How does social essentialism affect the development of ...sjleslie/Rhodes_et_al-2017-Development… · 2 of 15 | RHODES Et al. allow (Gelman & Rhodes, 2012; Shtulman & Schulz, 2008),

Developmental Science. 2017;e12509. wileyonlinelibrary.com/journal/desc | 1 of 15https://doi.org/10.1111/desc.12509

© 2017 John Wiley & Sons Ltd

Received:9July2015 | Accepted:24August2016DOI:10.1111/desc.12509

P A P E R

How does social essentialism affect the development of inter- group relations?

Marjorie Rhodes1 | Sarah-Jane Leslie2 | Katya Saunders1 | Yarrow Dunham3 | Andrei Cimpian1

1DepartmentofPsychology,NewYorkUniversity,NewYork,USA2DepartmentofPhilosophy,PrincetonUniversity,Princeton,NJ,USA3DepartmentofPsychology,YaleUniversity,NewHaven,CT,USA

CorrespondenceMarjorieRhodes,DepartmentofPsychology,NewYorkUniversity,6WashingtonPlace, rm301,NewYork,NY10003,USA.Email:[email protected]

Funding informationNSF,Grant/AwardNumber:BCS-1226942.

AbstractPsychologicalessentialismisapervasiveconceptualbiastoviewcategoriesasreflect-ing something deep, stable, and informative about their members. Scholars fromdiversedisciplineshave long theorized thatpsychologicalessentialismhasnegativeramificationsfor inter-grouprelations,yet littlepreviousempiricalworkhasexperi-mentallytestedthesocialimplicationsofessentialistbeliefs.Threestudies(N = 127,ages4.5–6)foundthatexperimentallyinducingessentialistbeliefsaboutanovelsocialcategory ledchildrentosharefewerresourceswithcategorymembers,butdidnotleadtotheout-groupdislikethatdefinessocialprejudice.Thesefindingsindicatethatessentialismnegativelyinfluencessomekeycomponentsofinter-grouprelations,butdoesnotleaddirectlytothedevelopmentofprejudice.

RESEARCH HIGHLIGHTS• Essentialism has been proposed to have negative socialramifications.

• Three studies experimentally tested this proposal among youngchildren.

• Essentialism led children to withhold resources from out-groupmembers.

• Essentialismdidnotleadtoout-groupdislike.

1 | INTRODUCTION

Expectingagentlebabytigerto inevitablygrowuptobeferocious,anappleseedtodevelopintoanappletreeevenifplantedinanolivegrove,orayounggirlgrowingupinahouseholdofboystopreferprin-cessestotrucks,allreflectaconceptualcommitmenttopsychologicalessentialism(Medin&Ortony,1989).Psychologicalessentialismisapervasiveconceptualbiastoviewcategories(e.g.,tigers,appletrees,girls)asreflectingsomethingdeep,stable,andinformativeabouttheirmembers, as highly predictive of individual development regardlessof other influences, and asmarking fundamental similarities amongmembersanddifferencesbetweenkinds(Gelman,2004).Essentialistbeliefsshapehowpeoplerepresentandreasonaboutmanytypesof

categories fromat least theearlypreschool yearsonward (Gelman,2003).

Essentialismhasmostoftenbeenstudiedinthecaseofbiologicalcategories, such as animal species. For example, essentialist beliefsabouttigersentailthinkingthatwhetherananimalisatigerisstableanddeterminedbybirth,thattigersarefundamentallysimilartoeachotheranddifferentfromnon-tigers,andthatananimal–onceborntotigerparents–will inevitably growup tobe ferocious, even if itlooksdifferent fromothertigers (e.g., iswhite insteadoforange)oris raised inanunusualenvironment (e.g., inazoo,where ithas fewopportunitiestolearnorpracticeferociousbehaviors;Gelman,2004;Medin&Ortony, 1989;Prentice&Miller, 2007).By at least age4,childrenmakeinferencesinlinewitheachofthesebeliefs(Gelman&Markman,1986,1987;Gelman&Wellman,1991;Waxman,Medin,&Ross,2007;forreviewseeGelman,2003).

In the case of animal categories, psychological essentialismmay help get conceptual development off the ground and facilitateknowledge acquisition by allowing children to overlook superficialdifferences(e.g.,betweenorangeandwhitetigers)andfocusonthepropertiesthatcategorymembersallshare(Gelman&Kalish,2006).Yet, essentialism reflects a biased, inaccurate picture of theworld;most categories have no real essences (Leslie, 2013; Mayr, 1991),specieschangeovertimeinwaysthatessentialistthinkingdoesnot

mailto:[email protected]

2 of 15 | RHODES Et al .

allow(Gelman&Rhodes,2012;Shtulman&Schulz,2008),andcate-gorymembersoftenvarymorewidelyfromeachotherthanessential-ismimplies.Indeed,essentialistthought–particularlyitsemphasisonwithin-categoryhomogeneityandstabilityovertime–interfereswithpeople’s understanding of the mechanisms that drive evolutionarychange(Shtulman&Schulz,2008)aswellaswithnormativereasoningregardinghowpropertiesaredistributedacrosscategories(Rhodes&Brickman,2010;Rhodes&Liebenson,2015).

Thenegativeconsequencesofessentialistbeliefsareevenclearerinthesocialdomain.Byages4or5,childrenholdessentialistbeliefsnotonlyaboutbiologicalspeciescategories,butaboutcertaintypesofhumansocialcategoriesaswell.Forexample,intheUnitedStates,preschool-age children hold essentialist beliefs about gender; theyexpect girls tobe fundamentally similar toeachother anddifferentfromboys(Gelman,Collman,&Maccoby,1986);thatbeingbornagirlmeans that a babywill inevitably growup to prefer tea sets to toytrucks(Taylor,1996;Taylor,Rhodes,&Gelman,2009),andthatgen-dercategoriesreflectnatural,objectivestructureintheworld(Rhodes&Gelman,2009;Rhodes,Gelman,&Karuza,2014).Byemphasizingwithin-category homogeneity, essentialism contributes to stereo-typing among young children and adults (Bastian & Haslam, 2006;Hoffman&Hurst,1990;Levy&Dweck,1999;Martin&Parker,1995;Pauker, Ambady, & Apfelbaum, 2010; Plaks, Stroessner, Dweck, &Sherman,2001;Prentice&Miller,2007).Essentialistbeliefsalsoleadpeopletoattributegroupdifferences(e.g.,menoutperformingwomenin higher-level mathematics; African Americans being convicted ofcrimesathigherratesthanwhites)tonatural,biologicalcausesinsteadoftoflexible,societalfactors(Leslie,inpress;Martin&Parker,1995;Salomon & Cimpian, 2014), thusmaking peoplemore accepting ofsocial differences and social hierarchies (Cimpian&Salomon,2014;Hussak&Cimpian,2015;Jayaratneetal.,2006).

Beyondcontributingtosocialstereotypingandattribution,socialessentialismhasalsobeen theorized tocontribute toothercompo-nentsofinter-grouprelations,includingprejudiceanddiscrimination(Allport,1954).Prejudiceanddiscriminationdifferfromsocialstereo-typing in that they involvenegative feelings towards (in the caseofprejudice)andnegativebehaviortowards(inthecaseofdiscrimination)a group, instead of, or addition to, expectations ofwithin-categorysimilarity(Brewer&Brown,1998).ConsistentwithAllport’stheoriz-ing,explicitlyessentialistbeliefs,suchasbeliefsthatmembershipinareligiousorethnicgroupisdeterminedbygenes,orthatsocialcatego-rieshavesharpboundariesthatcannotbecrossed,haveplayedcen-tralrolesinsomeofthemostperniciousinstancesofsocialprejudicethroughout human history (e.g., Hitler’sAryanism; thewhite powermovementintheUnitedStates;Holtz&Wagner,2009;Yzerbyt,Judd,&Corneille,2004).Inadditiontocontributingtotheout-groupdislikethatdefinesprejudice,socialessentialismhasalsobeenproposedtonegatively influence inter-group relationsby leadingpeople toviewcategory boundaries as discrete and inflexible, thus decreasing thelikelihoodthatpeoplewillchoosetointeractwithmembersofothergroups (Bernstein,Sacco,Young,Hugenberg,&Cook,2010;Levy&Dweck,1999;Noetal.,2008;Williams&Eberhardt,2008;Zagefka,Nigbur,Gonzalez,&Tip,2012).

Theproposalthatessentialismleadstothesenegativesocialcon-sequences has the potential to address interesting developmentalquestions about the origins of inter-group relations. By early child-hood,andperhapsevenearly in infancy (Kinzler,Dopoux,&Spelke,2007),childrenarehighlysensitivetoin-groupsandout-groups,andshowrobustpreferencesfortheirin-groupsbasedoneventrivialandarbitrary distinctions (e.g., randomly assigned shirt colors; Dunham,Baron,&Carey,2011;Patterson&Bigler,2006).Yet,notallin-grouppreferencesimplysocialprejudiceorleadtodiscriminatorybehaviors.Muchofthetime,childrenshowincreasedpositivefeelingsfortheirin-groupmembersintheseparadigms,butdonotholdmorenegativefeelingstowardstheout-group,orengageinmorenegativebehaviorstowardsout-groupmembers(rather,theyrespondtoout-groupmem-bersneutrally,Bigler,Jones,&Lobliner,1997;Dunhametal.,2011,Patterson&Bigler,2006;seealsoAboud,1988,2003;Brewer,1999;Nesdale, 2004, for discussion of the distinctions between in-grouppositivity andout-groupdislike). Somethingelse isneeded tomovefromin-grouppreferencestomorevirulentdislikeandnegativetreat-mentoftheout-group.

Based on the theorizing of Allport (1954) and others (Bigler& Liben, 2007;Hirschfeld, 1996; Leslie, in press; Prentice&Miller,2007;Rothbart&Taylor,1992),perhapspsychologicalessentialismisakey factor thatmovespeople frommildpreferences for in-groupstowardsthedevelopmentofnegativeattitudesandbehaviorstowardsout-groups.Althoughyoungchildrenandadultscanreadilylearnnewcriteriaforsocialcategories,peopleholdessentialistbeliefsregardingonlyasubsetofsocialcategories that theyencounter.Forexample,childrencanreadily learntogrouppeoplebasedonshirtcolorsandwillshowsomesocialpreferencesbasedonthisdistinction(Dunhametal.,2011),butdonotformessentialistbeliefsaboutsuchcatego-ries unless they are given additional information– that is, theywillnot expect color-based groups to be determined by birth or to bestable, to imply similarities across group members and differencesbetweengroups,ortoreflectfundamentalcausallypowerfulfeaturesofanindividual’sidentity(Rhodes&Brickman,2011;Rhodes,Leslie,&Tworek,2012;seealsoKalish,2012).Thereisalsovariationintheextent to which children hold essentialist beliefs about particularsocialcategoriesthattheyencounterintheireverydaylives;forexam-ple,whitechildrenintheUnitedStatesholdmoreessentialistbeliefsaboutgenderandlanguage-basedgroupsthanaboutrace(Kinzler&Dautel,2012;Rhodes&Gelman,2009),andIsraelichildrenholdmoreessentialistbeliefsaboutreligious-ethniccategoriesthanthosebasedonpersonalitytraits(Diesendruck&haLevi,2006).Whetherchildrendevelop essentialist beliefs about particular social groupings variesbasedonfeaturesoftheirculturalcontext,includingexposuretocer-tainformsoflanguage(Gelman&Heyman,1999;Rhodesetal.,2012;Segall, Birnbaum, Deeb, & Diesendruck, 2015), experiences withsocialdiversity(Deeb,Segall,Birnbaum,Ben-Eliyahu,&Diesendruck,2011), and the nature of existing social hierarchies (Mahalingam,2007; Mahalingam & Rodriguez, 2006). In short, whether peoplewithinacommunityholdessentialistbeliefsaboutparticulargroup-ings develops and changes across childhood (Astuti, Solomon, &Carey, 2004; Birnbaum, Deeb, Segall, Ben-Eliyahu, & Diesendruck,

| 3 of 15RHODES Et al .

