthus be an agent of evolutionary change assurely as the differential contribution of in-dividuals within groups. Total evolutionarychange is a combination of within- and be-tween-group selection, which can operate inthe same or opposite directions.

The rejection of group selection is theclaim that within-group selection is invariablystronger than between-group selection (Wil-liams 1966). This claim may have been rea-sonable in the 1960s, but it is no longer cred-ible based on theoretical and empiricaladvances that are reviewed elsewhere (Soberand Wilson 1998). Many evolutionary biolo-gists routinely partition selection into within-and between-group components (often usingstatistical methods such as the Price equation;Frank 1995) and conclude that traits evolveby group selection despite being selectivelyneutral or disadvantageous within groups. Itshould be noted that group selection remainscontroversial in part because of widespreadconfusion about the concept of selfish genes.Genes that evolve are called selfish becausethey replace the genes that did not evolve—regardless of whether the selective advantagewas within or between groups. The argumentfor selfish genes is not an argument againstgroup selection. In the terminology of selfishgene theory, it is the concept of vehicles, notreplicators, that is relevant to multilevel selec-tion. To say that the group is the vehicle ofselection is roughly equivalent to saying thatthe trait evolved by group selection (Dugatkinand Reeve 1994; Wilson and Sober 1994; So-ber and Wilson 1998).

These ideas provide insight into the con-ditions under which laughter and humorwere elaborated in humans. We argue thatthe benefits incurred through laughter andplay often occur at the group level as well asat the individual level, and between-group se-lection could thus have played a significant

role in the elaboration of laughter and hu-mor in humans. We also argue that bothwithin- and between-group selection wouldhave driven the elaboration of laughter andhumor in the same direction. To supportthese conclusions, we argue that laughter im-poses few or no within-group costs while alsopromoting between-group variation throughcontagion and within-group bond formation.These ideas will be used to evaluate one ofthe most recent and detailed evolutionarytheories of laughter proposed by Owren andBachorowski (2001).

A Historical Account of theElaboration of Laughter and Humor

Detailing an evolutionary scenario for agiven set of traits is a bold endeavor that runsthe risk of becoming an unsubstantiated “just-so story.” Nevertheless, there are a number ofclues amenable to scientific evaluation thatcan be utilized to triangulate the time frameand contexts of the evolutionary elaborationof laughter and humor.

establishing a time frameThe first clue comes from the phylogenetic

origins of laughter and humor mentionedearlier. The laughter-like vocalization accom-panying the play face in chimpanzees, goril-las, and orangutans suggests that the relaxedopen-mouth display was originally elaboratedin the hominoids sometime before the hom-inid line diverged around 6.5 mya (Fry 1994).Human laughter is moreover a highly derivedform of the generalized hominoid laughter(van Hooff 1972; Provine 2000). This indi-cates that the major changes to laughter sinceits emergence from the relaxed open-mouthdisplay occurred in the hominid line subse-quent to its split from the other hominoidsbetween 5 and 7 mya ( Jurmain et al. 2000).Similar reasoning applies to humor (Gamble2001).

A second line of evidence relates to bipe-dalism, which was fully developed in humanancestors (presumably the australopithe-cines) about 4 mya ( Jurmain et al. 2000).This is of great importance because bipedal-ism effectively lifted the morphological con-straints that allowed laughter to evolve as the


dynamic signal found in humans today (Prov-ine 2000). Provine’s Bipedal Theory (2000;see also Aiello 1996) maintains that bipedallocomotion freed the thorax of the mechan-ical demands of quadrupedal location andloosened the typical mammalian coupling ofvocalizations with breathing. This made lan-guage possible but also allowed laughter tobecome ritualized in humans as a stereotypedvocalization, in contrast to the one-syllable-per- breath play panting of chimpanzees(Provine 2000). Chimpanzees cannot laughlike humans because their anatomy inher-ently constrains their vocal abilities; theseconstraints vanished in hominids when theybegan walking on two legs (Provine 2000).

The brain case volumes of various homi-nids provide additional relevant evidence, es-pecially when combined with modern neu-robiological data. This convergence isparticularly relevant to the relationship oflaughter to language. While some have ar-gued that laughter evolved in tandem with orsubsequent to language (e.g., Caron 2002), itis more likely that the ritualization of laughterbegan shortly after the hominid transition tobipedalism and considerably before the evo-lution of language.

Language is wholly dependent upon thehigh degree of encephalization found in hu-mans (Deacon 1997). It is specifically the cor-tex—including frontal, prefrontal, parietal,and temporal cortices—that underlies hu-man language abilities. Yet the requisite in-creases in gross brain volume in these areasthat preceded (and coevolved with; Deacon1997) the evolution of language in humansdid not begin until around 2 mya, with Homohabilis (Deacon 1992). When these numbersare corroborated with current neurobiologi-cal findings relating to laughter and lan-guage, it becomes apparent that laughtermost likely did evolve independently of lan-guage, insofar as the cortical areas elaboratedin humans and necessary for language abili-ties are largely unnecessary for Duchennelaughter that stems from nonlanguage-drivenstimuli, such as play, tickling, and visual in-congruity (Deacon 1992; Wild, Rodden, et al.2003). Such laughter originates in subcorti-cal, limbic, and brain stem areas that arelargely shared with other primates. Du-

chenne laughter does not depend on thehigher cognitive centers that are uniquelyelaborated in humans. Human laughter ismuch more akin to primate calls than to hu-man language (Deacon 1992, 1997; Burling1993; Provine 2000; Rendall and Owren2002), and Deacon (1992) has proposed amodel that details the gradual transition fromprimate-like vocalizations to language inhominids. In this model, the major neural re-organizations that resulted in modern speechabilities did not begin until around 2 mya af-ter laughter and crying were already fully de-veloped (Deacon 1992; see also Burling1993). Further buttressing this idea is Fry’s(2002) suggestion that one can infer the evo-lutionary age of behaviors from the relativeevolutionary ages of the brain areas on whichthey depend. With Duchenne laughter origi-nating subcortically in the brain stem and lan-guage depending on highly derived corticalstructures, Fry’s logic supports the proposalthat laughter is evolutionarily more ancientthan language.

The developmental trajectory of laughterin human infants further supports this posi-tion. One of the first infant social vocaliza-tions, laughter develops at anywhere betweentwo and six months of age (Scheiner et al.2002), well before language (Deacon 1997)and even precedes the earliest syllabic imita-tion and experimentation (Scheiner et al.2002). Many agree that prelinguistic laughteris an integral tool of infant social interactionand self-regulation prior to the time whenlanguage can be utilized (Darwin 1872;McComas 1923; Sroufe and Waters 1976; Dea-con 1997; Panksepp and Burgdorf 2003). Theproximate causes of laughter viewed phylo-genetically and ontogenetically are also indic-ative of laughter’s prelinguistic origins. It isno doubt significant that ape laughter is elic-ited primarily through tactile stimulation(Provine 2000; Gamble 2001), since infantlaughter begins thus but progresses from in-creasingly visual and then auditory stimula-tion (Sroufe and Waters 1976) to humanadult laughter that occurs primarily duringsymbolic communication (Provine 2000). In-sofar as ontogeny does recapitulate phylog-eny among closely related species, this devel-opmental series indicates that vocalized


Duchenne laughter is more ancient than lan-guage.

Contrary to some of the prevailing evolu-tionary accounts (e.g., Caron 2002), Du-chenne laughter was likely already ritualizedwhen language abilities and their requisitecortical reorganization began to evolve inhominids around 2 mya. When this proposi-tion is combined with Provine’s (2000) argu-ment that bipedalism was a necessary precon-dition to the ritualization of laughter inhominids, a historical window is establishedbetween approximately 4 and 2 mya duringwhich laughter most likely evolved into theritualized vocalization found in humans to-day.

If laughter was elaborated between 4 and 2mya, then the number of variables that couldhave come to bear on the evolution of laugh-ter is greatly reduced. A discussion of the ini-tial ritualization of laughter need not takeinto account such factors as the volitionalcontrol of behaviors (like laughter), culturaloverlay, a Homo sapien-level Theory of Mind,and language, since these variables would nothave existed until after 2 mya. Deacon’s(1992;1997) model supports this proposition.Advanced culture and technology also beganno sooner than with H. habilis, the first spe-cies of hominid to fabricate the stone toolsthat are the antecedent to all other humancultural artifacts ( Jurmain et al. 2000). Thepossession by humans of a complete Theoryof Mind (Premack and Woodruff 1978), orthe ability to attribute a full range of mentalstates (including beliefs and desires) to our-selves and to others, is almost certainlyunique to the human species, which indicatesthat such an ability might not have evolveduntil as late as 40 kya (Baron-Cohen 1999).The evolutionary consequences that resultedfrom these variables actually interacting withlaughter post-2 mya will be considered afterwe establish the circumstances under whichDuchenne laughter was initially ritualized.

identifying the adaptive problem:form follows function and the eeaIf we tentatively accept that laughter was rit-

ualized to human levels following the foot-steps of bipedalism and preceding the evo-

lution of uniquely human cognitivecapacities, then the potential adaptive prob-lems that Duchenne laughter might havebeen elaborated to solve are greatly reduced.This is especially true if one considers that theform of a given adaptation tends to followthat adaptation’s function (Cosmides andTooby 1997). We have argued that the formof laughter is such that it can be consideredthe behavioral component of a positive emo-tional reaction to nonserious social incongru-ity, and that beyond being an overt behaviorit is specifically designed through ritualiza-tion to be a medium for playful emotionalcontagion. Another relevant feature of laugh-ter and play is that both are very dependenton levels of perceived seriousness and satiety(Fredrickson 1998; Biben and Champoux1999; Gamble 2001; Sharpe et al. 2002). Du-chenne laughter and play only occur whenindividuals are well nourished (Sharpe et al.2002) and do not perceive any threat to theirimmediate being (Gamble 2001). This is im-plicit in the nonserious qualification that ac-companies our general characterization ofthe Duchenne laughter stimulus. For exam-ple, differential safety levels determinewhether an infant will laugh or cry at an un-expected social event (Sroufe and Waters1976) or whether a tickler is menacing orplayful (Provine 2000). Also, we are prone tolaugh if someone trips and falls suddenly (so-cial incongruity), but if we recognize the ac-cident is serious, we rush to aid the personinstead. Thus, the necessity of nonseriousnessor positive emotion for laughter and playfulbehavior can be viewed as a design featurethat constrains their potential circumstancesof occurrence.

In her account of the evolutionary signifi-cance of positive emotions, Fredrickson(1998) argues that such an emotion programevolved in humans specifically in response tothe “momentto-moment fluctuations in safetyand satiation” experienced by our hominidancestors “as [they] made their way acrossland” (Fredrickson 1998:312) and to a prob-lem they would have thus encountered: “. . .when and how should individuals buildresources for survival?” (1998:313). Fredrick-son (1998) answers that during fleeting pe-riods of safety and satiation, when danger was


relatively low and needs were otherwise met,positive emotions broaden thought-actionrepertoires and build physical, cognitive, andsocial resources, thus promoting fitness.

