ANRV331-PS59-11 ARI 4November2007 20:27 Putting the Altruism y. AR Further nl o se Click here for quick links to Back into Altruism: al u Annual Reviews content online, n including: o ews.org0. For pers •••• TTOAooRtpph’s e dccrioot aewmrdtnpi calrolreetahiscd eilenends st hiavirest ivsceolealusrcmhe The Evolution of Empathy ualrevi4/29/1 Frans B.M. de Waal nals.annRY on 0 LEimvionrgyLUinnkivse,rYseitryk,eAstNlanattiao,nGalePorrgimiaa3te03R2e2s;eeamrcahilC:deenwteara,[email protected], ourRA om arjN LIB ded frELDI aK nloMC w DoK - 00. AR Annu.Rev.Psychol.2008.59:279–300 KeyWords 3P 79-GE FirstpublishedonlineasaReviewinAdvanceon perception-action,perspective-taking,prosocialbehavior, hol. 2008.59:2ND - COLLE hTJTuthhtnpiees:/A5a/r,npt2nsiyc0ulc0aehl7’s.Radneovniiu:ewalroefvPiesywchs.oolorggyisonlineat cAEovboospltuertraiaoctnitoanry theory postulates that altruistic behavior evolved for cA PsyYL 10.1146/annurev.psych.59.103006.093625 thereturn-benefitsitbearstheperformer.Forreturn-benefitstoplay ev. AR Copyright(cid:2)c 2008byAnnualReviews. amotivationalrole,however,theyneedtobeexperiencedbytheor- u. RF M Allrightsreserved ganism.Motivationalanalysesshouldrestrictthemselves,therefore, AnnY O 0066-4308/08/0203-0279$20.00 to the altruistic impulse and its knowable consequences. Empathy T is an ideal candidate mechanism to underlie so-called directed al- SI R truism,i.e.,altruisminresponsetoanother’spain,need,ordistress. E V NI Evidenceisaccumulatingthatthismechanismisphylogeneticallyan- U cient,probablyasoldasmammalsandbirds.Perceptionoftheemo- y b tionalstateofanotherautomaticallyactivatessharedrepresentations causingamatchingemotionalstateintheobserver.Withincreasing cognition,state-matchingevolvedintomorecomplexforms,includ- ingconcernfortheotherandperspective-taking.Empathy-induced altruismderivesitsstrengthfromtheemotionalstakeitoffersthe selfintheother’swelfare.Thedynamicsoftheempathymechanism agree with predictions from kin selection and reciprocal altruism theory. 279 ANRV331-PS59-11 ARI 4November2007 20:27 “Youbeginwiththeeffectofbehavioronac- Contents torsandrecipients;youdealwiththeproblem of internal motivation, which is a secondary INTRODUCTION................. 280 problem,afterward....[I]fyoustartwithmo- ORIGINOFEMPATHY............ 282 tivation, you have given up the evolutionary LEVELSOFEMPATHY........... 282 analysisattheoutset.” EmotionalContagion............. 282 Thisisaperfectlylegitimatestrategythat SympatheticConcern............. 283 y. has yielded profound insights into the evo- nl EmpathicPerspective-Taking..... 285 o lutionofaltruism(e.g.,Dugatkin2006).Un- e UNDERLYINGMECHANISMS... 286 s u fortunately,however,theseinsightshavenot al PerceptionActionMechanism.... 286 n come with a new terminology: Evolutionary so RussianDollModel.............. 287 nals.annualreviews.orgRY on 04/29/10. For per FERMOAESEPyLmmMAmTToppREtaHiatoUMhtYhniIcPeaSAltAPMiSCTcerEoCH.sn.VpoY.tenOa.ccgT.tLe.iiOorvV.nne.E.-..T.D...a...k.....i.n....g........................ 222288889898 bTlscaneeaioshoceltslokuleosstosdhgt,fbye“tewaoannpalhteceterfithrautosctehiirstsseiattfoirisocncntrot”ioneririnitsfsnudresoieecltiftgdna.blagtSlehtreoidnmedmleeboaifi“slectastsitnservooellaaryfifift,isdtrowaheennt”hclhaiieabeplrcteiehtteoregienaotrarhttmnreeodalisryr-st.. ourRA PROXIMATEMECHANISMOF The prototypical altruist is a honeybee that ded from arjELDIN LIB CODNICRLEUCTSIEODNA.L..T.R..U..I.S.M.................. 229912 stmeticontrgetshlaeinkheiilnvytera—ugdgeervree—snsistvahecortuihfigachninhbgeerhnemigrnoli.tfieTvathtoiiospnruosi--s aK 300. DownloPARK - MC Sympatrheyg.a.r.dceadnaSnsmoati,tshienl(fi1asn7h5yp9sr,eipnn.sce3ip,1bl7ee). atinegngeTt(ioSmhfoeetbhsehecriojt&anecflrkmWiicnstigsl“swsooeinftlfih1mst9hho9”et8iiav)rn.avdteiro“nnaaaltclruutleiasrrtmimci”enaoonfl--- 79-GE Empathy may be uniquely well suited for ogy by evolutionary biologists has been un- 008.59:2COLLE A(dbelitfiornuloiitgsimiocna)l: bisrmid,gsiinngcetihtehgasapthbeeptwroepeenrteygoofistmranasnfdoramltirnug- hpeerlpsfeu.lTfohrecowmaymtuoncicleaatirounpabthoeutcomnofutisviaotnioins hol. 2ND - binechraevaisoerstthhaet anotherperson’smisfortunfeeeliinntgooofndei’sstorwesns. teovodloutwiohnaatryTrainvealryssdesidrewqhueinrehtehdateceifdfeecdtsthbaet cA PsyYL recipient’sfitnessata Hoffman(1981a,p.133) considered separate from motivation. Con- ev. AR costtothe versely, motivational