2010;Diesendruck&Haber,2009;Kinzler&Dautel,2012;Rhodes&Gelman,2009;Taylor,1996).Thus,onepossibilityisthatvariationinessentialistbeliefsexplainswhichsocialgroupsbecomethetargetsof social prejudice and discrimination; in linewithAllport and oth-ers,perhapsthedevelopmentofessentialistbeliefsaboutparticulargroupsfostersthedevelopmentofmorevirulentdislikeofessential-izedout-groups.

Thereareseveralmechanismsbywhichessentialismcouldnega-tivelyinfluenceinter-grouprelations.Forexample,becauseessential-ism implies thatdifferencesbetweengroupsaretheresultofstableand inherent factors,perhapsessentialism leadspeople toview theout-groupasdifferingmoredramaticallyfromthein-group(Noetal.,2008).Seeingagroupasverydifferentfromone’sownin-grouphasbeenwelldocumentedtocontributetoprejudicedattitudes(Brewer,1999). Along the same lines, as essentialism indicates that socialgroupsreflectfundamentallydistinctkindsofentities(e.g.,inthepro-totypical case of animal categories – different species), such beliefscould lead people to think of out-groups as somehow less humanthantheirowngroup,adehumanizationprocessthatalsocontributestoprejudice inadultpopulations (Haslam,Bastian,Bain,&Kashima,2006; Leyens etal., 2001).Another pathway bywhich essentialismcould lead toprejudice involves influencinghowpeople respond tonegative information about individual groupmembers; for example,essentialismcouldmakepeoplemorelikelytodrawconclusionsabouttheinherentnatureofagroupbasedonthenegativeactionsofsin-gle individuals (Andreychick&Gill,2015;Leslie, inpress;Prentice&Miller,2007).

Alternatively (or inaddition), essentialismcouldnegatively influ-ence inter-group relations, but not directly via the processes thatAllport and others have proposed. Rather, essentialism could havenegative consequences for inter-group relations by leading peopletoviewgroupboundariesasdiscreteandinflexible,andthustoviewout-groupmembersaslessworthyofattentionandeffort(Bernsteinetal.,2010;Chao,Hong,&Chiu,2013;Levy&Dweck,1999;Noetal.,2008;Williams&Eberhardt,2008;Zagefkaetal.,2012).Ifso,essen-tialismwould decrease the likelihood of inter-group interactions orcooperation(Chaoetal.,2013;Noetal.,2008),withpossibledown-streamconsequencesforprejudice.

Despitethelonghistoryoftheorizingregardingtheimplicationsofessentialismforinter-grouprelations,however,empiricalworkonthisissuehasbeenlimitedandhasyieldedmixedresults.Amajorlimitationofprior researchon this topic is that ithasbeenconductedalmostentirelyamongadultpopulations;thereisverylittlepreviousempiricalworkexaminingtheimplicationsofessentialismforinter-grouprela-tionsinearlychildhood.Essentialismdoesappeartoplayaroleinthedevelopmentofsocialstereotypinginchildhood;forexample,Paukeretal.(2010)foundamongchildrenages3–10thatessentialistbeliefsaboutracepredictedchildren’suseofnegativeracialstereotypes(seealsoLevy&Dweck,1999).Also,DiesendruckandMenahem (2015)foundthat increasing thesalienceof Israelichildren’s (age6)essen-tialistbeliefsaboutethnicityledthemtodrawmembersofdifferentgroups farther apart (indicating perhaps that they perceived moresocialdistancebetweengroups)andtodrawin-groupmemberswith

morepositiveaffectthanout-groupmembers.Thesefindingsaresug-gestiveofanearlyemerginglinkbetweenessentialismandinter-grouprelations. Because this study used known groups aboutwhich chil-drenalreadyhadessentialistbeliefsandotherrelatedgroup-relevantknowledgeandexperience(Deebetal.,2011;Diesendruck,Goldfein-Elbas,Rhodes,Gelman,&Neumark,2013), however, it leavesopenthequestionofwhethertheformationofessentialistbeliefs–alone–causallyinfluenceshowchildrenfeelandbehavetowardsmembersof other groups. Experimentally testing whether the formation ofessentialistbeliefsinearlychildhoodholdsnegativeramificationsforinter-groupattitudesandbehaviorscan revealwhether suchbeliefshavecausalimplicationsforthedevelopmentofinter-grouprelations.

Althoughthereismoreworkonthisissueinadultpopulations,thisbodyofworkhasyieldedamixedpatternoffindings.Mostofthisworkhasexaminedtherelationbetweenessentialistbeliefsandprejudiceregardingsalientsocialcategoriesthatpeopleencounterintheirdailylives – testingwhether essentialist beliefs about race, for example,correlatewith increased prejudice towards racial out-groups. Somestudiestakingthatapproachhavefoundweakornorelationsbetweenmeasuresofessentialismandprejudice(Haslam&Levy,2006;Haslam,Rothschild,&Ernst,2002;Hodson&Skorska,2015),whereasothershaveindeedfoundsupportforsuchrelations(Jayaratneetal.,2006;Keller,2005).Experimentalworkonthisissuehasalsopredominantlyfocused on familiar social categories such as race (e.g., seeking toincreaseordecreasethesalienceofpeople’sessentialistbeliefsaboutracepriortocompletingmeasuresofinter-groupattitudesorbehavior)andhasalsorevealedamixedpictureoffindings(Andreychik&Gill,2015;Chaoetal.,2013;Diesendruck&Menahem,2015;Keller,2005;Levy,Stroessner,&Dweck,1998;Rangel&Keller,2011;Williams&Eberhardt,2008).

Tointerpretthesemixedpatterns,sometheoristshaveproposedthat essentialist beliefs do not causally influence inter-group rela-tionsperse,butinsteadareusedtojustifyexistingsocialattitudesorpractices(Haslam&Whelan,2008;Pettigrew,1979;Rangel&Keller,2011;Verkyuten,2003).Forexample,Morton,Hornsey,andPostmes(2009)foundthatwhetherpeopleendorsedessentialistbeliefsaboutparticular categories (e.g., ethnic differences) dependedonwhetheressentialism was being used as a justification for excluding an in-groupmemberoranout-groupmember,suggestingthatpeopleadoptessentialistbeliefsstrategicallytojustifytheiralready-heldbeliefsorattitudes.Inasimilarvein,Mahalingam(2003)foundthathighercasteIndians were more likely to endorse essentialist interpretations ofcastethanwerelowercasteIndians,presumablybecausethat inter-pretationvalidatedtheirownhigherstatus.

Insum,priorworkleavesopenquestionsregardingwhetheressen-tialistbeliefsplayacausalroleinthedevelopmentofnegativeinter-groupphenomena,whetheressentialismalone is sufficient tocausenegativesocialoutcomes (orwhether itdoes soonlywhen it inter-actswithother knowledge andexperience), andwhether thenega-tiveconsequencesofessentialismforinter-grouprelationsinchildrenextendbeyondincreasedstereotyping.Toaddressthesequestions,weadoptedanexperimentalapproachusingnovelgroupsforthepresentstudies.Byadoptingsuchanapproachamongyoungchildren,wecan


provideaclear testofhowessentialism influences inter-group rela-tions,duringthedevelopmentalperiodwhenbothessentialistbeliefsandprejudicedattitudesbegin toemerge.We focusedon theearlychildhoodyears(ages4.5–7)andcreatedanovelsocialgroup,aboutwhichchildrenhavenopriorknowledge,stereotypes,orattitudes.Wethenexperimentally inducedsomechildren (andnotothers) toholdessentialistbeliefsaboutthegroup,andtestedforeffectsonchildren’sfeelings and behavior towards out-group members. This paradigmthusallowsforastrong,directtestofwhetheressentialismcausallyinfluencesinter-grouprelations.

Given the paucity of previous developmental research on thistopic,aswellasthemixedfindingsfromstudiesamongadultpopu-lations,wenotedthattherewereseveralpatternsoffindingsthatwecouldobtain. In linewithAllport andothers (e.g.,Hirschfeld,1996),we could find that inducing essentialism in young children has far-reaching negative consequences for inter-group relations, includingleading to thedevelopmentofmorenegative feelings towardsout-groupmembers.Alternatively,wecouldfindthatessentialismaloneis not sufficient to initiate these processes, and therefore, that ourmanipulationinducesessentialismbuthasnoconsequencesforchil-dren’sinter-groupattitudes.

Athirdpossibilityisthatessentialismholdsimplicationsforinter-grouprelations,butinamorelimitedmannerandnotviatheactiva-tionofout-groupdislike.Asdescribed above, among adults, essen-tialist beliefs increase the extent to which people perceive groupboundariesasstrict,absolute,andinflexible(Noetal.,2008)andthusdecrease expectations of inter-group interactions (Bernstein etal.,2010;Williams&Eberhardt,2008;Zagefkaetal.,2012).Atleastsomeoftheseeffectsoccurindependentofassociationsbetweenessential-ismandprejudice(Williams&Eberhardt,2008).Thus, inadditiontomeasuresofchildren’sfeelingstowardsout-groupmembers,wealsoincludedameasureof inter-grouprelationsthatrelies lessonsocialpreferences, and more on expectations of inter-group interactions– resource allocation decisions. Resource allocation decisions oftendependmoreonpeople’sexpectationsof social reciprocity thanonfeelings regarding potential recipients among both children (Renno& Shutts, 2015) and adults (Yamagishi, Jim, & Kiyonari, 1999). Forexample,RennoandShutts(2015)foundthatpreschool-agechildren’stendency to givemore resources to racial in-group than out-groupmemberswas predicted by their expectations of social reciprocity;thatis,themorethatchildrenbelievedthataracialin-groupmemberwouldbemorelikelytohelpthemthanaracialout-groupmember,themorelikelytheyweretodistributemoreresourcestoracialin-groupmembers.Incontrast,children’sgivingbehaviorwasnotpredictedbytheirsocialpreferences;childrenwhosaidtheypreferredwhiteoverblackchildrenwerenomorelikelytoshowbiasintheirgivingbehav-ior.Also, Paulus andMoore (2014) found that, by age 5, children’ssharingbehaviorisstronglyrelatedtotheirexpectationsoftheextenttowhich theyexpectothers to sharewith them.Thus, essentialismcouldreducechildren’swillingnesstoshareresourceswithout-groupmembersbecauseitdecreasestheirexpectationsofreciprocityamongmembersofdifferentgroups,perhapswithoutactuallyleadingtoout-groupdislike.

2 | STUDY 1

TheaimofStudy1wastoprovidean initial testofwhether induc-ingessentialistbeliefstowardsanovelout-groupleadstothedevel-opmentofmorenegativeattitudesandbehaviorstowardsmembersofthatgroup.Todoso,weusedalanguagemanipulationtoinduceessentialistbeliefsaboutanovelgroupinsomechildrenandnototh-ers,andthentestedtheirattitudestowardsmembersofthegroup.Totestwhetheressentialismleadsdirectlytomorenegativeattitudesordoessoviathegeneralizationofnegativeinformationaboutindividualgroupmembers (asdescribedabove, seeAndreychick&Gill, 2015;Leslie, in press; Prentice & Miller, 2007), we assessed inter-groupattitudestwotimes–oncebeforechildrenwereexposedtonegativeactionscommittedbyindividualgroupmembers,andonceafterexpo-suretosuchinformation.