The living conditions that Fredrickson(1998) describes, of fluctuating food suppliesand constant dangers, are in accord with thecurrent hypotheses of the living conditionsthat hominids experienced following thetransition to bipedalism around 4 mya (Foley1996, 1999). The prevailing notion is thatalong with the emergence of bipedalism wentthe transition of hominids to more open, sa-vanna-type environments, as climate-inducedhabitat changes in Pliocene Africa createddrier habitats and more dispersed andpoorer-quality foods (Foley 1999; Owren andBachorowski 2001). This transition entailedincreased daily foraging times, distances trav-eled, and home range areas of hominidgroups, as well as increased predator pressureand inter-group competition (Foley 1996).Together these changes led to drastic in-creases in time stresses for early hominidgroups (Foley 1996, 1999), and most likelyalso in physical and psychological stresses. In-creased social stresses would also have oc-curred, as they would have resulted both fromother stressors (i.e., physical and psychologi-cal) and from the need for new types of socialorganization, including increased coopera-tion among unrelated individuals (Owrenand Bachorowski 2001).

One issue not stressed by Fredrickson(1998) is the means by which early hominidscould have recognized, albeit implicitly, thata given context was safe, and thus fit for theplaying and exploring. Another point is thatwithin-group variation would have undoubt-edly led some individuals to recognize propersafe contexts more readily, while others car-ried over the lingering effects of prior nega-tive emotions into times of safety and satietyand in effect dampened the potential forpositive emotions (see Biben and Champoux1999). When laughter and humor are placedwithin the context of Fredrickson’s (1998)conception of the role of positive emotionsfor early hominids in the face of increasedtime and body stresses, a novel connectionemerges: if positive emotions functioned dur-ing “free” time to prepare individuals for the

increased demands of bipedal life, and if suchfree times were fleeting and occurred unpre-dictably, then there would have been selec-tion on hominid behavior for an adaptivemechanism whereby positive emotions couldbe quickly and effectively elicited when timeswere safe. Additionally, if positive emotionscould be efficiently transmitted between in-dividuals then even those who did not rec-ognize the safety of the situation, or who hada lingering negative mood, could be re-cruited into playing through the elicitation ofpositive affect. At this point it should barelyneed suggesting that hominid humor andlaughter could potentially have functioned injust this fashion.

To make this scenario feasible, our analysiswill make explicit (1) the nature of the stimuliand signal of laughter as these would have ex-isted and been elaborated between 4 and 2mya, and (2) the fitness benefits that wouldhave thereby been conferred to the senderand the receiver(s) of such a signal and totheir group as well. Paramount is the relationof these considerations to the hypothesis thatlaughter became ritualized in hominids as amechanism whereby individuals and groupscould more readily capitalize on fleeting mo-ments of safety and satiety in order to“broaden and build” (Fredrickson 1998)their thoughts and resources through socialplay.

the roots of humorThe first issue at hand is the nature of the

stimuli that elicit Duchenne laughter. A num-ber of such stimuli have been identified inthis paper and elsewhere. We have arguedalong with Provine (2000) that tickling andsocial play constitute the origins of laughterstimuli as shared with nonhuman primatesand served as the base from which hominidhumor was elaborated. Social play amongchimpanzees is most prevalent among juve-niles (Fagan 1993) and has been shown to oc-cur most frequently leading up to feedingtime (Palagi et al. 2004). This suggests an in-strumental role in mediating stresses to fa-cilitate group functioning, and is supportedby the finding that play often co-occurs withgrooming, an activity often considered the


glue of primate societies (Dunbar 1996).Such specificity in chimpanzee social playcontrasts with modern human social play,which seemingly occurs at all times andthroughout the lifespan (Bjorklund and Pel-legrini 2002). With play nearly ubiquitousamong mammalian taxa (Spinka et al. 2001),social play and tickling would undoubtedlyhave been present among hominids, initiallyresembling chimpanzee play in context andfrequency but later approaching the charac-teristics of play in humans. How did such atransition occur?

Drawing from Fredrickson’s model (1998)we propose to bridge these interspecies dif-ferences by arguing that, following the tran-sition to bipedalism, there would have beentremendous selective pressures on hominidsfor the types of broadened thoughts and de-veloped resources that Fredrickson describes.As a result, there would have been selectivepressures for positive emotion and its proxi-mate causes, social play among them. All mo-ments that did not explicitly preclude socialplay (or that satisfied requisite conditions ofsafety and satiety) would have become highlyamenable to such play, with selection favoringthose individuals who could experience posi-tive emotions and play whenever possible.The role of protohumor in fulfilling such se-lective pressure is straightforward: events mir-roring social play and tickling in their quali-ties of nonserious social incongruity wouldhave inherently indicated that a situation wassafe and conducive to positive emotion andsocial play. The neuroanatomical mechanismthrough which play and tickling elicitedlaughter in apes was thereby evolutionarilybroadened in hominids, such that “playful”stimuli were decontextualized from the physi-cal act of social play itself, and Duchennelaughter could be triggered by a more gen-eralized type of nonserious social incongruity.Such a connection between stimulus and con-text need not have been present in the mindsof perceiving hominids, but the tendency toreact in such a way to a stimulus—the key tothe evolution of adaptive emotions (Cosmi-des and Tooby 2000)—would have been se-lected for and preserved by the correlation innature between nonserious social incongruityand safe situations. In short, hominids be-

came more playful through an evolutionaryprocess that made more aspects of experi-ence appreciable as “funny.” We can recog-nize this today in the extent to which play per-meates our social and developmental lives,earning humans the nickname “homo lu-dens” (Bjorkland and Pelligrini 2002).

To flesh out this idea, what we now call slap-stick humor (or physical mishaps that resultin no real harm) could very well haveemerged from this process. Witnessing an-other individual unexpectedly trip or slip(from an awkward bipedal gait?), while si-multaneously recognizing the nonseriousnessof the mishap, often elicits laughter in hu-mans today—observe the commercial successof The Three Stooges and America’s FunniestHome Videos. Such a mishap could have be-come a potent elicitor of laughter in earlyhominids as a result of Pliocene pressures forincreased social play. Ramachandran (1998)has recently proposed that in such a situation,laughter functioned as a signal to alert othersthat a potentially serious situation had in factturned out to be nothing of the sort. Rama-chandran (1998) has argued convincinglythat such nonserious social incongruity wouldhave been present in the hominid EEA char-acterized above, and it would have been a re-liable indicator of the safety of situations. Dueto the costs imposed on individuals who wereslow to utilize safe and satiated situations,such generalized nonserious social incongru-ity would have been selected to be a potentdistal elicitor of laughter, positive emotion,and play. The broadening of the laughter trig-ger thereby converted some stimuli—such asnear accidents, flatulence and excretion, andsexual mischief—from potential sources ofsocial stress to elicitors of social play and posi-tive emotion. This interpretation is in con-trast to the stimulus reward account of Weis-feld (1994), whose emphasis on “fitnessenhancing stimulation” has trouble account-ing for such protohumor.

the ritualization of laughterThe above account of the roots of humor

is meant to add coherence to our account butit could turn out to be an unnecessary adden-dum once the humor capacities of nonhu-


man apes are fully described. Such capacitieshave never been systematically studied acrossspecies, and it is possible that the apprecia-tion of humor in apes is little removed fromthat in humans (see Gamble 2001). In con-trast, laughter as a signal is inarguably elabo-rated in humans, and so accounting for itsritualization is of paramount importance. Inthe process of ritualization, a signal changesin structure so that it is more prominent andunmistakable, and thus more readily percep-tible (Eibl-Eibesfeldt 1989; Grammer andEibl-Eibesfeldt 1990). This occurs throughthe simplification of a signal via the additionof rhythmic repetition, an exaggerated am-plitude, and stereotypic structure. The releas-ing threshold of a signal is also lowered dur-ing ritualization. All of these changes tookplace during the evolutionary transition fromape play panting to human laughter (Gram-mer and Eibl-Eibesfeldt 1990).

The ritualization process of Duchennelaughter was surely intertwined with its affect-inducing properties that are now being de-scribed in humans (Provine 1992; Hatfield etal. 1994; Bachorowski and Owren 2001). Ar-guably the most potent stimulus in our EEAfor signaling safe times, eliciting in individ-uals playful affect, and undoing lingeringstress and negative emotional valence was thelaughter of others originating from socialplay or “false alarms.” Such laughter wouldhave signaled the playfulness of conspecificsand, if able to tap into the emotional systemsof receivers, would have elicited in them aplayful state. The ritualization process simplymade laughter more affective at affect induc-tion.

The shared manifold perspective describedearlier (Gallese 2003), in which the same sys-tem is used to perceive and send a signal,makes the evolutionary ritualization of laugh-ter all the more intelligible. Insofar as laugh-ter is contagious and can function as its ownreleaser, lowering its releasing thresholdwould have occurred with an exaggeration ofthe signal itself. Laughter, becoming moreperceptible, would have more readily elicitedlaughter in receivers by virtue of the receiver’sown laughter program also having been ex-aggerated through ritualization. Even a pro-cess as simple as laughter evolving to be max-

imally different from other signals wouldhave rendered the mirror matching of laugh-ter more effective (see Deacon 1997 for a de-scription of such “disruptive evolution” as ap-plied to laughter and crying). With sendersand receivers coevolving (Buck and Ginsburg1991), and variously being one and the sameindividual, exaggeration of the productionmechanism would have been equivalent toexaggerating the perception mechanism.

What selection would have driven this rit-ualization process? Fredrickson (1998) has al-ready argued for the adaptiveness of the emo-tion itself—to broaden thought-actionrepertoires and to build physical, psycholog-ical, and social resources, in this case throughsocial play. Positive emotions also undo nega-tively valenced emotions, which dampen ten-dencies to play and can have detrimental ef-fects on individuals over the long term(Fredrickson 1998; Tugade and Fredrickson2004). Laughing—that is, sending a signal inthe face of nonserious social incongruity—would have been adaptive by inducing in oth-ers a positive state, and thereby effectively re-cruiting them to engage in social play withthe laugher. It is in the evocation of this socialplay via emotional contagion that the essen-tial adaptive value of the signal of laughter (asapart from its affective concomitant) resides.Those who were more adept at becomingplayful during the appropriate times whilealso eliciting in others a playful state wouldhave found themselves with relatively in-creased within-group fitness through thebenefits of social play (Byers and Walker1995; Spinka et al. 2001). This function is notunlike that served by play panting in chim-panzees (Flack et al. 2004; Matsusaka 2004)and that proposed for laughter by Weisfeld(1994). It should be noted that in arguing forthe within-group adaptiveness of laughter viathe adaptiveness of social play, we are not es-pousing a particular function for social playbeyond the facilitation of development (Byersand Walker 1995; Spinka et al. 2001) and the“broadening and building” developed byFredrickson (1998).