analyses require us to RM performers Annu. Y OF Ugoltailm: athteebcaeunesefitosran INTRODUCTION kcoeenpsidmeroattiivoantiso.nItsiespnaortatfeorfnroomthienvgotlhuatitobniaorly- T Discussionsofaltruisticbehaviortendtosuf- ogistshammeronthedistinctionbetweenul- SI organismoritsclose ER kinderivefroma ferfromalackofdistinctionbetweenfunction timateandproximate(Mayr1961,Tinbergen V NI behavior,hencethe andmotivation.Thisisduetothecontrasting 1963). The ultimate cause refers to why a U probablereasonwhy emphasisofbiologistsandpsychologists,with behavior evolved over thousands of gener- by thebehaviorwas theformerfocusingonwhataparticularbe- ations, which depends on its fitness conse- favoredbynatural havior is good for, and the latter on how it quences. The proximate cause, on the other selection comesabout. hand, refers to the immediate situation that Proximatecause: Evolutionaryexplanationsarebuiltaround triggers behavior, and the role of learning, situationthat triggersbehaviorand theprinciplethatallthatnaturalselectioncan physiology, and neural processes—typically themechanism workwitharetheeffectsofbehavior—notthe thedomainofpsychologists. (psychological, motivationbehindit.Thismeansthereisonly Proximateandultimateviewpointsdoin- neural,physiological) onelogicalstartingpointforevolutionaryac- form each other, yet are not to be con- thatenablesit counts, as explained by Trivers (2002, p. 6): flated. For example, primate cooperation is 280 deWaal ANRV331-PS59-11 ARI 4November2007 20:27 promotedbysocialtolerance.Throughitsef- trinsically rewarding qualities in that it of- fectonfood-sharing,toleranceevensoutpay- fers the actor an emotional stake in the re- offdistributions(deWaal&Davis2003,Melis cipient’swell-being,i.e.,ifhelpingtheother Directedaltruism: et al. 2006). Tolerance likely is a proximate ameliorates the helper’s internal state (see helpingor mechanismthatevolvedtoservetheultimate Empathy as Evolved Proximate Mechanism, comfortingbehavior goalofcooperation,whichistoyieldbenefits below).Extrinsicrewards,ontheotherhand, directedatan forallcontributors. arelesslikelytoplayarole.Bydefinition,al- individualinneed, y. Cooperation and altruistic behavior are truismcarriesaninitialcost,andpositivecon- pain,ordistress nl o thoughttohaveevolvedtohelpfamilymem- sequencesoccuronlyafterasignificanttime Intentional e s altruism: the u bers and those inclined to return the favor interval (e.g., the recipient reciprocates) or al altruistdeliberately n (Hamilton1964,Trivers1971).Regardlessof notatall(e.g.,carefordependentkin),making o seekstobenefit s gper whetherthisisthewholeexplanationornot forratherpoorlearningconditions. eithertheother ews.or0. For (2s0e0e5S),obthere&poDinSt Wis itlhsoant u19lt9im8,atEeOacWcoiulsnotns quiIrnettehnetioacntaolrlytoseelxfipslhiciatlltyrueixspmectwootuhledrsrteo- (ailntrtuenisttiiocnaalltlryuism)or nualrevi04/29/1 sqtureensscerseftourrtnh-ebpenereffiotrsm, eir.ea.n,dp/oorsiittisvkeinc.oInnsaes-- rfoetrusrunchtheexfpaevcotra.tDioensspiinteatnhiemlaalcsk,tohfeeyvaidreenocfe- isteslefilsfh(ianltternutisiomn)ally ded from arjournals.anELDIN LIBRARY on mhisinfsioedmauewenc,dryehi,,.vmeptaeh.soar,e,iratnyiehlt,teprciosulslaeoriuyssbd.tneiTeilccsynlthberewiaeefisrhshtbssaaw.etvmhTcidooaahrmrtyievmarebeiesmsoectadeqlireivdueraaeiartttcitehiotforendwetdeaheellaeawrlcrytoosraelnuyiden--s,, tat(2seehtr0r.neag0uta.3et,sh,sasenKDuriemoamacgnwieapaldklynrs.ionTt2ccrs0iauhtr0lye1ey09ca,oa7asSlbm6taior,lmukmuFitoseoetnttaihicracvrellas.atp&i2rimeoe0ctn0iutF.e5hrisI)ans,t,ct-smhahbisbneuisasncmucceemohafianneetlr-sssl EaceaEcanamlmmtrporepatuppachbiaatiesothttmryhhrynyyt:wo-:fbhri(ttoaaehhm)slepebdaend 9-300. DownloaE PARK - MCK inw1.hittAceohlrtberdesuitgiresgedticincthegeidmolprapildnturiglsusteiars.enmsSdspmscoiaganrnyitnaacnlgosemoionruetrshae,beadocsitiusgiiotnh:n-t twmdhiiaaattthteelasayneilbfmfue-lgnalalesaifinptepantrrhgeeeacimginaedt.ieHoinenderlocpeoffcumihplomraowcoctnasiltf(oDewrxuiclighlmaaumtnkltgeiine-- atotohhfftfheeeaecnerrtemo’esatdoshsottebainroyt,sen(af,abnoa)ldrnsattdsshashteeaesrse chol. 2008.59:27AND - COLLEG 2. otoLifuoetanacnreonodmethdreeesasrpfltoionrrunptishsaemeinr.aeocHintrofenorl.eprecindegd. absyapcoosintidvie- 1hctnio9aemgvc9eitn7oi)rttt,hihbvteayuemtpthittcohhwareeliiltzryhoeortnruget,hramnine.a-eisbi.onm,eronibgceieiscfinyutoausrlnnosdlafoikcatteldh.tliryesTutaaithsonntiiitsmclcyobaanpeiln’---s (ooTeecxxtr)hhtihsieiedtsnesrer,ddniaensptdfiiebmfonreyspiyaptwtnoiieoninyctngdhtivhwtehihs.eat PsyYL 3. Intentionalaltruism.Helpbasedonthe pliestomostreciprocalaltruismintheanimal animals,buttheterm ev. AR prediction of behavioral effects. One kingdom. “empathy”inthe RM prediction could be that the help will presentreview Annu. TY OF bperordeucciperoacnaetetdb,ehneenficte.tShiantcethtehaectacwtoilrl mcoomOtievnsacdteiioscneoavnloanluvetecodtne,dombfreyoh,mai.veii.to,srituslotmifmtoetanitveaagtsioosaunlmsb.eeA-s acpriptelireisonev(ean)iisfmonelty RSI seeks to benefit itself, we may call this goodexampleissexualbehavior,whicharose Motivational E autonomy: V intentionally selfish altruism. The sec- NI ondpossibilityishelpbasedonanap- toservereproduction.Sinceanimalsare,asfar independenceof U asweknow,unawareofthelinkbetweensex motivationfrom by preciation of how one’s own behavior and reproduction, they must be engaging in ultimategoals willhelptheother.Sincetheactorseeks sex(asdohumansmuchofthetime)without tobenefittheother,wemaycallthisin- progenyinmind.Justassexcannotbemoti- tentionallyaltruisticaltruism. vated by unforeseen consequences, altruistic Some directed altruistic behavior is pro- behaviorcannotbemotivatedbyunforeseen moted by built-in rewards, such as the payoffs. oxytocin release during suckling that may Thealtruisticimpulseistobetakenvery underpin maternal care (Panksepp 1998). seriously, therefore, because even if altruis- Empathy-basedaltruismmayhavesimilarin- tic behavior were partially learned based on www.annualreviews.org • TheEvolutionofEmpathy 281 ANRV331-PS59-11 ARI 4November2007 20:27 short-termintrinsicrewardsorlong-termex- cognition is often critical, it is a secondary trinsicrewards,thisbynomeansrulesoutthe development. As noted by Hoffman (1981b, altruisticimpulse.Infact,itpresupposesthis p.79),“[H]umansmustbeequippedbiologi- Perception-action impulsegiventhatabehavior’sconsequences callytofunctioneffectivelyinmanysocialsit- mechanism(PAM): automaticallyand cannotbelearnedwithoutspontaneouslyen- uations without undue reliance on cognitive unconsciously gaginginitinthefirstplace. processes.” activatedneural This review seeks to restore the altruism The selection pressure to evolve rapid y. representationsof withinaltruismbyexploringtheroleofem- emotionalconnectednesslikelystartedinthe nl statesinthesubject o pathyinthedirectedaltruismofhumansand context of parental care long before our e similartothose s al u perceivedinthe other animals. Some definitions of empathy speciesevolved(Eibl-Eibesfeldt1974[1971], n object stressthesharingofemotions,whereasother MacLean1985).Signalingtheirstatethrough o s gper Emotional definitions stress the capacity to put oneself smilingandcrying,humaninfantsurgetheir ews.or0. For cemonottaiognioanl: ianretostohetoopt-hdeor’ws“ns,hhooews.e”vTerh,ethlaatttetrhdeeyfidniistcioonn-s cTahreogmivaenr 1to995c,omBeowilnbtyo 1a9c5ti8o)n. E(Aqcueibvaolen&t nals.annualreviRY on 04/29/1 ssotubabjtejeec-ctmtawticthhinangofa nWacdleuoecdtpiftneoigmnllgompwtaehtrheeayberbfmroooaotmdttoieomsintts-apulpopsosessniasbpsiblipetlirevdoieaatfiycnhntteoitciienoodstnhte,eneairtndss-.., mrwpeaaeprrcermhonaditnsnugicsatmolieofsrnttohrtpeeoelyireoaasntueodnngina.ffeAfaeelvdlciaitanenndgio,mrbcalymelsaatnhimnienmigwr,ahloaiinfacfnhd- ourRA Wefirstconsiderthevariouslevelsofempathy spring’s needs likely out-reproduced those ded from arjELDIN LIB iaPncrteiasontnoimnam&lsecdahneadnWitshamaelu((2nP0dA0eM2rlay))i.nAgpfrtopeperroctsehepdist,iowbnye- wcohuoOldrnebcmeeaaitpnhpeedlieeidnmdopifuafttehsriiedcnetc.tahpearceitayrinegxisctoendt,exitt aK nloMC explore the relation between empathy and and play a role in the wider network of so- 300. DowPARK - aisltrlAiukiesmlmya.tjoorhqauveestsieolenctiesdwehmepthaethryeavsolpurtiooxn- ctaaitainlthrdeeilsactrtoienosntsisvnhouicepadsl.izsTuathriveoinvfaasclintvttaohluaatedumoltfahmeommodpaalhstihnryet--s 9-E 7G imate mechanism to generate directed al- inducing signals. For example, primates of- 008.59:2COLLE ttrhueisdmir.ecDtiooenstheamtpeavtohlyuticohnaanrnyetlhaelotrryuiwsmouilnd t(Benoeliscckha1n9d92c)le,awnhtichheiwsosouncdristiocfalcfoonrspheecailfiincgs hol. 2ND - pkerepdtitcetm? Spoo,raervileynstehpoauragthemfroomtivaetvioolnutwioilnlabrye tthematpatdsutoltemntaeleramnaecwaqgureosuipnjoufrteedndteumripnograart-- cA