2.1 | Methods

Participants included 42 children (M age=5.8years, range=5.0–6.9years;16male,26female;36%White,19%multi-ethnicormulti-racial,7%Black,7%Latino,2%Asian,remainderdidnotprovidethisinformation)recruitedfromandtestedattheChildren’sMuseumofManhattan. An additional 13 children began testing but were notincludedinanalyses(fivebecausetheydidnotcompletetheresearchsession,fiveforexperimentererrors,andthreeforparentalinterfer-ence). To induce essentialism in some participants and not others,childrenwere randomlyassigned to theGeneric (n = 21)orSpecific(n = 21) conditions.Basedoneffect sizes frompreviousworkusingsimilarmanipulationsofessentialismandrelatedbeliefs(e.g.,Cimpian,Arce,Markman,&Dweck,2007;Gelman&Heyman,1999;Gelman,Ware,&Kleinberg,2010;Rhodesetal.,2012),weaimedtoincludeatleast14childrenperconditioninallstudies.Becauseweanticipatedthatsomeparticipantswouldneedtobeexcluded,however,weover-sampledbeyondthisnumbertoensurethatwewouldhavesufficientsamplesizes inourfinalanalyses.Stoppingrules fordatacollectionweredecidedbasedoncalendardatesthatwereselectedaprioritomeetorexceedtheminimumsamplesizeof14childrenpercondition.Exclusiondecisionsweremadeafteralldatawerecollectedandcodedfromvideos,butbeforedataanalysis.Wereportallusabledata,andnotethenumbersandreasonsforexclusionsforeachstudy.

2.1.1 | Manipulation of essentialist beliefs

Toexperimentallyinduceessentialistbeliefsforthenovelsocialcat-egory,we built on priorwork showing that generic language – lan-guagethatreferstoabstractkinds(e.g.,’tigershavestripes’)–causeschildren to apply essentialist beliefs to new categories that theyencounter(Rhodesetal.,2012;alsoGelmanetal.,2010).Thisprevi-ousworkdoesnot indicate thatgeneric languagecreates essential-ist thought. Essentialist beliefs reflect basic conceptual biases thatgo far beyond the content of generic language itself. For example,there isnoexplicitcontent inthesentence ’tigershavestripes’thatcommunicatesthatbeingatigerisamatterofinnateandimmutable


categorymembership.Yetchildrenconcludethatnewcategorieshavethosefeaturesafterfairlylimitedexposuretosuchgenerics(Gelmanetal.,2010;Rhodesetal.,2012).Fromthisperspective,childrenhaveabstract expectations that certain categories in their environmentreflectessentialkindsandrelyonlinguisticcuestodeterminewhich categorieshave this structure.Becausegeneric language communi-catesregularitiesregardingabstractkinds,childrenassumethatcat-egoriesdescribedwithgeneric languageare thekindsofcategoriesthatarecoherentandcausallypowerfulenoughtosupportsuchgen-eralizations(Cimpian&Markman,2011;Gelmanetal.,2010;Gelman&Heyman,1999; Leslie,2008;Rhodesetal., 2012). In thepresentcontext,genericlanguageprovidesausefulwaytoinduceessentialistthinkingforanovelcategorywithoutmanipulatinganyotheraspectofinter-grouprelations.

Inthepresentstudy,first,anexperimenterreadanillustratedbookthatpresentedanovelcategory(‘Zarpies’).ThebookswereidenticaltothoseusedinRhodesetal.(2012,Study1)andpresented16individ-ualpicturesofZarpies,oneperpage,eachdisplayingauniqueprop-erty.The16Zarpieswerediversewithrespecttorace,sex,andage,sothatchildrencouldnotmapthecategoryontoanygroupforwhichthey might already hold essentialist beliefs. By condition, childrenheard thepropertyoneachpagedescribedeitherwithgeneric lan-guage(e.g.,’LookatthisZarpie!Zarpiesclimbfences’)ornon-genericlanguage(e.g.,’LookatthisZarpie!ThisZarpieclimbsfences’).Noneofthepropertiesinvolvedanynegativequalities.Theexperimenterreadthebooktwicetothechild.

Using these materials, Rhodes etal. (2012) found that, in theSpecific condition, children did not hold essentialist beliefs aboutZarpiesafterexposuretothebook.Thatis,althoughtheylearnedthecategory ’Zarpie’, theydidnotexpectZarpiepropertiestobedeter-minedby birth, they did not expect individuals to carry out certainbehaviors becausetheyareZarpies,andtheydidnotexpectallZarpiesto share either thepropertiesmentioned in thebookorothernewproperties. In contrast, theGeneric conditionsignificantly increasedthelikelihoodoftheseessentialistbeliefsamongpreschool-agechil-dren,witheffectspersisting forat least severaldaysafterexposuretothegeneric language.Follow-upcontrolstudiesconfirmedthat itwasthegenericityofthetargetsentences–notsimplytheirsyntacticplurality–thatelicitedtheseeffects.

2.1.2 | Measures of essentialist beliefs

Afterthebook-reading,childrencompletedtwomeasuresofessen-tialistbeliefsusedbyRhodesetal. (2012)andGelmanetal. (2010).Theseincludedthreeexplanationitems,inwhichchildrenwereaskedtoexplainwhy individualZarpiesperformed specificbehaviors thathadbeen shown in thebook (e.g., ’Lookat thisZarpie.Why is thisZarpieclimbingatall fence?’).Responsesweretranscribedverbatimfromvideoandcodedbytwoindependentraters(agreement=93%across both studies,withdisagreements resolvedby a third coder).FollowingCimpianandMarkman(2009,2011),Gelmanetal.(2010),andRhodes etal. (2012), responseswere scored along twodimen-sions:(a)whethertheyreferredtointrinsic(e.g.,’becausehelovesto

climb’,scored’1’)orextrinsic(e.g.,’becauseitisniceoutside’,scored’0’) causal factors andwhether they referred to the category (e.g.,’becauseZarpies love to climb’, scored ’1’) or not (e.g., ’because he loves toclimb’, scored ’0’) toexplain thebehavior.Thus, scores foreachexplanationitemcouldrangefrom0to2essentialistresponses.

Next, children completed two inheritance items, in which theyweretoldthatababywasborntoaZarpiemombutraisedbyanon-Zarpie mom (Gelman & Wellman, 1991; Hirschfeld, 1995; Taylor,1996;Waxmanetal.,2007).Childrenwereaskedtopredictwhetherthegrown-upchildwouldexhibitapropertyheldbytheZarpiemom(e.g.,’flapsherarmswhensheishappy’,scored’1’)orthenon-Zarpiemom(e.g.,’clapsherhandswhensheishappy’,scored’0’).Theseitemstestwhetherchildren think thatpropertiesexhibitedbyZarpiesaredeterminedbybirthandstable(consistentwithessentialistthought)or determined by the environment (inconsistent with essentialistthought).FollowingRhodesetal.(2012),scoresforalloftheessential-ismitemsweretotaledtocreateonecomposite(numberofessentialistresponsesoutof8totalpossible).Weanalyzedthesedatausingbino-mialregressionmodelsandreportaverageprobabilitiesofessentialistresponses,accompaniedbyWald95%ConfidenceIntervals(CIs),withOddsRatios(andaccompanyingCIs)asindicatorsofeffectsize.

2.1.3 | First measure of inter- group attitudes

Aftertheessentialismmeasures,children’sattitudestowardsZarpieswere assessedwith four questions. For each, childrenwere showna new individual Zarpie, and asked a question aboutwhether theywouldwant toaffiliatewith theZarpie (‘Doyouwant to invite thisZarpietoyourbirthdayparty?’’DoyouwanttoplaywiththisZarpie?’’DoyouwanttoshareyourtoyswiththisZarpie?’’DoyouwanttositnexttothisZarpieatschool?’).Foreach,childrenrespondedwith’yes’or’no’andwerethenaskedafollow-upquestion.Forexample,iftheyrespondedaffirmatively,theywereasked,’doyousortofwanttoorreallywantto?’Eachitemreceivedascorerangingfrom0to3,withhighernumbersindicatingmorenegativeattitudes(0=reallywantto,1=sort ofwant to, 2=sort of don’twant to, 3=really don’twantto).Thefourindividualsthatwereshownwerematchedtothepar-ticipatingchild’sgenderandvariedbyraceacrossthefourquestions.Responsestothese itemswereaveragedandanalyzedviastandardanalysesofvariance.

2.1.4 | Exposure to negative information

Subsequently,childrenwereshownaseriesoffournegativeactionscompleted by new individual Zarpies. For example, children wereshownanindividualZarpiestealingacookie.Toavoidprovidingeithergenericorspecificlanguageregardingtheaction,childrenwereaskedtodescribethebehaviorthemselves(e.g.,’Look,aZarpie!Lookwhat’shappeninghere.What’sgoingonhere?’).Ifthechildrespondedwith’stealing a cookie’, the experimenter responded, ’Yes, that’s steal-ing a cookie.’ The experimenter described only the behavior; theexperimenterdidnotrepeatanyreferencestotheagent.Ifthechildresponded with a different action, the experimenter said, ’Right,


that looks like [child’s response].Youknowwhatelse?That’ssteal-ingacookie.Remindme,whatisthat?Right,that’sstealingacookie.’Childrencorrectlyrespondedwith’stealingacookie’,forexample,on87%oftrialseitherbeforeorafterthisprompting(thisratedidnotvarybycondition,p > .20).Thisprocedurewasrepeatedforfourbehaviors(each displayed by a different individual Zarpie): stealing a cookie,teasing,pushingsomeone,callingsomeoneameanname.Afterchil-drenwereshownallfourbehaviorsonce,theywerereviewedasec-ondtime.Foreach,theexperimentersaid,’Remindme,whatisthat?Right that’sstealingacookie.’Onthissecondrunthrough,childrengavetheintendeddescriptionson92%oftrials(thisratedidnotvarybycondition,p > .70).Exploratoryanalysesrevealedthatthenumberofaccuratebehavioraldescriptionsgivenbychildrenduringthisphasedidnotpredictanyof thedependentmeasures; responsestothesecomprehensionquestionswerenotconsideredfurther.

2.1.5 | Second measure of inter- group attitudes

Afterthenegativeactions,childrencompletedthemeasureassessingtheirattitudestowardsZarpiesasecondtime.Theywereaskedthesamefourquestionsasinthefirstmeasureofinter-groupattitudes,butwithdifferentindividualZarpiespictured(whetheraspecificpic-turewasshownbeforeorafterthenegativeinformationwascounter-balancedacrossparticipants).

2.1.6 | Resource allocation

Finally, participants completed a resource allocation task. Childrenwereshownapictureofanew,individualZarpie(notoneseenbeforeineitherthebook,measuresofinter-groupattitudes,orexposuretothenegativeinformation),givenfourstickers,andtold,’NowyougettogivesomestickerstoaZarpie.Youhavefourstickerstogive.Youcangiveasmanystickersasyouwant.’Childrenthendecidedwhethertogiveeachsticker.Anyremainingstickersweresimplyremovedfromthetable;theywerenotgiventothechildrenthemselves.Thus,givingtotheZarpiewasanon-costlydecisiononthepartofthechild.Weselectedthisapproachbecausepreviousresearchonchildren’sgroup-basedgivingbehavioralsousednon-costlysituations(Dunhametal.,2011;Renno&Shutts,2015).Thesedatawereanalyzedwithbinomialregressionmodels,withthedependentvariablebeingthenumberofdecisionstowithholdastickeroutoffourpossible.Tobeconsistentwiththemeasureofinter-groupattitudes,whichwasscoredsuchthathighernumbersreflectmorenegativeattitudes,resourceallocationsarereportedasprobabilitiesofwithholdingresourcesfromtheZarpie,such that higher numbers indicatemore negative treatment of theZarpie.