The receiver side of the laughter event ishighlighted by Ramachandran’s model(1998) and fits ideally into the above frame-work with one major amendment. Whereas it


makes sense that heard laughter functionedto recruit others into playing during safetimes, and that it would have been adaptivefor individuals to be influenced by the laugh-ter of others, Ramachandran (1998) ignoresthe emotional component of laughter. Whilehe treats laughter as a cognitive “false alarm,”we have instead argued that laughter is a me-dium for emotional contagion, a mechanismfor transmitting positive emotion and play-fulness. In fact the concept of automatic res-onance behavior (Rizzolatti et al. 1999) me-diated by the shared neural bases ofperception and action invokes no need forcognitive mediation. In this light, it wouldhave been adaptive for individuals within agroup to be influenced by the laughter of oth-ers as it would have allowed them to imme-diately capitalize on safe situations. The af-fect-inducing property of laughter could evenhave allowed the laughter of others to quicklymitigate the minor pain or stress experiencedby the subject of a laughter-inducing near-mishap, especially before hominids were ableto infer that they were being laughed at (seelater discussion). So through both the elicitedpositive emotion and the emergent socialplay, the fitness of the hearers of laughterwould have increased during periods of in-creased stress. While others have given thename “manipulation” (Krebs and Dawkins1984) to an influence over conspecifics suchas that which Duchenne laughter has, the evi-dence suggests that such a term would haveto be amended with “mutually beneficial” ifapplied to Duchenne laughter. Of course, thisdoes not have to hold for non-Duchennelaughter (see Owren and Bachorowski 2003).

The shared benefits incurred by both thesender and the receiver of Duchenne laugh-ter imply that there would have been robustwithin-group selection driving the ritualiza-tion of laughter towards increased percepti-bility, elicitation, and distinctiveness fromother signals. Since produced and perceivedlaughter potentially depend on a commonmechanism, as suggested by a mirror systemperspective, ritualization of the signal wouldhave been selected for by the fitness benefitsgained by individuals who were at differenttimes both sender and receiver.

A number of additional conclusions can be

drawn by examining these ideas from a mul-tilevel perspective. First, the benefits of laugh-ter that have been proposed are also oftengroup-level benefits. They benefit the groupof people unified and often defined by laugh-ter, compared to groups in which the traitsassociated with laughter are absent or lessprevalent. The individuals who possess thetraits benefit as members of their group, butit is not obvious that they benefit more thanother members or that their within-group ad-vantage is the driving force, compared to thecollective benefits. Everyone in a given groupbenefits from such signaling and coordina-tion mechanisms, not just the ones who pro-duce the signals. Bekoff’s (2001) argumentsabout social play agree with this argument.

Second, even if the traits associated withlaughter do not provide a strong within-group advantage, neither do they impose astrong within-group disadvantage. Sociobiol-ogy has been excessively concerned with theconcept of strong altruism, in which individ-uals can benefit others only at a cost to them-selves. This negative tradeoff sometimes exists(e.g., when you can save a drowning persononly by jumping into a river), but many traitssubstantially increase the fitness of everyonein a given group—including those possessingthe traits—at little or no private expense.These traits can evolve on the strength oftheir group-level benefits relatively unoppo-sed by selection within groups. Rather thanassuming a negative tradeoff between bene-fits to others and cost to self, it is useful todistinguish two potentially independent con-cepts: coordination and distribution (Wilson2001). The question of whether a group func-tions well or poorly is a matter of how mem-bers coordinate their activities. The questionof whether the traits evolve depends on howthe costs and benefits of coordination are dis-tributed among the group members. If coor-dination can be achieved with a relativelyequal distribution of costs and benefits, thengroup-level adaptations can straightforwardlyevolve. Many of the traits associated withlaughter appear to exist in this category.

Third, laughter favors group selection inthe same way that culture does. Culturalgroup selection is powerful because the socialtransmission of behaviors creates differences


among groups more easily than the genetictransmission of behaviors (Boyd and Richer-son 1985). For example, it is easy to imagineeveryone in a group adopting a culturally ac-quired behavior, which makes them very dif-ferent from other groups. The only way forcomparable variation among groups to arisefor genetically acquired traits is for thegroups to be clonal. The contagious nature oflaughter makes it similar to forms of culturaltransmission in its effects on behavioral uni-formity within groups and differences amonggroups. Even a single person who laughs canchange the mood and behavior of an entiregroup, thereby increasing between-group dif-ferences. If laughter increases the social playin a group, such a difference would increasethe fitness of that group relative to others (seeBekoff 2001), thereby driving group selec-tion.

These conclusions can be used to evaluateone of the most recent and detailed evolu-tionary theories of laughter by Owren andBachorowski (2001), who propose that laugh-ter evolved in response to the selfish exploi-tation of emotional signaling by cheaters.They envision a coevolutionary arms race inwhich: (1) cooperation was first fosteredamong unrelated individuals through anhonest emotional signal, namely, smiling; (2)smiling was then corrupted by the evolutionof cheaters who could volitionally mimic smil-ing and exploit the cooperation of others;and (3) laughter then evolved through anelaboration of the original smile mechanismin effect to replace smiling as a harder to fakesignal for promoting cooperation. The con-cept of honest signaling is being used to ex-plain a number of subjects besides laughter,such as religion (e.g., Irons 2001; Nesse 2001;Sosis 2004). In the case of laughter, however,there are a number of problems with Owren’sand Bachorowski’s account. First, the fact thatthey call it a “selfish gene” account says noth-ing about levels of selection, which requires acomparison of fitnesses within and amonggroups. Second, they assume that coopera-tion is always vulnerable to exploitation andcannot evolve in the absence of iron-cladhard to fake signals. This assumption is notbased on a consideration of laughter per se,but is made axiomatically for all forms of co-

operation. We have already criticized this as-sumption and suggested that Duchennelaughter occurs in contexts where coordina-tion can be achieved with little danger of ex-ploitation (i.e., safe and satiated contexts).Third, their account assumes that laughterevolved late in the coevolutionary sequence,after smiling was corrupted as a signal,whereas all the evidence that we have re-viewed points to the early evolution of laugh-ter, prior to volitional control, and with dif-ferent phylogenetic origins for smiling andlaughter (van Hooff 1972). We do not wish toentirely reject the concept of laughter as anhonest signal, but we think that Owren’s andBachorowski’s account fails as a primary ex-planation of laughter.

Both within- and between-group selectiondriven by the environmental conditions ofthe Pliocene help to explain the elaboratedcondition of protohumor and Duchennelaughter in humans. Modern multilevel selec-tion theory thus provides an excellent frame-work for the study of laughter. The dynamicsof laughter within single groups is a compli-cated subject in its own right. The manner inwhich groups form, interact, and disperseadds another layer of complexity. Multilevelselection theory provides a way of managingthis complexity by first focusing on singlegroups as evolutionary units and then focus-ing on the differential contribution of groupsto total evolutionary change. Given the stronggroup-level advantages, the weak disadvan-tages within groups, and the populationstructure of early hominid groups, there isample scope for laughter to have evolved onthe strength of group-level, in addition towithin-group, benefits.

Subsequent Co-optation and theDiversification of Functions

With Duchenne laughter and protohumorthus established in early hominids, the formand functional value of these traits wouldhave been present during subsequent evolu-tionary transitions through various species in-cluding H. habilis ,H. erectus, and eventuallyH. sapiens. This means that Duchenne laugh-ter and protohumor would have been presentin the prehuman behavioral repertoire as


various advances were made in successive spe-cies, including the evolution of volitionaloral-facial muscle control, the invention ofmaterial and cultural artifacts, language, theevolution of a Homo sapien-level Theory ofMind, and the emergence of fully modernhumans and their societies, belief systems,technology, and cultural variation. We suggestthat each of these evolutionary novelties in-teracted with the already-present laughterprogram resulting in the various forms oflaughter that we now find and their myriadfunctions in different social contexts. We pro-pose that it is from this process that non-Du-chenne laughter arose. We also maintain thatthere is a necessary continuity between alltypes of laughter, as they are all derived fromthe primate play display, hominoid proto-laughter, and hominid Duchenne laughter.

Essential to our account is treating Du-chenne laughter as a system that was co-optedby volitional motor circuits, allowing humansto use its affect-inducting properties to stra-tegically influence others. In this sense, it wasthe effect of Duchenne laughter on othersthat was co-opted. Whereas Duchenne laugh-ter originally evolved in the context of pro-tohumor and functioned as a medium of joy-ful emotional contagion, non-Duchennelaughter came to occur in aggressive, ner-vous, or hierarchical contexts, functioning tosignal, to appease, to manipulate, to deride,or to subvert. Such a transition is supportedby the existence of two neural circuits forlaughter (Wild, Rodden et al. 2003) and bythe work of Deacon (1992, 1997). Duchennelaughter was also expanded by new elicitors,including language and mental state attribu-tion. These stimuli came to elicit the full Du-chenne laughter program insofar as they con-formed to the structure of the alreadyestablished trigger mechanism, nonserioussocial incongruity.

For example, humans can now voluntarilyaccess the laughter program and utilize it fortheir own ends, including smoothing conver-sational interaction, appeasing others, induc-ing favorable stances in them, or downrightlaughing at people that are not liked. This isapparently the result of the evolution of vo-litional prefrontal motor circuits that have al-lowed humans to have a generalized kind of

willful control over oral-facial and respiratory-laryngeal motor systems (Deacon 1992). Theuses to which such non-Duchenne laughtercan be put stem from the established char-acteristics of Duchenne laughter, particularlyits direct and indirect effects on others.Laughter’s direct effects are the evolved af-fect-induction properties of the specific signalor its ability to emotionally influence pre-at-tuned receivers, while its indirect effects arederived from receivers’ learned associationsbetween laughter, play, and positive affect(Owren and Bachorowski 2003). In the caseof aggressive laughter, this connotation im-plies that the laugher derives pleasure fromthe misfortune of another. Laughter’s capac-ity to function metacommunicatively (Caron2002) as a conversational sign that frames be-havior with playful intention (Provine 2000)also arguably stems from the intrinsic playfuland pleasurable qualities of laughter andlaughter’s effects on others (Owren and Bach-orowski 2003). Even the convergence of smil-ing and laughter that can be observed inmodern humans (despite the different phy-logenetic origins of the two displays; vanHooff 1972; van Hooff and Preuschoft 2003)is arguably the result of learned mimicry ofthe two signals by these higher human cog-nitive abilities that creates a functional hybridutilized during conversation.

Language is a related evolutionary devel-opment made possible by such novel prefron-tal circuits (Deacon 1997). A linguistic systemenables symbolic representations of the worldto conceptually invoke the recognition ofnonserious social incongruity. This is the for-mal humor we know so well, which has cometo take myriad forms and have as its subjectmatter innumerable cultural symbols and in-stitutions (see Apte 1985). It is again signifi-cant that language can only spontaneouslytrigger Duchenne laughter if the structureand context of the statement or story embod-ies nonserious social incongruity and the al-ready-established laughter mechanism. Thisrelationship between old and new explainswhy certain constraints are imposed on thetypes of statements and stories that elicitlaughter—why a well-timed and precisebuildup of expectation is so necessary for ef-fective humor (Ramachandran 1998), and


why only a specific, playful, alternately intel-ligible type of punch line is humorous.