PsyYL considerations,intheendthetwowillmeet. ily return to their native group, where they ev. AR Empathymaybemotivationallyautonomous, aremorelikelytoreceivethisservice(Dittus RM u. F butitstillneedstoproduce—onaverageand &Ratnayeke1989). AnnY O inthelongrun—evolutionarilyadvantageous T outcomes. The central thesis to be argued SI LEVELSOFEMPATHY R here,then,isthatempathyevolvedinanimals E V NI asthemainproximatemechanismfordirected EmotionalContagion U altruism,andthatitcausesaltruismtobedis- y Thelowestcommondenominatorofallem- b pensed in accordance with predictions from pathic processes is that one party is affected kinselectionandreciprocalaltruismtheory. byanother’semotionalorarousalstate.This broad perspective on empathy, which goes ORIGINOFEMPATHY backasfarasLipps(1903),leadsonetorec- Empathyallowsonetoquicklyandautomat- ognizecontinuitybetweenhumansandother icallyrelatetotheemotionalstatesofothers, animals as well as between human adults whichisessentialfortheregulationofsocial and young children. Emotional connected- interactions, coordinated activity, and coop- nessinhumansissocommon,startssoearly eration toward shared goals. Even though in life (e.g., Hoffman 1975, Zahn-Waxler & 282 deWaal ANRV331-PS59-11 ARI 4November2007 20:27 Radke-Yarrow 1990), and shows neural and their own response to pain (Langford et al. physiological correlates (e.g., Adolphs et al. 2006). 1994, Decety & Chaminade 2003a, Rimm- Miller et al. (1959) published the first of Sympathetic Kaufman&Kagan1996)aswellasagenetic a series of pioneering studies on the trans- concern: concern substrate (Plomin et al. 1993), that it would mission of affect in rhesus macaques. These aboutanother’sstate bestrangeindeedifnocontinuitywithother monkeystendtoterminateprojectedpictures andattemptsto species existed. Evolutionary continuity be- of conspecifics in a fearful pose even more amelioratethisstate y. tween humans and apes is reflected in the rapidly than negatively conditioned stimuli. (e.g.,consolation) nl o similarityofemotionalcommunication(Parr Perhaps the most compelling evidence for Cognitiveempathy: e s empathycombined u &Waller2007)aswellassimilarchangesin emotional contagion came from Wechkin al withcontextual n brainandperipheralskintemperatureinre- etal.(1964)andMassermanetal.(1964),who o appraisalandan s gper sponse to emotionally charged images (Parr foundthatmonkeysrefusetopullachainthat understandingof ews.or0. For 200A1,flPoacrkr&ofHbiordpskitnaski2n0g01o)f.f all at once be- deleelcivtreircssfhooocdkttooathnedmtriigfgdeorisnpgaisnordeealcitvieornssainn wobhjaetctc’sauesmedottihoenal nualrevi04/29/1 creafluseex-loinkee,ahmigohnlygatdhaepmtiviesssptarertaldeidngshoofwfesaar acocnocmerpnanfoiornth.Weohthetehre(rsetehebierlosawc)rirfiecmearienflseucnt-s sPteatresonaldistress: ded from arjournals.anELDIN LIBRARY on twwotthoruhheatcsaetsrtncymrt(,ayHrtaiihrgnyooeggfronefembmroeeitasdcfniauantunlh1vlse9oaeou7lfivotn5ehon)it.mueiamaAnaaltytamirncetouhasnnncpeegdtriweecotorabnhsdro.eteiarmnSnnogisdmfsoibtnitafluhgardrtresielosstyde-sf, caSTlvyehoameird,npahenaoxcttweheeeovvofteilaruc,vtaeCisorsinoitvanmercyivegsithcreantpraioloscuocsbuaerrsoewxuphsalealn.ineemdoa-s sbwdeoiisltrfth-nrceaefsnrnsootmethreeedrm’sdpiastthreyss aK nloMC processesisadoption—inwholeorinpart— tional contagion is combined with appraisal 300. DowPARK - occoofnnattnaaoggtiioohnner’(siHseanmtofiotetlidaolnweatalyasslt.aat1e9,p9ai3.ses)..i,vEeemmpoorottiicooennsaasll, osWtfaatnhadle(ot1ht9he9ec6ra’)susssipetueoaafktisothnoefano“dtchoaegtrtn’esimteivmpetosettimoonupsna.tdhDeyre”- 9-E 7G though:Theobjectoftenaimstoemotionally when the empathic reaction includes such 008.59:2COLLE ateffmecptetrhetasnutbrujemcts,soufchyoausnthgeaepxetsrewmheelynntohiesyy conTtehxetupaslyacphporlaoigsaicl.alliteraturedistinguishes hol. 2ND - ahruemabneincghilrderjeencte(dPodteugrainlg20w0e0a)n,inthge.yLeikxe- sthymeirpasothcyialfcroomnsepqeuresnocneasladreisetraecshs,otwhheirc’shopin- cA PsyYL ploit emotional contagion to induce mater- posites. Sympathy is defined as “an affective ev. AR nal distress, which in turn may lead the response that consists of feelings of sorrow RM u. F mother to change her behavior to their or concern for a distressed or needy other AnnY O advantage. (ratherthansharingtheemotionoftheother). T Emotional responses to displays of emo- Sympathy is believed to