2.2 | Results

2.2.1 | Essentialism

Confirming that our manipulation successfully induced essential-ist beliefs, children in theGeneric condition gavemore essentialist

responses (probabilities of essentialist responses,M = .43, CI =.36,.50) than children in the Specific condition (M = .27, CI =.21, .35),Wald χ2(1)=8.72, p = .003, OR =1.99, CI =1.26–3.14. This findingreplicatesRhodesetal.(2012).

2.2.2 | Attitudes

There were no main or interactive effects of condition (Generic,Specific) or time (before the negative information, after the nega-tive information) on children’s attitudes towards Zarpies, ps > .30. Overall,children felt relativelypositively towardsZarpies (M = 1.38,SE =.16; possible range 0–3, with higher numbers indicating morenegative attitudes;Generic condition, before the negative informa-tion,M = 1.30,SE =.24;After,M = 1.16,SE =.23;Specificcondition,before,M = 1.55,SE =.24;After,M = 1.54,SE =.23).


Children in the Generic condition, however, withheld moreresources from the Zarpies (Probabilities of withholding resources,M generic=.45, CI =.35–.56;M specific=.30, CI =.21–.40), Waldχ2(1)=4.25, p = .04. The odds of withholding in the Generic con-dition were 1.95 times as high as those in the Specific condition(CI= 1.03–3.68).

WeconductedfurtheranalyseswiththeProcessProcedureinSPSSto testwhetheressentialismmediatedtheeffectofconditionontheproportion of resourceswithheld from Zarpies. Indeed, this analysisconfirmed that therewasa significant indirecteffectofconditionontheproportionof resourceswithheld throughessentialism (standard-ized indirecteffect=.12,Bootstrapped95%CI=.02–.29).Themedi-atoraccountedforroughlyhalfofthetotaleffectofconditionontheproportionofresourceswithheldfromZarpies(ratioofindirecttototaleffect= .52).Children’sdecisionstowithholdresourceswerenotcor-related with their responses on the attitude measure either before,r(40)=.24,p = .13,orafter,r(40)=.23,p = .14,thenegativeinformation.

2.3 | Discussion

InStudy1,wesuccessfullymanipulatedchildren’sessentialistbeliefsabout a novel out-group. Yet we found that increasing essentialistbeliefs towards thegroupdidnot leadchildren tohavemorenega-tiveattitudestowardsgroupmembers.Providingnegativeinformationabout individualgroupmembersalsodidnot lead tomorenegativefeelings in either condition, suggesting that children – even in theessentialismcondition–didnotgeneralizenegativeactionscommittedbyindividualstothegroupasawhole.Nevertheless,childreninducedviagenericlanguagetoholdessentialistbeliefswerelesslikelytoallo-cate resources tomembersof the essentializedout-group.Becausetheresourceallocationtaskwasalwaysadministeredafterexposureto the negative information, however, it remains unclear whetheressentialismleddirectlytotheseeffects,orwhetheritdidsoonlyoncechildrenwereexposedto informationaboutthenegativeactionsofindividualgroupmembers.ThisissueisaddressedinStudy2.


3 | STUDY 2

Study2had twomainaims.Thefirstwas to replicate theeffectofconditiononchildren’sresourceallocationdecisionsfoundinStudy1,andtotestwhetherthiseffectheldevenifchildrenwerenotexposedto negative information about any individual group members. Thesecondwastotestwhetheressentialismmightindeedinfluencechil-dren’s inter-groupattitudes (notonly their resourceallocationdeci-sions) if children receivemoreexposure togeneric languageoveralongerperiodoftimeandattitudesareassessedusingawiderrangeofmeasures.

3.1 | Methods

Participants included 52 children (M age=5.09, range=4.5–5.5;23male,29female;69%White,16%Asian,13%Latino,5%Black,remainderchosenottoreportthisinformation).Consistentwithourgoalofexaminingtheimplicationsofessentialismforinter-groupatti-tudesattheageatwhichessentialistbeliefsaboutsocialcategorieshavebeenfoundtoemerge,wesampledaslightlyyoungergroupofchildreninStudy2.Thelowerboundwassetat4.5years,however,because pilotwork conducted for Rhodes etal. (2012) found thatyoungerchildrenhaddifficultywithboththelengthoftheresearchsessionandtheverbaldemandsofthetestquestions.Anadditionalfive children began testing but were excluded (four because theirparents reported that they had not completed the assigned book-readingathomeprior to the research session, seebelow,andonebecause they did not complete the research session in the lab).Childrenwererecruitedfromadatabaseoffamilieswhohadsignedup toparticipate in research indevelopmental laboratoriesatNewYorkUniversity.

Familieswerecontactedbyphoneoremailandaskediftheywouldbeinterestedinparticipatinginaparent–childbook-readingstudy.Ifparentsexpressedinterestinthestudy,theywererandomlyassignedtoaGeneric(n = 27)orSpecificcondition(n = 25;asinStudy1)andtheZarpiebookspecifiedbyconditionwassenttothemviamail.Theywereaskedtoreadit8timesinthe10dayspriortotheirvisittothelaboratoryandgivenadiarytorecordtheirbook-readings;allfamiliesincludedinanalysesreportedhavingcompletedtheassignednumberofreadings.Familiesthenvisitedanon-campuslaboratorytocompletemeasuresofessentialismandinter-groupattitudes.

This ’read at home’ procedurewas also used by Gelman etal.(2010).Onepossible concernwith this approachwas that parentswouldformtheirownattitudesaboutZarpiesandthencommunicatethose attitudes to children; if children then adopted their parents’attitudes, this could perhapsmake it more difficult to find effectsof condition on themeasures of inter-group attitudes. To addressthis possibility, parents were also asked to complete measures oftheir attitudes towards Zarpies during the family’svisit to the lab;we found no correlation between parents’ and children’s attitudestowardsZarpiesonanymeasure(ps>.50).Wediscussthisissuefur-therbelow.


Children completed the samemeasures of essentialist beliefs usedinStudy1.Theresearchassistantwhoadministeredthesemeasureswasblindtowhichbookthefamilyhadreadathome(andthuswasblindtothechild’scondition).Thechildcompletedthesessionwiththeresearcherinoneroomwhiletheparentcompletedmeasuresonacomputerinanadjoiningroom.

3.1.2 | Inter- group attitudes

Children completed threemeasures of inter-group attitudes. Theseincluded twoquestions about liking (e.g., ’Look at thisZarpie! Let’spretendyouaregoingtothepark.DoyouwanttoinvitethisZarpieto come to theparkwithyou?’; scored ’0’ for ’yes’, ’1’ for ’no’; theother itemaskedwhetherthechildwould liketoplaywithaZarpieatrecess; itemsweretotaledandanalyzedwithbinomialregressionmodels); two questions about proximity (children were shown anarrayof chairswithaZarpie seatedatoneof themand told: ’Let’spretend you’re at themovies. Oh look, a Zarpie! Let’s pick a seat.Wheredoyouwanttosit?’Theseitemswerescoredasthenumberofseatsawayfromthetargetcharacterandtotaled;rightnexttotheZarpie=0, farthest seat away = 4; the other itemwas structurallyidenticalandaskedthechildtopickaseatinaschoolsetting;itemswereanalyzedwithPoisson regressionmodels); twooutcomeattri-butions, inwhichchildrenwereaskedtoguesswhetheraZarpieornon-Zarpieachievedapositiveoutcome (e.g., ’He isaZarpie.He isnotaZarpie.Look–someonebuiltthebestsandcastle!Itwaseitherthisone,or thisone!Whodoyou thinkbuilt thebest sandcastle?’;scored1=non-Zarpie,0=Zarpie;analyzedwithbinomialregressionmodels).AllitemswerescoredsothathighernumbersindicatemorenegativeattitudestowardsZarpies.


Children also completed two resource allocation trials, one wherethey were asked to allocate up to four stickers (e.g., ’Look at thisZarpie.GoaheadandgiveasmanystickersasyouwanttotheZarpie’)andtheotherwheretheywereaskedtoallocateuptosixtoyboats.DatafromthetwotrialsweretotaledandanalyzedasinStudy1.AsinStudy1,thesearereportedasprobabilitiesofwithholdingresourcesfromZarpies.

3.1.4 | MemoryAttheconclusionofthestudy,childrenwereshownabookcontain-ingtheZarpieillustrations,butwithnoaccompanyingtext,andwereasked to tell the story that they read at home back to the experi-menter.Children’sdescriptionsofeachpagewere scoredas ’1’ forcorrectlydescribingtheintendedpropertyand’0’forfailingtodoso.Forexample,fortheiteminwhichchildrenheard’Zarpieslovetosing’or’ThisZarpielovestosing’,childrenreceivedcreditforrememberingthepageiftheymentionedsinging(e.g.,’shelovestosing’,’shelikes


tosing’,’shesings’)butnotiftheymentionedadifferentbehavior(e.g.,’she’syellingreallyloud’).Thismeasureallowedustoconfirmthatchil-drenrememberedthebookequallywellinbothconditions.Memoryfor thedescribedpropertieswasexcellent (above98%correct)anddidnotvarybycondition.

3.2 | Results


AsinStudy1,childreninthegenericlanguageconditiongavemoreessentialistresponses(M = .55,CI =.48,.62)thanchildreninthespe-cific language condition (M = .40, CI =.33, .47), Wald χ2(1)=9.41,p = .002, OR =1.84, CI =1.25-2.72. These findings are very similartoStudy1andtoRhodesetal.(2012),suggestingthatalthoughtheprocedureheredifferedsomewhat(asparents,insteadofanexperi-menter,readthebooktochildren),themanipulationinfluencedchil-dren’sbeliefsinthemannerintended.

3.2.2 | Inter- group attitudes and resource allocation

Children’s responses to the threemeasuresof inter-groupattitudesand the resource allocation task are in Table1. Children’s feelingstowardsZarpiesdidnotdifferbyconditionasassessedbyanyatti-tudemeasure(invitations,proximity,oroutcomeattribution,ps >.20).Yet,asinStudy1,childrengavefewerresourcestotheZarpiesintheGenericconditionthanintheSpecificcondition;Waldχ2(1)=25.54,p < .001,OR =3.95,CI=2.32–6.73.As shown inTable2, the threemeasuresofinter-groupattitudes(theinvitationsmeasure,proximitymeasure, and outcomemeasure)were all inter-correlatedwith oneanother,butnonewerecorrelatedwithchildren’sresourceallocationdecisions.

3.2.3 | Participant variables

Toconsidertheroleofparticipantgenderandage,wecombineddataacrossStudies1and2toincreaseourpowertodetecteffectsofthesevariables. Even in the combined data set, there was not sufficientpowertoexamineeffectsofchildraceorethnicity;thisisanimpor-tant area for futurework. Although the procedures differed some-whatacrossthetwostudies,bothyieldedverysimilarfindings.Inbothstudies,exposuretogenericlanguageincreasedchildren’sessentialistbeliefs and decreased theirwillingness to allocate resources acrossgroup boundaries, but influenced no other measure of inter-groupattitudes.Herewecombinedthesedatatotestwhethertheeffectofthemanipulationon these twovariables (essentialismand resourceallocation)differedbyparticipantageorgender.

We first tested for effects of participant gender on children’sessentialist responsesandresourceallocationdecisions.Therewereno main or interactive effects of gender on children’s essentialismscores (ps>.15). However, girls (M = .24, CI=.20, .29) withheldmore resources thanboys (M = .16,CI =.12, .20),Waldχ2(1)=7.31,p = .007.Theeffectofconditionremainedsignificantinthisanalysis,Waldχ2(1)=21.16,p < .001(Generic,M = .28,CI=.23,.33;Specific,M = .13,CI =.10, .17),withnointeractionbetweengenderandcon-dition,p = .4.Further,theeffectofconditionwassignificantforbothboysandgirlsexaminedseparately,ps < .05.