Humans can also now find nonserious so-cial incongruity in the inferred subjectivestates of others, in the discrepancy betweentheir intentions and their actions or betweentheir beliefs and reality (Howe 2002; Jung2003). Such a reading of another’s mind canfacilitate the strategic use of humor, as onemust presumably understand what is ex-pected by others in order to present themwith the unexpected. This ability adds an ad-ditional new dynamic to laughter perceptionbecause it allows one to infer in a laugher themental states and intentions that elicited thelaughter, and as a result to attribute a newkind of situational and context-dependentmeaning to the observed act of laughter( Jung 2003). Quite possibly it is from thisability that much “aggressive” laughter arises,as the hearers of such laughter can therebyrecognize it as being aggressive. This abilityto read another’s mind allows people to rec-ognize that they are being laughed at and arethus part of the out-group. The sender ofsuch laughter is either intentionally laughing(non-Duchenne) to be aggressive or is simplylaughing (Duchenne) at what he or she per-ceives to be nonserious social incongruity.Such incongruity is in this case nonserious be-cause the aggressive laugher has no inherentconcern for the well-being of the target of thelaughter, and it is this lack of concern that thetarget can then infer. More generally, the in-teraction of “mind- reading” abilities withconversational laughter creates myriad op-portunities for nuanced information to beconveyed as variability in laughter acoustics iscorrelated with different social contexts (Vet-tin and Todt 2004). Laughter does functionto convey information under some circum-stances, as opposed to simply inducing affect(see Deacon 1997; Owren and Bachorowski2003; Vettin and Todt 2004), but arguablyonly insofar as conversational laughter inter-acts with a human-level understanding of oth-ers’ intentions.

One last elaborated example should makethe process of co-optation more concrete.Dunbar (1993, 1996; see also Provine andFischer 1989) has proposed that languageevolved in hominids to replace grooming as

the prime means of managing the prehumansocial environment. He proposed that selec-tive pressures for increased group size led lan-guage to evolve to manage social relation-ships through the sharing of socialinformation, or gossiping. Dunbar’s theoryhas been criticized on the grounds that lan-guage by itself could not have functioned likegrooming (Power 1998) because the func-tional basis of grooming is the honest timecommitment that it entails and the opiate re-lease that it facilitates. Language is a cheapsignal that can be used dishonestly and pro-vides no such pleasurable stimulation (Power1998), that is unless conveyed as humor orword play; only then can language elicit inothers what is effectively a parallel emotionalresponse to grooming. This idea is referredto as the “vocal grooming hypothesis” (Dun-bar 1996; see also Vaid 2002). That humorand laughter could have been utilized thus issupported by our argument that Duchennelaughter and protohumor were fully estab-lished when human language began toemerge. This idea further suggests that lan-guage might have necessarily evolved in re-lation to humor and laughter, with the capac-ity to conceptually mimic, using symbols, thestructure of protohumor.

In this scenario language and humor wouldno doubt have coevolved to some extent, withlanguage and speakers becoming more adeptat eliciting laughter and the laughter mech-anism becoming more susceptible to concep-tual triggering. But the instructive aspect ofthis hypothetical example is that only theoriginal properties of Duchenne laughterwould have allowed it to be evolutionarily har-nessed by language. Only insofar as languagewas able to conform to the structure of thealready established elicitor of laughter couldit trigger the laughter response. Thus, Du-chenne laughter was present to be co-optedby language (and by the mental abilities thatlanguage requires), but as such there neededto be some point of contact between the twosystems—in this case, the potential of lan-guage to represent nonserious social incon-gruity, and the affective properties of laugh-ter that constituted its functional potentialand endowed it with the capacity to be util-ized for vocal grooming.


There is a precedent for taking a co-opta-tive view of emotion systems. Rozin and col-leagues (e.g., Rozin et al. 1997) have detailedhow the ancient mammalian food rejectionsystem of disgust was co-opted by human cog-nition to play a role in cultural value systems.Previous laughter researchers have specu-lated along similar lines. Fry (1994:121) as-serted, “ . . . mirthful laughter is probably an-other of those archaic adventitious behaviorcomplexes which has been swept up in theflowing of time and evolution and carried farbeyond its original activity.” Likewise, Pank-sepp and Burgdorf (2003:541) said “ . . . thesubtleties of human adult laughter, so abun-dantly expressed during cognitive delights ofhumor, may highlight how certain ancientemotional processes interact with refinedcognitions within higher reaches of the brain-mind.” Caron (2002:267) argued in turn that“as hominids became humans, laughterwould have become useful as part of an in-creasingly more abstract language.” To ourknowledge, however, this is the first time thatsuch a notion regarding laughter and humorhas been incorporated within a formalizedtheoretical framework.

A final conceptual point needs to be keptin mind. The evolutionary scenario we havedetailed exists in two parts: (1) the initialelaboration of Duchenne laughter and pro-tohumor in hominids; and (2) the co-opta-tion of laughter and humor for novel func-tions. The first argument is more tenuous,attempting to pinpoint the actual conditionsunder which laughter evolved and the adap-tive functions that laughter originally served.Contentious is the point that laughter wasoriginally ritualized in hominids more than 2mya to serve as a mechanism for transmittingplayful emotion among members of a group.While supported by several lines of evidence,the idea is admittedly and perhaps inevitablyspeculative. Yet the second argument doesnot depend entirely on the validity of the first.While the history and initial characteristics ofDuchenne laughter are essential to treatinglaughter as a preadaptation, the specific cir-cumstances under which laughter was ritual-ized are partially irrelevant to laughter’s laterelaboration. This is as long as laughter wasfully and appropriately developed when the

implicated abilities began to evolve post-2mya. Regardless of the driving forces and cir-cumstances of Duchenne laughter’s historicalgenesis, viewing non-Duchenne laughter asan evolutionary derivative of Duchennelaughter makes sense of the variety of formsand functions that characterize laughter to-day. Without such an explicit recognition ofthe co-existence of Duchenne and non-Du-chenne laughter, and of playful, appeasing,aggressive, and dissociative functions, theclaim of Keltner and Bonanno (1997) holdstrue: the science of laughter is, and will con-tinue to be, swamped in contradiction andconfusion.

Predictions and Future DirectionsThe core objective of this paper has been

to synthesize the many data and specific hy-potheses of the evolution and functions oflaughter and humor into a single evolution-ary account that includes phylogeny and his-tory as well as proximate mechanisms andadaptive significance. Perhaps the most im-mediate criticism that can be made of ourevolutionary hypothesis is that it has all themakings of a “just-so story.” If this were thecase, we would have indeed woven a fine talebut it would be unfalsifiable and thereforenot scientific. We contend that it can be fal-sified in a number of ways, and we derive anumber of testable predictions that lend gen-uine scientific value to our account. In somecases these predictions diverge from or chal-lenge directly the work of other researchers,and where this is the case we have tried tonote the opposing perspectives.

The level of historical detail that we incor-porated into the first part of our argument iscontentious but potentially stands to be re-butted. This would be contingent upon acloser analysis by physical anthropologists ofhominoids and hominid fossils reliably datedfrom the middle Pliocene. While we incor-porated Provine’s (2000) Bipedal Theory intoour account, we did not discuss the specificfossil evidence of such a transition other thannoting the evidence for bipedalism by 4 mya.Neither did we discuss the larynx, to whichchanges might have needed to occur in orderfor laughter to go from pant-like to “voiced.”


Although laughter should not have requiredthe same types of morphological changes aslanguage (for instance alterations to thetongue, palate, and teeth to allow for the pro-duction of phonemes), some changes to thebasicranial structure of hominid skulls wouldperhaps still have been needed. Thesechanges, if in fact necessary to our account,should be observable in the fossil record. Wesay “if necessary” because the actual physicalrequirements of human laughter have notbeen established, and so a collaboration ofanatomists, biophysicists, and paleoanthro-pologists is apparently in order. It should benoted, however, that even the anatomical pre-requisites of speech have proven hard to inferfrom the fossil record (see Fitch 2000).

The neuroanatomical components of ouraccount also include concrete propositionsthat stand to be falsified by future research,the most revealing of which will probably re-quire advanced neuroimaging techniques.Our argument that the different proximatecauses of Duchenne laughter all share a com-mon form (nonserious social incongruity)leads to the prediction that the laughter fol-lowing from these different stimuli (infantlaughter, laughter from tickling and play, andthe various distally perceived forms of hu-mor) should share a singular mechanism onat least some level. The substrates of initialprocessing will no doubt differ (and indeeddo; Goel and Dolan 2001), but the last stageof higher-level perception, which triggers the“laughter-coordinating” center (Wild, Rod-den et al. 2003), should be singular and pre-sumably involve incongruity detection and/or resolution (see Fry 2002). Comparatively,our hypothesis predicts that the laughter ofapes (both tickle/play induced and symboli-cally elicited) should derive from a homo-logue of this cognitive “laugh generator”(Provine 2000). Following the work of vanHooff (1972), we would also predict that smil-ing and laughter spring from different neu-roanatomical bases except insofar as they ei-ther hybridize under volitional and/orlearned control or require similar motorcommands. This contrasts with the work ofOwren and Bachorowski (2001), which seeslaughter as an elaboration of smiling and so

predicts a singular mechanism for both sig-nals.

Our integration of a theoretical mirrormechanism predicts that the singular laugh-ter production mechanism should also be ac-tive during the perception of laughter, specif-ically Duchenne laughter that inducespositive affect. Such a finding would suggestthe existence of “echo-neurons” to go alongwith the visuomotor neurons uncovered al-ready (see Rizzolatti and Craighero 2004), al-though these neurons might well representvocal tract configurations and not sound perse (see Liberman and Whalen 2000; Rizzolattiand Arbib 1998). As discussed above, we im-plicated an area of the SMA as the locust ofthe laughter motor program, but this is by nomeans exclusive as much more research inthe area is warranted.

We predict that prefrontal and temporo-pa-rietal areas (Saxe et al. 2004) involved in The-ory of Mind tasks, such as belief attribution,should be activated during the perception oflaughter that induces negative affect, specifi-cally insofar as such negative affect followsfrom an interpretation that the laughter wasaggressive or derisive based on context andknowledge of the laugher. We also predictthat “mind-reading” abilities should not benecessarily required for the perception of alllaughter inducing stimuli, since we argue thatlaughter was first ritualized prior to the evo-lution of a full human Theory of Mind(Baron-Cohen 1999). We might predict some-thing of a “continuum of activation” in whichmind-reading abilities are progressively moreactivated as one moves from infant laughterto tickle- and play-induced laughter to physi-cal humor and finally to formal symbolic hu-mor. This stands in contrast to the work ofJung (2003), which predicts that mind-read-ing ability is involved in all laughter.

The perception of laughter is prime for fur-ther exploration using neuroimaging tech-niques (Sander and Scheich 2001; Shibataand Zhong 2001). As a fascinating new direc-tion for future research, the simultaneous im-aging of multiple subjects interacting or per-forming the same task (see Adolphs 2003)holds promise for fleshing out the complexnature of laughter production and percep-tion in varying contexts. Unfortunately neu-


roimaging research on the production oflaughter is in many cases not feasible with cur-rent technology due to motion artifacts(Iwase et al. 2002). Measuring motor-evokedpotentials (MEPs) from muscles in the mouthand tongue region (see Fadiga et al. 2002;Watkins et al. 2003) could assess the extent towhich motor areas that control laughter areactivated during the perception of laughter.