involve an other- SI R tion in others are so commonplace in ani- oriented, altruistic motivation” (Eisenberg E V NI mals (de Waal 2003, Plutchik 1987, Preston 2000,p.677).Personaldistress,ontheother U &deWaal2002b)thatDarwin(1982[1871, hand, makes the affected party selfishly seek y b p.77])alreadynotedthat“manyanimalscer- to alleviate its own distress, which mimics tainlysympathizewitheachother’sdistressor that of the object. Personal distress is not danger.” For example, rats and pigeons dis- concerned,therefore,withtheother(Batson playdistressinresponsetoperceiveddistress 1991).Astrikingnonhumanprimateexample inaconspecific,andtemporarilyinhibitcon- is how the continued screams of a punished ditioned behavior if it causes pain responses infantrhesusmonkeywillcauseotherinfants in others (Church 1959, Watanabe & Ono to embrace, mount, or even pile on top of 1986).Arecentexperimentdemonstratedthat thevictim.Thus,oneinfant’sdistressspreads mice perceiving other mice in pain intensify quicklytoitspeers,whichthenseektoreduce www.annualreviews.org • TheEvolutionofEmpathy 283 ANRV331-PS59-11 ARI 4November2007 20:27 their own negative arousal (de Waal 1996, his claims: “If I were to tell of his altruistic p.46). and obviously sympathetic behavior towards Concern for others is different in that it PanzeeIshouldbesuspectedofidealizingan Consolation: reliesonaseparationbetweeninternallyand ape.” Ladygina-Kohts (2001 [1935]) noticed comfortingbehavior directedata externally generated emotions. This separa- similartendenciesinheryounghome-reared distressedparty,such tion is observable in many mammals. In a chimpanzee.Shediscoveredthattheonlyway asarecentvictimof studythatsoughttodocumentchildren’sre- to get him off the roof of her house (better y. aggression sponsestofamilymembersinstructedtofeign thanrewardorthreatofpunishment)wasby nl o sadness (sobbing), pain (crying), or distress actingdistressed,hencebyinducingconcern e s u (choking), striking similarities emerged be- forherselfinhim. al n tweenthereactionsofone-year-oldchildren Perhapsthebest-documentedexampleof o s gper andpets,suchasdogsandcats.Thelatter,too, sympathetic concern is consolation, defined ews.or0. For sthheoiwrehdeacdoimnftohretilnagpoatfttehmep“tdsi,stsruecshseads”ppuetrtsionng astsarnedaesrsutroanocneeproofvtidheedcboymabnatuannitnsvoinlveadpbrye-- nualrevi04/29/1 (ZaYhenr-kWesax(l1e9r2e5t,apl..1294864)).reported how his vRioooussmaaglgenre1ss9i7ve9).inFcoirdeenxatm(dpele,Waathalird&pavratny nals.anRY on bfoornhoibsosi,cPklryinccheimCphainmz,esehcoowmepdansuiochn,cPoannczeeren, ganoeasrmovaerrotuontdhehliossoerrhoefrashfioguhltdaenrds(gFeingtulyrepu1t)s. ourRA thatthescientificestablishmentmightreject De Waal & van Roosmalen (1979) analyzed om arjN LIB ded frELDI aK nloMC w DoK - 00. AR 3P 9-E 7G 8.59:2OLLE 00C hol. 2ND - cA syL PY ev. AR RM u. F AnnY O T SI R Figure1 E V Consolationis NI U commonin y humansandapes, b butvirtuallyabsent inmonkeys.Herea juvenile chimpanzeeputs anarmarounda screamingadult male,whohasjust beendefeatedina fight.Photograph bytheauthor. 284 deWaal ANRV331-PS59-11 ARI 4November2007 20:27 hundredsofconsolationsinchimpanzees,and and best-known definitions by Smith (1759, de Waal & Aureli (1996) included an even p. 10) “changing places in fancy with the larger sample. These studies show that by- sufferer.” Empathic standers contact victims of aggression more Menzel(1974)wasthefirsttoinvestigate perspective-taking: often than they contact aggressors, and by- whetherchimpanzeesunderstandwhatothers thecapacitytotake standerscontactvictimsofseriousaggression know,settingthestageforstudiesofnonhu- another’s moreoftenthantheycontactthosewhohad man theory-of-mind and perspective-taking. perspective—e.g., y. receivedmildaggression. Afterseveralupsanddownsintheevidence, understanding nl another’sspecific o Subsequent studies have confirmed con- currentconsensusseemstobethatapes,but e situationandneeds s al u solation in captive apes (Cordoni et al. probably not monkeys, show some level of separatefromone’s n 2004; Fuentes et al. 2002; Koski & Sterck perspective-takingbothintheirspontaneous own—combined o gpers 2006; Mallavarapu et al. 2006; Palagi et al. social behavior (de Waal 1996, 1998 [1982]) withvicarious ews.or0. For 2C0a0s4tl,e2s020060)4,,wWilditcthigim&paBnozeesecsh(K20u0ts3u),kalakrege&- aentdal.un2d0e0r5;exHpaerreimeetnatla.l2c0o0n1d,it2io0n0s6;(BHria¨ruaetar