In the combined data set, we also tested for effects of partici-pantage(youngerchildren,n = 47,Mage=5.0,SD = .20 years; older children,n = 47,M = 5.8years,SD =.52years;foreaseof interpreta-tion,wedidamedian-splitonage,butexaminingageasacontinuouspredictorrevealedaverysimilarpatternofresults).Youngerchildren(M = .46, CI=.41, .51) gave more essentialist responses than olderchildren(M = .38,CI=.33,.43),Waldχ2(1)=4.47,p = .04.Theeffectofconditionremainedsignificantinthisanalysis,Waldχ2(1)=19.84,p < .001(Generic,M = .50,CI=.45,.55;Specific,M = .34,CI=.29,.39).Agedidnotinteractwithcondition,p > .7,andtheeffectofconditionwassignificantwithineachagegroupexaminedseparately,ps < .01.

For resource allocations, older children (M = .27, CI=.21, .32)withheldmoreresourcesthanyoungerchildren(M = .17,CI=.13,.21),Waldχ2(1)=8.62,p = .003.Theeffectofconditionremainedsignifi-cant in thisanalysis,Waldχ2 (1)=23.14,p < .001 (Generic,M = .30,CI=.26,.35;Specific,M = .14,CI=.11,.19).Agedidnotinteractwithcondition,p > .50,and theeffectofconditionwassignificantwithineachagegroupexaminedseparately,ps<.05.Overall,althoughtheseanalysesrevealedsomebaselinedifferences inchildren’sessentialist

TABLE 1 Children’sinter-groupattitudesandbehaviorsbycondition,Study2

Generic condition Specific condition

Probabilitiesof withholding invitations fromZarpies

M = .48 (CI=.35–.61)

M = .46 (CI=.33–.60)

Averagenumber ofseatsaway fromZarpie (outof4possible)

M = 1.46 (CI=1.14–1.87)

M = 1.74 (CI=1.39–2.18)

Probabilitiesofattributingthe positiveoutcome toanon-Zarpie

M = .61 (CI=.48–.73)

M = .70 (CI=.56–.81)

Probabilities ofwithholding resources fromZarpies*

M = .26 (CI=.21–.31)

M = .08 (CI=.05–.12)

*ProbabilitiesofwithholdingresourcesfromZarpiesdifferedbycondi-tion,p < .001.

TABLE 2 Correlationsamongmeasuresofinter-groupattitudesandbehaviors,Study2

Invitations Proximity Outcomes Resources

Invitations 1 .48** .36** .02

Proximity .48** 1 .25m −.08

Outcomes .36** .25m 1 −.09

Resources .02 −.08 −.09 1

**p < .01; mp = .07.


beliefsandresourceallocationdecisionsbygenderandage,theeffectsofconditiononchildren’sessentialistbeliefsandresourceallocationdecisionsdidnotdifferbytheseparticipantvariables.

3.2.4 | Further examination of children’s explanations

Inthecombineddataset,wealsousedtheexplanationsgeneratedbychildrentoconsideranalternativemechanismbywhichourconditionmanipulationcouldhaveinfluencedresourceallocationdecisions.Inparticular,weconsideredthatchildrenintheSpecificconditionmighthavebeenmorelikelytofocusontheZarpiesasindividualagents.Ifso,perhapschildrenaremorelikelytosharewithpeopletheythinkof as individual agents. As a preliminary test of this possibility,werecodedtheexplanationsthatchildrengeneratedaspartofouressen-tialismmeasurestoobtainanindicatoroftheextenttowhichchildrenthoughtofthecharactersasindividualagents.

Asdescribedearlier,ontheseitems,childrenwereshownapictureofaZarpieandwereaskedtoexplaintheZarpie’sbehavior(e.g.,’WhyisthisZarpieclimbingatallfence?’).Inourmaincoding(asreportedabove),the’scope’scorefocusedontheextenttowhichchildrenref-erencedthecategoryasexertingacausalforce(e.g.,’BecauseZarpieslovetoclimb’).Inthatscheme,everythingthatdidnotreferencethecategoryreceivedascoreof’0’for’scope’;thisincludedexplanationsthatreferencedtheindividualagent(e.g.,’becausehelovestoclimb’)andthosethatdidnot(e.g.,’becauseitisaniceday’)together.Thus,togetameasureoftheextenttowhichchildrenwerethinkingofthechar-actersasindividualactors,werecodedthesedataforwhetherchildrenreferencedtheindividualagentascausal.Childrenwereindeedmorelikely todoso in theSpecificcondition than theGeneric condition,Waldχ2(1)=30.34,p < .001(Generic,M = .48,CI=.40,.56;Specific,M = .80,CI=.73,.86).However,thisscoredidnotcorrelatewiththeirresourceallocationdecisions,r(93) =−.10,ns.Althoughapreliminarytestofthisissue,wethinkthesedataareconsistentwiththeaccountthat the reportedeffectson resourceallocationaredrivenmorebyincreased essentialism in the Generic condition than in increasedindividual-thinkingintheSpecificcondition.

3.3 | Discussion

Study 2 replicated several key findings of Study 1.We once againfoundthatexposuretogenericlanguageaboutanovelsocialcategory(a) increases essentialist beliefs and (b) decreases children’swilling-ness to allocate resources to members of that group. Further, theeffectofconditiononchildren’sresourceallocationdecisionsheldupevenwhen childrenwerenot exposed to anynegative informationaboutgroupmembers.Yet,evenwithmoreexposuretogenericlan-guageoveralongerperiodoftimeandawiderrangeofmeasuresofchildren’sinter-groupattitudes,weconsistentlyfoundthatchildrenintheGenericconditiondidnotholdmorenegativeattitudestowardsZarpies.Further,wefoundthatwhileourthreemeasuresofchildren’sattitudes towards Zarpies were correlated with one another, nonecorrelated with their resource allocation decisions. These findings

supporttheproposalthatresourceallocationdecisionsandchildren’ssocial attitudes depend on different underlying processes,with theprocessunderlyingchildren’sresourceallocationdecisionsbeingpar-ticularlysusceptibletoessentialistbeliefs.

For Study 2, we asked parents to read the book to childrenat home prior to the testing session, aswas done byGelman etal.(2010).AlthoughthismethoddifferedfromStudy1(andRhodesetal.,2012),wefoundsimilareffectsoflanguageconditiononessentialistbeliefs as in Study 1 andRhodes etal. (2012), suggesting that par-entsadministeredthemanipulationwithsufficientfidelitytoproducethe intendedeffects.Thevery similarpatternsacrossStudies1and2,aswellasthelackofcorrelationsbetweenchildren’sandparents’attitudestowardsZarpies(asdiscussedabove),indicatethattheread-at-homeproceduredidnotbiaschildren’sresponsesinaproblematicmanner.Further,theread-at-homeprocedurehadtheaddedbenefitthat the experimenter in Study 2was blind to the child’s conditionwhileadministeringthedependentmeasures (whichwasone limita-tionofStudy1).

4 | STUDY 3

InStudies1and2,wemanipulatedchildren’sessentialistbeliefsaboutanovelgroupofpeople–Zarpies.Zarpieswerepresentedasanout-group(aspresumablychildrenwouldrealizethattheythemselvesarenotZarpies),butwedidnothighlightthisout-groupstatus.Thus,oneopenpossibilityisthatessentialistbeliefswouldleadtomorenegativeattitudestowardsanovelgroupiftheout-groupstatusofthatgroupweremademoresalienttochildren. InStudy3,wetestedthispos-sibilitybyhighlightingtheout-groupstatusofZarpiesinbothspecificandgeneric languageconditions,byassigningchildrentotheirown,differentnovelin-group–Gorps.Childrenreceivedthesamegenericor specific input aboutZarpies as in Studies 1–2; thuswemanipu-latedessentialistbeliefsabouttheout-grouponly(notaboutthein-group). Children then completedmeasures of inter-group attitudesandbehaviorsimilartothoseinStudies1–2.

Consistentwithpriorwork,weexpectedchildrentoholdbiasedattitudesinfavoroftheirowngroupsinbothconditionsofthisstudy;that is,weexpectedthemtoprefer to invitemembersof theirowngroup,tosaytheywouldsitclosertomembersoftheirowngroup,andtosharemoreresourceswiththeirowngroupmembers.Previousworkhasshownthatchildrenoftheseagesreadilyshowsuchbiasesbasedonnew,arbitrarygroupmemberships towhich theyareassigned inexperimentalsettings(Dunhametal.,2011;Patterson&Bigler,2006).Ourkeyquestionwaswhetheressentialistbeliefs–inducedbygenericlanguage–wouldaccentuatethesegroupbiases.

InStudy3,wealsoaimedtopinpointmorepreciselythecompo-nentofessentialistbeliefs that leadschildrentowithhold resourcesfromout-groupmembers.Inparticular,wefocusedonthecomponentof essentialist beliefs that entails viewing the boundaries betweensocialgroupsasstrict,absolute,andinflexible(Noetal.,2008;Rhodes&Gelman, 2009).Among adults, this component of essentialism isthoughttodecreaseexpectationsofinter-groupinteractions(Zagefka


etal.,2012).Similarly,amongsomewhatolderchildren (age11), thetendencytoviewpersonalitiesasfixed(similartoanessentialistbelief)wasassociatedwith lessdesireto interactwithout-groupmembers(Levy&Dweck,1999).Asresourceallocationdecisionsoftendependonexpectationsofsocial reciprocity (Dunhametal.,2011;Renno&Shutts,2015;Yamagishietal.,1999),wehypothesizedthatthiscom-ponentofessentialism–viewingboundariesasstrict,absolute,andinflexible–wouldbeparticularlyimportantforexplainingtherelationbetween essentialism and resource allocation behavior observed inStudies 1 and 2. From this perspective, essentialismmakes peoplemore likely toview the boundary between their own group and anessentializedout-groupasabsolute,andthusdecreasesexpectationsof future opportunities for reciprocity, and consequently, decreaseswillingnesstoallocateresourcesacrossgroupboundaries.

InStudies1and2,however,onlyoneofthreeessentialismmea-sures – the ’scope’ score derived from the explanation measure –directlytapsthiscomponentofessentialism(asdescribedbyGelmanetal.2010).Theothertwomeasures(the’cause’scorederivedfromtheexplanationmeasureandtheinheritancequestions)insteadassessaspects of essentialist beliefs that relate toviewing category-linkedproperties as caused by intrinsic, innate qualities.Thus, in Study 3,weaddedanadditionalmeasureofchildren’sbeliefsaboutcategoryboundaries, so thatwe could statistically testwhich component ofessentialistbeliefsbestexplainstherelationbetweenessentialismandresourceallocationdecisions.

4.1 | Methods

Participantsincluded33children(10male,23female;Mage=5.16,range=4.5–6.0;58%White,12%Black,9%Hispanic,theremainderdid not provide this information) recruited from and tested at theChildren’sMuseumofManhattanoraprivatepreschool.

4.1.1 | Group introduction

First,childrenwere introducedtotwonovelgroupsandgiventheirgroupassignment,asfollows:

Today we are going to talk about two groups of kids called Zarpies and Gorps. Here are some Zarpies – they are wearing yellow. Here are some Gorps – they are wearing green. First we have to figure out which group you should go in. Let me ask you some questions to see which group you should go in. Which do you like better, cookies or cupcakes? (Regardless of answer): Okay, great, these kids who are Gorps like (child’s answer) better too. I think you should be in the Gorp group, but let me ask you one more question to be sure. Which do you like better, swimming or playing outside? (Regardless of answer): Okay, great, these kids who are Gorps like (child’s answer) better too. You are definitely in the Gorp group. Here, take this green sticker to show that you are in the Gorp group. And let me give you this green scarf to remind us that you are in the

Gorp group too. Here are some Zarpies and here are some Gorps. Can you point to someone who is in your group? And how about someone who is not in your group?