To take a different tack, children with au-tism might be investigated. Several groups ofautism researchers have recently proposedthat the disorder is at base an imitation deficitand that this primary inadequacy cascadesinto the other social and emotional deficitsthat have traditionally characterized autism,such as those involving emotion sharing, em-pathy, theory of mind, and social norms (seeWilliams et al. 2001; Rogers et al. 2003; Wil-liams et al. 2004). Mirror neuron dysfunctionhas even been explicitly implicated as theroot of autistic disorders (Williams et al.2001). If this characterization of autism is onthe right track, then one would derive fromour account a number of predictions regard-ing laughter and humor in autistic individ-uals. These include: (i) few problems in de-tecting or laughing at tactile or visualprotohumor (such as tickling, social play,peek-a-boo, and slapstick humor); (ii) re-duced reaction to intention- and norm-basedhumor; (iii) inappropriate or inexplicablelaughter because of disrupted social expec-tations and unusual notions of incongruity;(iv) laughter at serious social incongruities;(v) a lack of responsivity to the laughter ofothers and to humor based in facial expres-sions; and lastly (vi) a lack of effort to elicitlaughter from others since such laughterwould not feed back positively on an autisticchild via contagion. Remarkably, a recentstudy on laughter and humor in autism(Reddy et al. 2002) supported every one ofthese predictions, and none of the study’sfindings were left unexplained by our ac-count.

Laughter that occurs during everyday so-cial interaction in response to banal com-ments and humorless conversation is now be-ing studied (Provine 1993; LaGreca et al.1996; Vettin and Todt 2004). The unstated is-sue is whether such laughter is similar in kind

to laughter following from humor. Althoughconscious control of laughter in natural socialcontexts has been tentatively discredited (butnot systematically explored; Provine 2000;Owren and Bachorowski 2001), that does notpreclude learning and automatization fromplaying a major role in non- Duchenne laugh-ter. This type of laughter is probably learnedthrough practice and the conditioning ofnonconscious motor programs, and has onlyemerged as modern human cognitive abilitiesevolved post-2 mya. We have also implied thatit is this type of laughter that constitutes thebulk of conversational laughter, somethingnot discussed in the accounts of Owren andBachorowski (2003) and Vettin and Todt(2004). Our position thus raises a number ofnovel questions. Is conversational (non-Du-chenne) laughter typically learned? If so,what are the respective roles of conscious andunconscious processes in the utilization ofsuch laughter? Does such laughter ever pre-cede, coincide with, or follow from positiveaffect? Ontogenetically, how and when doessuch learned laughter develop in relation tothe Duchenne laughter of infants? What de-termines the differential effectiveness of in-dividuals at strategically utilizing such laugh-ter? Does such conversational laughter occurcross-culturally, and if so does its follow simi-lar patterns? We would predict greater cross-cultural variation in the contexts and physicalcharacteristics of conversational as opposedto humor-driven laughter, suggesting conver-sational laughter is essentially a learned skill.We would similarly expect the punctuation ef-fect (Provine 1993) to vary across cultures asit should apply most strongly to non-Du-chenne conversational laughter.

We have not discussed gender differencesin laughter production. The main reason forthis is straightforward: researchers have con-sistently found no differences in the averagenumber of laughs produced by males and fe-males in social contexts (Martin and Kuiper1999; Devereux and Ginsberg 2001; Bacho-rowski et al. 2003, as cited in Owren and Bach-orowski 2003; Smoski and Bachorowski2003a; but also see Provine 1993). The sec-ond reason is more complex. The only otherconsistent gender-related finding is that fe-males tend to laugh most around males in


conversational interaction (Bachorowski etal. 2003, as cited in Owren and Bachorowski2003; Grammer 1990; Provine 1993; Smoskiand Bachorowski 2003a). However, research-ers have documented this pattern without dis-tinguishing Duchenne from non-Duchennelaughter. It is therefore quite possible thatsuch a pattern describes only one type oflaughter or that it will not hold across cul-tures. We would predict that this pattern willbe robust across cultures if it does involve Du-chenne laughter but that it will prove to beunreliable if it merely involves non- Du-chenne laughter and culture-specific socialnorms.

These issues are of utmost importance asresearchers begin the task of distinguishingclassic stimulus-driven Duchenne laughterfrom non-Duchenne conversational laughter.It bodes well for our account that the findingsof Vettin and Todt (2004), described earlier,suggest the proximate basis for the functionaldifferences of Duchenne and non-Duchennelaughter and a means of telling them apart infuture observational studies. It should benoted, however, that the facial feedback hy-pothesis (and its claim that mimicking anemotional expression elicits the correspond-ing affect; see McIntosh 1996) complicates at-tempts to distinguish Duchenne from non-Duchenne laughter on the basis of emotionalvalence alone, but it also suggests that thepresence of positive affect following non-Du-chenne laughter does not prove it is similarin kind to Duchenne laughter.

We have argued that Duchenne laughter isessentially a medium for emotional contagionoperating through an intersubjective mirrorsystem. Such a perspective has implicationsfor treatments of all nonverbal behavior, lan-guage, and more generally, considerations ofhuman social evolution. Given a mechanisticbasis, like that proposed by Gallese (2003)and his colleagues, these phenomena providean organizational principle for humans as so-cial beings, facilitating the empathic under-standing of others, the emotional coupling ofgroup members, and the establishment of co-ordination among interacting individuals.The evidence indicates that an early form ofthe shared manifold was already present inhominoids when the hominid line diverged,

as mirror circuits were first discovered inmonkeys and chimpanzees show evidence ofprimitive emotional contagion (Parr andHopkins 2000). Hominid evolution thereforeinvolved not the evolution of a mirror match-ing system from scratch but (among otherthings) the evolution of elaborated motormovements within a pre-existing system. Sig-nals conjoined with emotional experiencesare ideal examples of such elaborated move-ments. Humans have a highly ritualized andconventionalized nonverbal communicationsystem (Eibl-Eibelfeldt 1989), particularly re-garding expressions of positive emotions (To-masello et al. in press), and ritualized Du-chenne laughter is no exception. A questioncan then be posed: If emotion is essential foraction (Damasio 1994), then might notshared group emotion be essential for groupaction, that is, for solving the problem of co-ordination (Wilson 2001; Spoor and Kelly2004)? If answered in the affirmative, thenlaughter and other elaborated emotional ex-pressions might well be a fountainhead of hu-man uniqueness. Empathy through the com-municatory coupling of individuals has in factbeen implicated in human cooperation andaltruism (Lakin et al. 2003). So while culturalinstitutions are thought to organize humansto the level of a “crude superorganism” (Rich-erson and Boyd 1999), the present argumentimplies that there is a much more fundamen-tal mechanism for human organization thancultural institutions—a mechanism that pre-dates such institutions and even language andcould have been a prerequisite to the emer-gence of such evolutionary novelties.

Future research can explore this possibilityin many ways. These include social experi-ments that measure the unique efficacy of hu-mor and laughter (and other nonverbal be-havior) at promoting relative group-levelfitness (see Banning 1987; Vinton 1989;Greatbatch and Clark 2003), and agent-basedmodeling techniques, which are gaining so-phistication and relevance in relation to hu-man social evolution (see Kohler and Gum-erman 2000). Perhaps most importantly, theabove discussion is founded on assumptionsof ape abilities. Future comparative studies ofthe cognitive, emotional, and expressive ca-pabilities of other primates, particularly chim-


panzees (e.g., Parr and Hopkins 2000; Parr2004), will shed tremendous light on suchconsiderations of human social evolution.

The main goal of this paper has been tointegrate theory and data across disciplines,thereby gaining insight while also exposingscientists on different fronts to novel researchand theory. We think that laughter research-ers have much to learn from and offer to so-cial neuroscience, positive and evolutionarypsychology, and multilevel selection theory.As a specialized subfield of research on non-verbal behavior and communication, thefindings of laughter researchers hold rele-vance to all fields that concern themselveswith human behavior, social interactions, so-cial organization, and both verbal and non-verbal communication. An increasingly im-

pressive catalogue of data is emerging fromstudies by social psychologists with ethologi-cal dispositions, and we feel that laughter re-search deserves to be integrated into thelarger fold as a serious topic of scientific studythat must be accounted for in discussions ofhuman behavior and evolution.

acknowledgments

We would like to thank K Sander, D Shibata, and twoanonymous reviewers for their constructive input. Wealso thank participants in the Evolutionary StudiesProgram at Binghamton University for invaluablecomments and discussion. We are especially gratefulto Anne B Clark and Michelle Kline for their feedbackand support. Portions of this work were presented tothe Evolution for Everyone class at Binghamton Uni-versity in December 2004.

REFERENCES

Adolphs R. 2003. Cognitive neuroscience of humansocial behaviour. Nature Reviews Neuroscience 4:165-178.

Aiello L C. 1996. Terrestriality, bipedalism and the or-igin of language. Pages 269-289 in Evolution of SocialBehaviour Patterns in Primates and Man: A Joint Dis-cussion Meeting of the Royal Society and the British Acad-emy, edited by W G Runciman, J Maynard Smith,and R I M Dunbar. New York: Oxford UniversityPress.

Anderson J R, Myowa-Yamakoshi M, Matsuzawa T.2004. Contagious yawning in chimpanzees. Proceed-ings of the Royal Society of London B 271(Supple-ment):S468-S470.

Apte M L. 1985. Humor and Laughter: An AnthropologicalApproach. Ithaca (NY ): Cornell University Press.

Bachorowski J-A, Owren M J. 2001. Not all laughs arealike: voiced but not unvoiced laughter readily elic-its positive affect. Psychological Science 12:252-257.

Bachorowski J-A, Owren M J. 2003. Sounds of emo-tion: production and perception of affect- relatedvocal acoustics. Annals of the New York Academy ofSciences 1000:244-265.

Bachorowski J-A, Smoski M J, Owren M J. 2001. Theacoustic features of human laughter. Journal ofAcoustic Society of America 110:1581-1597.

Baenninger R. 1997. On yawning and its functions.Psychonomic Bulletin and Review 4:198-207.

Banning M R, Nelson D L. 1987. The effects of activity-elicited humor and group structure on group co-hesion and affective responses. American Journal ofOccupational Therapy 41:510-514.

Baron-Cohen S. 1999. The evolution of a theory ofmind. Pages 261-277 in The Descent of Mind: Psycho-

logical Perspectives on Hominid Evolution, edited by MC Corballis and S E G Lea. Oxford and New York:Oxford University Press.

Bekoff M. 1995. Play signals as punctuation: the struc-ture of social play in canids. Behaviour 132:419-429.

Bekoff M. 2001. Social play behaviour: cooperation,fairness, trust, and the evolution of morality. Jour-nal of Consciousness Studies 8:81-90.

Biben M, Champoux M. 1999. Play and stress: cortisolas a negative correlate of play in Saimiri. Play Cul-ture Studies 2:191-208.