eTmarogtieotneadlhareolpuisnalg: nals.annualreviRY on 04/29/1 bcwthhrheaielidsnnaremdedneebWoi(rFbadusaseljir&(svSaaewAteiaudorneeelttpir(aao1ll.9t.o29c206o00)l0s7te6)ot,).oaduneHtdtetoochwtuaecpmvopeanlrny-, 2pin0egr0A,s6pw;ehScmihtciihvallejioi-tsrtohakeeitlmnpagalfi.nin2isfe0es-0ostt5u-ac)nt.aieoldlnetdotaoanfrgoethetemedrp’shasetphlpeic-- hoaoopntrehplspaeritrecau’csonaigsadtpitnoiceionatcnirivfieeocbfantsheeeedd ourRA solationinmonkeys,theyfailedtofindany,as cificsituationandgoals(deWaal1996).The ded from arjELDIN LIB dg(iai.dep.o,bthheutewmrsea(enWnsamatntosdnetkaaeplye.ss2)0a0nex0dt)e.tnThdehseHecvooemnnsiontlooaitditohenea ldpioaterurtbiacttuualrbaeorluyotntinhperaeipmxeissattee(nsecbeeeohFfarvtoaimorgreEtleemadvpheaestlhplyiitnttgloe, aK nloMC one situation where one would most expect Altruism,below).Amotherapewhoreturns 300. DowPARK - ct(Soocnchsoionmloaftoieortntaltt.ho2e0ior0c4oc)uw.rnO: M’oCffaoscnparnqienulegl’sma(f1to9etr9h5ea)rsficogfanhi-tl ttworaearedwtothhitmeheopnnereeixntthg—eybyoyouunswnggsatsyetirenrtgoishheterrlapopwiptenfrdtorimeneoatnnode- 9-E 7G tentanalysisofhundredsofreportsconfirms then drape her body between both trees— 008.59:2COLLE tchalatofreaapsesusryaentcrearoefidnismtroensskeedyso.tIhtesrtsililsnteyepdi-s greosepsobnesyeolnikdemlyeirnevocolvnecseernmfootriotnhaeloctohnetra.gHioenr hol. 2ND - tioorbaecteusatlalybldisoheesdr,ehdouwceevthere,dtihsatrtetshseisdbpeahrtayv’-s (sie.lev.,emswothheenrathpeeysohfeteanrbthreieirflyowffshpimrinpgerdtohesmo)-, cA PsyYL arousal. butaddsassessmentofthespecificreasonfor ev. AR theother’sdistressandtheother’sgoals.Tree RM Annu. Y OF EmpathicPerspective-Taking bwriitdhginmgoitsheardsairlyegouclcaurlryrenacneticinipoatriannggutthaneisr, T Psychologists usually speak of empathy only offspring’sneeds(vanSchaik2004,p.104). SI R when it involves perspective-taking. They For an individual to move beyond being E V NI emphasize understanding of the other, and sensitive to others toward an explicit other- U adoption of the other’s point of view. In orientation requires a shift in perspective. y b this view, then, empathy is a cognitive affair The emotional state induced in oneself by dependent on imagination and mental state the other now needs to be attributed to the attribution, which may explain the skepti- other instead of the self. A heightened self- cismaboutnonhumanempathy(Hauser2000, identityallowsasubjecttorelatetotheobject’s Povinelli1998).Perspective-takingbyitselfis, emotionalstatewithoutlosingsightoftheac- ofcourse,hardlyempathy:Itissoonlyincom- tualsourceofthisstate(Hoffman1982,Lewis bination with emotional engagement. The 2002). The required self-representation is latter here is called “empathic perspective- hardtoestablishindependently,butonecom- taking,” such as in one of the oldest monavenueistogaugereactionstoamirror. www.annualreviews.org • TheEvolutionofEmpathy 285 ANRV331-PS59-11 ARI 4November2007 20:27 The coemergence hypothesis predicts that 2005,Emery&Clayton2001).Thesereports mirrorself-recognition(MSR)andadvanced concern the finding or hiding of food, how- expressions of empathy appear together in ever,hencenotempathicperspective-taking. Mirror bothdevelopmentandphylogeny. Inthefuture,wemaybeabletoaddressthe self-recognition (MSR): recognizing Ontogenetically, the coemergence hy- self-other distinction more directly through thatone’sownbody pothesis is well-supported (Bischof-Ko¨hler neural investigation (Decety & Chaminade isreflectedinthe 1988, Johnson 1992, Zahn-Waxler et al. 2003b).Inhumans,therightinferiorparietal y. mirror 1992). The relation between MSR and the cortex, at the temporo-parietal junction, nl o development of empathic perspective-taking underpins empathy by helping distinguish e s u holds even after the data have been sta- between self- and other-produced actions al n tistically controlled for age (Bischof-Ko¨hler (Decety&Gre`zes2006). o s gper 1991).Gallup(1982)wasthefirsttopropose ews.or0. For pphiryicloalgleynseutipcpcoorteemdebrgyenthcee,caonprtreadsitctbioentweemen- UNDERLYINGMECHANISMS nualrevi04/29/1 mfoornMkeSyRs,ancodnaspoelas,tiwonit,hacnodmtpaerglleintegdevhiedlepnincge PerceptionActionMechanism nals.anRY on onlAypinaratpferso.mthegreatapes,theanimalsfor Pcorreestoofnth&edemeWpaatahlic(2c0a0p2aac)itpyrloiepsoasemtehcahtaantitshme ourRA which we have the most