Ifchildrenfailedtoaccuratelypointtoanin-grouporout-groupmem-ber,theexperimenterremindedthechild,’YouareintheGorpgroup.HerearesomeZarpiesandherearesomeGorps’andre-askedthesequestions.Subsequently,theexperimentersaid,’NowwearegoingtoreadastoryaboutsomeZarpies.Remindme,areyouaZarpie?That’sright,youarenotaZarpie,youareaGorp.Let’sreadthisbookaboutsomeZarpiesthenwe’llanswersomequestionsaboutsomeZarpiesandsomeGorps.’TheexperimenterthenreadtheZarpiebookusedin Studies 1–2 twice to the child (as specified by the child’s condi-tion,n = 19Generic,n = 14Specific).Afterthebook-reading,theex-perimenter said, ’Nowwe’regoing toanswersomequestionsaboutsomeZarpiesandsomeGorps.Remindme,areyouaZarpieoraGorp?That’sright,youareaGorp.’

4.1.2 | Essentialism measures

Children completed the measures of essentialist beliefs used inStudies 1–2. Also, to begin to pinpoint the features of essentialistbeliefsthatmightshapechildren’sresourceallocationdecisions,weincludedanadditionalmeasureofchildren’sbeliefsthatthebounda-riesbetweencategoriesarediscreteandobjective.Inparticular,chil-drenwereaskedanadditionalfollow-upquestionaftereachinherit-ance item, askingwhether theywould also endorse the alternativeofferedproperty.Forexample,ifthechildrespondedthatthegrown-upchildwouldflap itsarmswhen it ishappy, like theZarpiemom,thefollow-upquestionwouldbe, ’Doyouthinkthechildmightalsoclap its handswhen it is happy, like theGorpmom?’Rejecting thealternativepropertyontheseitems(scored’1’)indicatesthatchildrenreject the possibility that the grown-up child in the vignettemightdisplaythepropertiesofbothgroups,andthusthattheybelievetheboundariesbetweengroupsarediscreteandabsolute (Tayloretal.,2009).To increaseourpowertodetecteffectsof these judgments,weincreasedthenumberof inheritance itemsforthisstudyfrom2to3.Foranalyses,wefirstexaminedtheessentialismcompositeusedinStudies1–2, includingtheexplanationitemsandinitialresponsestotheinheritancequestions,inordertofacilitatecomparisonsacrossstudies.Subsequently,wetesttheextenttowhichacompositescoremeasuringchildren’sbeliefsaboutcategoryboundaries(consistingofthe sumof the scope score and the newmeasure of beliefs aboutthe discreteness of categories) and a composite score measuringchildren’sbeliefsaboutintrinsiccauses(consistingofthesumofthecause score and the inheritancemeasure) account for theeffectofconditiononchildren’sbehavior.

4.1.3 | Inter- group attitudes and resource allocation measures

Next,childrencompletedadaptedversionsoftwooftheinter-groupattitudesmeasuresusedinStudy2–theinvitationsmeasureandthe


proximitymeasure–alongwithameasureoftheirresourcealloca-tiondecisions.Childrenwereaskedabouttwoitemseachforallthreemeasures, and foreachone,wereasked tochoosebetweenan in-groupmemberandanout-groupmember; forexample, Invitations:’DoyouwanttoinvitethisGorptogotothecircuswithyou,ordoyouwanttoinvitethisZarpietogotothecircuswithyou?’;Resourceallocation:’LookatthisGorp,andlookatthisZarpie.GoaheadandgiveasmanytrainsasyouwanttotheGorpandasmanytrainsasyouwanttotheZarpie.’Seatdistance:’Let’spretendyou’reoutside.Ohlook,aGorpandaZarpie!Let’spickaseat.Wheredoyouwanttosit?’Fortheresourceallocationtask,childrencompletedtwotri-als,onewheretheywereaskedtodistributefourresources(e.g.,toyboats)andoneinwhichtheywereaskedtodistributesixresources(e.g.,toytrains).Althoughchildrenwerenotrequiredtodistributeallresourcestoonegrouportheother,theychosetodosoinpractice.Theseitemsarereportedasprobabilitiesofitemswithheldfromtheout-group(asinpreviousstudies);inthiscase,then,higherprobabili-tieswithheldfromtheout-groupmeansthatmoreitemsweregiventothein-group.

4.2 | Results


AsinStudies1–2,childrenintheGenericconditiongavemoreessen-tialist responses (M = .60,CI=.53, .67) thanchildren in theSpecificcondition (M = .33, CI=.25, .41), Wald χ2 (1)=21.62, p < .001,OR =3.14,CI =1.94–5.09.

4.2.2 | Inter- group attitudes

Children showed in-group biases on both measures of inter-groupattitudes,buttheextentofthesebiasesdidnotvarybylanguagecon-dition.Childrenreliablyinvitedtheirin-groupmemberovertheirout-groupmember (M = .78,CI=.66, .87),Waldχ2 (1)=15.74,p < .001 (OR =3.13,CI=1.78,5.49;responsesdidnotvarybycondition,Waldχ2(1)=2.51,p = .11,Generic,M = .68,CI=.52,.81;Specific,M = .86,CI=.68,.95).Theyalsochosetositfartherawayfromtheout-groupmember(andthusclosertothein-groupmember)thanexpectedbychance (M = 1.35, SD=.59, t(32)=3.38, p < .002; again, responsesdid not vary by condition, t(31)=1.44, p = .16; Generic,M = 1.47,SD=.49;Specific,M = 1.18,SD=.70).


Overall,childrengavefewerresourcestotheout-group(andthereforedesignatedmoretothein-group)thanexpectedbychance(M = .39,CI=.34, .45),Waldχ2 (1)=14.62,p < .001.ConsistentwithStudies1and2,however,children’s resourceallocationsdifferedbycondi-tion.ChildrengavefewerresourcestotheZarpie(andthusmoretotheGorp)intheGenericcondition(M = .35,CI=.28,.42)thanintheSpecificcondition(M = .46,CI=.38,.54),Waldχ2(1)=4.05,p = .04.(OR=1.58,CI=1.01,2.47).

Further,analysesusingtheProcessProcedureinSPSSconfirmedthatchildren’sbeliefsaboutthediscretenessofcategoryboundariesmediatedtheeffectofconditionontheproportionofresourceswith-heldfromZarpies,whereastheirbeliefsabouttheinnateorintrinsicnatureofcategory-linkedpropertiesdidnot.Inparticular,theseanaly-sesconfirmedasignificantindirecteffectofconditiononwithholdingresources from Zarpies via the category boundariesmeasure (stan-dardizedindirecteffect=.17,CI=.0003,.48),withcategoryboundar-iesaccountingforapproximately70%ofthevariance(ratioofindirecteffecttototaleffect=.71).Incontrast,theindirecteffectviabeliefsabout the innate nature of category propertieswas not significant(standardizedindirecteffect=−.06,CI=−.25,.03).

4.3 | Discussion

Study3replicatedthefindingsofStudies1–2indicatingthatexposuretogeneric language increasesessentialistbeliefsaboutanewsocialcategory(asevidencedbyincreasedessentialismintheGenericcom-pared to theSpecific condition), extending thisfinding to instanceswhentheout-groupstatusofthisgroupismademoresalienttochil-dren.Wealsoreplicatedthefindingthatinducingessentialistbeliefsleads children to share fewer resources with out-group members.Further, we conceptually replicated the findings of Dunham etal.(2011) and others, who have found that children rapidly developin-grouppreferences,evenfornovelsocialgroups–childreninthisstudy preferred to affiliate with members of their own groups (onthe’invitations’measure),satclosertotheir in-groupmembersthanout-groupmembers(onthe’proximity’measure),andalsodesignatedmore resources to their in-group members when asked to choosebetween allocating to in-group and out-group members. That wefoundin-groupbiasonallofthemeasures,includingtwoofthemeas-uresofsocialpreferenceusedinStudy2,showsthatallofthemeas-uresusedherewereindeedsensitivetovariationinchildren’sbeliefs.Nevertheless,essentialisminfluencedonlyresponsesontheresourceallocationtask–weonceagainreplicatedthenullresultsfromStudies1and2regardingtheeffectofessentialismonchildren’sinter-groupattitudes,evenwhentheout-groupstatusofZarpieswashighlighted.Study3alsorevealedthatvariation in theextent towhichchildrenperceivedtheboundariesbetweenZarpiesandGorpsasabsoluteanddiscreteaccountedfortherelationbetweentheirassignedconditionandtheirresourceallocationdecisions.

5 | GENERAL DISCUSSION

Thesestudiesexaminedhowessentialistbeliefs influencechildren’sinter-group attitudes and behaviors. In three studies, essentialistbeliefs related to children’s resource allocation decisions. In Study1, children induced to hold essentialist beliefs about a novel groupshared fewer resources with group members after hearing aboutinstancesofnegativebehavioronthepartofspecificindividualgroupmembers. In Study 2, children induced to hold essentialist beliefswithheldresourceseventhoughtheyhadnotbeenexposedtosuch


instancesofnegativebehavior.InStudy3,childrenwithessentialistbeliefsagainwithheldmoreresourceswhenthegroupinquestionwaspresentedasanexplicitout-group.Yet,innostudydidchildrenseemtofeelmorenegativelytowardsZarpiesasafunctionofessentialism.Thispatternisconsistentwiththepossibilitythatessentialisminflu-ences inter-groupbehaviorby increasingbeliefsaboutthediscrete-ness of category boundaries (seeGaither etal., 2014), thusmakingchildren more likely to consider a person’s out-group status whenmakingjudgmentsaboutsocialreciprocity(seeDunhametal.,2011;Paulus&Moore,2014;Yamagishietal.,1999).

Thesedataarethefirsttorevealthatthedevelopmentofessen-tialistbeliefsaboutanentirelynovelgrouphasimmediatenegativeconsequencesforchildren’sinter-groupbehavior.Previousresearchhasdocumentedcorrelationsbetweenchildren’sessentialistbeliefsabout familiar social categories and some other aspects of inter-group cognition (e.g., correlations between racial essentialism andendorsing racial stereotypes; Levy & Dweck, 1999; Pauker etal.,2010),andDiesendruckandMenahem(2015)foundthatincreasingchildren’sessentialistbeliefsaboutfamiliargroups influencedtheirimplicit attitudes in some cases.Whilevery informative, this priorwork leaves open questions regardingwhether essentialist beliefsplayacausal role inthedevelopmentofthesenegativefeaturesofsocial cognition, whether essentialism alone is sufficient to causenegativesocialoutcomes(orwhetheritdoessoonlywhenitinter-acts with other aspects of children’s knowledge and experienceswith familiar categories), andwhether the negative consequencesof essentialism for inter-group relations extend beyond increasedstereotyping.Ourexperimentalapproachwithnovelgroupsallowedustoaddressthesequestions.Wefoundthat inducingessentialistbeliefs in children – even if theywere providedwith no negativeinformationaboutanewgroup–immediatelyledchildrentobelesswilling to share resources across groupboundaries.Thus,wepro-vide the first evidence that essentialismplays a causal role in thedevelopment of some negative inter-group phenomena, that suchphenomenaextendbeyondstereotypingtoinfluenceresourceallo-cationdecisions,andthatessentialismaloneissufficienttoproducetheseeffects.