Bjorklund D F, Pellegrini A D. 2002. The Origins of Hu-man Nature: Evolutionary Developmental Psychology.Washington (DC): American Psychological Asso-ciation.

Bock W J. 1959. Preadaptation and multiple evolution-ary pathways. Evolution 13:194-211.

Boehm C. 1999. Hierarchy in the Forest: The Evolution ofEgalitarian Behavior. Cambridge (MA): HarvardUniversity Press.

Boyd R, Richerson P J. 1985. Culture and the Evolution-ary Process. Chicago (IL): University of ChicagoPress.

Buck R, Ginsburg B. 1991. Spontaneous communica-tion and altruism: the communicative gene hy-pothesis. Pages 149-175 in Review of Personality andSocial Psychology, Volume 12, edited by M Clark.Newbury Park (CA): Sage Publications.

Buck R, VanLear C A. 2002. Verbal and nonverbalcommunication: distinguishing symbolic, sponta-neous, and pseudo-spontaneous nonverbal behav-ior. Journal of Communication 52:522-541.

Burling R. 1993. Primate calls, human language, andnonverbal communication. Current Anthropology34:25-53.


Butcher J, Whissell C. 1984. Laughter as a function ofaudience size, sex of the audience, and segmentsof the short film “Duck Soup.” Perceptual and MotorSkills 59:949-950.

Byers J A, Walker C. 1995. Refining the motor traininghypothesis for the evolution of play. American Nat-uralist 146:25-40.

Caron J E. 2002. From ethology to aesthetics: evolu-tion as a theoretical paradigm for research onlaughter, humor, and other comic phenomena.Humor 15:245-281.

Carr L, Iacoboni M, Dubeau M-C, Mazziotta J C, LenziG L. 2003. Neural mechanisms of empathy in hu-mans: a relay from neural systems of imitation tolimbic areas. Proceedings of the National Academy ofSciences USA 100:5497-5502.

Chapman A J. 1983. Humor and laughter in social in-teraction and some implications for humor re-search. Pages 135-157 in Handbook of Humor Re-search, Volume 1: Basic Issues, edited by P E McGheeand J H Goldstein. New York: Springer-Verlag.

Chartrand T L, Bargh J A. 1999. The chameleon ef-fect: the perception-behavior link and social inter-action. Journal of Personality and Social Psychology76:893-910.

Cosmides L, Tooby J. 1997. Evolutionary psychology:a primer. http://www.psych.ucsb.edu/research/cep/primer.html.

Cosmides L, Tooby J. 2000. Evolutionary psychologyand the emotions. Pages 91-115 in Handbook of Emo-tions, Second Edition, edited by M Lewis and J MHaviland-Jones. New York: The Guilford Press.

Damasio A R. 1994. Descarte’s Error: Emotion, Rea-son, and the Human Brain. New York: Putman.

Darwin C. 1872. The Expression of the Emotions in Manand Animals. London: John Murray. [Reprint.1972. New York: AMS Press.]

de Oliveira C R, Ruiz-Miranda C R, Kleiman D G, BeckB B. 2003. Play behavior in juvenile golden liontamerins (Callitrichidae: Primates): organizationin relation to costs. Ethology 109:593-612.

Deacon T. 1992. The neural circuitry underlying pri-mate calls and human language. Pages 121- 162 inLanguage Origin: A Multidisciplinary Approach, ed-ited by J Wind, B Chiarelli, B Bichakjian, and ANocentini. Dordrecht (The Netherlands) and Bos-ton (MA): Kluwer Academic Publishers.

Deacon T W. 1997. The Symbolic Species: The Co-evolutionof Language and the Brain. New York: W W Norton.

Devereux P G, Ginsburg G P. 2001. Sociality effects onthe production of laughter. Journal of General Psy-chology 128:227-240.

Donaldson T M, Newberry R C, Spinka M, Cloutier, S.2002. Effects of early play experience on play be-haviour of piglets after weaning. Applied Animal Be-haviour Science 79:221-231.

Drea C M, Hawk J E, Glickman S E. 1996. Aggression

decreases as play emerges in infant spotted hyae-nas: preparation for joining the clan. Animal Be-haviour 51:1323-1336.

Dugatkin L A, Reeve H K. 1994. Behavioral ecologyand levels of selection: dissolving the group selec-tion controversy. Pages 101-133 in Advances in theStudy of Behavior, Volume 23, edited by P J B Slater,J S Rosenblatt, C T Snowdon, and M Milinski. SanDiego (CA): Academic Press.

Dunbar R I M. 1993. Co-evolution of neocortex size,group size and language in humans. Behavioral andBrain Sciences 16:681-694.

Dunbar R I M. 1996. Grooming, Gossip, and the Evolutionof Language. Cambridge (MA): Harvard UniversityPress.

Edmonson M S. 1983. Notes on laughter. Anthropolog-ical Linguistics 29:23-33.

Eibl-Eibesfeldt I. 1989. Human Ethology. New York: Al-dine de Gruyter.

Ekman P. 1999. Basic emotions. Pages 45-60 in Hand-book of Cognition and Emotion, edited by T Dalgleishand M Power. New York: Wiley.

Fadiga L, Craighero L, Buccino G, Rizzolatti G. 2002.Short communication: speech listening specificallymodulates the excitability of tongue muscles: aTMS study. European Journal of Neuroscience 15:399-402.

Fagen R. 1993. Primate juveniles and primate play.Pages 182-196 in Juvenile Primates: Life History, De-velopment, and Behavior, edited by M E Pereira andL A Fairbanks. New York: Oxford University Press.

Fagen R, Fagen J. 2004. Juvenile survival and benefitsof play behaviour in brown bears, Ursus arctos. Evo-lutionary Ecology Research 6:89-102.

Ferrari P F, Gallese V, Rizzolatti G, Fogassi L. 2003.Mirror neurons responding to the observation ofingestive and communicative mouth actions in themonkey ventral premotor cortex. European Journalof Neuroscience 17:1703-1714.

Fitch W T. 2000. The evolution of speech: a compar-ative review. Trends in Cognitive Sciences 4:258-267.

Flack J C, Jeannotte L A, de Waal F B M. 2004. Playsignaling and the perception of social rules by ju-venile chimpanzees (Pan troglodytes). Journal of Com-parative Psychology 118:149-159.

Fogel A, Dickson K L, Hsu H, Messinger D, Nelson-Goens G C, Nwokah E. 1997. Communication ofsmiling and laughter in mother-infant play: re-search on emotion from a dynamic systems per-spective. Pages 5-24 in The Communication of Emo-tion: Current Research from Diverse Perspectives, editedby K C Barrett. San Francisco (CA): Jossey-BassPublishers.

Foley R A. 1996. An evolutionary and chronologicalframework for human social behaviour. Pages 95-117 in Evolution of Social Behaviour Patterns in Pri-mates and Man: A Joint Discussion Meeting of the Royal


Society and the British Academy, edited by W G Run-ciman, J Maynard Smith, and R I M Dunbar. Ox-ford and New York: Oxford University Press.

Foley R A. 1999. Hominid behavioural evolution: miss-ing links in comparative primate socioecology.Pages 363-386 in Comparative Primate Socioecology ed-ited by P C Lee. New York: Cambridge UniversityPress.

Frank S A. 1995. George Price’s contributions to evo-lutionary genetics. Journal of Theoretical Biology175:373-388.

Fredrickson B L. 1998. What good are positive emo-tions? Review of General Psychology 2:300-319.

Fridlund A J, Loftis J M. 1990. Relations between tick-ling and humorous laughter: preliminary supportfor the Darwin-Hecker hypothesis. Biological Psy-chology 30:141-150.

Fried I, Wilson C L, MacDonald K A, Behnke E J. 1998.Electric current stimulates laughter. Nature391:650.

Fry D P. 2005. Rough-and-tumble social play in hu-mans. Pages 54-85 in The Nature of Play: Great Apesand Humans, edited by A D Pellegrini and P KSmith. New York: Guilford Press.

Fry W F. 1977. The appeasement function of mirthfullaughter. Pages 23-26 in It’s a Funny Thing, Humour,edited by A J Chapman and H C Foot. Oxford andNew York: Pergamon Press.

Fry W F. 1994. The biology of humor. Humor 14:163-179.

Fry W F. 2002. Humor and the brain: a selective review.Humor 15:305-333.

Fuhr M. 2002. Coping humor in early adolescence.Humor 15:283-304.

Gallese V. 2003. The roots of empathy: the sharedmanifold hypothesis and the neural basis of inter-subjectivity. Psychopathology 36:171-180.

Gamble J. 2001. Humor in apes. Humor 14:163- 179.Gazzaniga M S, Smylie C S. 1990. Hemispheric mech-

anisms controlling voluntary and spontaneous fa-cial expressions. Journal of Cognitive Neuroscience2:239-245.

Goel V, Dolan R J. 2001. The functional anatomy ofhumor: segregating cognitive and affective com-ponents. Nature Neuroscience 4:237-238.

Goldman A I, Sripada C S. 2005. Simulationist modelsof face-based emotion recognition. Cognition94:193-213.

Goodson F E. 2003. The Evolution and Function of Cog-nition. Mahwah (NJ): Lawrence Erlbaum Associ-ates.

Gould S J, Vrba E S. 1982. Exaptation: a missing termin the science of form. Paleobiology 8:4-15.

Grammer K. 1990. Strangers meet: laughter and non-verbal signs of interest in opposite-sex encounters.Journal of Nonverbal Behavior 14:209-236.

Grammer K, Eibl-Eibesfeldt I. 1990. The ritualization

of laughter. Pages 192-214 in Naturlichkeit der spra-che und der kultur: Acta colloguii, edited by W Koch.Bochum (Germany): Brockmeyer.

Greatbatch D, Clark T. 2003. Displaying group cohe-siveness: humour and laughter in the public lec-tures of management gurus. Human Relations56:1515-1544.

Guard H J, Newman J D, Roberts R L. 2002. Morphineadministration selectively facilitates social play incommon marmosets. Developmental Psychobiology41:37-49.

Hammerstein P. 2003. Dahlem Workshop on Genetic andCultural Evolution of Cooperation (90). Cambridge(MA): MIT Press.

Hatfield E, Cacioppo J T, Rapson R L. 1994. EmotionalContagion. Cambridge and New York: CambridgeUniversity Press.

Harris C R. 1999. The mystery of ticklish laughter.American Scientist 87:344-352.

Hauser M D. 1996. The Evolution of Communication.Cambridge (MA): MIT Press.

Howe N E. 2002. The origin of humor. Medical Hy-potheses 59:252-254.

Irons,W. 2001. Religion as a hard-to-fake sign of com-mitment. Pages 292-309 in Evolution and the Capac-ity for Commitment, edited by R Nesse. New York:Russell Sage Foundation.

Iwase M, Ouchi Y, Okada H, Yokoyama C, NobezawaS, et al. 2002. Neural substrates of human facialexpression of pleasant emotion induced by comicfilms: a PET study. NeuroImage 17:758-768.

Jung W E. 2003. The inner eye theory of laughter:mindreader signals cooperator value. EvolutionaryPsychology 1:214-253.