striking accounts that provides an observer (the subject) with ded from arjELDIN LIB opAfhlticrnousnisasmonl,dabteieollenopwha)na.ndGttsaalr(lsgueepete(Fd1r9oh8me3lp)Eihnmagdpaaartlehreydaodtloy- aobcbocjdeeisclsyt)trtoehptrrhoeeusgesnhutbatjhteiecotnsiuvs.ebWjsetcahtt’esenoowtfhnaennsouetbuhjreearclt(atanhtde- aK nloMC predictedMSRindolphinsandelephants,and tendstotheobject’sstate,thesubject’sneural 300. DowPARK - tbthoyetlsoheceapmtreeadrakicmtteiaosrtnk,sinohnwavhietiscenhlofawtnhibantedeiintvicdcaounnanlfionrtmeeseeddes risciemaplrliyleasarennadntadutinsoocnocsinaoslcflyisoicmulsoillsyaeratcswttiaovtaeitnsedadir.veiTdahuueatolmsmaoarrete-, 9-E 7G withoutamirror(Plotniketal.2006,Reiss& theeasierthesubject’sidentificationwiththe 008.59:2COLLE Minatrhienore2st00o1f)t.hMeSanRimisablekliinegvdedomto(Abendaebrsseonnt oinbgjemcto,tworhaicnhdaeunthoannocmesicthreespsuobnjseecst.’sTmhiastlceht-s hol. 2ND - &GItaslhluopul1d9b9e9)a.ddedthatself-representation tjehcet,subbojdeiclyt gshetar“inugndietsrethmeotsikoinns”aonfdthneeeodbs-, cA PsyYL is unlikely to have appeared de novo in a whichinturnmayfostersympathyandhelp- ev. AR few large-brained animals. The framework ing. Preston & de Waal’s (2002a) PAM fits RM u. F ofdevelopmentalpsychologists,accordingto Damasio’s(1994)somaticmarkerhypothesis AnnY O whichself-representationemergesinsmallin- of emotions as well as evidence for a link at T crementalsteps(Lewis&Brooks-Gunn1979, thecellularlevelbetweenperceptionandac- SI R Rochat 2003), may apply also to phylogeny. tion,suchasthemirrorneuronsdiscoveredin E V NI Insteadofadheringtoanall-or-nothingdivi- macaquesbydiPellegrinoetal.(1992). U sionofself-representation,someanimalsmay Humandatasuggestthatasimilarphysio- y b reach an intermediate stage similar to that logicalsubstrateunderliesbothobservingand of pre-MSR human infants (de Waal et al. experiencinganemotion(Adolphsetal.1997, 2005). 2000), and that affect communication cre- Possibly, the link between MSR and ates matching physiological states in subject perspective-taking is relatively loose. and object (Dimberg 1982, 1990; Levenson Perspective-takinghasrecentlybeenreported & Reuf 1992). Recent investigations of the for species that appear to lack MSR, both neural basis of human empathy confirm the mammals (Kuroshima et al. 2003, Vira´nyi PAMinthattheyreportneuralsimilaritybe- et al. 2005) and birds (Bugnyar & Heinrich tweenself-generatedandvicariousemotions 286 deWaal ANRV331-PS59-11 ARI 4November2007 20:27 (Carretal.2003,Decety&Chaminade2003a, involvedinbothimitationandempathy,one Decety&Jackson2006,deGelderetal.2004, expectscorrelationsbetweenbothcapacities. Singer et al. 2004), such as activation of the Highlyempathicpersonsareindeedmorein- anteriorventralinsulabothwhenwearedis- clinedtounconsciousmimicry(Chartrand& gusted and when we see another person ex- Bargh 1999) and humans with autism spec- pressingdisgust(Wickeretal.2003). trum disorder are not only deficient in em- Theideathatperceptionandactionshare pathy but also imitation (Charman 2002, y. representationsisanythingbutnew.Accord- Charman et al. 1997). Functional magnetic nl o ingly,empathyisarapidroutine,asconfirmed resonance imaging studies neurally connect e s u byelectromyographicstudiesofmusclecon- motormimicry,suchascontagiousyawning, al n tractionsinthehumanfaceinresponsetopic- withempathicmodeling(Plateketal.2005). o s gper tures of facial expressions, even if presented Other primates, too, yawn when they ews.or0. For scoeibvreidefl(Dyitmhabtetrhgeeytcaaln.2n0o0t0b)e.Acocncoscuinotussolyfpemer-- sPeaeukconnersp&eciAfincsdeyraswonn(2A0n0d6e).rsIonnfeatcta,l.b2eh00av4-, nualrevi04/29/1 pauatthomyaastiaccroeganctiitoivnes,pwrohciecshsaorfetefnarnteogolercatpsiductho ioofratlhceoppyriimngat(e“sa.piSnogc”i)alisfapcriolintaotuionncedexipnerail-l nals.anRY on beundervoluntarycontrol. minegnatgsasihnowwhtehnattshaetyiatseedeportihmerasteesabte(gAidndeeasts-i ourRA &Visalberghi2001,Dindo&deWaal2006), ded from arjELDIN LIB REinmudispvsaiditahunyal’DcsoeovmellrosMtiaolonldatlehsletawteayasffeinctswahniochtheorn’se, sstacelrlvaiemtcshit(atNhtieaomknasyseiamlmveialsawr2ht0oe0n4th)o,atthanoerfdshssuchmroawatcnhnintehfoaennmat--s aK nloMC withsimplemechanismsatitscoreandmore (Bard2006,Ferrarietal.2006).Novelbehav- 300. DowPARK - calaobymielirpteiledesxnamasteuicrtsheaoonfuisttmehreslcaaanypdearcsp.ietirBesespceiancutvisovelev-oetfadk,tihwniges