These data also reveal, however, that essentialismmight play amorelimitedroleinthedevelopmentofsomeinter-groupphenomenathan suggested by previous theorizing (Allport, 1954). In particular,across three studies andmultiplemeasures,we found no evidencethatessentialismledchildrentoholdnegativeattitudestowardsmem-bers of essentialized out-groups.Thus, essentialism alone does notseemsufficienttoleadtosocialprejudice,atleastinearlychildhood.Thesefindingsdonotprecludethepossibility,however,thatessential-ismplaysacausalroleinthedevelopmentofprejudicedattitudesviamorecomplexsocialordevelopmentalphenomena thanweconsid-eredhere.Forexample,essentialismcouldcontributetoprejudiceinthecontextofstatusdifferencesbetweengroups,oncechildrenaccu-mulate personal experiencewith in-group and out-groupmembers,orifchildrenareexplicitlyexposedtonegativestereotypes(Bigler&Liben, 2007; Birnbaumetal., 2010;Cimpian& Salomon, 2014; seealsoRutland,Killen,&Abrams,2010).Ourfindingsonlysuggestthat

essentialistbeliefs–ontheirown–donot immediately leadtothedevelopmentofnegativefeelingstowardsout-groupmembers;theydonotprecludethepossibilitythatessentialismcontributestosocialprejudiceinthecontextofchildren’sbroaderexperiences(seeLeslie,in press). Experimentally testing how essentialism interacts withothertypesofinformation(e.g.,informationaboutstatusdifferences,negativestereotypes,andsoon)isanimportantdirectionforfutureresearch.

Revealing that essentialist beliefs – alone – influence resourceallocationdecisionsbutdonotelicitnegative feelings towardsout-groupmembersprovidessomeinsightintothemechanismsunderly-ingthesevariousinter-groupphenomena.Inparticular,thesefindingsare consistent with proposals that resource allocation decisions ininter-groupcontextsdonotreflectsimplesocialpreferences(Renno&Shutts,2015;Yamagishi etal., 1999). In three studies, ouressen-tialism manipulation consistently affected resource allocation deci-sionsbutnotsocialpreferences,suggestingthatdifferentprocessesunderliethesephenomena.Moredirectly,althoughourmultiplemea-suresofsocialpreferences(e.g.,choicesaboutproximity,decisionstoissueinvitations,andoutcomepredictions)werecorrelatedwithoneanother,innostudydidanyoftheseattitudemeasurescorrelatewithchildren’s allocationdecisions.Rather,wepropose that essentialisminfluencessuchdecisionsby increasingtheextenttowhichchildrenview the boundaries between groups as discrete and absolute, andtherefore the likelihood that they incorporate such boundaries intotheirexpectationsofsocialreciprocity.

Further, these data may help to distinguish various theoreticalmodels of why essentialism often correlates with prejudice amongadultpopulations.Oneperspectiveinadultsocialpsychologyhassug-gestedthatessentialismleadstoprejudicebecause it increasesper-ceptionsofgroupdifferences,perhapsleadingpeopletoviewin-groupandout-groupmembersasfundamentallydistinctkindsofpeopleandthus to dehumanize out-groups (Leyens etal., 2001) and attributegroup differences to biological or immutable factors (Keller, 2005).Althoughwedidnotdirectlymeasureperceptionsofgroupdifferenceshere,thepresentdataaredifficulttoreconcilewiththisperspective,which predicts that essentialism should lead directly to prejudice.Childreninthepresentstudieswhowereinducedtoholdessentialistbeliefsviewedcategory-linkedpropertiesasinnatelydeterminedandimmutable,expectedsuchpropertiestobesharedacrossgroupmem-bers,andviewedintrinsicfactorsasresponsibleforcategory-typicalbehaviors.Yet,thesebeliefsalonedidnotleadthemtofeelmoreneg-ativelytowardsmembersoftheessentializedgroup.Thus,essential-ismdoesnotappeartoleaddirectlytoprejudiceviamechanismsthatrelatesolelytoemphasizinggroupdifferenceorattributingsuchdif-ferences to immutableorbiologicalcauses.Alternativeperspectivesonwhyessentialismmightrelatetosocialprejudiceamongadultpop-ulationsdependmoreonhowessentialismrelates toother featuresofsocialexperience,oronhowessentialismmightbeusedtoexplainstatus differences or other features of the social environment, andaremorecommensuratewiththepresentdata (Mortonetal.,2009;Pettigrew,1979;Rangel&Keller,2011;Verkyuten,2003).Heretoo,though, thepresent data suggest amore complex story thanmight


have been assumed. Evenwhen children essentialized a social out-groupandwerepresentedwithevidencethatmembersofthatgroupengageinnegativebehaviorssuchasstealing,essentialismstilldidnotpredictincreasedprejudice.Whilenotyetdefinitive,thissuggeststhatchildrendonotassumethatallbehaviorsassociatedwithanessential-izedgrouparedeeppropertiesofthegroupsuchthattheywouldbeexpectedtogeneralizebroadly.

The present research suggests several key directions for futurework. For example, no prior work has examined how essentialistbeliefs relate to children’s social attitudes (e.g., prejudice) or inter-group behavior (e.g., resource allocations) for categories that theyencounter in their daily lives (prior work in this area has insteadfocusedonstereotyping;Levy&Dweck,1999;Paukeretal.,2010).In futurework, itwouldbeuseful to combineexperimental studies–likethoseperformedhere–withstudiesexamininghowchildren’spre-existingessentialistbeliefsaboutparticulargroups(e.g.,theirlev-elsofracialessentialism)predictsimilarbehaviors.Suchanapproach,combinedwithmoredetailedexaminationoftheunderlyingprocessesthanwas conducted here, could eventually reveal the mechanismsthat contribute to the development of the most virulent forms ofsocialprejudice.Further,itwouldbeusefultoexpandtheagerangesincluded, to consider how the processes that underlie inter-groupattitudesandbehaviorschangeacrossage.Finally, itwouldbeveryhelpfultoincludemorediversesubjectpopulations,astherelationofessentialistbeliefs tovarious inter-groupphenomena likelydependsonchildren’sownexperiences inessentializedgroups (seeKinzler&Dautel,2012;Noetal.,2008).

ACKNOWLEDGEMENTS

Wethank theChildren’sMuseumofManhattan forparticipating inthisresearch.FundingwasprovidedbyNSFgrantBCS-1226942toRhodesandLeslie.

REFERENCES

Aboud,F.E.(1988).Children and prejudice.Cambridge:BasilBlackwell.Aboud, F.E. (2003).The formation of in-group favoritism and out-group

prejudiceinyoungchildren:Aretheydistinctattitudes?Developmental Psychology,39,48–69.

Allport,G.(1954).The nature of prejudice.Cambridge:Addison-Wesley.Andreychik, M., & Gill, M. (2015). Do natural kind beliefs about social

groupscontribute toprejudice?Distinguishingbio-somaticessential-ismfrombio-behavioralessentialism,andbothofthesefromentitativ-ity.Group Processes & Intergroup Relations,18,454–474.

Astuti,R.,Solomon,G.E.,&Carey,S.(2004).Constraintsonconceptualde-velopment.Monographs of the Society for Research in Child Development,69,vii–135.

Bastian,B.,&Haslam,N.(2006).Psychologicalessentialismandstereotypeendorsement.Journal of Experimental Social Psychology,42,228–235.

Bernstein,M.,Sacco,D.,Young,S.,Hugenberg,K.,&Cook,E.(2010).Being’in’withthein-crowd:Theeffectsofsocialexclusionandinclusionareenhanced by the perceived essentialism of ingroups and outgroups.Personality and Social Psychology Bulletin,36,999–1009.

Bigler,R.S.,Jones,L.C.,&Lobliner,D.B. (1997).Socialcategorizationandthe formationof intergroup attitudes in children.Child Development,68,530–543.

Bigler, R.S., & Liben, L.S. (2007). Developmental intergroup theory:Explaining and reducing children’s social stereotyping and prejudice.Current Directions in Psychological Science,16,162–166.

Birnbaum,D.,Deeb,I.,Segall,G.,Ben-Eliyahu,A.,&Diesendruck,G.(2010).Thedevelopmentofsocialessentialism:Thecaseof Israelichildren’sinferencesaboutJewsandArabs.Child Development,81,757–777.

Brewer,M.B.(1999).Thepsychologyofprejudice:Ingrouploveoroutgrouphate?Journal of Social Issues,55,429–444.

Brewer,M.,&Brown,R.(1998).Intergrouprelations.InD.Gilbert,S.Fiske&G.Lindzey(Eds.),The handbook of social psychology,Vol.1(4thedn.,pp.554–594).NewYork:McGraw-Hill.

Chao,M.,Hong,Y.,&Chiu,C.(2013).Essentializingrace:Itsimplicationsonracialcategorization.Journal of Personality and Social Psychology,104,619–634.

Cimpian,A.,Arce,H.C.,Markman,E.M.,&Dweck,C.S.(2007).Subtlelin-guistic cues affect children’s motivation. Psychological Science, 18,314–316.

Cimpian,A.,&Markman, E.M. (2009). Information learned from genericlanguagebecomescentral tochildren’sbiologicalconcepts:Evidencefromtheiropen-endedexplanations.Cognition,113,14–25.

Cimpian,A.,&Markman,E.M.(2011).Thegeneric/nongenericdistinctioninfluenceshowchildreninterpretnewinformationaboutsocialothers.Child Development,82,471–492.

Cimpian,A., & Salomon, E. (2014).The inherence heuristic:An intuitivemeansofmakingsenseoftheworld,andapotentialprecursortopsy-chologicalessentialism.Behavioral and Brain Sciences,37,461–480.

Deeb,I.,Segall,G.,Birnbaum,D.,Ben-Eliyahu,A.,&Diesendruck,G.(2011).Seeingisn’tbelieving:Theeffectofintergroupexposureonchildren’sessentialistbeliefs aboutethnic categories. Journal of Personality and Social Psychology,101,1139–1156.

Diesendruck,G.,Goldfein-Elbaz,R.,Rhodes,M.,Gelman,S.A.,&Neumark,N.(2013).Cross-culturaldifferencesinchildren’sbeliefsabouttheob-jectivityofsocialcategories.Child Development,84,1906–1917.

Diesendruck,G.,&Haber,L.(2009).God’scategories:Theeffectofreligi-osityonchildren’steleologicalandessentialistbeliefsaboutcategories.Cognition,110,100–114.

Diesendruck, G., & haLevi, H. (2006). The role of language, appear-ance, and culture in children’s social category-based induction.Child Development,77,539–753.

Diesendruck,G.,&Menahem,R.(2015).Essentialismpromoteschildren’sinter-ethnicbias.Frontiers in Psychology,6,1180.

Dunham,Y., Baron,A.S., & Carey, S. (2011). Consequences of ’minimal’groupaffiliationsinchildren.Child Development,82,793–811.

Gaither, S., Schulz,J., Pauker,K., Sommers, S.,Maddox,K.,&Ambady,N. (2014). Essentialist thinking predicts decrements in children’smemoryforraciallyambiguousfaces.Developmental Psychology,50,482–488.

Gelman,S.A. (2003).The essential child: Origins of essentialism in everyday thought.NewYork:OxfordUniversityPress.

Gelman, S.A. (2004). Psychological essentialism in children. Trends in Cognitive Sciences,8,404–409.

Gelman,S.A.,Collman,P.,&Maccoby,E.E.(1986).Inferringpropertiesfromcategoriesversusinferringcategoriesfromproperties:Thecaseofgen-der. Child Development,57,396–404.

Gelman,S.A.,&Heyman,G.D.(1999).Carrot-eatersandcreature-believers:Theeffectsoflexicalizationonchildren’sinferencesaboutsocialcate-gories.Psychological Science,10,489–493.

Gelman,S.A.,&Kalish,C.W.(2006).Conceptualdevelopment.InD.Kuhn&R.Siegler(Eds.),Handbook of child psychology, Vol. 2: Cognition, percep-tion and language(pp.687–733).NewYork:Wiley.

Gelman,S.A.,&Markman,E.M.(1986).Categoriesandinductioninyoungchildren. Cognition,23,183–209.

Gelman, S.A., & Markman, E.M. (1987). Young children’s inductionsfrom natural kinds: The role of categories and appearances. Child Development,58,1532–1541.