Jurmain R, Nelson H, Kilgore L, Trevathan W. 2000.Introduction to Physical Anthropology, Eighth Edition.Belmont (CA): Wadsworth.

Kalinowski J, Saltuklaroglu T. 2003. Choral speech: theamelioration of stuttering via imitation and themirror neuronal system. Neuroscience and Biobehav-ioral Reviews 27:339-347.

Keltner D, Bonanno G A. 1997. A study of laughterand dissociation: distinct correlates of laughterand smiling during bereavement. Journal of Person-ality and Social Psychology 73:687-702.

Kipper S, Todt D. 2001. Variation of sound parametersaffects the evaluation of human laughter. Behaviour138:1161-1178.

Kipper S, Todt D. 2002. The use of vocal signals in thesocial play of Barbary macaques. Primates 43:3-17.

Kohler E, Keysers C, Umilta M A, Fogassi L, Gallese V,Rizzolatti G. 2002. Hearing sounds, understandingactions: action representation in mirror neurons.Science 297:846-48.

Kohler T A, Gumerman G J, editors. 2000. Dynamicsin Human and Primate Societies: Agent- based Modeling


of Social and Spatial Processes. New York: Oxford Uni-versity Press.

Koski L, Iacoboni M, Dubeau M-C, Woods R P, Maz-ziotta J C. 2003. Modulation of cortical activity dur-ing different imitative behaviors. Journal of Neuro-physiology 89:460-471.

Krebs J A, Dawkins R. 1984. Animal signals: mind-read-ing and manipulation. Pages 380-402 in Behav-ioural Ecology: An Evolutionary Approach, SecondEdition, edited by J R Krebs and N B Davies. Sun-derland (MA): Sinauer Associates.

Kuiper N A, Martin R A. 1998. Laughter and stress indaily life: relation to positive and negative affect.Motivation and Emotion 22:133-153.

LaGreca S, Laforest M, Alioua A, Vincent D. 1996.Laughter as interaction strategy: discursive andphonetic aspects. Pages 313-332 in SociolinguisticVariation: Data, Theory, and Analysis: Selected Papersfrom NWAV 23 at Stanford, edited by J Arnold. Stan-ford (CA): Center for the Study of Language andInformation.

Lakin J L, Jefferis V E, Cheng C M, Chartrand T L.2003. The chameleon effect as social glue: evi-dence for the evolutionary significance of noncon-scious mimicry. Journal of Nonverbal Behavior 27:145-162.

Lefcourt R M. 2000. Humor: The Psychology of LivingBuoyantly. New York: Plenum Publishers.

Leslie K R, Johnson-Frey S H, Grafton S T. 2004. Func-tional imaging of face and hand imitation: towardsa motor theory of empathy. NeuroImage 21:601-607.

Levy S G, Fenley W F, Jr. 1979. Audience size and thelikelihood and intensity of response during a hu-morous movie. Bulletin of the Psychonomic Society13:409-412.

Liberman A M, Whalen D H. 2000. On the relation ofspeech to language. Trends in Cognitive Sciences4:187-196.

Lloyd E A, Feldman M W. 2002. Evolutionary psy-chology: a view from evolutionary biology. Psycho-logical Inquiry 13:150-156.

Martin R A, Kuiper N A. 1999. Daily occurance oflaughter: Relationships with age, gender, and TypeA personality. Humor 12:355-384.

Maryanski A M, Molnar P, Segerstrale U, VelichkovskyB M. 1997. The social and biological foundationsof human communication. Pages 181-200 in Hu-man By Nature: Between Biology and the Social Sciences,edited by P Weingart, S D Mitchell, P J Richardson,and S Maasen. Mahwah (NJ): Lawrence ErlbaumAssociates.

Matsusaka T. 2004. When does play panting occur dur-ing social play in wild chimpanzees? Primates45:221-229.

McComas H C. 1923. The origin of laughter. Psycho-logical Review 30:45-55.

McGhee P E. 1976. A model of the origins and early

development of incongruity-based humour. Pages27-36 in It’s a Funny Thing, Humour, edited by A JChapman and H C Foot. Oxford and New York:Pergamon Press.

McGhee P E. 1979. Humor: Its Origin and Development.San Francisco (CA): W H Freeman.

McIntosh, D. N. 1996. Facial feedback hypotheses: Evi-dence, implications, and directions. Motivation andEmotion 20: 121-147.

Miller G. 2000. The Mating Mind: How Sexual ChoiceShaped the Evolution of Human Nature. New York: An-chor Books.

Mobbs D, Greicius M D, Abdel-Azim E, Menon V, ReissA L. 2003. Humor modulates the mesolimbic re-ward centers. Neuron 40:1041-1048.

Moran J M, Wig G S, Adams R B, Jr, Janata P, KelleyW M. 2004. Neural correlates of humor detectionand appreciation. NeuroImage 21:1055-1060.

Nesse R, editor. 2001. Evolution and the Capacity for Com-mitment. New York: Russell Sage Foundation Pub-lications.

Neuhoff C C, Schaefer C. 2002. Effects of laughing,smiling, and howling on mood. Psychological Reports91:1079-1080.

Osaka N, Osaka M, Kondo H, Morishita M, FukuyamaH, Shibasaki H. 2003. An emotion-based facial ex-pression word activates laughter module in the hu-man brain: a functional magnetic resonance im-aging study. Neuroscience Letters 340:127-130.

Owren M J, Bachorowski J-A. 2001. The evolution ofemotional expression: A “selfish gene” account ofsmiling and laughter in early hominids and hu-mans. Pages 152-191 in Emotions: Current Issues andFuture Directions, edited by T J Mayne and G A Bon-anno. New York: Guilford Press

Owren M J, Bachorowski J-A. 2003. Reconsidering theevolution of nonlinguistic communication: thecase of laughter. Journal of Nonverbal Behavior27:183-200.

Palagi E, Cordoni G, Borgognini Tarli S M. 2004. Im-mediate and delayed benefits of play behavior: newevidence from chimpanzees (Pan troglodytes). Ethol-ogy 110:949-962.

Panksepp J. 1998. Affective Neuroscience: The Foundationsof Human and Animal Emotions. New York: OxfordUniversity Press.

Panskeep J. 2000. The riddle of laughter: neural andpsychoevolutionary underpinnings of joy. CurrentDirections in Psychological Science 9:183-186.

Panksepp J, Burgdorf J. 2003. “Laughing” rats and theevolutionary antecedents of human joy?. Physiologyand Behavior 79:533-547.

Panksepp J, Panksepp J B. 2000. The seven sins of evo-lutionary psychology. Evolution and Cognition 6:108-131.

Paquette D, Carbonneau R, Dubeau D, Bigras M,Trembley R E. 2003. Prevalence of father- child


rough-and-tumble play and physical aggression inpreschool children. European Journal of Psychology ofEducation 18:171-189.

Parr L A. 2004. Perceptual biases for multimodal cuesin chimpanzee (Pan troglodytes) affect recognition.Animal Cognition 7:171-178.

Parr L A, Hopkins W D. 2000. Brain temperatureasymmetries and emotional perception in chim-panzees, Pan troglodytes. Physiology and Behavior71:363-371.

Pellis S M, Iwaniuk A N. 2000. Adult-adult play in pri-mates: comparative analyses of its origin, distribu-tion and evolution. Ethology 106:1083-1104.

Pellis S M, Pellis V C. 1996. On knowing it’s only play:the role of play signals in play fighting. Aggressionand Violent Behavior 1:249-268.

Pinker S. 1997. How the Mind Works. New York: W WNorton.

Platek S M, Mohamed F B, Gallup G G, Jr. 2005. Con-tagious yawning and the brain. Cognitive Brain Re-search 23:448-452.

Power C. 1998. Old wives’ tales: the gossip hypothesisand the reliability of cheap signals. Pages 111-129in Approaches to the Evolution of Language: Social andCognitive Bases, edited by J R Hurford, M Studdert-Kennedy, and C Knight. Cambridge: CambridgeUniversity Press.

Premack D, Woodruff G. 1978. Does the chimpanzeehave a “theory of mind”? Behavioral and Brain Sci-ences 4:515-526.

Preston S D, de Waal F B M. 2002. Empathy: its ulti-mate and proximate bases. Behavioral and Brain Sci-ences 25:1-72.

Preuschoft S, van Hooff J A R A M. 1997. The socialfunction of “smile” and “laughter”: variationsacross primate species and societies. Pages 171-189in Nonverbal Communication: Where Nature Meets Cul-ture, edited by U Segerstrale and P Molnar. Mah-wah (NJ): Lawrence Erlbaum Associates.

Provine R R. 1986. Yawning as stereotyped action pat-tern and releasing stimulus. Ethology 72:448-455.

Provine R R. 1991. Laughter: a stereotyped human vo-calization. Ethology 89:115-124.

Provine R R. 1992. Contagious laughter: laughter is asufficient stimulus for laughs and smiles. Bulletin ofthe Psychonomic Society 30:1-4.

Provine R R. 1993. Laughter punctuates speech: lin-guistic, social, and gender contexts of laughter.Ethology 95:291-298.

Provine R R. 1996. Laughter. American Scientist 84:38-45.

Provine R R. 2000. Laughter: A Scientific Investigation.New York: Viking.

Provine R R, Fischer K R. 1989. Laughing, smiling, andtalking: relation to sleeping and social context inhumans. Ethology 83:295-305.

Ramachandran V S. 1998. The neurology and evolu-tion of humor, laughter, and smiling: the Falsealarm theory. Medical Hypotheses 51:351-354.

Reddy V, Williams E, Vaughan A. 2002. Sharing hu-mour and laughter in autism and Down’s syn-drome. British Journal of Psychology 93:219-242.

Rendall D, Owren M J. 2002. Animal vocal communi-cation: say what? Pages 307-313 in The Cognitive Ani-mal: Empirical and Theoretical Perspectives on AnimalCognition, edited by M Bekoff, C Allen, G M Bur-ghardt. Cambridge (MA): MIT Press.

Richerson P J, Boyd R. 1999. Complex societies: theevolutionary origins of a crude superorganism.Human Nature 10:253-289.

Richerson P J, Boyd R. 2004. Not by Genes Alone: HowCulture Transformed Human Evolution. Chicago (IL):University of Chicago Press.

Rizzolatti G, Arbib M A. 1998. Language within ourgrasp. Trends in Neuroscience 21:188-194.

Rizzolatti G, Craighero L. 2004. The mirror-neuronsystem. Annual Review of Neuroscience 27:169-192.

Rizzolatti G, Craighero L, Fadiga L. 2002. The mirrorsystem in humans. Pages 38-59 in Mirror Neuronsand the Evolution of Brain and Language, edited byM I Stamenov and V Gallese. Philadelphia (PA):John Benjamins Publishing.

Rizzolatti G, Fadiga L, Fogassi L, Gallese V. 1999. Res-onance behaviors and mirror neurons. Archives It-aliennes de Biologie 137:85-100.

Rogers S J, Hepburn S L, Stackhouse T, Wehner E.2003. Imitation performance in toddlers with au-tism and those with other developmental disor-ders. Journal of Child Psychology and Psychiatry44:763-781.