ipooltrehseisarrsceo(djpueiveWedn,aitlaeolso1i,m9a9ti8tlae[t1ains9tg8b2thy],etKhpeoe¨chaulpelierasr1.w9E2ax5lak)moa-sf 9-E 7G speak of the Russian doll model, in which wellassuccessfuldo-as-I-doexperimentswith 008.59:2COLLE haigfihrmer,cohgarndit-iwveirleedveblsaosifs,emsupcahthaysbtuhiledPupAoMn hYuammaaknomshoid&elsM(Catussutzaanwceae1t9a9l9.)1.995,Myowa- hol. 2ND - (bdye iWtsealafl 2e0xp03la)i.nTshseymclpaiamtheisticnoctotnhcaetrPnAMor of tBhoedislaymseimgielanrditeyr—asnudchspaesciwesit—hlmikeelmybeenr-s cA PsyYL perspective-taking,butthatitunderpinsthese hances shared representation and identifica- ev. AR cognitivelymoreadvancedformsofempathy, tion,whichhasbeenproposedasthebasisof RM u. F and serves to motivate behavioral outcomes. trueimitation(deWaal1998,2001),suchas AnnY O Without emotional engagement induced by seen in the apes (Horner & Whiten 2005). T state-matching, perspective-taking would be Thetendencyofnonhumanprimatestocopy SI R a cold phenomenon that could just as easily eachotherisasspontaneousastheempathic E V NI lead to torture as to helping (Deacon 1997, response. Thus, mirror neurons fire auto- U deWaal2005). maticallytoobservedactions,evenintentions y b Perception-action mechanisms are well (Fogassietal.2005),andmonkeysrequireno known for motor perception (Prinz & extrinsicrewardstocopyeachother’sbehav- Hommel2002,Wolpertetal.2001),sothat ior(Bonnie&deWaal2006). we may assume PAM to underlie not only In accordance with the PAM (Preston & emotional state matching but also motor de Waal 2002a), the motivational structure mimicry. This means that the Russian Doll ofbothimitationandempathythereforein- also relates to doing as others do, including cludes(a)sharedrepresentations;(b)identifi- bodily synchronization, coordination, imita- cationwithothersbasedonphysicalsimilarity, tion, and emulation (Figure 2). If PAM is shared experience, and social closeness; and www.annualreviews.org • TheEvolutionofEmpathy 287 ANRV331-PS59-11 ARI 4November2007 20:27 Imitation Empathy n True imitation, Perspective- o cti emulation taking, targeted n helping y. sti nl Di se o er nal u Oth Sympathetic nals.annualreviews.orgRY on 04/29/10. For perso Increased Self- MCshooaotroredrd imn gaiomtiaoiclnsr,y PAM Ecccooomnnnoctsateoigolraninot,ainoln ourRA om arjN LIB ded frELDI FTihgeuRreus2siandollmodelofempathyandimitation.Empathy(right)inducesasimilaremotionalstatein wnloaMCK tlahyeesrus,bsjueccthaansdsythmepoabthjeectti,cwcoitnhcaetrnitsancdorpeetrhsepepcetrivcee-pttaikoinn-ga,cbtiuoinldmuepcohnanthisimsh(aPrAdM-w)i.rTedhseodcoioll-’asfofeuctteirve DoK - basis.Sharingthesamemechanism,thedoll’simitationside(left)correlateswiththeempathyside.Here, 00. AR thePAMunderliesmotormimicry,coordination,sharedgoals,andtrueimitation.Eventhoughthedoll’s 9-3E P outerlayersdependonprefrontalfunctioningandanincreasingself-otherdistinction,theseouterlayers 7G remainconnectedtoitsinnercore. 008.59:2COLLE (c)automaticityandspontaneity.Allofthisap- empathy.Agoodexampleseemstheintensi- hol. 2ND - pthlieesmtoortehecocmorpelmexeochutaenrislmay,enrostonfetcheessRaurislsyiatno fiinepdapinai(nLraensgpfoonrdseetoaflm.2i0c0e6s)e.eEinmgootitohnearlmcoince- cA syL dollmodel,whichdevelopininteractionwith tagionmayleadindividualsfrightenedbythe PY ev. AR theenvironment. alarmofotherstohideorflee,amotherdis- RM u. F tressedbyheroffspring’sdistresstoreassure AnnY O FROMEMPATHYTOALTRUISM bothherselfandheroffspringbywarmingor T nursing them, or inhibit an individual from SI R Not all altruistic behavior requires empa- inflictingpainuponanotherbecauseofthevi- E V thy. When animals alert others to an out- cariousnegativearousalinducedbytheother’s NI U sidethreat,worktogetherforimmediateself- distresscalls.Thus,simpleempathicreactions y b reward,orvocallyattractotherstodiscovered may benefit both the actor and individuals food,biologistsmayspeakofaltruismorco- closetothem. operation,butthisbehaviorisunlikelytobe Behavioral copying, too, often produces motivatedbyempathywiththebeneficiary. adaptive outcomes. Imagine a group of ani- malsinwhicheverymemberwastoeat,sleep, forage,orplayindependently:Thiswouldbe EmotionalContagion impossiblefornomadicanimals,suchaspri- Self-centeredvicariousarousal,knownasper- mates.Beinginsyncisoftenamatteroflife sonal distress, represents the oldest kind of ordeath(Boinski&Garber2000). 288 deWaal