Gelman,S.A.,&Rhodes,M. (2012). ’Two-thousandyearsof stasis’:Howpsychological essentialism impedes evolutionary understanding. InK.R.Rosengren,S.Brem,E.M.Evans&G.Sinartra(Eds.),Evolution chal-lenges: Integrating research and practice in teaching and learning about evolution(pp.3–21).NewYork:OxfordUniversityPress.

Gelman,S.A.,Ware,E.A.,&Kleinberg,F.(2010).Effectsofgenericlanguageoncategorycontentandstructure.Cognitive Psychology,61,273–301.

Gelman,S.A.,&Wellman,H.M.(1991).Insidesandessences:Earlyunder-standingsofthenonobvious.Cognition,38,213–244.

Haslam,N.,Bastian,B.,Bain,P.,&Kashima,Y.(2006).Psychologicalessen-tialism,implicittheories,andintergrouprelations.Group Processes and Intergroup Relations,9,63–76.

Haslam,N.,Rothschild,L.,&Ernst,D.(2002).Areessentialistbeliefsasso-ciatedwithprejudice?British Journal of Social Psychology,4,87–100.

Haslam,N.,&Whelan,J.(2008).Humannatures:Psychologicalessential-isminthinkingaboutdifferencesbetweenpeople.Social and Personality Psychology Compass,2,1297–1312.

Hirschfeld, L. (1995). Do children have a theory of race?Cognition, 54,209–252.

Hirschfeld,L.A.(1996).Race in the making. Cambridge.MA:MITPress.Hodson, G., & Skorska, M. (2015). Tapping generalized essentialism to

predict outgroup prejudice. British Journal of Social Psychology, 54,371–382.

Hoffman,C.,&Hurst,N.(1990).Genderstereotypes:Perceptionorratio-nalizations?Journal of Personality and Social Psychology,58,197–208.

Holtz,P.,&Wagner,W.(2009).Essentialismandattributionofmonstros-ity in racist discourse: Right-wing internet postings aboutAfricansand Jews. Journal of Community & Applied Social Psychology, 19,411–425.

Hussak,L.J.,&Cimpian,A.(2015).Anearly-emergingexplanatoryheuristicpromotessupportforthestatusquo.Journal of Personality and Social Psychology,109,739–752.

Jayaratne,T.E.,Ybarra,O.,Sheldon,J.P.,Brown,T.N.,Feldbaum,M.,Pfeffer,C.A.,&Petty, E.M. (2006).WhiteAmericans’ genetic lay theories ofrace differences and sexual orientation:Their relationshipwith prej-udice towards Blacks, gay men, and lesbians. Group Processes and Intergroup Relations,9,77–94.

Kalish,C.(2012).Generalizingnormsandpreferenceswithinsocialcatego-ries and individuals. Developmental Psychology,48,1133–1143.

Keller,J.(2005).Ingeneswetrust:Thebiologicalcomponentofpsycholog-icalessentialismanditsrelationshiptomechanismsofmotivatedsocialcognition.Journal of Personality and Social Psychology,88,686–702.

Kinzler,K.,&Dautel,J.(2012).Children’sessentialistreasoningaboutlan-guageandrace.Developmental Science,15,131–138.

Leslie,S.J.(2008).Generics:Cognitionandacquisition.Philosophical Review,117,1–47.

Leslie, S.J. (2013). Essence and natural kinds:When sciencemeets pre-schoolerintuition.Oxford Studies in Epistemology,4,108–165.

Leslie,S.J.(inpress).Theoriginalsinofcognition:Fear,prejudice,andgen-eralization.Journal of Philosophy.

Levy, S.R., & Dweck, C.S. (1999). The impact of children’s static ver-sus dynamic conceptions of people on stereotype formation. Child Development,70,1163–1180.

Levy,S.R.,Stroessner,S.J.,&Dweck,C.S.(1998).Stereotypeformationandendorsement:The roleof implicit theories. Journal of Personality and Social Psychology,74,1421–1436.

Leyens,J.P.,Rodriguez-Perez,A.,Rodriguez-Torres,R.,Gaunt,R.,Paladino,P.M.,Vaes,J.,&Demoulin,S.(2001).Psychologicalessentialismandthedifferential attribution of uniquely human emotions to ingroups andoutgroups.European Journal of Social Psychology,31,395–411.

Mahalingam,R.(2003).Essentialism,culture,andpower:Representationsofsocialclass.Journal of Social Issues,59,733–749.

Mahalingam,R.(2007).Essentialism,power,andtherepresentationofso-cialcategories:Afolksociologyperspective.Human Development,50,300–319.

Mahalingam,R.,&Rodriguez,J.(2006).Culture,braintransplantsandim-plicittheoriesofidentity.Journal of Cognition and Culture,6,453–462.

Martin,C.,&Parker,S. (1995).Folktheoriesofsexandracedifferences.Personality and Social Psychology Bulletin,21,45–57.

Mayr,E.(1991).One long argument: Charles Darwin and the genesis of mod-ern evolutionary thought.Cambridge,MA:HarvardUniversityPress.

Medin,D.,&Ortony,A.(1989).Psychologicalessentialism.InS.Vosniadou&A.Ortony (Eds.), Similarity and analogical reasoning (pp. 179–195).NewYork:CambridgeUniversityPress.

Morton,T.,Hornsey,M.,&Postmes,T.(2009).Shiftingground:Thevariableuseofessentialismincontextsofinclusionandexclusion.British Journal of Social Psychology,48,35–59.

Nesdale,D. (2004).Social identityprocessesandchildren’sethnicpreju-dice.InM.Bennett&F.Sani(Eds.),The development of the social self(pp.219–246).London:PsychologyPress.

No,S.,Hong,Y.,Liao,H.,Lee,K.,Wood,D.,&Chao,M. (2008).Laythe-oryofraceaffectsandmoderatesAsianAmericans’responsestowardsAmerican culture. Journal of Personality and Social Psychology, 95,991–1004.

Patterson,M.M., & Bigler, R.S. (2006). Preschool children’s attention toenvironmentalmessagesaboutgroups:Socialcategorizationand theoriginsofintergroupbias.Child Development,77,847–860.

Pauker,K.,Ambady,N.,&Apfelbaum,E.(2010).Racesalienceandessen-tialistthinkinginracialstereotypedevelopment.Child Development,81,1799–1813.

Paulus,M.,&Moore,C.(2014).Thedevelopmentofrecipient-dependentsharing behavior and sharing expectations in preschool children.Developmental Psychology,50,914–921.

Pettigrew,T.F. (1979).The ultimate attribution error: ExtendingAllport’scognitiveanalysisofprejudice.Personality and Social Psychology Bulletin,5,461–476.

Plaks,J.,Stroessner,S.,Dweck,C.,&Sherman,J.(2001).Persontheoriesandattentionallocation:Preferencesforstereotypicversuscounterst-ereotypicinformation.Journal of Personality and Social Psychology,80,876–893.

Prentice,D.A.,&Miller,D.T. (2007).Psychologicalessentialismofhumancategories.Current Directions in Psychological Science,16,202–206.

Rangel,U.,&Keller,J.(2011).Essentialismgoessocial:Beliefinsocialde-terminism as a component of psychological essentialism. Journal of Personality and Social Psychology,100,1056–1078.

Renno,M.,&Shutts,K.(2015).Children’ssocialcategory-basedgivinganditscorrelates:Expectationsandpreferences.Developmental Psychology,51,533–543.

Rhodes,M.,&Brickman,D.(2010).Theroleofwithin-categoryvariabilityincategory-basedinduction:Adevelopmentalstudy.Cognitive Science,34,1561–1573.

Rhodes,M.,&Brickman,D.(2011).Theinfluenceofcompetitiononchil-dren’s social categories. Journal of Cognition and Development, 12,194–221.

Rhodes,M.,&Gelman,S.A. (2009).Adevelopmentalexaminationof theconceptual structureofanimal, artifact, andhumansocial categoriesacrosstwoculturalcontexts.Cognitive Psychology,59,244–274.

Rhodes,M.,Gelman,S.A.,&Karuza,J.C.(2014).Preschoolontology:Therole of beliefs about category boundaries in early categorization.Journal of Cognition and Development,15,78–93.

Rhodes,M.,Leslie,S.J.,&Tworek,C. (2012).Culturaltransmissionofso-cialessentialism.Proceedings of the National Academy of Sciences of the United States of America,109,13526–13531.

Rhodes,M.,&Liebenson,P. (2015).Continuityandchange inthedevel-opmentofcategory-basedinduction:Thetestcaseofdiversity-basedreasoning.Cognitive Psychology,82,74–95.

Rothbart,M.,&Taylor,M.(1992).Categorylabelsandsocialreality:Doweviewsocialcategoriesasnaturalkinds?InG.Semin&K.Fiedler(Eds.),Language, interaction and social cognition(pp.11–36).ThousandOaks,CA:SagePublications.


Rutland,A.,Killen,M.,&Abrams,D.(2010).Anewsocial-cognitivedevel-opmentalperspectiveonprejudice:Theinterplaybetweenmoralityandgroupidentity.Perspectives on Psychological Science,5,279–291.

Salomon, E., & Cimpian, A. (2014). The inherence heuristic as a sourceof essentialist thought. Personality and Social Psychology Bulletin,40,1297–1315.

Segall,G.,Birnbaum,D.,Deeb,I.,&Diesendruck,G.(2015).Theintergen-erationaltransmissionofethnicessentialism:Howparentstalkcountsthemost.Developmental Science,18,543–555.

Shtulman,A.,&Schulz,L.(2008).Therelationbetweenessentialistbeliefsandevolutionaryreasoning.Cognitive Science,32,1049–1062.

Taylor,M.G.(1996).Thedevelopmentofchildren’sbeliefsaboutsocialandbi-ologicalaspectsofgenderdifferences.Child Development,67,1555–1571.

Taylor,M.G.,Rhodes,M.,&Gelman,S.A.(2009).Boyswillbeboys,cowswillbecows:Children’sessentialistreasoningaboutgenderandanimaldevelopment.Child Development,80,461–481.

Verkuyten, M. (2003). Discourses about ethnic group (de-)essentialism:Oppressiveandprogressiveaspects.British Journal of Social Psychology,42,371–391.

Waxman,S.R.,Medin,D.,&Ross,N.(2007).Folkbiologicalreasoningfromacross-culturaldevelopmentalperspective:Earlyessentialistnotionsareshapedbyculturalbeliefs.Developmental Psychology,43,294–308.

Williams,M.,&Eberhardt,J.(2008).Biologicalconceptionsofraceandthemotivationtocrossracialboundaries.Journal of Personality and Social Psychology,94,1033–1047.

Yamagishi,T.,Jin,N.,&Kiyonari,T.(1999).Boundedgeneralizedreciprocity:Ingroupboastingandingroupfavoritism.Advances in Group Processes,16,161–197.

Yzerbyt,V.,Judd,C.,&Corneille,O. (2004).The psychology of group per-ception: Perceived variability, entitativity, and essentialism. New York:PsychologyPress.

Zagefka,H.,Nigbur,D.,Gonzalez,R.,&Tip,L.(2012).Whydoesingroupes-sentialismincreaseprejudiceagainstminoritymembers? International Journal of Psychology,48,60–68.

How to cite this article:RhodesM,LeslieS-J,SaundersK,DunhamY,CimpianA.Howdoessocialessentialismaffectthedevelopmentofinter-grouprelations?.Dev Sci. 2017;00:e12509.https://doi.org/10.1111/desc.12509

https://doi.org/10.1111/desc.12509

Documents

How does social essentialism affect the development of ...sjleslie/Rhodes_et_al-2017-Development… · 2 of 15 | RHODES Et al. allow (Gelman & Rhodes, 2012; Shtulman & Schulz, 2008),