Rosner F. 2002. Therapeutic efficacy of laughter inmedicine. Cancer Investigation 20:434-436.

Rozin P, Haidt J, McCauley C, Imada S. 1997. The cul-tural evolution of disgust. Pages 65-82 in Food Pref-erences and Taste: Continuity and Change, edited byH Macbeth. Providence (RI): Berghahn Books.

Ruch W. 1993. Exhilaration and humor. Pages 605-616in Handbook of Emotions, edited by M Lewis and JM Haviland. New York: Guilford Press.

Russell J A, Bachorowski J-A, Fernandez-Dols J-M.2003. Facial and vocal expressions of emotion. An-nual Review of Psychology 54:329-349.

Sander K, Scheich H. 2001. Auditory perception oflaughing and crying activates human amygdala re-gardless of attentional state. Cognitive Brain Research12:181-198.

Saxe R, Carey S, Kanwisher N. 2004. Understandingother minds: linking developmental psychologyand functional neuroimaging. Annual Review of Psy-chology 55:87-124.

Scheiner E, Hammerschmidt K, Jurgens U, Zwirner P.2002. Acoustic analyses of developmental changes


and emotional expression in the preverbal vocali-zations of infants. Journal of Voice 16:509-529.

Schurmann M, Hesse M D, Stephan K E, Saarela M,Zilles K, Hari R, Fink G R. 2005. Yearning to yawn:the neural basis of contagious yawning. NeuroImage24:1260-1264.

Seligman M E P, Csikszentmihalyi M. 2000. Positivepsychology: an introduction. American Psychologist55:5-14.

Sharpe L L, Clutton-Brock T H, Brotherton P N M,Cameron E Z, Cherry M I. 2002. Experimental pro-visioning increases play in free-ranging meerkats.Animal Behaviour 64:113-121.

Shibata D, Zhong J. 2001. Humour and Laughter: lo-calization with fMRI. NeuroImage 13:S476.

Shultz T R. 1976. A cross-cultural study of the structureof humour. Pages 175-180 in It’s a Funny Thing,Humour, edited by A J Chapman and H C Foot.Oxford and New York: Pergamon Press.

Smoski M J, Bachorowski J-A. 2003a. Antiphonallaughter between friends and strangers. Cognitionand Emotion 17:327-340.

Smoski M J, Bachorowski J-A. 2003b. Antiphonallaughter in developing friendships. Annals of theNew York Academy of Sciences 1000:300-303.

Sober E, Wilson D S. 1998. Unto Others: The Evolutionand Psychology of Unselfish Behavior. Cambridge(MA): Harvard University Press.

Sosis R. 2004. The adaptive value of religious ritual.American Scientist 92:166-172.

Spinka M, Newberry R C, Bekoff M. 2001. Mammalianplay: training for the unexpected. Quarterly Reviewof Biology 76:141-168.

Spoor J R, Kelly J R. 2004. The evolutionary signifi-cance of affect in groups: communication andgroup bonding. Group Processes and Intergroup Re-lations 7:398-412.

Sroufe L A, Waters E. 1976. The ontogenesis of smilingand laughter: a perspective on the organizationand development in infancy. Psychological Review83:173-189.

Stamenov M I, Gallese V, editors. 2002. Mirror Neuronsand the Evolution of Brain and Language. Philadel-phia (PA): John Benjamins Publishing.

Storey R. 2003. Humor and sexual selection. HumanNature 14:319-336.

Tomasello M, Carpenter M, Call J, Behne T, Moll H.in press. Understanding and sharing intentions:T�the origins of cultural cognition. Behavioral andBrain Sciences.

Tooby J, Cosmides, L. 1990. The past explains the pres-ent: emotional adaptations and the structure of an-cestral environments. Ethology and Sociobiology11:375-424.

Tugade M M, Fredrickson B L. 2004. Resilient individ-uals use positive emotions to bounce back from

negative emotional experiences. Journal of Person-ality and Social Psychology 86:320-333.

Vaid J. 2002. Humor and laughter. Pages 505-516 inEncyclopedia of the Human Brain, Volume 2, editedby V S Ramachandran. San Diego (CA): AcademicPress.

Van den Berg C L, Van Ree J M, Spruijt B M. 2000.Morphine attenuates the effects of juvenile isola-tion in rats. Neuropharmacology 39:969-976.

van Hooff J A R A M. 1972. A comparative approachto the phylogeny of laughter and smiling. Pages209-243 in Nonverbal Communication, edited by R AHinde. Cambridge: Cambridge University Press.

van Hooff J A R A M, Preuschoft S. 2003. Laughterand smiling: the intertwining of nature and cul-ture. Pages 261-287 in Animal Social Complexity: In-telligence, Culture, and Individualized Societies, editedby F B M de Waal and P L Tyack. Cambridge (MA):Harvard University Press.

Vettin J, Todt D. 2004. Laughter in conversation: fea-tures of occurrence and acoustic structure. Journalof Nonverbal Behavior 28:93-115.

Vinton K L. 1989. Humor in the work place: is it moretelling than jokes? Small Group Behavior 20:151-166.

Watkins K E, Strafella A P, Paus T. 2003. Seeing andhearing speech excites the motor system involvedin speech production. Neuropsychologia 41:989-994.

Weisfeld G E. 1994. The adaptive value of humor andlaughter. Ethology and Sociobiology 14:141-169.

Wicker B, Keysers C, Plailly J, Royet J-P, Gallese V, Riz-zolatti G. 2003. Both of us disgusted in my insula:the common neural basis of seeing and feeling dis-gust. Neuron 40:655-664.

Wild B, Erb M, Eyb M, Bartels M, Grodd W. 2003. Whyare smiles contagious? An fMRI study of the inter-action between perception of facial affect and fa-cial movements. Psychiatry Research: Neuroimaging123:17-36.

Wild B, Rodden F A, Grodd W, Ruch W. 2003. Neuralcorrelates of laughter and humour. Brain126:2121-2138.

Williams G C. 1966. Adaptation and Natural Selection: ACritique of Some Current Evolutionary Thought. Prince-ton (NJ): Princeton University Press.

Williams J H G, Whiten A, Singh T. 2004. A systematicreview of action imitation in autistic spectrum dis-order. Journal of Autism and Developmental Disorders34:285-299.

Williams J H G, Whiten A, Suddendorf T, Perrett D I.2001. Imitation, mirror neurons, and autism. Neu-roscience and Biobehavioral Reviews 25:287-295.

Wilson D S. 2001. Cooperation and altruism. Pages222-231 in Evolutionary Ecology: Concepts and CaseStudies, edited by C W Fox, D A Roff, D J Fairbairn.Oxford: Oxford University Press.


Wilson D S. 2004. What is wrong with absolute indi-vidual fitness? Trends in Ecology and Evolution19:245-248.

Wilson D S, Sober E. 1994. Reintroducing group se-lection to the human behavioral sciences. Behav-ioral and Brain Sciences 17:585-654.

Wilson E O. 1998. Consilience: The Unity of Knowledge.Thorndike (ME): Thorndike Press.

Wilson S M, Saygin A P, Sereno M I, Iacoboni, M. 2004.Listening to speech activates motor areas involvedin speech production. Nature Neuroscience 7:701-702.

THE QUARTERLY REVIEW OF BIOLOGYPage Charge and Offprint Order Form

This form must be returned with your manuscript and a corrected set of proofs within 48 hours of receipt. Authors are entitled

to 10 complimentary copies of the journal issue or a year’s subscription (or renewal) free of charge; however, we cannot send

them until we have received this form. Please see the reverse of this form for a list of offprint prices.

PAGE CHARGESAuthors who attribute support to funding agencies will be expected to meet charges of $40 per journal page for each of the first

20 journal pages, and $80 for each journal page in excess of 20. Authors not supported by funding agencies and authors of

invited articles are eligible for 20 free journal pages (approximately 50 double-spaced typewritten manuscript pages or their

equivalent in bibliographic or illustrative matter). When such articles are longer than 20 journal pages, authors will be charged

$80 for each page above the allotment of 20 free journal pages. These charges will be used for direct support of the issue in

which the article appears.

Author(s)

Article title: Page length of article:

In order to determine the length of your article, please count the number of paged galleys. The cost of author alterations is

chargeable to the author and will be invoiced after the article is published, along with the cost of offprints and page charges.

OFFPRINTS

Check Indicate Quantity* Check only one:

( ) offprints without covers ( ) 10 free copies of the journal issue

( ) offprints with covers OR

( ) I do not want offprints ( ) free year’s subscription or renewal

*Order offprints in multiples of 50

SHIP TO: BILL TO (for page charges and offprints):

PURCHASE ORDER NUMBER: ( ) ATTACHED ( ) WILL FOLLOW

Signature:

Daytime telephone number: FAX:

OFFPRINT

PRICE LIST

Offprint prices are in dollar amounts.

The cost for covers (one color, half-title) is in addition to the cost of offprints.

For longer articles, combine rates; e.g., 30 pages = 24 + 6

SHIPPINGOffprints are shipped within 21 days of the printing of the issue. The above prices include postage for domestic orders.

CANADIAN ORDERS: For each 50 copies, add $3.50 to the total. OTHER FOREIGN ORDERS: For each 50 copies, add $5.50

to the total.

LATE ORDERSThe prices listed above apply only to orders placed by authors before publication of the journal issue. After the journal is

published, estimates for reprints in multiples of 100 may be obtained from Journals Production, The University of Chicago

Press, 1427 E. 60th

Street, Chicago, IL 60637, Attention: Reprints Desk.

INSTITUTIONAL ORDERSPlease list VENDOR as: The Quarterly Review of Biology, The University of Chicago Press, 11030 South Langley Avenue,

Chicago, IL 60628. If a purchase order is sent separately from this order form, it must be received before the publication of the

journal issue and must be marked “Confirming.” All purchase orders must include journal name, issue date, author’s name, and

the number of offprints and covers desired. Institutional orders without a purchase order will not be processed.

CREDIT CARD ORDERSPayment for page charges and offprints can be made with Mastercard or Visa. Please complete all information requested below

and include your signature and the expiration date of your card.

Cardholder name as it appears on card:

Card number:

Expiration date:

Signature:

Your Mastercard or Visa account will be charged after shipment of your offprints.

1-4 5-8 9-12 13-16 17-20 21-24 add’l 4 pages COVERS

25 $26 $30 $35 $41 $46 $50 $6 $14.00

50 $34 $40 $46 $53 $59 $66 $7 $18.00

100 $52 $65 $77 $91 $103 $116 $13 $35.00

150 $67 $86 $104 $125 $143 $163 $19 $53.00

200 $87 $111 $135 $162 $174 $202 $25 $70.00

250 $116 $147 $177 $211 $241 $276 $31 $88.00

300 $138 $176 $212 $253 $289 $330 $37 $105.00

350 $161 $204 $243 $294 $336 $373 $43 $123.00

400 $183 $233 $329 $337 $385 $440 $49 $140.00

Number of pages

Qu

anti

ty

volume no. 2005 the quarterly review - semantic scholarthe quarterly review of biology, december...

Documents